Tobacco Documents Online

+ ENVIRONMENTAL, TOBACCO SMOKE - Measuring Exposures . and Assessung Health Effects Committee on Passive Smoking Board on Environmental Studies and'Ibuticology National Research Council NATIONAL ACADEMY PRESS - - -- --- -- - - Washington, U G. 1986

National Academy Press * 2101 Constitution Avenue, NW ' Washington, DC 20418 NOTICE: The project that is the subject of this report was approved by - the Governing Board of the National Research Council, ncil, whose members are -lrawn from the councils of the National Academy of Sciences, the National Academy of Engineering, and the Institute of Medicine. The membsrs of the committee responsible for the report were chosen for their special competences and with regard for appropriate balance. This report has been reviewed by y a group other than the authors according to procedures approved by a Report Review Committee consisting of inembers of the National Academy of Sciences, the National Academy of Lngineering, and the Institute of Medicine. The National Research Council was established by the National Academy of Sciences In 1916 to associate the broad community of science and technol- ogy with the Academy's purposes of furthering knowledge and of advising the federal government. The Council operates in accordance with general policies - - determined by the Academy under the -authority of its congressional charter of 1863, which establishes the Academy as a private, nonprofit, seif-governing membership corporation. The Council has become the principal operating - agency of both the National Academy of Sciences and the National Academy - of Engineering In the cortduct of their services to the government, the public, and the scientific and engineering communities. It In administered jointly by both Academies and the Institute of Medicine. The National Academy of Engineering and the Institute of Medicine were established In 1964 and 1970, respectively, under the charter of the National Academy of Sciences. This study was prepared under EPA Contract #68-02-4073 and De- partment of Health and Human Services, Public Health Services Grant #ASU000001-00-S1. The content of this publication does not necessarily reflect the views or policies of the U.S. Environmental Protection Agency or the Department of Health and Human Services, and an official endorsement -- should not be inferred. - INTERNATIONAL STANDARD BOOK NUMBER 0-309-03730-1 LIBRARY OF CONGRESS CATALOG CARD NUMBER 86-28622 Copyright Q 1986 by the National Academy of Sciences No part of this book may be reproduced by any mechanical, photographic, or electronic process, or In the form of a phonographic recording, nor may it be stored In a retrieval system, transmitted, or otherwise copied for public or private use, without written permission from the publisher, except for the purposes of official use by the United States Government. Printed In the United States of America 46R9e44e BOARD ON ENVIRONMENTAL STUI)Ih";.? . AND TOXICOLOGY DONALD HORNIG, Harvard University, Boston, Massachusetts, -- - Chairman ALVIN L. ALM, Thermal Analytical, Inc., Waltham, Massachusetts RICHARD N. L. ANDREWS, University of North Carolina, Chapel Hill, North Carolina WILLIAM E. COOPER, Michigan State University, East Lansing, Michigan JOHN DOULL, University of Kansas Medical Center, Kansas City, Kansas EMMANUEL FARBER, University of Toronto, Toronto, Ontario, Canada BENJAMIN a. FERRIS, Harvard School of Public Health, Boston, Maseachusetts PHILIP LANDRIGAN, Mt. Sinai Medical Center, New York,.New York RAYMOND O. LOEHR, University of Texas, Austin, Texas ROGER MINEAR, University of Illinois, Urbana, Illinois PHILIP A. PALMER, E.I. DuPont de Nemours & Co., Wilmington, - Delaware EMIL PFITZER, Hoffman-La Roche, Inc., Nutley, New Jersey PAUL PORTNEY, Resources for the Future, Washington, D.C. PAUL RISSER, Illinois Natural History Survey, Champaign, Illinois WILLIAM H. RODGERS, University of Washington, Seattle, Washington F. SHERWOOD ROWLAND, University of California, Irvine, California LIANE B. RUSSELL, Oak Ridge National Laboratory, Oak Ridge, Tennessee ELLEN SILBERaELD, Environmental Defense Fund, Washington, D.C. - PETER SPENCER, Albert Einstein College of Medicine, Bronx, - New York National Research Council Staff DEVRA LEE DAVIS, Acting Director, BEST JACQUELINE PRINCE, Staff Associate tu

COMMITTEE ON PASSIVE SMOKING BARBARA S. HULKA, University of North Carolina, Chapel Hill, North Carolina, Chairman 0LAV AXELSON, University Hospital, Linkoping, Sweden JOSEPH BRAIN, Harvard School of Public Health, Boston, Massachusetts PATRIOIA BUFFLER, University of Texas at Houston, Houston, Texas A. SONIA BUIST, Oregon Health Sciences University, Portland, Oregon DIETRICH HOFFMANN, American Health Foundation, Valhalla, New York BRIAN LEADERER, Yale University, New Haven, Connecticut aENEVIEVE MATANOSKI, Johns Hopkins University, Baltimore, Maryland JAMES ROBINS, Harvard School of Public Health, Boston, Massachusetts JOHN SPENC3LER, Harvard School of Public Health, Boston, Massachusetts NICHOLAS WALD, Medical College of St. Bartholomew's Hespital, - - London, England National Research Council Staff DEVRA LEE DAVIS, Acting Director, BEST DIANE K. WAGENER, Project Director MARVIN SOHNEIDERMAN, Senior Sta-ff- Officer RIOHARD E. MORRIS, Editor EDNA W. PAULSON, Information Specialist MARY ELLEN SOHENKENBAOH, Staff Assistant JULIETTE L. WALKER, Senior Secretary Pr-eface The Office of Air and Radiation of the Environmental Pro- ---- - - tection Agency and the Office on Smoking and Health of the - Department of Health and Human Services asked the National - Research Council to evaluate methods for assessing exposure to --- environmental tobacco smoke and to review the literature on the - health consequences from such exposures. The National Research Council responded to this request by appointing 11 scientists to serve on the Committee on Passive Smoking, in the Board on Envi- - ronmental Studies and lbxicology, under the Commission on Life - - Sciences+ The committee membership berehip represented the disciplines of toxicology, biochemistry, atmospheric science, epidemiology, - - 'bioatatistica, and pulmonary physiology. The committee's charge was to review the existing scientific - literature and to identify the current state of knowledge with respect to known facts and areas of uncertainty. Many more - --- - of the latter were found than the former. Ti the extent that ---- - they could be justified scientifically, conclusions have been stated --- and recommendations proposed. Many of the recommendations are for future research, rather than for public policy. The latter were for the most part avoided on two grounds: the data were - - frequently not sufficiently secure and the charge to the committee was primarily for scientific review. The committee conducted a public hearing on scientific stud- ies relevant to its charge on January 29, 1986. Furthermore, it reviewed the published ecientific literature and received testimony from professional societies; medical, industry, consumer, and pub- lic interest groups; academic scientists; and others involved in the generation and interpretation of scientific evidence on the health 96999449 iv

consequences of exposure to cigarette smoking. Pursuance of these activities was followed by the preparation of individual chapters by - committee members and consultants. Thereafter, chapters were discussed, revised, and integrated with each other for the full report. In producing this report, the committee confronted a complex charge under severe time constraints.. That it completed its task well and on time is a credit both to its members and the scientific staff of the National Research Council. I would like to express my personal appreciation to every one of the committee members, all of whom donated their time, intellect, and knowledge to the sub- stance stance of this report. Dr. Diane Wagener of the National Research Council assumed the difficult task of coordinating, translating, and negotiating ideas and insights among committee members, -- - consultants, and reviewers. Drs. Devra Davis and Marvin Schnei- derman worked with Dr. Wagener in ensuring the thoughtful and derman - timely completion of this report. While the committee restricted itself to analysis of the sci- entific data, it was not unmindful of the fact of modern life that smokers and nonsmokers have taken strong positions regarding the right to smoke on the one hand and a rejection of being exposed to other people's smoke on the other. Persons on each side of the issue may wish to infer information from this report that the committee did not intend. Our strategy has been to synthesize in- formation, present judgments and conclusions wherever possible, and to recognize inadequacies in existing data in order to provide - - a focus for future research. We have not taken the stance of a public policy board that necessarily has to make decisions on less- than-adequate information. Rather, we have chosen to prepare a scientifically responsible report that will be intelligible to a lay audience and useful to a scientific one. ' BARBARA S. IiULKA, Chairman - Committee on Passive Smoking I Acknowledgments The preparation of this report by the Committee on Passive Smoking vaould not have been possible without assistance from a large number of people. Thq committee consulted with a number of experts about var- ious topics. We would like to thank the Office on Smoking and ' Health, particularly Clarisse Brown, who provided us with the many statistics and data that were requested by various mem- bers of the committee. We would also like to thank William Cain and Edward LaVoie for their contributions. Other individuals who gave special assistance in the preparation of the report in- clude-Lesl'te Waters Barger, Kiran Nanchahal, Simon Thompson, Christopher Frost, and Don Blevins. The committee thanks all the peer reviewers of the report: Their constructive remarks contributed to the improvement of presentations of technical information and its readability. We would like to express our thanks to the NRC staff for their work in supporting the committee. We would especially like to thank Edna W. Paulson and the staff of the Toxicology Information Center, who were of great assistance. 66~9~~g ~~ vii

Contents flf}69QlZR I EXECUTIVE $UMMARY ........................................ 1 Introduction, I Environmental Tobacco Smoke, 2 Measures of Exposure, 3 In Vive and In Vitro Studies, 7 Health Effeets, 7 1 INTRODUCTION ........................................... .13_ DeHnitions, 14 Trends in Cigarette Usage, 15 Organization, 20 References, 21 Part I I'NYSICOCHEMICAL AND TOXICOLOGICAL STUDIES OF ENVIRONMENTAL TOBACCO SMOKE 2 THE PHYSICOCHEMICAL NATURE OF SIDESTREAM SMOKE AND ENVIRONMENTAL TOBACCO SMO_ KE ........................................ . . 25_ - Introduction, 25 Sidestream Smoke, 28 - - Principal Chemical Constituents of Environmental Tobacco Smoke, 36 Radioactivity of Environmental Tobacco Smoke, 37 - - Toxic and Carcinogenic Agents in Tobacco Smoke, 44 Summary and Recommendations, 45 References, 48

3 IN VIVO AND IN VITRO ASSAYS TO ASSESS THE HEALTH EFFECTS OF ENVIRONMENTAL TOBACCO ACCO SMOKE .................64 Introduction, 54 In Vivo Assays on Environmental Tobacco Smoke, 55 In Vitro Assays on Environmental Tobacco Smoke, 58 Summary and Recommendations, 59 References, 61 Part II ASSESSING EXPOSURES TO - ENVIRONMENTAL TOBACCO SMOKE 4 INTRODUCTION ........................................... 6b ASSESSING EXPOSURES TO ENVIRONMENTAL TOBACCO SMOKE IN THE EXTERNAL ENVIRONMENT ........................................... .69 Tracers for Environmental Tobacco Smoke, 70 Personal Monitoring, 76 Concentrations of Environmental Tobacco Smoke in Indoor Environments, 79 Modeling, 81 Summary and Recommendations, 94 References, 97 6 ASSESSING EXPOSURES TO ENVIRONMENTAL TOBACCO SMOKE USING QUESTIONNAIRES .......101 Exposure Histories Derived from Questionnaires, 102 Environmental Tobacco Smoke Exposure Data for Studies of Acute and Chronic Health Effects, 107 Data Quality, 108 Other Variables, 115 Summary and Recommendations,116 References, 118 EXPOSURE-DOSE RELATIONSHIPS FOR ENVIRONMENTAL TOBACCO SMOIfE ................ 120 Estimating Dose, 120 Particle Size, 121 - Breathing Pattern, 122 Deposition of Cigarette Smoke Particles, 123 Particle Retention in the Lungs, 126 Gases in Environmental Tobacco Smoke, 127 Summary and Recommendations, 129 References, 131 8 ASSESSING EXPOSURES TO ENVIRONMENTAL TOBACCO SMOKE USING BIOLOGICAL MARKERS ..... . .. .. .... ... . . . ............ 1gg Biological Markers in Physiological Fluids, 134 Genotoxicity of the Urine, 148 Future Needs, 152 Summary and Recommendations, 152 References, 154 Part III HEALTH EFFECTS POSSIBLY ASSOCIATED WITH EXPOSURE TO ENVIRONMENTAL TOBACCO SMOKE BY NONSMOKERS INTRODUCTION ......................................... 163 10 SENSORY REACTIONS TO AND IRRITATION EFFECTS OF ENVIRONMENTAL TOBACCO SMOKE .......................................166 Odor, --- 166 Irritation, 172 Hypersensitive Individuals, 176 Summary and Recommendations, 177 References, 179 xi

11 EFFECTS OF EXPOSURE TO ENVIRONMENTAL TOBACCO SMOKE ON LUNG FUNCTION AND RESPIRATORY SYMPTOMS ............................ _ 182 Lung Function and Symptoms in Active Smokers, 182 Plausibility for an Effect Due to Passive Smoking, 194 Methodologic Considerations for Epidemiologic Studies, 185 Cross-sectional Studies, 188 Longitudinal Studies of Lung Function in Children and Adults, 200 The Effect of Passive Smoking on Respiratory Infections, 202 When Do Pulmonary Effects of Passive Smoking Occur?, 209 Studies of Acute Pulmonary Effects, 212 Summary and Recommendations, 216 References, 218 12_ EXPOSURE TO ENVIRONMENTAL TOBACCO SMOKE AND LUNG CANCER ........................... 223 Using Biological Markers to Estimate Risk, 224 Assessing the Risk From Epidemiologic Studies of Lung Cancer and Exposure to ETS, 227 Corrections to Estimates for Systematic Errors, 231 Other Considerations, 242 Summary and Recommendations, 245 References, 246 13 CANCERS OTHER THAN LUNG CANCER ............ 250 Smoking-Related Cancers, 250 Cancers Not Related to Smoking, 252 Interpretation, 254 Summary and Recornmendations, 255 References, 255 14 CARDIOVASCULAR SYSZ'-EM ........................... _ 25q Acute Cardiovascular Effects of Environmental Tobacco Smoke Exposure, 257 Cardiovascular Disease Morbidity and Mortality, 262 Summary and Recommendations, 265 References, 266 ZO~'i99449 Xli 15 OTHER HEALTH CONSIDERATIONS IN CHILDREN ................................................ 2_ 269 Environmental Tobacco Smoke Exposure by Nonsmoking - Pregnant Women, 269_ Growth in Children, 271 Chronic Ear Infections, 272 - Summary and Recommendations, 273 • References, 274 + APPENDIXES A. Guidelines for Public and Occupational Chemical Exposures to Materials That Are Also Found in Environmental Tobacco Smoke ............................ . 279 B. Method of Combining Data From Studies of Environmental Tobacco Smoke Exposure and Lung Cancer ................................................284 Case-Control Studies, 284 Prospective (or Cohort) Studies, 286 Summing Over Studies, 287 - References, 288 C. Adjustments to Epidemiologic Estimates of Excess Lung Cancer in Persons Exposed to Environmental Tobacco Smoke , ............................ 289 Using Cotinine Measurements to Correct Misreporting, 290 References, 293 D. Risk Assessment-Exposure to Environmental Tobacco Smoke and Lung Cancer ..........................294 James Ro6ins Introduction, 294 D-1 Estimation of the Ttue Relative Risk, 297 D-2 The Carcinogen-Equivalent Number of Actively Smoked Cigarettes Inhaled Daily by Passive Smokers: Comparisons of Epidemiologic with Dosimetric Estimates, 301 D-3 Estimating the Number of Lung Cancer Deaths in Nonsmokers in 1985 Attributable to ETS, 304 D-4 Lifetime Risk of Death From Lung Cancer Attributable to ETS, 306 Discussion, 311 Technical Discussions, 313 References, 336

, Executive Summary , INTRODUCTION A Committee of the National Research Council's (NRC's) Board on Environmental Studies and Toxicology prepared this re- port in response to requests from two federal government agencies, the Office of Air and Radiation of the Environmental Protection Agency (EPA) and the Office of Smoking and Health of the De- partment -- - partment of Health and Human Services. The report evaluates methodologies in epidemiologic and related studies for obtaining measurements of exposure to environmental tobacco smoke (ETS) by nonsmokers and also outlines the possible health effects of such exposures as reported in the published literature. This committee was asked to review original research data and identify research needs but was not charg,ed with preparing policy statements or rec- ommendations for public health actions. In particular, the NRC . was asked to: e review the chemical and physical characterizations of the - -- - constituents of ETS; - --- * include a toxicological profile of sidestream and environ- mental tobacco smoke; - -- --- * review the epidemiologic and related literature on the health effects of exposure to ETS; 'and * recommend future exposure monitoring, modeling, and -- epidemiologic research. To address these and related issues, the NRC fnrmed the - Committee on Passive Smoking in the Board on Environmental Studies and Toxicology of the Commission on Life Sciences. The ~0698~~8 ! . i

i committee consists of professionals in a variety of fields, includ- ing epidemiology, toxicology, biochemistry, atmospheric science, biostatistics, and pulmonary physiology. The subject of the committee's report is the use of epidemiol- ogy and related disciplines for the study of possible health effects of exposure to ETS by nonsmokers. Smokers are also exposed to ETS, but the health effects of this exposure, which are likely to be less intense than those of active smoking, are not the subject of - this report. The primary goal of the studies reviewed in this report is to determine whether there is a relationship between health out- comes in hum- an populations and ETS-exposure of nonsmokers. It is a formidable task to assess exposure to the complex mixture of ETS with enough precision to permit use in analytic studies, including quantitative risk estimation. For some health outcomes the relevant durAtion of exposure may be minutes, for others it may be decades. Numerous factors, in addition to exposure to smoke, can influence the risk of illness. These other factors must be taken into account if the magnitude of the effects of exposure - - to ETS is to be evaluated. ENVIRONMENTAL TOBACCO SMOKE More than 3,800 compounds have been identified in cigarette smoke. The major source, by far, for ETS is sidestream smoke - (SS) which is emitted from the burning end of a cigarette in be- tween - _ -- tween puffs. The remainder of ETS consists of exhaled mainstream smoke (MS), smoke which escapes from the burning end during puff-drawing, and gases which diffuse during smoking through the cigarette paper. Each of the mixtures, MS, SS, and ETS, is an aerosol consisting of a particulate phase and a vapor phase. How- ever, the smokes of MS, SS, and ETS differ, as the result of changes in the concentrations of individual constituents, the phase (partic- ulate or vapor) in which the constituents are present, and various secondary reactions that chemically and physically alter ("age") the composition of the smoke. Undiluted SS contains higher con- centrations of some toxic compounds than undiluted MS, including ammonia, volatile aminea, volatile nitrosamines, nicotine decom- position products, and aromatic amines. However, concentrations of these SS emissions are considerably diluted in the indoor space where ETS exposures take place. 'The hydrophobic vapor phase V0G991,49 constituents of ETS are likely to enter the lung of the exposed indi- vidual, while the hydrophilic vapor phase constituents are likely to be absorbed in the upper respiratory tract. Particles <2.5 Nm (in this report referred to as respirable suspended particulates [RSP)) dominate the particulate phaee of ETS and can be inhaled deeply into the lung. Standard laboratory procedures have been established to as- sess the physicochemical properties of SS and MS. Research is needed to standardize both the collection and evaluation of ETS so that the effects of ETS can be studied in laboratories and in human populations. The changes in distribution of particular constituents of ETS as the smoke ages in the indoor environment are largely unknown. For example, it is known that almost all of the nicotine shifts from the particulate phase in MS and fresh SS to the vapor phase in ETS. Consequently, indoor air-cleaning systems designed to remove particles will not greatly alter the nicotine exposure, but may alter the concentrations of other noxious or toxic components. Research is needed to determine the distribution of constituents in the particulate and vapor phases of aged ETS. Also, the efficiency of air-cleaning systefns in removing the constituents needs to be studied. Indoor radon comes from sources in the environment and decays to short-lived radon daughters, which may become bound to the RSP in ETS. However, some long-lived radon daughters come from tobacco itself. Research should be conducted on possible interactions between ETS and radon daughters, especially as radon n daughters can adhere to RSP and increase the potential hazard of ETS. MEASURES OF EXPOSURE There are currently no direct measures of the dose absorbed of ETS in a population under study. Exposures to ETS, however, can be assessed by questionnaires, air monitoring, modeling of concen- trations, or biological markers. Future epidernioloyic studies should incorporate into their design several of these exposure assessment metAods in order to assess exposures to ETS more accurately and to estimate dose.

Questionnaires The simplest measure of E"fS exposure is contained in the reply to the questions: "Are you a cigarette smoker?" and "lf you are a nonsmoker, do you live with, or work with, or have regular contact with persons who are smokers?" There are great , difficulties in developing uniform questions that elicit unambigu2 ous replies and, more particularly, in using these replies to make`` firm quantitative estimates of exposure. They can be used, how- ever, as a-b- asis for classifying individusls into broad categories of exposure, recognizing the problems such as incorrectly estimating exposure through errors in reporting of current smoking habits, neglecting exposure to ETS in other environments like workplaces or public places, and reporting an exsmoker as a nonsmoker. Re- po#ts of whether or not the subject has smoked can be obtained with reasonable reliability from surrogate respondents. However, quantification of integrated exposure over many years is not likely to be fully reliable or precise. At best, such quantification pro- vides an approximation of exposure, whether the information is obtained from the individual himself or from a surrogate. To esti- mate integrated exposure to ETS, future studies need to estimate a long-term ETS exposure history, including what fraction of the day is spent in the presence of ETS and at what ages these exposures occurred. The data from such a history should be entered into a specific time-place model, from which cumulative exposure can be estimated. Monitoring The use of air monitoring (personal or indoor space) is hand- icapped by the lack of a clear definition of the physicochemical nature of ETS and the identification of the individual, or target, constituents of ETS associated with the health or comfort effects under study. Proxy, or surrogate, constituents have been mea- sured in a number of studies as indicators of ETS exposure in both personal and indoor space monitoring. RSP, carbon monox- ide, nicotine, nitrogen oxides, acrolein, nitroso-compounds, and benzo(alpyrene are some of the compounds or classes of air con- taminants that have been measured under field conditions as in- dicators of ETS exposure. While some of the ETS constituents, particularly nicotine and RSP, have proved to be useful surrogates S©G98US 5 for ETS, no single measure has completely met all the criteria for an ideal ETS surrogate. To facilitate the study of the health effects of ETS exposure, an ideal marker or tracer of exposure to ETS should be unique (or nearly unique) to tobacco smoke, should be a constituent of tobacco smoke, that is present in su -fficient quantity so it can be measured even at low ETS levels, and should stand in a fairly constant ratio across brands of cigarettes to other tobacco smoke constituents (or contaminants) of interest. Reliable infor- mation needs to be obtained on the quantity, transport, and fate of such chemicals in ordinary indoor environments. A majority of field studies have used RSP as an indicator of exposure to ETS because of the substantial emission of RSP in indoor spaces from tobacco combustion. ETS is the dominant con- tributor to the indoor levels of RSP. The total RSP, as measured by personal monitors, has been found to be substantially elevated for individuals who reported being exposed to M as compared with those who reported no such exposure. Both air monitoring and modeling clearly indicate that RSP concentrations will be elevated over background levels in indoor spaces when even low smoking rates occur. The importance of variation in_ the_ input parameters- such as room size, temperature, humidity, air exchange rate, and numbers of cigarettes smohed--ahould be noted when interpreting the data on the constituents of ETS obtained from personal moni- tors and indoor space monitora. Biological Markers In theory, dose of ETS to the tissues or organs could be mea- sured directly through the use of biological markers that accurately indicate uptake in the tissues or organs. Optimal assessment of ex- posure to ETS should derive from measures made on physiological fluids of exposed persons. Several chemicals found in such fluids may be able to serve as biological markers of recent exposures. The criteria for acceptable biological markers are similar to those for measuring ETS in the external environment. The biological markers that have been most ost useful for as-g sessing recent exposures to ETS are nicotine and its metabolite, - -- - -- -- - cotinine. Nicotine and cotinine derive virtually exclusively from - -- - tobacco products, of which tobacco smoke ke is the most important ---- -- - direct source. They can be identified and quantified in saliva, -- -- - --- - - blood, or urine. Generally, the mean concentrations of nicotine

, 6 and cotinine in the plasma or urine of nonsmokers exposed to ETS - are about 1 percent of the mean values observed in active smokers. ' Several studies have indicated that urinary cotinine concentrations ' in infants and children increase as the numbers of reported smok- ers ers increase in the home. At present, there may be difficulty in - interpreting the relative cotinine levels in.nonsmokers compared with smokers because of the reported slower clearance of eotinine in nonsmokers. Absorption, metabolism, and excretion of ETS constituents, including nicotine, need to be carefully studied in or- der to evaluate whether there are differences between smokers and nonsmokers in these Jaetors. Ptirther epidemiologic studies using biological markers are needed to quantify exposure-dose relation- ships in nonsmokers. Thiocyanate, as measured in saliva, serum, or urine, does not appear to be sufficiently sensitive as an Indicator of ETS exposure. Similarly, exhaled carbon monoxide and carboxyhemoglobin are not sufficiently sensitive to moderate or low levels of ETS exposure - and thus are not particularly useful biological markers for expo- sure to ETS, except in experimental, acute exposure situations. There are several other sources of carbon monoxide in the environ- ment that equal or exceed the concentrations of carbon monoxide attributable to ETS. Other suggested biological markers of exposure are N-nitroso- proline, nitrosothioproline, and some of the aromatic amines that are present in high concentrations in SS. However, data on sensi- tivity and reliability of laboratory procedures for these markers are not eufl'icient to recommend their use at this time in epidemiologic studies of ETS. Laboratory assays have shown mutagenic activity in the urine of smokers and ETS-exposed nonsmokers. The mutagenicity of urine is a function of many factors--such as dietary constituents, occupational exposures, and other environmental factors-which render any findings of mutagenicity nonspecific. Research is needed to clarify the appropriate methods for estimating mutagenicity and to isolate and identify the active agents in body fluids o f ETS- exposed nonsmokers. DNA adducts derived from tobacco-related chemicals can be measured in the blood. However, -these chemicals, such as b®nzo[a]pyrene, are not unique to ETS. Studies are needed that can measure adducts of tobacco-specific chemicals. 90G99448 IN VIVO AND IN VITRO STUDIES Laboratory studies can contribute to a better understanding of the factors and mechanisms involved in the induction of disease by environmental agents. There have been numerous bioassays conducted on MS. In examining the effects of MS, many research workers have used condensates of the smoke painted on the shaved skin of mice. This contrasts with the human exposure that is mainly in the respiratory tract. Nonetheless, these skin-painting studies have been useful in examining the carcinogenicity of dif_ ferent tobacco constituents and thus advancing knowledge of the actions of MS on a gross exposure level. Similar work with skin painting has not been done with ETS and would be of value for assessing the differential toxicity ojETS and MS. In constrast to MS exposure, ETS exposure involves propor- tionately more exposure to gas phase than to particulate phase ecinstituents. There have not, however, been studies of the effects of exposure to aged ETS. The relative in vivo toxicity of MS, SS, and ETS needs to be assessed. Some studies have attempted to evaluate the gas phase of MS. SS, and ETS in short-term, in vitro assays. A solution of the gas phase of MS has been shown to induce dose-dependent increases -- in sister-chromatid exchanges in cultured human lymphocytes. Mutagenic activity has been found in the particulate matter of SS --- -- - - - an.d in condensates of ETS. However, the work done to date is too sparse to permit any estimates of the mutagenicity of ETS per ae, - - even though most of ETS consists of SS. Further in vitro assays of ETS are needed. HEALTH EFFECTS This report reviews both chronic and acute health effects as- sociated with ETS exposure in nonsmokers. Most epidemiologic studies of chronic health effects have been conducted on persons who have had long-term exposures to ETS from household mem- bers. The studies do not directly address chronic,health effects in individuals who are exposed at work or have occasional exposures in the home or elsewhere. Hecause the physicochemical nature of ETS, MS, and SS dif- fer, the extrapolation of health effects from studies of MS or of

i a active smokers to nonsmokers exposed to ETS may not be appro- priate. However, chemicals known to be toxic and carcinogenic in MS are also present in ETS. Laboratory studies in conjunction with epidemiologic investigations are needed to help clarify possible - - health effects of exposure to ETS in nonsmokers. Acute, Noxious Effects The most common acute effects associated with exposure to ETS are eye, nose, and throat irritation, and objectionable smell of tobacco smoke. Tobacco smoke has a distinct and persistent - odor, making control through ventilation particularly difficult. In closed rooms where smoking is allowed, a ventilation rate of greater than 50 cubic feet per minute per occupant is necessary to achieve air quality that is acceptable to more than 80% of adults entering the room as contrasted with rates of less than 10 cubic - feet per minute per occupant when there is no smoking or other pollution. Annoyance with noxious tobacco odor largely governs the reactions of visitors, while occupants of smoky rooms are more likely to complain about Irritating effects to the eye, nose, or throat. Particle filtration appears to lead to little or no decline in odor and irritation, suggesting that the effects are produced by gas-phase constituents. During exposure to ETS, eye blink rate is correlated with sensory irritation, such as burning eyes and nasal irritation. For some persons, eye tearing can be eo intense as to be incapacitating. There is some evidence that nonsmokers are more sensitive to the noxious qualities of cigarette smoke than are smokers. Objective physiological or biochemical indices should be sought to validate reports ojnosious reactions and chronic irritation associated with ETS. Smoke contains immunogens, that is, substances that can ac- tivate the immune system. Approximately half of atopic (allergy prone) individuals react to various extracts of tobacco leaf or amoke presented in skin tests.. However, the components of the extract that are responsible for this reaction have not been iso- lated. There is little correlation between positive reactions to skin tests and self reported complaints of tobacco smoke sensitivity. Research is needed to evaluate the medical importance in stopic persons of these positive reactions to skin tests using ETS extracts and to relate immune response on skin tests to subjective complaints about the noxious, irritating properties of tobacco smoke. 9 Respiratory Symptoms and Lung Function Respiratory symptoms, such as wheezing, coughing, and spu- tum production, are increased in children of smoking parents. These symptoms are more common in children of smokers than children of nonsmokers. The largest studies place the increased risk of 20 to 80%, depending on the symptom being assessed and number of smokers in the household. Also, respiratory infections manifested as pneumonia and bronchitis are significantly increased in infants of smoking parents.. Some studies have reported that in- fants fants of smoking parents are hospitalized for respiratory infections more frequently than children of nonsmokers. Among children aged under 1 year, studies are remarkably consistent in showing an increased risk of respiratory infections among children living in homes where parents smoke. There is a dose-response relationship that relates more to maternal smoking than paternal smoking. The association persists after allowing for possible confounding factors such as occupational data, respiratory illness in the parents, and birthweight. The mechanisms of the increased risk may either be - - a direct effect of ETS or due to a higher risk of cross-infection in such homes. Regardless of the mechanism, the exposure of small children to smoking in the home appears_ to put them at risk of - --- - respiratory illness. Since children exposed to ETS from parental smoking have an increased frequency of pulmonary symptoms and respiratory infections, it is prudent to eliminate ETS exposure from the envi- ronments of small children. There is some evidence that parental smoking may affect the - rate of lung growth in children. In children with one or more par- ents who smoke, lung function increase, which is a normal growth phenomenon, shows a small decrease in the rate of growth. An important issue currently unresolved is whether a child who is affected by exposure to ETS from parental smoking may be at an increased risk for the development of chronic airflow obstruction in adult life. In all studies of children, it is difficult to distin- guish - guish between the role of ETS exposure in utero and postnatally. Research is needed to address the issues of ETS exposure during childhood and fetal life and its possible relationship with airway hyperresponsiveness and pulmonary diseases in adult life. 4OG9gLtae

10 Three studies have shown a small reduction in pulmonary - function in normal adults exposed to ETS. Interpretation of these findings is difficult because pulmonary effects in normal adults are likely to reflect the cumulative burden of many environmental and occupational exposures and other insults to the lung. Thus, the effects of ETS on the lungs of adults are likely to be confounded by many other factors, making it difficult to attribute any portion of the effect solely to ETS. In some studies of asthmatics, in whom pulmonary reactions to ETS should be more readily produced, no effects on lung function were reported. In other studies, asthmatics reported complaints upon exposure to ETS and showed significant pulmonary func- tion changes after experimental smoke exposure. .Future studies of asthmatics exposed to ET3 should be designed so as to limit the distortion produced by heterogeneous patient groups, varying medication schedules, and psychogenic effects of ETS. Lung Cancer Considering the evidence as a whole, exposure to ETS in- creases the incidence of lung cancer in nonsmokers. Estimates of the magnitude of the increased risk vary. Among studies of var- ious populations in Europe, Asia, and North America, the risk of lung cancer is roughly 30% higher for nonsmoking spouses of smokers than it is for nonsmoking spouses of nonsmokers. There is consistency among the studies in that all of the studies indi- vidually include the 30% increased risk within the 95% confidence intervals. Patterns and extent of exposure may vary in different communities and countries. Based on presently available epidemi- ologic ologic data, the estimate of the increased risk from the American studies is lower than the average for all the studies, though not sig- _- nificantly so. These estimates are almost exclusively derived from the comparison of persons identified as exposed, or unexposed, on the basis of their spouse's smoking habits. Certain errors in the reporting of smoking habits have proba- bly contributed to the risks observed in the epidemiologic studies. Misclassification of current or examokers as nonsmokers would tend to produce an observed relative risk that is larger than the true risk. This effect was studied in detail using estimates of the ex- tent of the errors involved and judged to contribute only a portion tent - - of the excess risk. Underestimation of the increased risk might also 80698448 11 be introduced because ause the supposedly unexposed population had some exposure to ETS, although they were classified as unexposed in the studies. Taking both types of errors into account produces an estimate of the excess lung cancer risk for nonsmokers married to smokers compared with completely unexposed individuals that is similar to the relative risk observed in the epidemiologic studies considered. Since carcinogenic agents contained in ETS are inhaled by nonsmokers, in the absence of a threshold for carcinogenic effects, gn increased risk of lung cancer due to ETS exposure is biologi- cally plausible. Laboratory studies would be important in determin- ing the concentrations of carcinogenic constituents of ETS present in typical daily environments. The use of biological markers in epidemiologic studies is recommended to more precisely. quantify dose-response relationships between ETS exposure and lung cancer occurrence. Other Cancers There have been few studies of risk for cancers other than lung in nonsmokers exposed to ETS. Some of the sites considered have been brain, hematopoetic, and all sites combined. The results of these studies have been inconsistent. Whether or not there is an association between ETS exposure and cancers of any site other than lung is an important topic for future epidemiologic inquiries. Cardiovascular Disease Since active smoking has an adverse effect on cardiovascular disease morbidity and mortality, ETS exposure has also become - suspect. Reports have noted an excess risk of cardiovascular r dis- ease in ETS-exposed nonsmokers; however, methodologic prob- lems in the designs and analyses of these studies preclude any firm conclusions about the results. Studies reporting that ETS can precipitate the onset of angina pectoris among people who already have this condition are subject to the same-precautionary note. - Exposure to ETS produced no statistically significant effects on heart rate or blood pressure in school-aged children or healthy adult subjects, either during exercise or at rest. Data are not available as to possible adverse cardiovascular effects in suscepti- ble ble populations, such as infants, elderly, or diseased individuals.

12 Further esperimental and observational studies ahould be conducted to assess the effect of long-term and acute ETS exposure on cardiac function, blood pressure, and angina in nonsmokera. Other Health Considerations in Children Several other health outcomes have been studied that relate to the growth and health of children. For all postnatal outcomes, it is often not possible to differentiate the effect of in utero exposure to ETS from subsequent childhood exposures to ETS. Nonsmoking pregnant women exposed to smoking spouses have been reported to produce babies of lower birthweight than nonsmoking women with nonsmoking spouses. Some studies have noted a dose-response relationship between the number of ciga- rettes smoked by fathers and birthweight of the offspring. Ad- ditional studies of intrauterine fetal growth retardation associated with ETS exposure of nonsmoking mothers need to be conducted with better assessments of the magnitude of ETS exposure. Several studies have examined possible relationships between chronic exposure to ETS by children and parameters of growth and development. Growth is an.especially difficult phenomenon to study since many factors-such as genetics, nutrition, social class, and ethnicity-play important roles. It is difficult to assign proportional causality to each factor. Moreover, height and weight ratios and other growth measures are not reliably obtained in standard pediatric surveys. A few studies have shown that children of smokers have reduced growth and development, and one study reported a dose-response relationship between reduced height and increasing numbers of cigarettes smoked in the home by either the mother or the father. Further work is needed to determine the nature of this association. Otitis media is a common occurrence in young children. In several studies, parental smoking, along with several other risk factors, has been linked to increased risk of chronic ear infections in children. Further work is needed to determine whether the asso• ciation is causal. 1 Introduction Environmental tobacco smoke (ETS) occurs in homes, at - workplaces, and in public places. The acute irritating and nox- ious effects of involuntary exposure to ETS, or °paasive smoking," are well established. Based in part on these irritating proper- ties of ETS, a recent report of the NRC recommended a ban on - - smoking in the small enclosed spaces of airliner cabins (National Research Council, 1986). More than 20 states and numerous lo- cal governments have enacted legislation and policies restricting smoking (1985 information obtained from the Office on Smoking and Health, personal communications). Such public information campaigns and other actions have convinced a large portion of the population that active cigarette smoking is dangerous to health. To many, this also implies that exposure to ETS can affect health. This report, in part, evaluates whether the latter beliefs are war- ranted. It also makes recommendations for future exposure moni- toring and epidemiologic research. The issues are complex. In some cases the conclusions are - uncertain, because much of the scientific data necessary to shed light on these concerns does not exist. Thi4 report addresses the following major issues pertaining to ETS: • The nature of the smoke. What constitutes ETS? What are the chemicals in ETS and what are the dilutions therein? There are two physical phases of smoke: particulate phase and vapor phase. What chemicals are in each phase? Are any of these chemicals carcinogenic or toxic, as determined in bioassays? E'iQ698L49 ~ 13

14 • Factors affecting exposure and the assessment of exposure. To what extent is the nonsmoker exposed to harmful chemicals that can be measured in ETbZ How can we measure exposure to _- ETS? Can ambient monitoring be used in epidemiological studies7 How reliable is questionnaire information? What constitutes the dose a person may receive? Are there objective measures of dose received, such as tobacco-smoke-specific biological markers? What choices and reasons for choice are there among the markers? • Effects of exposure. What are the health effects, if any, - consequent to exposure to ETS? Are these health effects related to discomfort or irritant effects only, or more serious disease? Are the potential health effects reversible when exposure ceases? What are the data from human studies? Do interactions with other environmental agents at workplaces or in homes need to be considered? Are there biologically plausible explanations for the various effects ascribed to ETS exposure? The report considers sensitive populations such as children, pregnant women, older persons, and those with persisting respira- tory illnesses. It does not consider the established effects on the fetus carried by a pregnant, smoking woman because this is not an instance in which a nonsmoking individual breathes ETS gen- erated by other people. However, a pregnant, nonsmoking woman might be affected by exposure to ETS, as may her fetus. The health effects considered include respiratory symptoms and- lung function, and other respiratory ailments (especially in children), such as asthma and allergic responses, cancer at various sites, and cardiovascular disease, among others. Some attention is paid to irritation, annoyance, and associated responses. DEFINITIONS Environmental tobacco smoke (ETS) originates from the smol- dering end of the tobacco product in between puffs, known as sidestream smoke' (SS), and from the smoker's exhaled smoke. [T-he smoke that the smoker inhales is known as mainstream smoke (MS).] Other contributors to ETS include minor amounts of smoke that escape during the puff-drawing from the burning cone and some vapor-phase components that diffuse through the cigarette paper into the environment. These various components are re- leased into the environment and are diluted by ambient air. They 15 may also aggregate with pollutants already in the environment and thereby change character. The composition of this complex mix- ture, known as ETS, has different physicochemical characteristics than the MS. There.are various terms in the literature that refer to the inhalation of ETS by nonamokers, e.g., "passive smoking," "in- voluntary smoking," and "breathing other people's smoke." We will refer to the inhalation of ETS by using the terms "passive smoking" and "exposure to ETS by nonsmokers" interchangeably. TRENDS IN CIGARETTE USAGE Exposure of nonsmokers to ETS is a function of several vari- ables, one of which is the number of active smokers with whom the nonsmoker comes into contact throu hout some g period of time. The percent of the population who smoke steadily increased over the first two-thirds of this century but has declined more recently. - - In 1980, 32% of the adult population considered themselves to be - - - cigarette smokers (U.S. Department of Commerce, 1984). This percentage, now roughly equal for men and for women, reflects ts a reduction of almost one-third in men since the publication of the first Surgeon General's Report on Smoking and Health in 1964 (U.S. Public Health Service, 1964). Figure 1-1 shows the trends in cigarette usage between 1955 and 1985 for males and females. Table 1-1 gives cigarette consumption since 190U. Table 1-2 il- lustrates an overall increase in cigar and pipe smoking, followed by a decline during the past decade. The actual probability of exposure to ETS is complex, affected by ventilation rates, size of houses, restrictions on where tobacco products may be smoked, and changes in the cigarette itself. The consequence of Figure 1-1 is that the general probability of being exposed to some_ ETS for the nonsmoker has increased until quite recently. The magnitude of exposure to ETS will depend upon the number of cigarettes and/or cigars and pipes smoked in a given environment, as well as other factors such as ventilation. Light smokers are more likely to stop smoking than heavy smokers, which might explain why over the past 30 years the number of - cigarettes per smoker and the total consumption (Figure 1-2) have not declined as rapidly as the percentage of people who smoke (see also cigar and loose tobacco consumption in Table 1-2). From a peak consumption in the early 1960s, there has been a decline of 0169g449

18 . 50 Mele. 40 30 rT 20 10 0 f I 1 1 I I 1 I1 1955 1960 1'88S 1970 1975 1980 1985 YEAR FIGURE 1-1 Percentage of current smokers In the United States. Adult population, by .ex, 1966-1983. From Shopland and Brown (1986). 20°rli in the per capita (U.S.) consumption of cigarettes (Shopland and Brown, 1985). These data, however, are averaged over the total U.S. population, including smokers and nonsmokers. Among persons who consider themselves smokers, the cigarette consump- tion per adult smoker actually has increased from 27.3 to 30.0 cigarettes per day. Table 1-a demonstrates that, for both sexes, the percent of smokers who are heavy smokers hes steadily in- creased over the past 30 years. Therefore, the consumption per active smoker indicates that the nonsmoker who has close contact with a smoker may be exposed to greater amounts of smoke in 1985 than in 1955, although the total number of hours a nonsmoker is exposed to ETS would have declined. Counteracting this trend of increased exposure has been the trend of reduction in amount of tobacco used to fill each cigarette. Physical changes of the leaf due to modern ntethods of processing, the use of filter tips (United States, >90% of all cigarettes since 17 TASI,E 1-1 U.S. Cigarette Consumption, 1900 to 1983a ear otal Billions Number Per Capita, 18 Years and Older ear otal BHlions Number_ Capita, 18 Years and Older ear otal Billions Number - Per Per Capita,. 18 Years and Older 1900 2 5 - - . 54 1930 119.3 1,485 1960 484.4 4 171 1901 2.5 S3 1931 114.0 1,399 1961 502.5 , 4 266 ' 1902 2.8 60 ' 1932 102.8 1,245 1962 508.4 , 4 265 1903 3.1 64 1933 111.6 1,334 1963 523.9 , 4.345 1904 3.3 66 1934 125.7 1,483 1964 511.3 4,195 1905 3.6 70 193S 134.4 1,564 1965 528.8 4,259 1906 4.5 86 1936 152.7 1.754 1966 541.] 4,287 1907 5.3 99 1937 162.8 1,847 1967 549.3 4,280 1908 5.7 105 1938 163.4 1,830 1968 545.6 4 186 1909 7.0 125 1939 172.1 1,900 1969 528.9 , J,993 1910 8.6 151 1940 181.9 1,976 1970 536.5 3,985 1911 10.1 173 1941 208.9 2,236 1971 555.1 4 037 1912 13.2 223 1942 245.0 2,585 1972 566.8 , 4,043 1913 15.8 260 1943 284.3 2,956 1973 589.7 4,148 1914 16.5 267 1944 296.3 3,039 1974 599.0 4,141 1915 17.9 285 1945 340.6 3,449 197S 607.2 4,123 1916 25.2 39S 1946 344.3 3,446 1976 613.5 4,092 1917 35.7 SS1 1947 345.4 3.416 1977 617.0 4,051 1918 45.6 697 1948 358.9 3.S0S 1978 616.0 3,967 1919 48.0 727 1949 360.9 3,480 1979 621,5 3,861 1920 44.6 66S 1950 369.8 3,522 1980 631.5 3,851 1921 50.7 742 19S1 397.1 3.744 1981 640 0 3 840 1922 53.4 770 1952 416.0 3.886 1982 . 634.0 , 3 753 1923 64.4 911 1953 408.2 3,778 1983 600.0 , 3 502 1924 71.0 9s2 1954 387.0 3,546 1984 600.41 , 3,461k 1925 79.8 1,085 1955 396.4 3.597 1985 595.09 3,384` 1926 89.1 1,191 1956 406.5 3,650 1927 97.5 1,279 1957 422.5 3,755 1928 106.0 1.366 1958 448.9 3,953 1929 118.6 1,504 - 1959 467.5 4,073 "lncludes o.ereees forcer, 1917-1919 and 1940 to date. Commodity Economics Division, Economic Research Service, USDA. 6SubJect to revision. `Estimated. SOURCE: U.S. Department of Agriculture, 1985, TI6954L8

i8 650 600 400 350 1955 1960 196$ 1970 1975 1980 1985 YEAR FIQURE 1-2 Total cigare«9 consnmptioa (domestic salee), 1955- 1985, 1978; Griese, 1984), and variationa in the compOsition oI tobacco blends for cigarettes (Norman, 1982) have made this reduction possible. In 1956, the U.S. average ~~ce then, tar n and ln'a ot e yie de rng and 2.69 mg, respectl/ y_ nicotine in have steadily decreased to 13.2 mg tar and 0.95 mg 1980 (The Tobacco inetitute', 1981). liowever, tar and nicotine yields in the SS of cigarettes have not eignificantly changed except 19 TABLE 1-2 U.S. Consumption of Cigars and Tobacco for Pipes and Hand-rolled Cigarettes Year Cigars, Tobacco, millions Mn.lb• Year Cigars, millions Tobacco, Mn.lb' Year Cigars, millions 1920 8,609 1950 . 5.608 104.3 1980 5,386 1921 7.435 1951 5,778 97.4 1981 5,231 1922 7,527 1952 6,037 92.9 1982 4,901 1923 7,505 1953 6.107 84.3 1983 4,884 1924 7,189 1954 6,024 81.2 1925 66.949 1955 6,078 77.8 1926 7,008 1956 6,039 70.0 1927 7,008 1957 6.194 68.9 1928 6,874 1958 6,586 74.4 1929 6,972 1959 7,377 71,9 1930 6,272 1960 7.434 72.2 1931 5,656 196t 7,083 72.7 1932 4,724 t962 7,103 69.8 1933 4.553 1963 7,434 69.7 1934 4.818 1964 9,899 81.7 1935 4,943 1965 8,949 69.8 1936 5,362 1966 8,610 68.6 1937 5,516 1967 8.403 66.4 1938 55.294 1968 8.331 69.6 1939 5,469 1969 8.579 68.3 1940 5,491 1970 8,881 74.0 1941 5,933 1971 8,830 69.5 194; 6.339 1972 11,125 66.8 1943 5,350 1973 11,126 59.5 1944 4,878 1974 9,339 1945 5,027 1975 8,663 1946 S,924 1976 7.492 1947 5,706 1977 6,792 1948 5,860 1978 6,231 1949 5,625 1979 5,706 'Tobaccn for pipes and hand-rolled dgarettes not available prior to 1950. SOURCES: Lee, 1975; Tobacco Reporur, 1984. i

20 TABLB 1-3 Number of Cigarettes Smoked per Day, as a Percentage of Current Smokers, by Sex Less Than 15 IS-24 25 or More M.les 1965 30.1 45.7 24.1 1976 24.9 44.4 30.7 1980 24.2 41.7 34.2 1983 23.5 42.9 33.6 Females 1965 46.2 40.8 13.0 1976 37.6 43.4 19.0 1980 34.7 42.0 23.0 1983 33.8 45.6 20.6 SOURCE: Shopland and Brown, 1985. in the case of cigarettes designed for ultralow yields of tar and nicotine. Certain other components, in particular volatile, toxic components, are released into SS in significantly greater amounts than into rlS. Furthermore, ETS contains significantly smaller _ particles than MS, and nicotine, and perhaps other smoke con- stituents, is volatilized to a greater extent in SS than in MS. This means that the gas-phase composition of SS differs substantially frum that of NiS. The health implications to nonsmokers of exposure to ETS_ may not be a simple extrapolation from the studies of active smokers. The complexities of such extrapolations will be discussed. Children represent a large population of nonsmokers who may be exposed to environmental smoke. Several cohort studies of chil- dren are reviewed in Chapter 11. Although there is some variation among these studies, they indicate, mainly through questionnaires, that between 50 and 6-5 percent of the children have been exposed to tobacco smoke in the home during the past 20 years. Health implications of this exposure for the developing child will be dis- cussed. cussed. ORGANIZATION This report begins with a discussion of the components of ETS (Chapter 2) and what in vivo and in vitro studies have determined about ETS (Chapter 3). Various methods of exposure assessment 21 are considered in Chapters 4 through 8, including physical effects, questionnaires, and biological markers. Chapters 9 through 15 review epidemiologic studies of possible health effects of these ex- posures. The health consequences examined range from irritation - and allergic reactions to cancer and cardiovascular disease. Only - studies that assess exposures under experimental conditions or in - the home are included. ETS potentially interacts with constituents of the ambient air. This makes the evaluation of possible health effects due to workplace exposure complex and specific to each situation because of the varying nature of contaminants. Each chapter concludes with a summary of what is currently known, the strength of that knowledge, and what additional information would further clarify the relationship of ETS and possible health effects. Some recommendations for additional research are also given. REFERENCES ,tiriese, V.N. Market growth of reduced tar cigarettes. Recent Adv. Tob. Sci. 10:4-14, 1984. Lee, P.N., Fd. Tobacco Consumption In Various Countries, pp. 82-84. Lon- don, England: Tobacco Research Council, 1975. National Research Council, Committee on Airliner Cabin Air Quality. Air- liner Cabin Environment: Air Quality and Safety. Washington, D.C.: National Academy Press, 1986. 303 pp. Norman, V. Changes In smoke chemistry of modern day cigarettes. Recent Adv. Tob. Sci. 8:141-177, 1982. Shopiand, D R., and C. Brown. Changes in cigarette smoking prevalence in - - the U.S.: 1966-1983. Ann. Behav. Med. 7;6-8, 1986, The Tobacco Institute. U.S. tar/nicotine levels dropping. The Tob. Observ. 6:1, 1981. Tobacco Reporter. Cigars In the U.B.: Is the upturn real? Tob. Rep. 111:45-48,1984. U.S. Department of Agriculture. Tobacco: Outlook and Situation Report. DOA Publ. No. TS-129. Washington, D.O : U.S. Government Printing OtRce,1986. U.S. Department of Commerce. Statistical Abstract of the United States: 1986. Washington, D.C.: U.S. Department of Commerce, Bureau of the Census, 1984. 119 pp. CTf 9g44A

I PHYSICOCHEMICAL AND TOXICOLOGICAL STUDIES OF ENVIRONMENTAL TOBACCO SMOKE PU9R449

2 The Physicochemical Nature of Sidestream Smoke and Environmental Tobacco Smoke INTRODUCTION Mainstream smoke (MS) is the aerosol drawn into the mouth of a smoker from a cigarette, cigar, or pipe. Sidestream smoke (SS) , is the aerosol emitted in the surrounding air from a smoldering tobacco product between puff-drawing. SS is a major source of environmental tobacco smoke (ETS), i.e., air pollution caused by the burning of tobacco products. Other contributors to ETS are the exhaled portion of MS and the smoke that escapes from the burning part of a tobacco product during puff-drawing. In -- - addition, some volatile components (e.g., carbon monoxide) diffuse through cigarette paper and contribute to ETS. _ Tobacco smoke aerosols are diluted with air by the time they are inhaled as ETS air pollutants. Furthermore, the physical char- acteristics and chemical composition of ETS change as the pollu- tants "age" : nicotine is volatilized; particle sizes decrease; nitrogen - -- oxide gradually oxidizes to nitrogen dioxide; various components of the ambient air (e.g., radon daughters) can be adsorbed on the - - - - particles; and other physicochemical changes can occur. In the scientific literature, the terms "passive smoke," "passive smoking," and "involuntary sm4king" are used often. These terms do not adequately describe ETS and its inhalation, but they are used interchangeably with "ETS" in this report. Most of the reported data on MS, SS, and ETS pertain to cigarette smoking. Few comparative data on smoke pollutants from other tobacco products are available. ST69gZ48 25

26 In the_ laboratory, cigarettes, cigars, and pipes are smoked by machines under standardized conditions (Wynder and Hoffmann, 1967) to obtain reproducible data for the determination of various individual constituents of undiluted MS and SS. Such data pro- vide a.scientiHc basis for comparing tobacco products and brands. The standardized machine-smoking conditions were developed 3 decades ago to simulate human smoking behavior (Wartman at at., 1959). However, these can differ substantially from those of today's cigarette smokers, especially in the case of filter-tipped products that are designed to deliver low yields of tar and nicotine (Herning et al., 1981). For cigarettes and cigarette-like cigars weighing up to 1.5 g, the most widely used machine-smoking conditions in the test laboratory are as follows: one 35-m1 puff lasting 2 seconds taken once a.minute. The butt length for nonfilter cigarettes is 23 mm. For filter-tipped cigarettes, the total length is increased 3 mm for filter tip plus overwrap (Pillsbury et al., 1969; Brunnemann at al., 1976). For cigars, the conditions are as follows: a 30-m1 puff taken once every 40 seconds and a butt length of 33 mm (International Committee for Cigar Smoke Study, 1974). For pipe smoking the test calls for a bowl filled with 1 g of tobacco and for a 50-mi puff lasting 2 seconds to be taken every 12 seconds (Miller, 1964). Several devices have been used for generating SS from ciga- rettes and cigars (Dube and Greene, 1982). Among them, the Neurath and Ehmke chamber or modification thereof have been used for chemical analytic work on SS (Neurath and Ehmke, 1964; Brunnemann and Hoffmann, 1974). When SS is generated, a stream of air is sent through a chamber at 25 ml/second. At this rate, the tar and nicotine yields in the MS of cigarettes and cigars smoked in the chamber are similar to those obtained by smoking cigarettes or cigars in the open air. However, the velocity of the airstream through the chamber has considerable influence on the yields of individual compound4 in SS (Buhl at al., 1980; Klus and Kuhn, 1982). In order to collect the particulate matter of MS and SS, the aerosols are directed through a glass-fiber filter that traps more than 99% of all the particles with diameters of 0.1 pm or more (Wartrnan et a1.,1959). The portion of the smoke that passes through the filter is designated as the vapor phase. This arbitrary separation into particulate phase and vapor phase does not neces- sarily reflect the physicochemical conditions prevailing in MS and 3I~9g~~,g 27 SS. However, it does reflect specific trapping systems and analytic - - methods that have been developed for the standardized determi- nation of individual components or groups of components in MS or SS (Brunnemann and Hoffmann, 1982; Dube and Gr-eene,1982). Standardized machine-smoking conditions do not exactly du- plicate the smoking patterns of an individual, which depend on many factors. For example, low nicotine delivery in cigarette smoke generally induces a smoker to puff more frequently (up to 5 puffs/minute), to draw larger volumes (up to 55 mt/puff), and to - inhale more deeply. Puffing more frequently increases the amount of tobacco consumed nsumed during generation of MS and thus diminishes the amount of tobacco burned between puffs. This, in turn, af fecte the release of combustion_ products in_ SS, so an increase in puff frequency diminishes the production of SS and ETS. Also, smoking behavior appears to depend strongly on the blood con- centration of nicotine that the smoker desires to reach (Krasnegor, 1979; Grabowski and Bell, 1983). The smoker, because of proximity to the source, usually in- hales more of the SS and ETS originating from the burning of the - tobacco product than a nonsmoker; however, we do not know the exact amount and we do not know the degree to which inhaled SS and ETS aerosols are retained in the smoker's respiratory tract. Model studies with MS have shown that more than 90% of some hydrophilic volatile components (e.g., acetaldehyde) is retained after inhalation by the smoker (Dalham et al., 1968a). There- fore, one may assume that a large proportion of the hydrophilic agents in the vapor phase of SS and ETS is also retained when smoke-polluted ambient air is inhaled. In the case of hydrophobic components of the vapor phase of MS (e.g., carbon monoxide), the retained fraction depends on the depth of inhalation, but it hardly ever exceeds 50°,6 (Dalham et al., 1968b). An active smoker gen- erally retains 90% or more of MS particles (Dalharn et al., 1968b; Hiller, 1984), whereas a nonsmoker exposed to ETS appears to retain a smaller percentage of ETS particles. It has been calcu- lated that, depending on the degree of SS pollution, a nonsmoker exposed to ETS can retain 0.014 to 1.6 mg of particles per day from ETS (Hiller, 1984).

28 SIDESTREAM SMOKE The SS generated between puffs originates from a strongly re- ducing atmosphere. Therefore, undiluted SS contains more com- bustion products that result from oxygen deficiency and thermal cracking of molecules than does MS. In addition, SS formation involves generation of higher amounts of compounds from nitrosa- tion reactions. Consequently, $S differs substantially from MS. Table 2-1 compares MS and SS from nonfilter cigarettes. Dur- ing the consumption of one whole cigarette under standard smokg ing conditions, the formation of cigarette MS generated during 10 ' puffs (each 2 seconds) of a blended nonfilter cigarette requires 20 a and consumes 347 mg of tobacco. The formation of SS from the same cigarette smoldering requires 550 seconds and consumes 411 mg of tobacco. However, as shown with experimental cigarettes, the amounts of tobacco consumed during and between puffs de- pend greatly on the type of tobacco (Johnson et al., 1973a). In addition, MS and SS are generated at different temperatures. For example, under laminar atmospheric conditions, the SS of a smol : dering cigarette enters the surrounding atmosphere about 3 mm in front of the paper burn line, at about 350°e (Baker, 1984). The pH of the MS of a blended American cigarette ranges from 6.0 to 6.5, whereas the pH of SS is 6.7 to 7.5. Above a pH of 8.0, the proportion of unprotonated nicotine in undiluted smoke increases; therefore, SS contains more free nicotine in the gas phase than MS. The pH of SS of cigars is 7.5 to 8.7; pH values for pipe smoke have not been reported (Brunnemann and Holfmann, 1974). Under conditions prevailing in MS,•SS, and ETS, unprotonated nicotine -- - is primarily present in the vapor phase; its absorption through the mucous membranes is faster; thus, its pharmacologic effect is different from that of unprotonated nicotine in the particulate matter (Armitage and Turner, 1970). About 300-400 of the more than 3,800 compounds identified in tobacco smoke have been measured in MS and SS. Table 1-2 lists the amounts of selected substances reported to occur in the MS and in SS from the burning of a whole nonfilter cigarette and the range of the ratio of their amounts in SS/MS. A ratio greater than unity ity means that more of a substance is released in SS than in MS.. The separation of the compounds in Table 2-2 into vapor phase and particulate phase constituents reflects the conditions 1a~6~8la~~ i 29 TABLB 2-1 Some Physiocochemical Characteristics of Fresh, Undiluted Mainstream and Sidestream Smoke from a Nonfilter Cigarette° Characteristics MS SS Reference Duration of smoke production, a 20 . 550 Neurath and Horstmann, 1973 Tobacco burned, mg 347 411 Neurath and Peak temperature during formation, 'C 900 600 Hor,tmann, 1973 Wynder and Ho(fmann, 1967 pH 6.0-6.2 6.4-6.6 Brunnemann and Number of particles per cigarette Particle size, pm 10.5 X 10t= 0.1-1.0 3.5 x tOt= 0.01-0.8 Hoffmann, 1974 Scassel latt i• Sforzol ine and Savino, 1968 Carter and Hasegawa, 1975; Hllkr et al., 1982 Partkk mean diameter, pm 0.4 0.32 Carter and Hasegawa, 1975; Hilkr et al.. 1982 Gas concentration, vol.% Carbon rrrono:ide 3-5 2-3 Keith and Derrick. 1960 Carbon dioxide 8-li 4-6 Wynder and Oxygen 12-16 1.5-2 Hoffmann, 1967 Baker, 1984 Hydrogen 3-I5 0.8-1.0 Hoffmaan et al., 1984a,b 'Data were obtained under standard laboratory smoking conditions of one puff per min- ute, lasting 2 s, and having volume of 35 ml. Mainstream smoke collected directly from end of cigarette. Sldeslrcam smoke was measured 4 mm from burnln cone 350°C). g (gas temperature. prevailing in MS and does not apply to the distribution of these compounds in the vapor phase and particulate phase of SS. The ratio of the amount of tobacco burned during SS genera- tion to that burned during MS generation is 1.2:1 to 1.5:1 (see Ta- ble 2-1 for data on nonfilter cigarettes). Therefore, if one assumed that the combustion process is the, satne during the generation of the two kinds of smoke, the ratios of their various constituents would also be between 1.2:1 and 1.5:1. That is not the case, as indicated by the higher bS/MS values in Table 2-2. For instance, in the first part of Table 2-2, which lists volatile compounds, the ratios for carbon monoxide range from 2.5 to 4.7, for carbon diox- ide from 8 to 11, for acrolein from 8 to 15, and for benzene about 10.

30 TAeLE?.2 Distribution of Constituents in Fresh, Uudiluted Mainstream Smoke and Diluted Sidestream Smoke from Nonfilter Cigarettes• Constituent Amount In MS Range In SSlMS Vapor phase! 7 5-4 2 bon monoxide C 10-23 mg . . ar Carbon dioxide 20-40 mg 8-11 03-0.13 0 18-42 pt . Carbonyl sulfide 5-10 ` 12-48 pg Benzene 5.6-8.3 Toluene 100-200 pg t-a50 0 Formaldehyde 70-100 pt . 8-15 60-100 pt AcrQlein 2-5 100-250 pg Acetone 5-20 6 16-40 pg . Pyridine 3-13 3-Mcthylpyridine 12-36 pg 20-40 3rVinylpyrtdine 11-30 pg 25 1-0 0 en cyanide dro H 400-500 pg . . g y ~ 32 ng 3 liydrazine 10-170 50-130 pg Ammonia 2-6.4 4 Methyiamine II.S-28.7 pg . 1 7-5. 3 Dhnethylamine 7.8-10 pg _ . 4-10 Nitrogen oxides 100-600 pg 20-100 N-Nitrosodimethylamine• 10-40 ng _ <40 N•Nitrowdlethylamind yrrolidine N-Nitros ND-25 ng 6-30 ng 6-30 op mic aeid F 2i0-49u pg 1.4-1.6 or tic acid A 330-8t0 pg 1.9-3.6 ce Methyl chloride 150-600 pt 1.7-3.3 Particulate phasea 1 3-1 9 rticulate mattert P 15-40 mg . . a 1-23 mg 2.6-3.3 Nicotine 1-0.5 <0 Anatabine 2-20 pg . 0 6-3 1 Phenol 60-140 pt . . 6-0.9 0 chol C t (00-.360 pg . a e Hydroguinone 110-300 pt 360 ng 0.7-0.9 30 Aniline 19 Toluidine 2 160 n` - 2-Naphthylamine 1.7 ng 30 Aminobiphenyl` 4 4.6 ng 31 - zinianthracene B 20-70 ng 2-4 en BenzolalPyrcme 20-40 ng 2.5-3.5 0 9 lesterol Ch 22 pg . o ry-Butyrolactone• 10-22 pg 3.6-5.0 8-11 0.5-2 pg Quinoline 7-1.7 0 Harmani 1.7-3.1 pg . 5-3 0 Nilrosonornleotine N'- 2D0-3,000 ng . . 100-1,000 ng 1-4 NNKR 2 1 N-Nitrosodiethanolamine-` 20-70 ng . 31 TABLB 2-2 Continued Constituent Amount in MS Range of SS/MS Cadmium 100 ng 7.2 Nicket0 20-80 ng 13-.30 Zinc 60 ng 6.7 Polonium-210` 0.04-0.1 pCi 1.0-4.0 Benzoic acid 14-28 pg 0.67-0.95 Lactic acid s 63-174 pg 0.5-0.7 Glycolic acid 37-126 pg 0.6-0.95 Succinic acid 1 l0-NO pg 0.43-0.62 'Data from Elliot and Rowe (1975); Schmeltz et al. (1979); Hoffmann el al. (1983); Klu: and Kuhn (1982); Sakuma et al. (1983, 1984a,b); Hiller et al. (1982). Diluted iluled SS Is collected with airflow of 25 mlfs, which Is passed over the burning cone. -- aSeparation into vapor and particulate phases reflects conditions prevailing In MS and - - does not necessarily Imply same separation in SS. -- 'Human carcinogen (U.S. Department of Health and Human Services, 1983). dSuspected human carcinogen (U.S. Department of Health and Human Services, 198J). 'Animal carcinogen (Vainlo at al., 1985). fl-melhyl-9H-pyridoi7,4•bl-indole. ONNK = 4-(N-methyl-N-nitrosamino)-1-(3-pyridyq-l-butanone. The high SS/MS values of carbon monoxide and carbon diox- ide show that more of each of these constituents is generated in the oxygen-deficient cone during smoldering than during puff drawing. After passing briefly through the hot cone, most of the carbon monoxide is oxidized to carbon dioxide, probably because of the high temperature gradient and sudden exposure to air. The high SS/MS values of volatile pyridines are thought'to be due to the fact that these compounds are formed from the alkaloids during smoldering (Schmeltz et al., 1979). Hydrogen cyanide is formed primarily from protein at temperatures above 700°C (Johnson and Karg, 1971). Thus, smoldering of tobacco at - 600°C does not favor the pyrosynthesis of hydrogen cyanide to the extent that it occurs during MS generation. With regard to the carcinogenic potential of SS, it is impor- tant to consider the SS/MS ratio of NO,-4 to 10. More than 95% of the NO, inhaled by the smoker is in the form of nitric oxide, and only a small portion is oxidized to the powerful ni- trosating agent, nitrogen dioxide.. Only a small fraction of nitric oxide is expected to be retained in the respiratory system by being bound to hemoglobin. NOs released into the environment in SS

32 is partially oxidized to nitrogen dioxide (Vilicins and Lephardt, 1975). Thus, environments polluted with SS are expected to con- tain increased concentrations of the hydrophilic nitrosating agent, nitrogen dioxide. Perhaps the most remarkable data in this portion of Table 2-2 are the very high SS(MS values of ammonia, nitrogen oxide, and the volatile N-nitrosamines. &tudies with ('eN]nitrate have shown that, during burning of tobacco, nitrate is reduced to ammonia, which is released to a greater extent in SS than in MS during puff- drawing (Just et at., 1972). An extreme example ip the case of a cigarette made exclusively from burley tobacco, a variety generally rich in nitrate (2.0-5.0% in U.S. survey); ammonia is released in SS at 8,500 µg/cigarette (SSj1MS 170, according to Johnson et al., 1973b). (In the case of a blended cigarette, the greater generation of ammonia in SS causes an increased pH, which can be above 7, whereas the pH of MS is about 6.) The ranges of high SS/MS ratios of the highly carcinogenic-' volatile N-nitroaamines (such as N-nitrosodimethylamine-2© to 100) have been well established (Brunnemann et al., 1977, 1980; Riihl et al., 1980). The second part of Table 2-2 lists some constituents of par- ticulate matter, their amounts reported to occur in MS during the burning of one cigarette, and ianges of the relative amounts in SS/MS. The increases in SS of tobacco-specific N-nitrosamines, such as 4-(-N-methyl--N-nitrosan~ino)-i-(O~-pyridyl)-1-butanone (NNK), N-nitrosodiethanolamine, and N'-nitrosonornicotine, are up to fourfold. Presently we do not know whether the tobacco- specific N-nitrosamines are present in the particulate phase or in the vapor phase of ETS (Hoffmann and Hecht, 1985). Constituents of the vapor phase would be less likely to settle with the smoke particles, but would remain in the ambient air for longer spans of time. Research is needed to evaluate this distribution, which is important with respuct to the carcinogenic potential of SS. The meaning of the abundant release of amines in SS (SS/MS, to 30-fold)-as indicated by the data In aniline, 2-toluidine, and the alkaloids-should also be examined. Some amines are readily nitrosated to N-nitrosantines, but analytic data on secondary reactions of amines in polluted environments are lacking. 6tG9g44g 33 TABLE Z-3 Relative Concentrations (SS/MS) of Selected Components In ---- - Fresh, Undiluted Smoke of Four 85-mm Commercial American Cigarettes" fonstituene Constituent Concentrations in Smokch Cigarette A, Cigarette B. Cigarette C. Cigarette D. NF F F PF SS SS/MS SS SS/MS SS SS/MS SS SS/MS Ter, mg/g 22.6 1.1 24.4 1.6 20.0 2.9 14.1 15.6 Nlrnline, mg/g C 4.6 2.2 4.0 2.7 3.4 4.2 3.0 20.0 tT mg/g NH 28.3 2.1 36.6 2.7 33.2 -- - 3.5 26.8 14.9 ,, mg/t 524 7.0 893 46 213.1 6.3 2J6 5.8 Catechol, P8/8 58.2 1.4 89.8 1.3 69,5 2.6 117 13.9 aP, ng/8 67 2.6 45.7 2.6 51.7 4.2 448 20.4 NDMA, ng/g - 735 23.6 597 139 611 50.4 685 167 NPYR, ng/g N 177 2.7 139 13.6 233 7.1 234 17.7 NM , ngig 857 0.85 307 0.63 185 0.68 338 5.1 'Data from Adams_ et al. (1985). Tar values for MS: cigarette A, 20.1 mg; cigarette B, 15.6 mg; cigarette C, 6.8 mg; cigarette D, 0.9 mg. hNF = nonfilter cigarette; F = filter cigarette; PF = cigarette with•perforated filter tip; BqP = bensolalPyrene. `NDMA s N-nitrosa8methylamine; NPYR = N•nitro.opyrroHdine; NNN = N'-nitro- sonornicoline. To comprehend the data in Table 2-2 fully, some aspects should be emphasized. First, the data are based on analyses of nonfilter cigarettes that were smoked under standard labora- tory conditions. Second, those conditions, established according to smoking patterns observed 3 decades ago, have been shown not to reflect today's smoking behavior. The difference is especially evident in the case of filter cigarettes designed for low smoke yields. Most consumers inhale the smoke of such cigarettes more intensely than the smoke of nonlilter cigarettes (Hill and Marquardt, 1980; lierning et al., 1981). This difference affects the yield of SS. Con- ventional cigarette filter tips primarily influence the yield of MS, but have little impact on SS yield. However, highly active filter tips, especially those with perforations, also affect the yield of SS (Aderns et al., 1985). It is apparent in Table 2-3 that for all cigarettes studied the SSIMS values are greater than 1 for many toxic and carcinogenic constituents. L' ,

Tw= 2-+4 Measured Concentrations: of!Carbon Iwionoaide'in EfI'S° Carbon 'N4onoxide Concentrations. ppm , Nonsmoke Controls Location ?obacco Burned Ventilation ! Mean Range Mean Range References Rooms Train 1-18smokers Submuines 157 eigarettes/tlaD ~Natnral Yes 40 4:3-9 0-M0 2.2 t 0.98 0:4-4':5 Coburn at at.. 1965 - Harmsea and Etfenberger, 1957 Catw at al., 1970 (i66 ml) 78 milita*y aircratt 8 commercial airatatt pooms 14 public ptaem Psteq boat Theaterlfoyer IIntereit,y bus 2aonietmtae morns Otfiee A-uwttab(le 9 night clubs 14 testaurants 45 retauosts 94-1103 cigarettes/day - 3 cigarettes 3 eigarettea - 2 smokers (4 cigarettes) 'Yes Yes - 5 ehanges/h 1SdhanQes/h 8 changes/h 236 m3/h Natural Natural Meshaaical Varied <40 <2-S <2 <10 18.4 t8:7 3:4t0.8 32 18 - 3.4 9.9 +t S.S 83t 2.2 2'5 (peak) <2.5-4:6 <2:9-9:0 42i(peak) 32 (peak) 6.5-41:9 :l rt 1.7 U.S. Departmeot of Ti ransportation. 1971 U.S.12qpaittnent of lfran;pottation, 1971 Pottlrcin. 1971 Perry. 1973 3.0 t 2.4 Godin at at.. 11972 1.4 t 0,8 Godin at il.. 1972 5eif1. 1973 1-2 Slarin and Hettz. '1975 Harke, 1974 - 1135 (peak) Harke and'Petets. 1974 - 15.0 (peak) Sebben: et aL. 1977 9.2 (outifooc) 3.0-35 Sebben at a1.,11977 Sebben et al.. '1977 Co 33 uores - - 10.0~t 4.2 111.5 t 6.5 11.5;t6.5 Sebben at at., 1977 3 hospital lobbies - - - 4.8 Sebben at it.. 1977 6 coHeehouses Varied - 2-23 - Badre at al., 1978 Room 18 smokets - 50 Badte at al.. 1978 Hospital ldbby 12-30 smokets - 5 Badte at al. 1978 3 trsia compartments 2-3 smokers - - 4-5 , Bache et a1.. 1978 Automobile 3 smokers 2 sttwkees Natnra-, open Na;ural. ¢baed 14 20 Badte et al.. 1978 10 ofEias - - 2.5 t 10 1:5-1:0 2:5t1s0 13t4:5 Chappoll and Parker. 1977 15 restaurants - - 4.0 t23 1.0-9S 2:5 t 1.5 1.0-5:0 Chappell and Parker, 1977 14 night ehrbs and taverns 13.0 t 7.0 10-29L0 Chappell and Parker, 3.0 ± 2.0 1:0-5.0 1977 ?arern JArtificial Now 8.5 - ak) 36 ( Chappell and Patker, 1977 co o Ol'fice Natural,open 1.0 pe 10.0 (peak) Chappell and Parker, ~ 1977 Restaurant - imechanioal 5A 2.1-9.9 4.8(outdoot3) - Fischer et al.. !1978 Restaurant - Natural 2.6 1.4-3.4 1.5 (outdoors) - Fiacher et YI„ '1978 Bar - ! Natnnl. open 4.8 2.4-9,6 1.7 ~(outdootss) - 'VYeber et al„ '1976 Cafeteria - ffi 44' 11 dhanges/b 1.2 0.7-1.7 0:4(outdooes) - Weberet al., 1976 o ces - - 1.1 6:5!(ma8) - - VVrtber. 1984 25 offices - - 3 78 f 1:42 - 3 59 ± 3.33 - Szadkowtki at al., 1976 Tavern - 16 changes/h 11.5 110-12 I(outdoors) - 'Cuddeback at aL, !1976 Tavern - 1-2 changes/h 12.0 3-22 - Cuddeback at al.. I1976 'Tme-weighted average (TWA) of carbon monokide. 50,ppm,(55 mgg/m3). TWA = average concentration towhich worker may be exposed continu- ously for 8 h without damage to!heahh (National Inatitute forOccupational Safety and Health. 11971).

36 pRINCIpAL CHEMICAL CONSTITUENTS OF ENVIRONIvIENTAL TOBACCO SMOKE Air dilution physicochemically changes bS and other contrib- . - utors to ETS. Depending on the degree of air dilution of SS, the concentration of particles in ETS can range from a few rmicrograms a A high degree of air dilution can reduce this to 8~-5~ m~~m ' er cubic meter. At the same time, yield to a few micrograms p_ the median diameter of the particles will decrease from 0.32 pm to pm Keith and Derrick,1960; Wynder and Heffmann, 0.14or0.098 µ_ ( ebrethsen and Sears, 1985). Another change caused by 1987; Ing. , air dilution of SS i at ex lusively in the vsportphase (Eudy ettsl, tine is present almo _ 1985). In addition, ~ a count for the presence oftothertsemiSola ile to air dilution mtgh lack data on such chemicals in the vapor phase of ETS, but we _ effects. The scientific literatFuTSc U~ Public 1[eal h$ervi ea 1979; indoor air pollution by . . . ( - National Research Council, 1981). We limit our review here to measurements made under field conditions and have excluded data from experimental studies: Most of the published data, summa- rized in Tables 2-4 through 2-9, do not exclude the possibility that, even though the respiratory environments analyzed were polluted largelyby ETS, some other sources of pollution contributed to the - reported concentrations of individual ageflts_ , (Many studies have dealt with the b~ ou~~ ~et efinp~ po er 5 di ~ sses the measurement ttuted by to of particulate matter, and the results of the studies are summa- rized in Table 5-1.) Table 24 shows concentrations of carbon monoxide measured in a variety of indoor spaces with and without occ-upancy by smokers. Carbon monoxide concentrations were generally higher in spaces where smoke was present. They were highly variable, however, and collected data on each space were insufficient (e.g., number of cigarettes smoked and volume of space) to show a Nico- consistent relationship. Tobacco is tibesoecfic indicator for tobaccotsmoke P llution. tiana alkaloid is a P - Nicotine concentrations in smoke-poll_uted toon$ were generally nt~ in found to be 5-50 pg/ms and much higher (up to 500 µg~ ) ~~- f 37 heavily polluted environments (Table 2-5). In small interior com- partments, partments, such as automobiles, occupants who smoke tobacco have generated nicotine concentrations of 1,010 µg/ms (Badre et al., 1978). Freshly generated tobacco emoke contains nitric oxide, but not nitrogen dioxide. On release into the environment, nitric oxide is gradually oxidized to nitrogen dioxide. The estimated half-life of nitric oxide is 10-20 minutes, depending on the degree of air dilution.. Table 2-6 shows concentrations of nitric oxide and nitro- gen dioxide in smoke-polluted environments and indicates means ranging from 9 to 195 ppb for nitric oxide and 21 to 76 ppb for nitrogen dioxide. Generally, the nitrogen oxide values reported in Table 2-6 are signif cantly in excess of those observed for out- door atmospheres. However, some severe air pollution episodes in industrial areas have reportedly caused levels of 100 ppb, which persisted over several hours or even for several days (Goldsmith and Friberg, 1977). As a constituent of the respiratory envi- ronment, nitrogen dioxide conceivably contributes to endogenous nitrosation, which leads to the presence of nitrosamines in ex- posed subjects. Whereas it has been clearly demonstrated that inhaled cigarette smoke increases the endogenous formation of N_ nitrosamtnes (Hoffmann and Brunnemann, 1983; Ladd et al., 1984; Lu et al., 1986; Tsuda et al., 1986), the endogenous formation of N-nitrosaminee in nonsmokers exposed to ETS has so far not been demonstrated (Brunnemann et al., 1984). Tables 2-7 and 2-8 show concentrations of acrolein and acetone in ETS. These volatile carbonyl compounds are known to affect mucociliary function and thus inhtbit the clearance of smoke par- -- - ticles from the lung (Wynder and Hoffmann, 1967). Table 2-9 shows concentrations of some additional toxic agents in ETS. Benaene, N-nitrosodimethylamine, N-nitrosodiethyl- amine; and the polynuclear aromatic hydrocarbons, represented by benso(ajpyrene, are of concern, because they are known car- cinogens (Vainio et al., 1985). RADIOACTIVITY OF ENVIRONMENTAL TOBACCO SMOKE The radioactive isotopes of lead (Pb-210), bismuth (Bi-210), and polonium (Po-210), known as long-lived radon daughters in the decay chain of uranium via radium and radon (Radford and

zz(;95448 zAB,E Z,5 '1+rir.astlrad Concemtradoms of Niaot'sne'in ETS• Nicotine Corteentratioas. pg/ml Location Tobacao Burned Ventilation iNean Range Raferences Train _ Nataral. doscN - 0.7-3.1 Hasmsen ana Effenberger.11957 6 coffee houses smokars .ariea - - 25-52 Batlre; ct l1.. 1978 Roan '18 sawkcrs - ~ - Badto at a1..197!8 ' • 12-30 smokers . - itrllobby 'Hos $7 1978 Badse et ill.. , p 2 traia compattmeats Y_3 smokess - - 36=50 Badre a al.. 1978 Antomobile 3 smokeas hlatnral, open Natnsa'1 eloned I 165 1.010 Hatlre a al.. 4979 Submarines 157 et8arseta/tl;7 Yes ~ - Caao a al„ 1970 (d6 ,u) 44-aos ai8arattes/e;7 Yes 15-35 - _ .4.9 - Hinds and'FuA, 1975 Train _ 6.7 - H'mds aed' First. !1975 - Bus waitin8, room - - 1.0 - H'+nds and First. 1975 1975 s d F Airline.raitinf toom - - 3.1 - ir t. Hinds an ' 1975 S F pustaurant - - 5.2 - ust, Hinds ao 10.3 - Hinds and Fust. 1975 ~i~A~ - - Stodmt lounge 2:8 FTinds and Fust.'1975 44 offices - - 0.9 t :13.8I(peak) Weber and Fm¢har. 1980 Tuna-weiRhteB arere8e (FV1!D1) of aicoiine, 5001e8/m3. TWA s average concentration to w;hich worker may be exposed eontinuoauly for 8 d witbout tlama8e to!beslth (National lnstitute for'OecnQatiooal'Safay anllHealth. 1971). TA= 2-6 Measured Conctntratioas of Nitrogen Oxides im' ETS• Nrtrogen Orides Conasatratioas, i PPb Location Tobaaoo; Bnrned ' Ventiiation M Nonsmoke Cantroi. Leestanrsnt - Mcrhani¢ai ean NO= 76 NO 12D Range ~15 M'a° ~(O~°D~) Refeieaeea Fisclmr a a1.. 1978 Resmucsnt - N t il 36-219 115 (outdoors) a ur NO= 63 NO 80 2" 50 (outdoors) Webet.' 1984 Bar - 114-,121 11 (outdoots) . w Naturfl. open N0: 21 NO 195 1-61 66-41N4 48, (cutdoois) MYkber a aI„ '1979 m ~~~ _ 11 h ~ (oatdoas) c anges/h ~NO 35-103 27 Weber et al.. 11979 44 aff"uxs i d V 9 2-38 5 ar e Mrried 'NO= 2N t 22 1~ NO 32, t 60 200 (pew ~ - Weber and Fisc6ec. 1980 contin~tu .~8hted a.eraQa (THV!AtY aitrio oaiide.' 25 PPaK; nitrogeo' diozide. l ousl7~ Ibo r'81 h without duaage~tb health ~(NationaC In~telfor Oceu ~ TWA= average oon~ to.rhich wadkar may be, expcraed Pstine 'Safery' aod' Headt. 1971),

rtA=2-7 ~I+vltasured ~Co:naantradians~ of Acrolein !in ErTS• Acrdkin Concentrations Location Tobacco Burned Vantilation Meao RanQe Refetarass Cofice housa Varied . _ - 0M-0:10 tnB/m' Batlte et al„' 1978 Bppu 1B, smokets - 0:185 mB/m1 - Batire, et tt1., 1978 ffoipitalJobby 12-30, smokas - 0.02 :mp/pil' - Badte et a1..11978 2 train compartments 2-3 smokers - - 0:02•0.'12 mY'/en3 Bsdre at a1..1978 0 ;Automwbik 3 smokers Nrtural. open 0.03 mB/ma - i Badre et at.. 1978 Z smokers Natural. eiasetl 0.3 mQ/m3 - Restaurant - MedanipJ 7 pPb - Fischer at a1., '1978 - Natural B ppb gu - 'Matusal. open 10,ppb - .• Weber et a1.. , 1976 Cnfeteria - 11 ahanges/h 6 ppb - Weber at ri..;1976 ''i"imeaweiahted a.ersBc'(TWA) of acrolein. 0.1 ppm. TWA s average ooncentsauon, to which wroiker may be exposed oontinuouiN'for B h.rithout aamaye to hesith (Nationai Institute for Occupational Safety and Health, 1971). aAS1.E 2=$ Mearured Conceatrafions of Acetone in ETS, Location ~'CtutrationS, mj/m~ 6 coffee houses BOOfn Tobacco Burneti Ventilation Varied 18smokets - lulesn Range - 0:91-5.88 Pef"MM ~Badteata'1.,'1978 Ho;pital lobby 2train compaxtnaeets 12-30: smokas 2.3 1~6 - >laClrr ar al., 1978 r. r Awtumobik 3 saaokes ~ Niaturall open 32 0 36-0:7S 0 Badre et ai., 1978 , 2 staokas histutw, closed . 1.20 .. B&dt'° °t a1•. 1978 Tme-.reiPhteJ! sxrajr (71MA) nf acetrnte ' = ' TW . ppm, A : ~ f~-8 h~rttbout'tlamaQe to~ health I(Mlatiooil tnsrjqctcfarOernPatiooa) Safety, and Eiealtlh~1971~). ~k~~y ~ ~~

42 ~ a a Y .Y A ~ ~~ ~ 8 Yy_l. ~ ~ ~~~ uc3UC3c`i ,« s ~ 43 Hunt, 1964; Martell, 1075; Hill, 1982), are present in tobacco and therefore appear in tobacco smoke. Furthermore, when radon is present in the air, aerosol particles, including those of tobacco smoke, tend to adsorb the earlier decay products of radon, namely the so-called short-lived daughters (Po-218, Pb-214, Bi-214, and Po-214), i.e, those preceding the long-lived daughters in the decay chain (Raabe, 1969; Kruger and Nothing, 1979; Bergman and Axelson, 1983). The presence of Pb-210 and subsequent decay products in tobacco might derive from uptake of Pb-210 from the soil, espe- cially if radium-rich phosphate fertilizers have been used (Tso et - al., 1966). It may also result from adsorption of short-lived radon - daughters on the leaves of the tobacco plant when phosphate fer- tilizers are used and the leakage of radon from the ground is therefore increased. This adsorption applies to short-lived daugh- ters, which then decay to the long-lived Pb-210, and subsequent nuclides found in the tobacco when phosphate fertilizers, contain- ing radium-226, are used (Fleischer and Parungo, 1974; Martell, 1975). 7.'he origin of these decay products could also be due to the general occurrence of radon in the atmosphere (Hill, 1982). In recent years, relatively high concentrations of radon and short-lived radon daughters have been found in indoor air in homes in several countries (Nero et al.,19.85). In clean air, the short-lived radon daughters tend to be more unattached to aerosol particles and therefore are more easily deposited on walls, furniture, etc., es- pecially through electrostatic forces. In the presence of an aerosol like tobacco smoke, some of the short-lived radon daughters are attached to particles, and therefore remain available for inhalation - to a much,greater extent than would otherwise be the case. Indoor radon-daughter concentration can more than double in the pres- ence of tobacco smoke (Bergman and Axelson, 1983). Since radon daughter exposure is a well-7cnown cause of lung cancer in miners, the described attachment of radon daughters to cigarette smoke would contribute to the carcinogenic,potential of ETS (Little et al., 1965; Rak1ewsky and Stahlhofen, 1966; Radford and Martell, 1978). Given the presence of appreciable amounts of radon in indoor air, irradiation of the bronchial tract from radon daughters attached - to smoke aerosol could be more important than the irradiation from the long-lived daughters in the tobacco itself. This subject .needs further research, especially in light of recent reports on the widespread prevalence of indoor radon throughout the world. VZS9e44e

44 TOXIC AND CARCINOGENIC AGENTS IN TOBACCO SMOKE Combustion products of cigarettes are the main contributors of ETS. Therefore, compari (Tables ~ 4 through 2-9) w th E r~respond- and carcinogens in ETS (Ta- ing concentrations in MS are relevant. rQ riate However, comparisons ofM~s a,ffE enees in chemical m- only if one considers the important ( g~ p position (including pH) and physicochemical nature e. . article sise, air dilution factors, b~~ en thettwooaeroe ls.b~ m°re, and particulate phases) anied b ollutants in ETS in indoor environments is often accomP - y i' the work environment or derived from other sources, such as cook- ing stoves and space heatet~11T~edre hala g°$~oneentratedesmoke between inhaling amblen aerosol during puff drawing. Finally, chemical and physieochem- ma- ical characteristics basedll co~ lafsble wml those of compounds chine smoking are not f- y- P eciahy in generated when a smoker Inhales cigarette smoke. Eep p e~ the case of low-yield cigarettes, the yields of constituents a p.- to be different between machine smoking and human smoking (Herning et al., 1981). Table 2-10 compares concentrations of some smoke constitu- ents in the MS generated in the laboratory from one cigarette to those inhaled by a nonsmoker exposed to ETS for 1 hour t The physical and chemical changes that occur in reactive smoke emission constituents during aging of the compounds after their le n't r'e into the environment mus durnglsmoking and Is ehemicslly oxide is generated in a cigarette ------------ - 's The Eomputations for exposures to nonsmokers for 1 hour in Table 2-10 - - - are made using the equation: mg/- ° mg/me x i0-ams/L X 600L/h, usuming an average respiratorY rate of 10 L/minute. To convert from pPm i. used: ti on g equa (or ppb) to mg/ms, the followin s ppm X molecular weight . mg/ m ° - RT 0 where RT at 2000 is 24.46. 5Z698L4g 45 intact when it leaves the cigarette in MS about 2 seconds later. - However, once emitted into SS and diluted to become an ETS component, nitric oxide is partially oxidized to nitrogen dioxide and progressively more oxidized as more time elapses, producing a potent hydrophilic nitrosating agent. SUM11rIAIt.Y. AND ItECO31rIl4tENDATIONS The smoldering of tobacco between puffs generates SS. Undi- luted - luted SS contains some toxic compounds in much higher concen- trations than MS, especially ammonia, volatile amines, volatile ni- -- trosamines, some nicotine decomposition products, and aromatic - amines. Furthermore, decay products of radon from the tobacco and frolri other sources adsorbed on some particles in indoor air might contribute to the carcinogenic potential of ETS. SS is a major contributor to ETS. Respiratory environments that are polluted with SS contain measurable amounts of nicotine and other toxic agents, including carcinogens. We lack data on the presence and concentrations of many of the knowia SS components in polluted, enclosed environments. The concentrations of toxic agents of ETS are governed primarily by the amount of tobacco smoked, the degree of ventilation, and the volatility of the agents. Future studies should concentrate on the analysis of toxic and carcinogenic agents in smoke-polluted environments. What Is Known 1. SS is the aerosol that is freely emitted into the air from the -- smo dering tobacco products between puffs. 2. ETS consists of diluted SS, exhaled MS, smoke that escapes from the burning cone during puff-drawing, and vapor-phase com- ponents (such as carbon monoxide) that diffuse through cigarette - paper into the environment. However, secondary reactions can oc- cur before a nonsmoker inhales ETS, such as aging, volatilization of nicotine, and adsorption of radon daughters on particles.. 3. Undiluted SS contains higher concentrations of some toxic compounds than undiluted MS, including ammonia, volatile aminee, volatile nitroeamines, nicotine decomposition products, and aromatic amines. 4. Conventional cigarette filter tips primarily influence the yield- of MS, but have little impact on the yield of SS. Highly

46 47 active filter tips, especially perforated ones, also affect the yield of components in SS. 5. Radioactive decay products in tobacco itself, for instance, Pb-210 and Po-210, and short-lived radon daughters adsorbed on _ _ smoke particles in indoor air can contribute to the carcinogenic potential of ETS. 6. ETS in indoor environments Is accompanied by pollutants, such as nitrogen oxides and carbon monoxide, derived from other sources, including cooking stoves and space heaters. ETS contains measurable amounts of nicotine and other toxic agents, including carcinogens. The concentrations of toxic agents of ETS are gov- erned - erned primarily by the amount of tobacco smoked, the degree of ventilation, and the volatility of the agents. 7. Nicotine, found in MS primarily in the particulate phase, occurs in ETS pFimarily in the vapor phase. Therefore, filters designed to reduce particles in the air will not substantially alter the nicotine concentration. What Scientific Information Is Missing 1. We lack data on the presence and concentrations of toxic and carcinogenic components in tobaeEo-smoke-polluted enclosed environments. 2. The distributions of various agents in vapor and particu- late phases of ETS are not well characterized. Further, the effect of air-cleaning systems on these distributions has not been stud- ied. Distributions are important with respect to the carcinogenic - potential of ETS. 3. We need to examine the importance of the abundant release of amines into ETS. We lack analytic data on secondary reactions of amines in polluted air, such as N-nitrosation and condensation with other ETS components. 4. The transfer of constituents other than nicotine from the particulate phase of SS to the vapor phase of ETS could be impor- • tant with respect to the retention of ETS in the_ respiratory tract of nonsmokers. mokers. 5. We do not know the extent to which nitrogen dioxide can contribute to endegenous nitrosation in nonsmokers as a con- stituent of the respiratory environment. Endogenous nitrosation - leads to nitrosamines in exposed subjecte. 93ZG98448

48 6. We need studiee to determine the extent to which ETS dif:s fers from M8 in ways related to health and their relative toxicities. 7. We should analyze toxic and carcinogenic agents in smoke- polluted environments, especially enclosed natural environments, and their uptake by nonsmokers. 8. Research should be conducted on interactions between ETS and radon daughters, especially as radon daughters can adhere to - - RSP, and can thereby enter the lung.. REFERENCES Adams, J.D., K.J. O'Mara-Adams, aed D. Hoffmann. On the mainstream- sidestream distribution of cigarette smoke components. Presented at the 39th Tobacco Chemists' Research Conference, Montreal, Canada, Oct. Z-6, 1986. American Conference of Governmental Industrial Hygienists (ACGIH). TLV Threshold Limit Values and Biological and Experimental Indices of 1986-1988. Cincinnati, Ohio: ACGIH, 1986. 114 pp. Armitage, A.K., and D.M. Turner. Absorption of nicotine in cigarette and cigar smoke through the oral musco. Nature 226:1231-1232, 1970. Badre, R., R. Ouilierme, N. Abram, M. Bourdin, and 0. Dumas. Pollution atmospherique par la_ fumde de tabac. Ann. Pharm. Fr. 38p443-462, 1978. Baker, R.R. The effect of ventilation on cigarette combustion mechanisms. Recent A_dv. Tob. 8ci. 10:88-160, 1984. Bergman, H., and 0. Axelson. Passive smoking and indoor radon daughter concentrations. Lancet 2:1308-1309, 1983. Brunnemann, K.D., and D. Hoffmann. The pH of tobacco smoke. Food Cosmet. Toxicol. 12:116-124, 1974. B_ runnemann, K.D., and D. Hoffmann. Pyrolytic origins of major gas phase constituents of cigarette smoke. Recent Adv. Tob. Sci. 8:103-140 1982. Brunnemann, K.D., D. Hoffmann, E.L. Wynder, and G.B. Gorl. Determi= nation of tar, nicotine, and carbon monoxide In cigarette smoke. A comparison of international smoking conditions. XXXVII of Chemical Studies on Tbbacco 8moke, pp. 441-449. In E.4 Wynder, D. Hoffmann, and G.B. f3ori, Eds. Smoking and Health i. Modifying the Risk for the Smoker. Proceedings of the 3rd World Conference on Smoking and Health. NIH Publ. No. 76-1221. Bethesda, Maryland: U.S. Department of Health, Education, and Welfare, Public Health Service, National Cancer Institute, 1976. _Br_ unnemann, K.D., L. Yu, and D. Hoffmann. Assessment of carcinogenic volatile N-nitrosamines in tobacco and in mainstream and sidestream smoke from cigarettes. Cancer Res. 37:3218-3222, 1977. Brunnemann, K.D., W. Fink, and F. Moser. Analysis of volatile N- nitrosamines in mainstream and sidestream smoke from cigarettes by (3 LC-TEA. Oncology 37:217-222, 1980. 49 Brunnemann, K.D., J.O. Scott, N.J. Haley, and D. Hoffmann. Endogeneous formation of N-nitrosoproline upon cigarette amoke inhalation. IARQ Sci. Publ. 67:819-828, 1984. Cano, J.P., J. Catalfn, It.. Badre, 0. Dumas, A. Viala, and It. Guillerme. Determination ination do la nicotine par chromatographie en phase gaseuse. Il. Applications. Ann. Pharm. Fr. 28:633-640, 1970. Carter, W.L., and 1. Hasegawa. Fixation of tobacco smoke aerosols for sise distribution studies. J. Colbid Interface Sei. 53:134-141, 1976. Uhappell, S.B., and R.J.. Parker. Smoking and carbon monoxide levels in enclosed public places in Now Brunswick. Can. J. Public. Health - 68:159-161, 1977. Coburn, R.F., 1t.E. Forster, and P.B. Kane. Considerations on the physio- logical variables that determine the blood carboxyhemoglobin concen- trations In man. J. Olin. Invest 44:1899-1910, 1986. - f3uddeback, J.E., J.R. Donovan, and W.R. Burg. Occupational aspects of passive smoking. Am. Ind. Hyg. Assoc. J. 37:283-287, 1976. Dalham, T., M: L. Edfors, and R. Rylander. Mouth absorption of various compounds in cigarette smoke. Arch. Environ. Health 18:831-836, 1988a. Dalham, T., M.-L. Edfors, and R. Rylander. Retention of cigarette smoke components In human lungs. Arch. Environ. Health 17:748-748, 1968b. Dube, M.F., and O.R. Green. Methods of collection of smoke for_ analytical purposes. Recent Adv. Tob. Scl. 8:42-102, 1982. Elliot, L.P., and D.R. Rowe Air quality during public gatherings. J. Air Pollut. Control Assoc. 26:B36-638, 1976. Eudy, L.W., F.A. Thome, D.L. Heavner, O.R. Green, and B.J. Ingebrethsen. Studies on the vapor-particulate phase distribution on environmental nicotine by selected trapping and detection methods. Presented at the 39th Tobacco Chemists' Research Conference, Montreal, Canada, Oct. 2-6, 1986. Fischer, T., A. Weber, and E. Clrandjean. Air pollution due to tobacco smoke In restaurants. Iat. Arch. Occup. Environ. Health 41:287-280, -- -- ----- -- - --- 1978 . , Fleischer, R.L., and F P. Parungo. Aerosol particles on tobacco trichomes. Nature 260:168-169, 1974, Galuskinova, V. S,4-Benspyrene determination In the smoky atmosphere of social meeting rooms -and restaurants: A contribution to the problem - - - - - - --- --_ o t e nox ousness of so-called passive smoking. Neoplasm 11:466-468, • - 1964. Godin, d., 0. Wright, and R.J. Shephard. Urban exposure to carbon monoxide. Arch. Environ. Health 26:306-313, 1973. Goldsmith, J.R., and L.I. Friberg. Effects of air pollution on human health, pp. 468-E10. In A.O. Stern, Ed. Air Pollution, Vol. Il, 3rd ed. New York: Academic, 1977. Grabowski, J., and C.S. Bell, Eds. M_ ea.urement_ in the Analysis and Treatment of Smoking Behavior. PnbL No. DHH9%PUB/ADM-83- 1286. Rockville, Maryland: U.S. Department of Health and Human Services, National Institute on Drug Abuse, 1983. 132 pp. Grimmer, Q., H. BShnke, and H.P. Harke. Passive smoking: Intake of poly- cyclic aromatic hydrocarbons by breathing of cigarette.moke containing air. Int. Arch. Occup. Environ. Health 40:93-99, 1977. 4zs9s4L8

50 Harke, H: P. Znm Problem ds Passivrauchens. 1. Uber den Ein9nss dee Rauehens aut die CO-Konsentration in BQrortumen• lnt. Arch. Are beitsmed. 33:199-g06, 1974. Harke, H: P., and H. Peters. Zum Problem des Passivranchens. Ill. Ober den EinBuss des Rauchens auf die CO-Konsentration im Kraftfahrseug bel Fahrten im Stadtgebiet. Int. Arch. Arbeitsmed. 33:221-229, 1974. Harmsen, H., and E. Elfenberger. T-abakranch in Verkehnmitteln, WohB-und Arbeitsr6umen. Arch. Hyg. Bakterial. 1e1s383-400, 1957. Herning, R.I., R.T. Jones, J. Bachman, and A. Mines. Puff volume inerbases when low-nicotine cigarettes are smoked. Br. Med. J. 283:187-189, 1981. Hill, C.R. Radioactivity in cigarette smoke. N. Englf ,i Me: ~i g different Hill, P., and H. Marquardt. Plasma and urine change brands of cigarettes. Olin. Pharmacol. Ther. 27:85R-858, 1980. Hiller, F.C. Deposition of sidestream smoke In the human respiratory tract. - - - Prey. Med. 13:602-607, 1984. Hilier, F.C., K.T. McCnsker, M.K. Masumder, J.D. Wilson, and R.C, Bone. Deposition of sidestream cigarette smoke in the human respiratory tract. Am. Rev. Resp. Dia. 125:406-408, 1982. Hinds, W.C., and M.W. First. Concentrations of nicotine and tobacco smoke in public places. N. Eng1. J. Mad. R94:844-846, 1975. Ho/Emann, D., and K.D. Brunnemann. Endogeneous formation of N-. nitrosoproline in cigarette smokers. Cancer lies. 43:5670-6574, 1983. Hoffmann, D., and S.S. Hecht. Nicotine-derived N-nitrosamines and tobacco- related cancer: Current status and future directions. Cancer Ree. 45:935-944, 1986. Hoffmann, D., N.J. Haley, K.D. Brunnemann, J.D. Adams, and E.L. Wynder. Cigarette sidestream smokes Formation, analysi w model of Lung the uptake by nonsmokers. U.S: Japan Meeting, .. Cancer; Honolulu, Hawaii, Mar. 21-23, 1983. Hoffmann, D., K.D. Hrunnemana, J.D. Adams, and N.J. Haley. Indoor air pollution by tobacco smoke: Model studies on the uptake by non.ruok- ers, pp. 313-318. In B. Berglund, T. Lindvall, and J. Sundell, Eds. . Indoor Air, Vol. 2. Radon, Passive Smoking, Particulates, and Hous- ing Epidemiology. Stockholm, Swedeh: Swedish Council for Building Research, 1984a. Hoffmann, D., N.J. Haley, J.D. Adams, and K.D. Brunnemann. Tobacco sideetream smoke: Uptake by nonsmokers. Prey. Med. 13s808-817, 1984b. Ingebrethsen, B.J., and S.B. Sears. Particle sise distribution measurements of sidestream cigarette smoke. Presented at the 39th Tobacco Chemists' Research Conference, Montreal, Canada, Oct. 2-6, 1985. International Committee for Cigar Smoke Study. Machine smoking of cigan. Coresta Inform. Bull. 1:31-34, 1974. Johnson, W•R " anQ C• ainheeacidss• dorelatedg ompou d.toJ Org. from the PY Ysis of am _ _ Chem. 36:189-192, 1971. ]ohn_son, W.R., R.W. Hale, J.W. Nedlock, H.J. Grubbs, and D.H. Powell. The distribution of products between mainstream and sidastream smol<e• Tob. Sei. 17:141-144, 1973a. 51 Johnson, W.R., R.W. Hale, S.C. Cough, and P_H. Chen. Chemistry of the conversion of nitrate nitrogen to .moke products. Nature 343:223-Y25, 1973b. Just, .1., M. Borkowska, and S. Masiarka. Zaniecsysscsenie dymem tyto- niowym powietrsa kawlarn warssawskich. (Tobacco smoke in the air of - Warsaw coffee rooms.) Rocs. Panestw. Zakh. Hyg. 43:1Z9-135, 1972. Keith, O.H., and J.C. Derrick. Messurement of the particle sise distribution and concentration of cigarette smoke by the econifuge_' J. Colloid Sci. 16:340-358, 1960. lflus, H., and H. Kuhn. Distribution of various tobacco smoke constituents in main and sidestream smoke. A review. Beitr. Tabakforach. lnt. 11,929-285,1984. Krasnegor, N.A., Ed. Cigarette Smoking as a Dependence Process. Publ. No. - -- DHEW/PUB/ADM-79/800. Rockville, Marylandt U.S. Department of Health and Human Services, National Institute on Drug Abuse, 1979. 92 pp. Kruger, J., and J.F. Nothiing. A comparison of the attachment of the decay products of radon-220 and radon-222 to monodispersed aerosols. J. Aerosol. Sci. 10:671-679, 1979. Ladd, K.F., H.L. Newmark, and M.C. Archer. N-nitrosation of proline in smokers and nonsmokers. J. Natl. Cancer Inet. 73:83-87, 1984. Little, J.B., M.P. Radford, and H.L. McComb. Distribution of polonium-410 in pulmonary tissue of cigarette smokers. N. Engl. J. Med. 273:1343- 1361, 1986. Lu, P.H., Y.Z. Hong, N.a. Shi, W.D. Zhang, C.S. Dai, J.W. Huang, X.X. - -- qin, M.Z. Kin,-an_ d D.H. 'Iong. Radiographic findings in•cotton textile workers and the relationship to cigarette smoking. Regul. Toxicol. Pharmacol. 6:80-65, 1986. Martell, E.A. Tobacco radioactivity and cancer in smokera. Am. Sci. 83:404- 412,1976. Miller, J.E. Determination of the components of pipe tobacco and cigar smoke by means.of a new smoking machine, pp. $84-696. Proceedings of the Third World Tobacco Scientific Congress, University College nf Rhodesia and Nyasaland, Salisbury, Southern Rhodesia, Feb. 18-58, 1963. Salisbury: Salisbury Printers, Ltd., 1964. National Institute for Occupational Safety and Health. Health Aspects of Smoking in Tfansport Aircraft. Publ. No. NIDA/RD-79/028, Research Monograph Series No. 23. Rookville, Maryland: National Institute for - Occupational Safety and Health and Federal Aviation Administration, Washington, D.C., 1971. 92 pp. National Research Council, Committee on Indoor Pollutants. Tobacco smoke, pp. 149-168. In Indoor Pollutanb. Washington, D.C.: Na- tional Academy Preu, 1981. Nero, A.V., R.G. Sextro, S.M. Doyle, B.A. Moed_ , W.W. Nasaroff, K.L. Revsan, and M.D. Schwehr. Characterising the sources, range and environmental in8uence of radon-222 and its decay products. Sei. Total Eaviron. d6:Y33-244, 1985. Neurath, U., and H. Ehmke. Apparatur sur Untersuchung des Nebenetrom- rauchee. Beitr. Tabakforsch. Y:117-1Y1, 1964. I QZ69g4L9

52 Nenrath, a., and H. Horstmann. Einfluu des F.uchtigkeits gehaites ven Cigaretten anf die Zusammnensetsung sles Ranches und die E3lutsenen- temperaturen. Beltr. T-abakforsch. Int. 2:93-100, 1973. Perry, J. Fasten your seatbeltst Non smoking. B.C. Med. J. 16:904-306, 1973. ' Pil_lsbury, H.C., 0.0. Bright, K.J. O'Connor, and• F.W. Irish. Tar and nicotine In cigarette smoke. J. Assoc. Off. Anal. Chem. 62c468-48R, 1969. Porthein, F. Zum Problem des 'Passiv-Ranchen..' Msnch. Med. Wochen- achF. 118:707-709, 1971. Raabe, O.G. Concerning the interactions that occur between radon decay products and aerosols. Health Phys. 17:177-186, 1969. Radford, E.P., Jr., and V.R. Hunt. Polonium-210r A volatile radioalement in cigarettes. Science 143:247-249, 1964. Radford, E.R, and E.A. Martell. Polonium-210: Lead-210 ratios as an Index of residence times of insolubk particles from cigarette smoke in bronchial hial epithelium, pp. 687-681. In W.H. Wdlton and B. McGovern, Eds. Inhaled Particles IV, Part 2. New Yorkt Pergamon, 1978. Rajawsky, B., and W. Stahihofen. Polonium-210 activity In the lungs of cigarette smokers. Nature 209s1312-1315, 1966. R.$hl, 0., J.D. Adams, and D. Hoffmann. Chemical studies on tobacco smoke. LXVI. Comparative assessment of volatile and tebacco-speci8e N-nitrosamines in the smoke of selected cigarettes from the UBA, West Germany and France. J. Anal. Toxicol. 4:266-269, 1980. Sakuma, H., M. Kusama, S. Munakata, T. Ohsuml, and S. Sngawara. The distribution of cigarette smoke components between mainstream - - and sidestream smoke. i. Acidic components. Beltr. Tabakforsch. Int. 12e83-71, 1983. Sakuma, H., M. Kusama, K. Yamaguchi, T. Matsuki, and S. Sugawara. The distribution of cigarette smoke components between mainstream and sidestream smoke. 11. Bases. Beitr. Tabakforsch. lat. 12:199-209, 1984a. - - - Sakuma, H., M. Kusama, K. Yamaguchi, and S. Sugawara. The distribution of cigarette smoke components between mainstream and sidestrsam smoke. Ili. Middle and higher boiling components. Beitr. Tabakforsch.. int. 11sZ61-268, 1984b. Scasseilsti-Sforselini, (1., apd A. Sa.ino. Evaluation of a rapid inde: of ambient contamination by cigarette smoke in rslation to the composition of gas phases of the smoke. Riv. ltal. Ig. 28:43-66, 1968. Schmelts, I., A. Wenger, D. Hoffmann, and T.C. Tso. Chemical studies on tobacco smoke. 63. On the fate of nicotine during pyrolysis and (n a_ • burning cigarette. J. Agric. Food Chem. 27:802-808, 1979. Sebben, J., P. Pimm, and R.7. Shephard. Cigarette smoke in enclosed public facilities. Arch. Eaviron. Health 32:63-68, 1977. Seiff, H.E. Carbon Monoxide as an•Indicator of Cigarette-Caused Pollution Levels in intercity Buses. Pubi. No. BMCS-IHS-73-1. Washington, D.C.: U.S. Department of Transportation, Bureau of Motor Carrier Safety, 1973. 13 pp. S1_avin, R.(3., and M. Herts. Indoor air pollutiont A study of the thirtieth annual meeting of the American Academy of Allergy. 30th_ Annual Meeting Am. Acad. Allergy, San Diego, Feb. 16-19, 1976. 53 Steh__Iik, Ci., O. Richter, and H. Altmana. Conc.ntration of dimethylni- - - -- - - trosamine in the air -of smoke-Hlled room., t;cotoxicol. Environ. Saf. 8:496-600, 1982. Ssadkow4ki, D., H.-P. Harks, and J. Angerer. Kohknmonoxidbelastung durch Passivrauchen in Biiraori<umen. Zentralbl. Prak. Inn. Med. 3:310-313, 1978. Tso, TA., N. Harley, and L.T. Alexander. Source of Iead-210 and polonium- 210 in tobacco. Science 163:880-882, 1966. Tsuda, M., j. Niitusuma, S. Sato, T. Hirayama, T. Kakisog, and T. Sugimura. Increase in the l.veie of N-nitro.oproline, N-nitrothioproline, and H- nitre-s-methylthiepreline in human urine by cigarette smoking. Cancer hstt. 80:117-124, 1986. U.S. Department of Health and Human Services. Third Annual Report on Carcinogens. National Toxicology Program. Research Triangle Park, North fSarolina: U.S. Department of Health and Human Services, 1983. 137 pp. U.S. Department of Transportation snd U.S. Department of Health, Educa- tion, and Welfare. Health Aspects of Smoking in Transport Aircraft. U.S. Department of Transportation, Federal Aviation Administration, and U.S: Department of ITealth, Education, tion, and_ Welfare, Washington; D.C.: NIOSH, 1971. 86 pp. U.S. Public Health Service. Smoking and Health. A Report of the Surgeon General. DHEW Pubi. No. (PHS) 79-60088, Washington, D.C.: U.S. Department of Health, Education, and Welfare, Public Health Service, 1979. Vainio, H., K. Hemminki, and J. Wilbourn. Data on the carcinogenicity of chemicals in the IARC Monographs programme. Carcinogenesis (Lond.) 8:1863-1886, 1986. Vilcins, ©., and ,1.tQ. I,ephardt. Aging process of cigarette smoke:_ Formation rm_ ation of inethyi nitrite. Ohem. Ind. (Lot:d.) 22:974-975, 1976. Wartman, W.B., Jr., B.C. Cogbill, and E.S. Harlow. Determination of particulate matter in concentrated aerosols. Application to analysis of cigarette smoke. Anal. f3hem. 31:1706-1709, 1969. Weber, A. Annoyance and irritation by passiYe smoking. Prey. Med. 13:818- 826, 1984. Weber, A., and T. Fischer. Passive smoking at work.. lnt. Arch. Occup. Environ. Health 47:209-221, 1980. Weber, A., O. Jermini, and B. Grandjean. Irritating effects en man of air pollution dee to cigarette smoke. Am. J. Public Health 88;872-878, 1976. Weber, A., T. Fischer, and I.. (Irandjean. Passive smoking: Irritating effects of the total smoke and gas phase. Int. Arch. Occup. Environ. Health 43:183-193, 1979. Wynder, E.I,., and D. Hoffmann. Tobaeco and Tobacco Smoke: Studies in Experimental Carcinogenesi.. New York: Academic Press, 1967. 730 PP. (;Z,c;9gl.4R

66 3 In Vivo and In Vitro Assays to Assess the Health Effects of Environmental Tobacco Smoke INTRODUCTION Suitable methods for assessing the potential for adverse health effects resulting from exposure to environmental tobacco smoke (ETS) are limited by the complexity of the composition of the mixture. In vivo and in vitro assays are commonly used to establish - - carcinogenicity and in some casee to extrapolate risks to humans. - For complex mixtures such as ETS, these assays may be done on the mixture itself or on individual chemical constituents. Many properties of ETS change as the smoke "ages" after its initial generation. Aging probably affects the bioavailability, as well as physicochemical characteristics, of the smoke. As inhalation is the primary route by which humans are ex - posed to tobacco smoke, it is obviously the preferred method of administration in animal models for evaluating the toxicologi- cal properties of both cigarette smoke and ETS. While extensive inhalation studies have been performed on the toxicological prop- _--- __ erties of mainstream cigarette smoke (MS), far fewer studies have been performed on sidestream smoke (SS) and ETS.''he selection of appropriate animal models requires familiarity with exposure systems, as well as with basic anatomical differences between the model and human respiratory tracts. Methods other than inhalation, such as in vitro assays, have been developed for the evaluation of MS. A few of these methods have been applied to the assessment of the relative toxicological properties of SS versus MS. These methods are frequently crit- icized because of differences in the way the smoke constituents 0C69gLL.9 54 are presented to the test system as compared with that which oc- curs in the human situation. Despite these limitations, the use of cigarette smoke condensate (CSC) from MS has provided insight into the relative carcinogenic potential of various constituents in - the MS of cigarettes. Similar studies using suitable condensates -- - - from SS and aged ETS could provide additional data on the effects - --- - of ETS.. IIY- VIVO ASSAYS ON ENVIRONMENTAL - TOBACCO SMOKE Exposure Methods in Laboratory Research Several methods are available to evaluate the potential health effects of inhaled pollutants. Some common ones are whole-body exposure, head-only exposure, nose- or mouth-only exposure, lung- only exposure, or partial-lung exposure. Since the primary objec- tive of an inhalation experiment is_ to determine the effects of the test substances or mixture on the respiratory system, it is prefer- able to eliminate or limit exposure through the skin or through ingestion (such as through contact with materials deposited on the fur or contaminated food and water). Three methods have been used to determine the amount of material deposited in the respiratory tract (Phalen, 1984): di- rect measurement, calculations using airborne concentrations and uptake models, and calibration of the exposure apparatus using tracer substances. Direct measurement requires analysis of major components and their metabolites in tissues as well as in urine and feces or measurement of the amounts of material in the in- spired and expired air. Aside from calculating dose based upon particle aerodynamic size and physiological data on lung function of experimental animals, tracers can provide reasonable estimates. of exposure. Inhalation exposure chambers are used for those studies in which whole-body exposure is desired. The ability to expose a large number of animals at one time and the absence of a need to restrain or anesthetize the animals are among the advantages in using this approach. There are, however, several major disad- vantages. The animals are exposed through skin absorption and mouth ingestion and, in prolonged instances, by food and possi- bly water contamination. Animals tend to avoid exposure in such

68 chambers by huddling together or covering their noses with their own fur. Losses of particulate aerosols to the interior walls of the chambers are also frequently a problem. Head-only exposure systems eliminate many of these prob- lems. The disadvantages of these systems are that the animal must be restrained and is stressed or anesthetized, and there is difficulty in forming an adequate seal. Nose- or mouth-only exposure systems further limit exposure to the oral cavity and the respiratory tract. Masks or the use of catheters in the nose are generally used with larger animals. Lung and partial-lung-only exposure systems such as endotracheal tubes are employed to bypass the upper respiratory tract and to directly expose the lung. Most of these methods require that the animal be anesthetized, which may alter normal respiration. Other disadvantages include disruption of normal airflow by the presence of tubes in the airways and the lose of normal humidification and thermal regulation of the inspired air caused by bypassing the upper respiratory tract. Intratracheal instillation is an alternative to inhalation for evaluating the effects of individual compounds on the respiratory system. While there are several advantages in employing tliis bioassay technique, it is also known that the distribution of test material to respiratory tissue may differ from that which would be obtained by actual inhalation exposures. Instiltation of an aqueous suspension of radiolabeled particles resulted in a less uniform deposition than inhalation (Brain et al., 1976). Animal Models In Inhalation Studies The selection of an appropriate animal model for inhalation studies with potentially toxic agents is compounded by the fact that one of the major functions of the mainmalian pensory ap- paratus is to limit the exposure to toxic agents either by alter- ing breathirig or by producing avoidance behavior (Alarie, 1973; Wood, 1978). Also, the selection of animal species and strains for inhalation exposure studies requires thorough evaluation. The use of several (at least three) animal species, several dose levels, and animals that metabolize the suspect toxin iu a similar manner to humans is recommended for those studies that attempt to evalu- ate human hazards (Stuart, 1976). The appropriate animal model should have (1) a similarity to the human respiratory tract with TesseZ,4e 57 respect to anatomy, physiology, and susceptibility; (2) a life span appropriate for the proposed study; (3) a sensitivity to certain -- - - - classes of toxic agents; (4) anatomical or physiological properties ---- - - that could lead to increased precision in empirical measurements; (5) an existing data base; (6) a documented history of appropriate procedures; and (7) an adaptability for generating data that might be used for mathematically modeling the animal system and its responses to airborne particulates. Results of Inhalation Studies Inhalation studies on the careinogenicity of MS have been per- formed on a variety of laboratory animals. The early studies with rodents have been previously reviewed (Wynder and Hoffmann, 1967; Mohr and Resnik, 1978). More recent studies verify these findings for several animal species exposed to whole smoke or MS. A few studies have exposed mice to the vapor phase of fresh MS, and one (see below) exposed mice to the vapor phase of flue-cured MS. Because commonly utilized filter systems do not remove many of the vapor-phase constituents, studies contrasting the effects of exposure to whole smoke with the effects of exposure to the gas phase should throw some light on the possible health effects of ETS. Male and female C57B1 mice (100 in each group) were ex- posed nose only for 12 minutes daily to the gas phase of smoke of cigarettes prepared from tiue-cured tobaccos (IIarris et al., 1974). The treated mice had lung tumors and emphysema, independent of the tumors, which were not found in control mice. A total of 219 C&7Bl and 186 BLH mice were exposed to the gas phase of cigarette MS. The particulate matter was removed by passing the smoke through a Cambridge filter. The animals were exposed to the gas phase of 12 cigarettes for 90 minutes - daily over 27 months. The percentages of mice with lung adeno- mas - - mas were 5.5% and 32% in the smoke-exposed C57B1 and BLH mice, as compared with 3.4% and 22% for their respective controls - (Otto and Elmenhorst,1987). Therefore, it appears that there'are carcinogenic constituents in the vapor phase of the smoke. Using Snell's mice, similar studies evaluated the toxicological properties of whole MS and the gas phase of MS. In these stud- ies, the animals were housed in individual chambers during the exposure (Leuchtenberger and Leuchtenberger, 1970). There was

58 a significant difference (p < 0.1) in the incidence of pulmonary tumors between the animals exposed to whole smoke and control animals. The difference was greater (p = 0.005) for animals ex- posed to only the gas phase of cigarette smoke as compared with - - the same controls, so that the rate of tumors among the gas-phase- exposed animals was greater than among the whole-smoke exposed ,animals.. In Vivo Bloassays Other Than Inhalation Alternative methods have been used to assess the relative chronic toxicity of cigarette MS in an attempt to reduce the cost and technical difficulties associated with inhalation experiments. The most common approach has been to use the CSC in bioassay procedures. In preparing the condensate, many of the volatile and aemivolatile components are lost. In addition, it is not known how the aging of the CSC may affect chemical composition and biological iological activity. To date, only one study has examined the carcinogenic po- tential of the condensate of SS of cigarettes (Wynder and Hoff mann, 1967; International Agency for Research on Cancer, 1986). Cigarette "tar" from the SS of nonfilter cigarettes, which had set- tled on the funnel covering a multiple-unit smoking machine, was - - suspended in acetone and applied to mouse skin for 15 months.. Fourteen of 30 Swiss-ICR mice developed benign skin tumors, and 3 had carcinomas. In a parallel assay of MS from the same source, .a 50% CSC:acetone suspension applied to deliver a comparable dose of CSC to 100 Swiss-ICR female mice led to benign skin tu- mors in 24 mice and malignant skin tumors in 6. This indicates - that the smoke condensate of SS has greater tumorigenicity per equivalent dose on mouse skin than MS "tar" (p < 0.05, Wynder and Hoffmann, 1967). IN VITRO ASSAYS ON ENVIRONMENTAL TOBACCO SMOKE Several short-term bioassays have been performed to evaluate the genotoxicity of cigarette MS. These studies have been the subject of two recent reviews (DeMarini, 1981; Obe et al., 1984). While most of them have evaluated the effects of CSC, some have attempted to evaluate either the gas phase or the whole smoke. 59 The most commonly employed assay for mutagenic activity employs various strains of Salmonello typhimurium. Whole smoke as well as CSC from four types of tobacco were found to be mu- tagenic in S. typhimurium TA1538 (Basrur et al., 1977). Recent studies have shown that SS is also mutagenic in a system where the smoke was tested directly on the bacterial plates (Ong et al., 1984). They support extensive assays performed on CSC that in- dicate that tobacco smoke has significant mutagenic potential and show that the particulate matter of SS is likely to be a significant contributor to the mutagenic activity of indoor air particulate mat- ter -- ---- ter (Bos et al, 1983; Lofr6th et al., 1983). Thus, similar mutagenic activity for the CSC of SS would be expected. In another study (Lewtas et al, in press), condensate from air polluted with ETS for 10 hours was used in an assay employing - -- - --- - S. typhimurium. The average indoor air mutagenicity per cubic meter was significantly correlated with the number of cigarettes smoked. - Another in vitro assay measures the number of sister-chroma- tid exchanges (SCEs) in human lymphocytes. Valadand-Barrieu and Izard (1979) used a solution of the gas phase from cigarette - MS. They showed that this solution induced a significant dose- related increase in SCEs. SUMMARY AND RECOMMENDATIONS Sufficient data are not available to assess the relative genotox- icity ---- - icity and toxicity of whole ETS. A few isolated reports have dealt - - - - with the genotoxicity of SS and ETS, and the relative toxicity of MS and SS. In order to evaluate ETS, it is suggested that in vitro genotoxicity assays in at least two systems should be done with ETS per se as well as with its particulate matter. These assays under controlled and, subsequently, under field conditions should - - not be limited to freshly generated ETS, but should also attempt - - - to determine effects of various degrees of air dilution and aging. In a comprehensive analytical approach, data should be generated - to determine systematically the concentrations of toxic and tu- - --- - morigenic agents in various milieus with ETS. At the same time, it may be useful to examine the uptake of tobacco-specific agents as well as the mutagenicity of the urine of nonsmokers exposed to - - - - ETS. All of these measures should be considered in the context of - - detailed exposure histories.

61 What Is Known 1. The lungs of various species have different physiological properties, making each of them the experimental species of choice only for certain situations, depending on the objective of the re- search study. 2. ETS and SS have been shown to be inutagenic in a system where the smoke was tested directly on bacterial plates. . 3. The extensive studies of MS can serve as a guideline for the evaluation of ETS. Many of the constituents in the smokes are similar. Despite the limitations of extrapolating from various bioassays to man, the use of CSC from MS has provided insight as to the contribution of various components to the carcinogenic - potential of MS from cigarettes. 4. In the only study reported to date using SS condensate, SS condensate was shown to be more carcinogenic than MS conden- sate. What Scientific Information Is Mssing 1. Only a few laboratory methods have been applied toward the assessment of the relative toxicological and genotoxic prop- erties of SS generated from cigarettes and, more importantly, of ETS. Research is needed to clarify the appropriate methods for es- timating genotoxicity and to isolate and identify the active agents in body fluids of ETS-exposed nonsmokers. 2. Comparative inhalation studies with MS, SS, and ETS are still needed. Such assays, while not duplicating human exposure patterns, would provide more definitive information about the - relative carcinogenic potential of SS in comparison to the MS of the same cigarettes. 3. The aging of the atmosphere in which ETS occurs can havee a profound effect on its chemical composition, physical c-harac-, teristice, and overall biological effects. Therefore, studies of aged ETS are needed. - 4. Where exposure histories can be specified clearly, valida- tion and quantitative determination of genotoxic markers for sub- stances in ETS that also occur in the environment would be of value for measuring dose of ETS. 6. In examining the effects of MS, many research workers have used condensates of the smoke painted on the shaved skin of mice. 1~1:s9eZ44e Similar work with skin painting has not been done with ETS and would be of value for assessing the differential toxicity of ETS and MS. 8.. In vitro assays are needed for estimation of the tumor promotion and eocarcinogenic effect of E'!'S. In vitro tests are quicker than in vivo teste, and enough material can not be collected to do in vivo tests. , REFERENCES Alarle, Y. Sensory Irritation by airborne chemicals. C3RQ Grit. Rev. Toxicol. 2:299-363, 1973. Basrur, P.K., S. McClure, and B. Ziikey. A comparison of short term bioauay results with careinogenicity of experimental cigarettes, pp. 2041-2048. In H.E. Nieburg., Ed. Prevention and Detection of Cancer, Vol. 1. New York: Marcel Dekker, 1977. Ho., R.P., J.L.E3. Theuws, and P.Th. Henderson. ndenon. Excretion of mutagens in human urine after passive smoking. Cancer Lett. 19:86-90, 1983. Brain, J.D., D.E. Kundson, S.P. Sorokin, and M.A. Davis. Pulmonary distribution of particles given by intratracheal intillation or by aerosol inhalation. Environ. Res: 11:13-33, 1976. D_ eMarini, D.M.'Mutagenicity of fractions of cigarette smoke condensate in Nuer»eporo enues and Solmonelln lyphimuriwn. Mutat. Res. 88:363-374, 1981. Harris, R.J., a. Negroni, S. Ludgate, O.R. Pick, P.O. Chesterman, and B.J. Maidment. The incidence of lung tumours In c6781 mice exposed to cigarette smoke: Air mixtures for prolonged periods. Int. J. Cancer 14:130-138, 1974. International Agency for Research on Cancer (IARO) Monographs: Evalua- tion of Carcinogenic Risk of Chemicals to Humans, Vol. 38, pp. 163-314. Tobacco Smoking. Lyons: IARC, 1988.. 421 pp. Lenchtenberger, O., and R. Leuchtenberger. Effects of chronic Inhalation of whole fresh cigarette smoke and of its gas phase on pulmonary tumarigenesis In Snell's mice, pp. 329-346. In P. Nettesheim, M.fi.. Hanna, Jr., and J.W. Deatherage, Jr., Eds. Morphology of Experimental Respiratory Carcinogenesis. Proceedings of a Biology Division, Oak Ridge National Laboratory, Conference, Gatlinburg, Tenn., May 13-18, - 1970. Washington, D.C.: U.S. Atomic Energy Commission, 1970. - - L__ewtas, J., S. (]oto, K. Williams, J.C. Chuang, B.A. Petersen, and N.K. Wilson. The mutagenecity of indoor air particles In a residential pilot fieid study. Atmos. Environ., In press. Lofroth, E3., L. Nilsson, and 1. Alfhelm. Passive smoking and urban air pollution: S.Imonella/microsome mutagenicity assay of simultaneously collected indoor and outdoor particulate matter, pp. 61b-626. In M.D. Waters, S.S. Saadhu, J. Lewtas, L. Claxton, N. Chernoff, and S. Nesnow, Eds. Short-Term Bioassays in the Analysis of Complex Environmental - Mixtures. ill. New York: Plenum, 1983.

62 Mohr, U., and 0. liesnick. Tobacco carclnogenasis, pp. 263-367. In O.C. - Harris, Ed. Pathogenosis and Therapy of Lung Cancer. Now Yorkt Marcel Dskker, 1978. Obe, fl., W.-D. Holler, and H.,.i. Vogt. Mutagenie activity of cigarbtte smoke, pp. 223-246. In a. Obe; Ed. Mutations In Man. Berlin: Springer-Verlag, 1984. Ong, T., J. Stewart, and W.Z. Whong. A simple In situ mutagenicity test system for detection of mutagenic air pollutants. Mutat. Ras. 139:177-181, 1984. Otto, H., and H. Elmenhorst. Experimentelle Untersnchnngen sur Ta- morinduktion mit der Gasphase des Zigarettenraucks. Z. Krebsfersch. 70:46-47, 1967. Phalen, R.F. Inhalation Studies; Foundations and Techniques. Boca Raton, Florida: CRE) Press, 1984. Stuart, B:O. Selection of animal models for evaluation of Inhalation hasards In man, pp. 288-288. In E.F. Aharonsen, S. lte_ n-David, and Ivt.A. Kiingberg, Ed., Air Pollution and the Lung. New York: John Wiley de Sons, 1976. Vsladsud-Barrien, D., and O. Isard. Action de la phase gaseuse de fum6e de cigarette sur le taux d'echangee des chromatides-soeurs du lymphocyte humain In vitro. O.R.. Acad. Scl. (Paris) 488:899-901, 1979. Wood, R.W. Stimulus properties of inhafed substances. Environ. Health ' Perspect. Z8:89-78, 1978. Wynder, E.L., and D. Ho/[mann. Tobacco and Tobacco Smoke: Studies In Experimental ental Carcinogenesis. New York: Academic Press, 1967. 730 PFe - _ -- II . ASSESSING EXPOSURES TO ENVIRONMENTAL TOBACCO - SMOKE VCG98448

4 Introduction SEC9B44Q I Exposure to a variety of air contaminants has been shown to produce adverse health and discomfort responses in humans. In another report from the National Academy of Sciences (NRC, 1985), the methodological issues of studying exposures to air pol- lutants and subsequent health effects are discussed in detail. This part of the report considers issues relevant to assessing exposure to ETS. Ideally, evidence for health effects in humans should be demonstrated in epidemiologic studies that are consistent with a plausible hypothesis across a range of exposures or doses. However, many epidemiologic studies have substantial uncertainties associ- ated with exposure variables. A framework for assessing exposures - - to environmental tobacco smoke (ETS) is discussed below. A va- riety of approaches to current and historic exposures to ETS, such as personal monitoring, locational monitoring, questionnaires, and biologic monitoring, are presented. Concentrations of air contaminants exhibit pronounced spa- tial and temporal variations, regardless of the microenvironments in which they are found (outdoors, residential, industrial, etc.). Ideally, identifying the air- contaminant or class of contaminants implicated in producing adverse health or comfort effects is essen- tial in designing an air-monitoring program. In practice, however, it is often necessary to monitor a clas$ of contaminants (for in- stance, total mass of respirable particles) or a proxy contaminant (for instance, nicotine), when the specific air contaminant produc- ing the adverse impact can not be identified or easily measured. The air -contaminants associated with ETS are comprised of a 65

66 b.road range of many vapor- and particle-phase inorganic and or- ganic chemicals noted in Chapter 2, some of which can undergo pronounced physicochemical changes. Assessing impact on human health and comfort requires the identification of proxy air eontam- . inants for ETS that will permit a determination of exposure in a background of contaminants from other sources (see Chapter 5). In epidemiologic studies of air contaminants, it is important to specify exposure to specific particulates or gases on a time scale corresponding to the health or comfort effect sought. The impact of exposure to an air contaminant should, ideally, be evaluated in terms of the biologic dose of the contaminant or its metabolites received by the target tissue. In most casea, this is not practi- cal. The uptake, distribution, metabolism, and sito and mode of action of the contaminant in humans is neither well understood nor easily measured. Moreover, dose cannot be directly assessed. Factors affecting the uptake of air contaminants include physical characteristics of the contaminant," as well as physiological char- - acteristics and activity levels of the exposed person (see Chapter 7). In the absence of an ability to measure or specify the dose of a contaminant received, exposures to air contaminant(s) are as- sessed by either using biological markers, measured in the subject population, or by measuring the air-contaminant concentrations in the physical environment (Figure 4-1). Exposures to airborne contaminants can be assessed by three basic approaches (Figure 4-1): ~ • personal air-contaminant monitoring, • modeling, based on air sampling, time-activity patterns, and questionnaires, and e biological markers. Personal monitoring employs samplers (worn by subjects) that record the integrated concentration individuala are exposed to in the course of their normal activity for time periods of several hours to several days (see Chapter 5). The modeling approach employs the use of stationary moni- tora to measure the air-contaminant concentrations in a number of - -- --- - - - microenvironments. These measured concentrations are combined - - with time activity patterns (time budgets) to determine the aver- age exposure of an individual as the sum or the concentrations in each environment weighed by the time spent in that environment. 67 AM1 CONTA ENI/YIONYEIIrAL TOBACCO {YOKE {UPR08ATE{ OR MODELf EEPf1{ e WBOOR YICRO- EMVBIONYEMI{ YwYBdaN Y~BII~~Be~ c.ne«w.Bo.. BIOLOGICAL MARKER{ FIGURE 4-1 Flow dlaEram of components for tuessin` human exposure. to air contaminants from environmental tobacco smoke. Questionnaires are employed in two capacities: (1) to provide in- formation on the physical properties of each environment, includ- ing --- - ing source use parameters; in order to model the concentration of air contaminants in the microenvironment, thus permitting a pre- diction of air-contaminant concentrations in spaces not'monitored; and (2) to provide a simple categorization of exposure levels, such as exposed versus unexposed or none versus low versus high. Questionnaires have been used to categorize subjects' expo- sure to ETS in all studies of risk of chronic lung disease reported to date. Chapter 6 discusses the use of questionnaires to categorize ETS exposures. Chapter 7 reviews assumptions required to estimate exposure- dose relationships for ETS and gives an approximation to the dose received under a specific situation. sess944Q

68 Chapter 8 examines the use of biologicnl markers, such, as urinary cotinine, as indices of exposure to ETS.. There are several factors (Figure 4-1) that determine the com- position position and level of ETS air contaminants in the indoor environ- ments. Determining the range of values for each of these factors will lead to an understanding of their impact on ET$ exposures. Efforts to modify or eliminate exposures to ETS must focus on - the factors that control the concentrations in the physical environ- ment, ment, since these factors result in the exposure that relates to the adverse health or comfort effect. REFERENCE National Research Council, Committee on the Epidemiology of Air Pollu- tants. Epidemiobgy and Air Pollution. Wuhington, D.C.: National Academy Press, 1986. 224 pp. 5 Assessing Exposures to Environmental Tobacco Smoke in the External Environment Environmental tobacco smoke (ETS) is composed of more than 3,800 compounds. The emitted compounds are found in vapor or particulate phases, or in some cases both. Volatile mate- rial may evaporate from particles within seconds to minutes after emission (e.g., nicotine, see Chapter 2). ETS has not yet been adequately characterized such that its chemical and physical na- ture can be clearly defined. The concentration of any individual or group of E1S constituents in an enclosed space is a function of: a) the generation rate of the contaminant(s) from the tobacco, ~b) the source consumption rate, (c) the ventilation or infiltra- tion rate, (d) the concentration of the contaminant(s) of interest in the ventilation or infiltration air, (e) the dogree to which the air is mixed, (f) the removal of the contaminant(s) by surfaces or chemical transformations, and (g) the effectiveness of any air- cleaning devices that may be in use. Exposure to ETS takes place in many settings-such as public, industrial, nonindustrial occu- pational, and residential buildings-and is a function of the time an individual spends in a microenvironment and the concentration of the ETS constituents in that environment. ETS exposures can be determined either by extrapolation from Rxed-location moni- toring survey instruments that are portable or by direct personal monitoring, using lightweight pumps and filters worn by subjects. - This chapter will conaid"er the methodology and data available - - for assessing human exposures to ETS in the physical (external) environment, including the suitability of proposed tracers or proxy air contaminants that would be representative of ETS, available- - data on ETS exposure from personal monitoring and monitoring of LC69g448 1 69

. 70 indoor environments, and the application of modeling to assessirig ETS exposures, TRACERS FOR ENVIRONMENTAL _ ---- - TOBACCO SMOKH' It is difficult to assess the ETS contribution to exposures because it usually exists in a complex mix of air contaminants from other sources. It is not practical, or pcs;sible, to monitor the full range of air contaminants associated with ETS, even under laboratory conditions. Chamber and field studies of ETS have monitored proxy contaminants as indicators of ETS. Most studies to date have been less than ideal because the component that was measured did not meet all the following criteria for an ETS tracer. A marker or tracer for quantifying ETS concentrations should be: . unique or nearly unique to the tobacco smoke so that other sources are minor in comparison, - a constituent of the tobacco smoke present in sufficient quantity such that concentrations of it can be easily detected in air, even at low smoking rates, - • similar in emission rates for a variety of tobacco. products, and • in a fairly consistent ratio to the individual contaminant of interest or category of contaminants of interest (e.g., suspended - particulates) under a range of environmental conditions encoun- tered tered and for a variety of tobacco products. While a variety of measures have been used as proxies or tracers of ETS, no single measure has met all the criteria out- lined above, nor has any measure been universally accepted or - - recognized as representing ETS exposure. Carbon monoxide (CO) has been measured extensively both in chamber studies (Bridge and Corn, 1972; Hoegg, 1972; Penkala and De Oliveira, 1975; Weber et al., 1976, 1979a,b; Weber and Fisher 1980; Weber, 1984; Muramatau et al., 1983; Leaderer et - al., 1984; Winneke et al., 1984; Clausen, et al., 1985) and in oc- cupied public and nonindustrial occupational indoor spaces (see Table 2-4) to represent ETS levels. Under steady-state conditions in chamber studies, where outdoor CO levels are known and the tobacco brands and smoking protocols constant, CO can be a rea- sonably reproducible indicator of ETS exposure. The variability r 71 of CO production from tobacco combustion ustion is not well known and may vary considerably as a function of a number of variables (puff volume, puff duration, temperature, etc.). The ratio of CO, a nonreactive contaminant, to the more reactive gas-phase contam- inants in ETS and to reactive,suspended particulate mass is not well established, particularly in the dynamic phase of smoking, that is, the non-steady-state phase. Chamber and -field studies have indicated that, under realistic smoking conditions that would be encountered in residences or offices, the typical smoking and - ventilation rates would produce CO levels well within the levels observed in the outdoor air or in the indoor air generated from - the indoor sources, such as kerosene heater, gas stove, etc. Con- sequently, it is difficult to factor out the contribution of CO from ETS in any specific, uncontrolled situation. In areas -where heavy smoking is experienced, and where other sources of CO do not exist, CO may provide a rough measure of ETS exposure because the CO produced by the tobacco combustion will dominate. - Both chamber and field studies (Table 5-1) have demonstrated that tobacco combustion has s a major impact on the mass of sus- pended particulate matter in occupied spaces in the size range <2.5 pm, defined in this report as respirable suspended particu- lates (RSP). Suspended particulate mass is a major component of environmentally emitted tobacco smoke. Even under conditions - of low smoking rates, easily measurable increases in RSP have been recorded above background levels (Table 5-1). The term RSP, however, encompasses a broad range of particulates of vary- ing chemical composition and size emanating from a number of sources (outdoors, cooking indoors, etc). Smoking is not the only source of particulate matter sus- pended - - pended in the indoor air. The apportionment of the measured -- - RSP to tobacco combustion in an occupied space will not be ac- curate unless the RSP emission rates for a variety of brands of tobacco are similar under a variety of conditions and source use --- - information is obtained. The variability of RSP emissions into - the environment for a variety of brands of tobacco needs to be investigated, as does the relationship between the vapor and par- ticulate phases of tobacco-combustion emissions under a variety of environmental conditions, such as different humidities, and under a variety of smoking conditions, such as subject smokers versus smoking machines. 8C69944A

TAe1.E 54 iParticulate Levels Measured in ilndoor Environments, Including Smoking and Nonsmoking Occupancy Concentrations Type of Volume. ' Ventilation Wloniroaing Niean'(range). Study Ptemise Occupancy tn' Type/Rate Type/Tune µg/m3 Comments Brunckreef and 4 residences NS - N/- Boleij, 1982 7 tesidenees S = 1 - N/- 14 residences 'S' = 2 - N/- 1 tasidettee S= 3 - ~N/- Outdnots - - - Cutltlebaek 2 taverns S a'~4p - N.M/!1-6 a¢h en al.. 11976 I NS = 5-260 T :',10-300 Elliot and 3' an:nas NS „ - - Rowe. 11975 3 aeeoas S - M'/- T =~ 2.000- 14.277 Fnxt. 1984 1 schoui NS - M>- $ public S - N.M/- buiidings Hawthorne il residences NS ]50-674 M/0:18-0.96 et aL. '1984 8 resi'dences NS 150,674 M/0.26-I1'.98 2 residences S 150-,674 M/037-1:47 Lxadera 3'BubOe ' NS et nL, buildings personal 7 public 1:7-4 576 imotnmuni- I builtliop T - 2-6 btiolt 163=026 M/0.37-5.V 16"00 Mroa"i-;.5'Y G/2 mo 55 (20-90) G/2: me 125 (60-250> G'/2 mo 152 i(60-340) 0/2 mo 335 (-) 0/2 mo - (41-73) G/9'h 446 (233-986) G/24 h 55'(42=92) G/0.3' h 350'(148-620) P/= X (-) P/- 260 (40-660) QCMI/5-.15 min 9-40 (r) (otra 6,h) QC,%fl/5,15 min 12-46 1 (trner 6 h) QCM1/5=15 min '' 96-106 (o.er 6'h) G/4-21 b 17.80:1=32.2) G/2-24'h 205.1 (58452) TSP. repeat trteacuras 10.2, mg TSP: sensitivity TSP sensitivity TSP sensitivity TSP.emilstinn estimated TSP TSP TSP TSP RSP. .rirtter'humrtter- no'soutces RSP. .anter/summer- sourees;` RSP, winter/summer- rottras` +' eig. TSP, teperY nttutsutes, all Tar. Measured (160.0 peak) Moschandreas Outdoors Gn4h 17A,(-) RSP. TSP also et ai.. 11981 measured 2 ofRces G/24 h 16.8-20.2 (53 peak) RS'P. TSP also measutr:d 5 teddenaa NS N/0.5-1:3 tuh G/24 h 49.4-4s01 RS'P.' TSP aiso T = 2-6 (1'18.9;peait) rneasurtad . 5 tesitlencies S N/0.Sr1.3 ach iG/24 tt 36:9-99'.9 RSP: TSP also measured Nitschke !Otitdopes !G/168 h 111.3 t 6.0'(1-28) RSP et al.. i1985 1'9 tesitlencea NS 315-1,021 ; N/- G/1'68 h 26.0 t 22:6i(6-88) RSP, tspeat, measures. source mitt` 11 essiiiencea ' S 290-800 NY- G11'68h 59.2 ± 38:8i(,10-144) iRSP, nepestinreawrea, t+ource miu` Parker et ai., 1 tssidenoe NS - NY0:2-1 .9 ach 0/241t K 10 (-) TSP 1984 T=3 2 tssiden¢es S = 1-2 N'/0.2=0.7 ach 0/24 h ~ 10-46 () TSP T ='3-4 .~ Repaa and l.awreq. Outdoors - P/2 min 42.9 (22-63) RSP,:avarage of 2-min samples ea 1980. '1962 27 Public buildings 0.13-3.5y M/- P/2 min 278 (86-1,0+00) RSP. sMersge of 2-min', samples Sexton et al., Outdtwrs - - G/24 h 17.0 t I Lb (b-23) RSP,,tepeat sampks 1984 19 homes 24 residennes NS` N/- G/24i h ~ 25.0 ± 11 L0 (13-63) Used fireplaces Spen>)kr Outdoors - G/241h 21.11 t 111:9 (-) RSP, repeat tneatatras et al.. 1981 35 nsidenees NS N/- G/24; h 24.11 t '111:6 (-,) RSP, repeat tnessuies 15 residences S = 1 IN/- G/24 h 36.5 ± 14.5 RSP, repeat measures 5 ttsidences S = 2 ,N/- :G/24 h 70.4 ± 42.9 (-) RSP, repeat measures Spengler Outdoors - iN/- G/24 h 18 ± 2.1 (-) RSP, irepeat measures et al.. 1985 73;residencea NS iG/2'4 h 28 t :1.1 (-) RSP.repeartneasures 28 residences S G/24 h 74 ± 6.6 R9P, repeat measures Sterling and il,o({ice S;etinstr. ,G(?)/- 25.5 (15-36) . TSP Sterling. 1983 22 offkes S G(T)/- 31.7(-) TSP

I .1 A I I I ~ e I I 01 o C4 4 ry aa (} yi k v1 F- w m ~ 3 $o 74 1 1 a . 76 Nicotine exhibits many of the properties necessary to serve - - - as a potential marker for ETS. It is unique to tobacco smoke, is a major constituent of the smoke, and occurs in environmental concentrations that are easily measurable. It has been used as a marker for ETS in several studies (Table 2-5). The major problems - with using nicotine are: (a) the ratio of nicotine [recently found to be in vapor phase in ETS (Eudy et al., 1985)] to other ETS constituents (RSP, in particular) for a variety of brands of tobacco ' is not known, (b) the_ reactivity rate (removal rate) of nicotine relative to 'other ETS constituents is not known, (c) particulate- or or vapor-phase nicotine once deposited on surfaces may be re- emitted, and (d) until recently sampling methods for nicotine have not been efficient in collecting total nicotine (both vapor and particulate phase). Two new air-sampling methods for nicotine (Muramatsu et al., 1984; Hammond et al., in press) hold promise for obtaining total nicotine concentrations with the sensitivity and_ accuracy required for environmental air monitoring. A number of aromatic hydrocarbons (benzene, toluene, benzo-. [a]pyrene, pyrene, etc.) have been measured in field studies (Galuskinova, 1964; Just et al., 1972; Perry, 1973; Elliot and Rowe, 1975; Badre at al., 1978) investigating the impact of smoking on indoor air quality. Many of these air contaminants have other important sources, indoors and outdoors, that make measured levels difficult to interpret. Therefore, the aromatic hydrocarbons generally are poor indicators of ETS alone. Controlled chamber studies that elevate the variability of emission of the compounds - - - from a variety of brands of tobacco have not been carried out, and - - the ratios of these compohnds to categories of ETS contaminants (for instance, RSP) have not been established. Tobacco-specific nitrosamines and nitrogen oxides (Tables 2-6 and 2-9), acrolein and acetone (Tables 2-7 and 2-8), and polonium- 210 have been measured as indicators of ETS. The low environ- mental concentrations, existence of other sources, reactivity of the tracer contaminants, and lack of data on the ratios of these con-s taminants to ETS contaminants for a variety of brands of tobacco limit their usefulness as indicators of ETS in indoor spaces. Research efforts need to be directed toward identifying a tracer or proxy air contaminant for ETS that meets the four criteria out- lined above. At present, RSP is widely used as a general measure of ETS exposure indoors, particularly if the measurements are limited to locations where the levels of RSP from other sources. QV698449

76 are known and present at a low background concentration. The variability of RSP emissions for a number of brands of cigarettes, however, has yet to be evaluated. PERSONAL iklQNITORING Measurements of concentrations of air contaminants in the immediate breathing zone of an individual provide information on personal exposure. Personal monitoring can be accomplished with active samplers that integrate concentrations across a va- riety of locations or conditions using filters or vapor traps with subsequent laboratory analysis. Continuous portable monitoring instruments are available but have not been widely used. Particles are measured by light-ecattering principles or frequency change as mass is deposited on a vibrating quartz crystal. For the most - part, gases are measured using IR absorption or electrochemical reactions. Continuous-recording instruments have been utilized ---- itiore for characterizing microenvironments than for direct meat surcmcn4.g of personal exposures. Passive personal monitoring utilizes diffusion and permeation to concentrate gases on a e411ec- tion mediutn for subsequent laboratory analysis. 13oth active and passive monitors have been employed in assessing an ind•ividual's total exposure to individual or general categories of air contam- inants. A discussion of the type, application, and usefulness of passive monitors to assess air contaminant exposures can be found in Elliott and Rowe (1975) and Wallace and Ott (1982). A relatively small number of studies have utilized personal monitors to determine total exposures to ETS (Muramatsu et al., 1984; Schenker et al., 1984; Sexton et al., 1984; Spengler et al., 1985; Hammond et al., in press). In one study (Spengler et al., 1985) indoor (residential), outdoor, and personal 24-hour concentrations of RSP (measured in this study as particles with a 50% cut point of 3.5 µm) were obtained for a sample of 101 nonsmoking individuals living in Roane County, 2lennessee. In the sample, 28 of those monitored reported some exposure to ETS in either the home or workplace (nonindustrial), while 73 reported no such exposure. Each participant was sampled on 3 nonconsecutive days. Personal exposures tq respirable particles for the subgroup exposed to ETS and the subgroup not exposed to ETS are shown in Figure 5-1. Personal exposures to ItSP were dominated by indoor levels of ETS. Those reporting passive emoke exposure had 77 W 100 < g ~ ~ 40 20 . ---- N16Nn1 ....•... Pasowl, Non-3molc.-EYPowd - • ~ PusoeM. _ smok.-Expoeea 40 so 120 ieo 200 240 RESPIRABLE PARTICULATE CONCENTRATION (py/m3) FIGtURE 6-1 Qnmulative frequency distributions of RSP concentrations from central site ambient and personal monitoring of smoke-exposed and nonsmoke-exposed Individuals. Reprinted with permission from Spengler et al. (198b). mean personal respirable particulate levels 28 µg/m3 higher than those without passive smoke exposure. Particulate levels for those exposed to ETS showed a large variation, with approximately 25% of the personal samples having RSP levels in excess of the EPA - -- - ambient standard for outdoor total suspended particles. The EPA standard, however, includes particles up to approximately 50 µm and does not specify chemical composition. A direct comparison - - with the EPA standard requires a consideration of average time exposed as well as concentration. Sexton et al. (1984) conducted U-hour personal monitoring for RSP for 48 nonsmoking individuals for 24 different residences. - - Samples were collected every other day, for a 2-week period, during a heating season in Vermont. Those individuals reporting exposure - to ETS for more than 2 hours per day had RSP levels 18.4 µg/m' - higher than those who reported no exposure (5Q.1 pg/ms versus 81.7pg/mS). - In demonstrating a new method for the collection and analy- sis of nicotine in air, one study (Muramatsu et al., 1984) obtained personal-monitoring samples of nicotine for one nonsmoker in 53 --- nonindustrial indoor microenvironments, including offices, houses, - - restaurants, cnrs, buses, etc. The samples were collected over a ---- 1-hour to 8-hour time period in each space and were specific for nicotine. A wide range of nicotine concentrations were reported, Tv698Lfa9

78 from 1.76 µg/m3 in a laboratory to 83.13 pg/in3 in a car. It is dif_ ficult to interpret these results in terms of an integrated exposure for a large segment of the population, since the sampling scheme - did not explicitly provide population estimates of exposure-that is, personal samples were obtained on a microenvironment basis for only one individual. Lacking good data for the ratio of total nicotine to RSP in ETS, it is difficult to estimate the RSP exposure levels. The data, however, do demonstrate that the variability of nicotine concentrations and the occurrence of high concentrations of other ETS components catr be found in various microenviron- ments. In an epidemiologic study of the health effects of diesel ex- haust on railroad workers (Schenker et al., 1984), which included a control group of railroad office workers who were not exposed to diesel exhaust but were exposed to ETS, E'1'S was recognized as an important component of the respirable particulate exposures. Hammond et al. (in press) used a newly developed air-sampling and analytical method for measuring total nicotine in collected RSP personal samples to determine the contribution of ETS to the RSP levels measured. Their results indicated that the major portion of the office workers' RSP exposure is due to ETS. Ratios of nicotine to RSP for a variety of brands of tobacco need to be established before absolute ETS exposures can be assessed. Personal monitoring can provide a useful measure of an indi-s vidual's exposure to an air contaminant or class of contaminants over a period of several hours to several days. The usefulness of personal monitors for assessing ETS exposure would be greatly enhanced if the personal monitor were passive in nature and in- expensive. Personal and portable monitors, however, need to be evaluated to determine their usefulness in establishing ETS ex- posures associated with long-term adverse health outcomes, such as cancer. They may be useful in establishing ETS exposures, in a background of confounding air contaminants, associated with short-term effects. A variety of sample collection and analysis methods has been used to monitor individual constituents ancl categories of contam- inants found in ETS for both personal monitoring and air mon- itoring of spaces. While this report does not offer a review and evaluation of the mo- nitoring methods that have been employed and are available, it should be clearly noted that the specificity and sensitivity of the measurement method must be evaluated to 79 assess the uncertainties in the measured concentrations. The con- - -- stituents of ETS will exhibit a pronounced spatial and temporal distribution in an indoor space and among indoor environments, - - due to variations in smoking rates and building characteristics. - In interpreting measured concentrations of ETS constituents, one - - must recognize the potential for pronounced spatial and temporal - variations. CONCENTRATIONS OF ' ENVIRONMENTAL TOBACCO SMOKE IN INDOOR ENVIRONMENTS Various Environmental Tobacco Smoke Constituents There is a sizable body of literature reporting on measure- ments of various constituents (acrolein, aromatic hydrocarbons, carbon monoxide, nicotine, etc.) of ETS in a variety of microen- -- - yironments. These studies have reported a wide range of con- centrations of ETS-related air contaminants under conditions of normal space use (Tables 2-4 to 2-9). However, the majority of the measurements are of very limited use in assessing actual human exposures to ETS for a large segment of the population for the following reasons: • the representativeness of the air contaminant measured to the total ETS in the space is unknown; a the proxy air contaminant measured may have a variety of other potential sources that were not accounted or controlled for; • data were not collected on smoking rates or numbers of - - smokers; and • important building characteristics such as infiltration or volume were not recorded. While these studies have indicated the range of concentrations of several ETS-ralated air contaminants that can be found indoors, they do not provide a sufficient basis_ o_ _n_ which ETS indoor expo- sure estimates can be made. Particulate Levels and Smoking Occupancy The most extefisive and suitable data base for modeling ETS is the RSP (<2.5 µm) associated with ETS. This RSP comprises ZV698Z449

80 o ~ ©utdoor • r . Indoor, no smokers ~ib i • Indoor, 1 saaker '++ 110- •• tndoor, > 1 smoker ~. J - ~ 100 • ~ E ~ ~eo `/ y! Q ° ~ ~ d < SO ~~ ~ Z 10 ac 3o - ~ Z w o 20 2 V to oL - Nov 6.c Jor FEb Mar Aw Mny ,An. ,Al7W+9 S.P. Oct Mow D*e Jan.F.D. Mur.Apf. 1976 1977 1978 FIGURE 6-2 Monthly mean RSP eoncentrations fn .ix U.S. citie.. Ra printed with permission from Spengler (1981). a major general category of ETS contaminants aud is produced in concentrations that are easily measured in occupied spaces where smoking occurs. In a survey of more than 80 homes in six U.S. cities (Spengler et al., 1981), 24-hour gravimetric samples of RSP were collected every sixth day for up to 2 years in stationary samplers in each home and outdoors. The resulting data (monthly RSP means) aggregated by the number of smokers are shown in Table 5-1 and Figure 5-2. Homes without smokers exhibited RSP levels roughly equal to outdoor levels and followed outdoor trends. The presence of just one smoker in a home had a pronounced impact on RSP levels. Using regression analysis, the authors estimated that the impact on overall average RSP levels in a residence of a pack-per- day smoker was approximately 20 µg/0. The impact of smoking in a home with central air conditioning was effectively doubled, presumably due to reduced air exchange. Table 5-1 presents the range of RSP levels measured in a variety of indoor microenvironments for smoking and nonQmok- ing occupancies. It also indicates whether direct measures of the variables necessary for the model outlined in Equation 5-1 below, or necessary to explain the RSP levels meaeured, were recorded. These variables include, among others, vt:ntilation, mixing, re- moval by surfaces, and smoking occupancy. Outdoor levels of 81 RSP or total suspended particulates are generally lower than or equal to indoor levels in homes without smoking. Smoking occupancy is strongly associated with elevated levels of RSP in a variety of indoor microenvironments at levels well above outdoor concentrations and indoor concentrations where there is nonsmoking occupancy. However, few studies have directly recorded the data on the parameters that are necessary to validate models for predicting RSP levels due to smoking'occupancy (see section below). Even so, using a number of assumptions, data in Table 5-1 have been used for model validation by some studies (Repace and Lowrey, 1980, 1982). ' MODELING The process of assessing exposure and attributing it to vari- ous microenvironments requires knowledge of the time-individuals spend in such microenvironments and the typical air-contaminant levels (average and peak) occurring in them. The nature of the health or comfort effect under study determines the time-average concentration of concern. A number of microenvironments have been identified (Moschandreas, 1981), and time-budget surveys have shown that most individuals spend more than 00% of their time indoors, that is, in residential, industrial, and nonindustrial occupational environments (Szalai, 1972). The,indoor residential and nonindustrial occupational environments represent the major microenvironmenta in which exposure to ETS takes place. Tobacco combustion is a major source category that affects the quality of the air indoors. The air-contaminant concentra- tions in an enclosed space resulting from tobacco combustion, - - and hence human exposures, are the result of a complex interac- tion of several interrelated variables (Figure 4-l), including source air-contaminant emission characteristics and source use, building characteristics, infiltration or ventilation rates, air mixing, loss terms (removal by surfaces or chemical transformations), and the efIiciency, of air-cleaning equipment. The interaction of these vari- ables in determining the resultant indoor concentrations of ETS has typically been evaluated in both controlled laboratory (cham- ber and test house) studies and field studies within the theoretical framework of the general mass-balance equation (Turk,1963; Shair and Heitner, 1974; Eamen, 1978; Ishizu, 1980; Repace and Lowrey, 1980; Leaderer et al., 1984). CK9844.9

82 The mass-balance equation may be applied to tobacco smoke as either an equilibrium model (time-indepondent) or a dynamic model (time-dependent). Equilibrium models rely on the assump- tion that many of the input parameters-such as source rates, removal or loss rates, and ventilation rates-are constant, even though these parameters in actuality may vary considerably in time._ These models are useful in developing air-contaminant emis- sion factors for ETS in controlled laboratory studies and in as- sessing long-term average exposures in given indoor microenviron- mente. Dynamic models are usually more flexible than equilibrium models and can provide information on short-term concentrations. They may be used to compare the sensitivity of results to varia- tione of input parameters. Equilibrium models, when applied to field studies of ETS, require average information on the impact variables, while dynamic models require a great amount of de- tailed information obtained as a function of time. Dynamic and equilibrium models are useful in laboratory studies; equilibrium models are best suited to evaluating and predicting ETS concen- trations in field studies, particularly when average concentrations over a period of days or longer are of interest. Equilibrium Models for It.SP Laboratory and field studies typically utilize some form of a single-compartment equilibrium model to evaluate the input parameters to the mass-balance equation, to evaluate field-study data, and to project RSP concentrations from ETS indoors. These studies have reduced the general single-compartment mass-balance d f lif orm. ie equation to the following simp a •, s f'!i(R~ '}' Ro) V (5-1) where C,q is the equilibrium concentration of RSP in a sllace expressed as micrograms per cubic metcr (µg/me) due to ETS, G is the RSP generation rate from, tobacco combustion into the space in micrograms per hour (pg/hour), a, is the ventilation or filtration rate in air changes per hour (ach), n, is the loss rate of RSP due to surface removal in a space in air changes per hour, V is the volume of the space in cubic meters (0), and rn is the mixing rate expressed as a fraction. The above model assumes no air-cleaning devices, either in the space or recirculated air. 83 Under laboratory conditions, the input parameters rs can be controlled and evaluated. In conductin field studies or in esti- - - mating -- - mating past itSP levels indoors, the values on the right side of - -- -- - - - Equation 5-2 have to be determined from available data. It should - - -- --- -- - be emphasized that this equation assumes equilibrium 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. Generation Rate The generation rate of RSP for ETS is a function of the number of cigarettes smoked and the emission rate of RSP per cigarette. Few studies have investigated the RSP emission rate for SS plus exhaled MS, i.e., contributions to ETS. One recent study (Rick- ert et al., 1984) examined sidestream and mainstream ernisdions of tar, nicotine, and carbon monoxide hi 15 brands of Canadian cigarettes with a range of advertised mainstream tar deliveries (0.7 to 17.0 mg tar/cigarette). The experiments utilized a single- port smoking machine and collected mainstream emissions and sidestream emissions, from a small chamber, onto Cambridge fi1- tere. The subsequently measured sidestream emissions of tar were found to average 24.1 mg/cigarette with a range of 15.8 to 36.0 mg/cigarette. These emissions were independent of mainstream emissions, which averaged 11.4 mg of tar per cigarette with a range of 2.5 to 19.4 mg/cigarette. Sidestream emissions were higher for ventilated branda. RSP emission rates were developed for 10 brands of U.S. cigarettes with rated tar deliveries from 1.0 to 23.0 mg and for one standard cigarette (University of Kentucky #1RSF). The study (B.P. Leaderer, S.K. Hammond, and T. Tosun, personal commu- nication) utilized a 34-ms charnber in which the cigarettes_ were smoked by occupants at a prescribed rate in an effort to create realistic environmental conditions. RSP measurements easurements were made over a 4-hour period during equilibrium conditions via collection of well-mixed room air on filters with subsequent gravimetric anal- ysis. ysis. RSP emission levels were found to range from 18.0 to 35.4 mg/cigarette, with an average of 26.9 * 4.8 mg/cigarette. g/cigarette. Three brands of British cigarettes (very low tar, 1.5 mg; low tar, 12 mg; and medium tar, 18 mg) were evaluated for both mainstream wid sidestream emissions of tar (U.K. Government, VVG99448

84 1980). The average RSP sidestream emission rate measured was 24.5 mg/cigarette, with a range of 23.0 to 26.4 tug. This study utilized a small test chamber and a smoking machine. The emission by-products were collected onto Cambridge filters. Only one study (Hoegg, 1972) rreports mainstream and aida stream RSP levels for cigarettes prior to 1970. This test chamber study reported a sidestream particle emission rate of 25.8 mg/cig-e arette and mainstream particle emission rate of 36.2 mg/cigarette for one brand of cigarette. Current data would suggest an overall RSI' emission rate from ETS in the range of approximately 20 to 36 mg tar/cigarette. An accurate estimate of an average emission rate for modeling purposes requires the weighing of the above emission data by the sales-weighted average cigarette brand sold. Data are not available to estimate the historical trend, if any, in the RSP emissions for ETS. Equation 5-1 assumes a constant, or near constant, source emitting over a sufficiently long time period to reach and maintain equilibrium conditions. In controlled experiments a constant rate of tobacco combustion is easily obtained. In practice, however, tobacco combustion rates in terms of numbors of cigarettes con- sumed over some period of time in different indoor environments is sumed variable. In the absence of detailed data on cigarette consumption in a space, such as number of cigarettes smoked, time smoked, to- tal weight of tobacco consumed, etc., estimates are required. For example, one smoker in a household smoking at a national average rate of two cigarettes/hour and 10 minutes/cigarette constitutes an intermittent source [G(t)dt]. A continuous source would be the smoking of six cigarettes/hour. Using a 26-mg/cigarette emie- sion rate, the estimated total RSP emissions from the intermittent source, i.e., 52 mg/hour, would be represented as being emitted uniformly over a 1-hour period for the full averaging time consid- ered. In large occupied spaces where smoking is permitted, such as nonindustrial occupational environments, estimates (Bridge and Corn, 1972; Jaffe, 1978; Repace and Lowrey, 1980) would indicate that, at any given time, 11% of the population would be smoking (one-third of the U.S. population are smokers, who are smoking at the rate of two cigarettes/hour and 10 minutes/cigarette). This would constitute a continous source. In practice, the smoking rate is probably highly variable in time. The RiP emissions from this I 85 example would also be averaged over time to produce uniform av- erage emission rates per hour. For the estimat_ion of equilibrium conditions in Equation 5-1, G= NE, where N is equal to the number of cigarettes consumed per hour in a space and E is the number of milligrams of RSP emitted into the environment per - - cigarette. The assumption of a eontinuous source introduces errors in the estimated RSP concentration. It is also important to note that the equilibrium model assumes that the source will be emitting contaminants over a sufficiently long period of time to achieve balance with the removal mecha- nisms (ventilation, removal by surfaces, and air cleaning). If G', is the concentration at time t (in hours) then: Oe -= Ceq(1- e-etv). (5-2) where G,q is the equilibrium concentration and N is the effective removal rate (N = n -- .+ a,). If the total impact of the removal rate, i.e., ventilation plus loss to surfaces, is equivalent to one ach, 85% of the equilibrium concentration would be obtained in 2 hours. Thus, to the extent that the source emissions do not combine over long periods of time, the equilibrium concentration will not be reached and maintained, introducing errors into the estunated or modeled RSP. VentilationlInfiltration The supply of fresh air to a space by ventilation (air supplied by mechanical systems) or by infiltration (uncontrolled movement of air through cracks and unintentional openings in the building envelope) serves to reduce the levels of air contaminants generated by an indoor source. Building codes adopted and enforced by local, state, and fed- eral government agencies generally specify minimal acceptable ventilation criteria to be maintained in buildings. These codes are usually derived from standards that have been promulgated by authoritative bodies (American National Standards Institute, American Society of Heating, Refrigerating and Air-Conditioning Engineers, etc.). These standards are usually developed by con- sensus and are generally voluntary until adopted by municipal or state governments. The studies of Yaglou et al. (1936) formed the basis of minimum required ventilation rates that persisted un- til 1979. These studies reported ventilation rates (fresh odor-free s'V69S44Q

86 air) in cubic feet per minute per person (c€rn/person) necessary to provide an acceptable odor environment. They found that, as occupant density increased, so did the required cGn/pefeon venti- lation rate. Because of the odor level, smoking occupancy required significantly more ventilation air. A minimurn ventilation rate of approximately 10 cfm/person for nonsmoking occupancy and an occupant density of 400 ft3/person was recommended. Prior to 1936, minimum recommended ventilation rates were as high as 30 cfm/person. In 1973, the American Society of Heating, Itefrigerating and Air-Conditioning Engineers (ASHRAE) adopted and published .$tandard 62-73 (Standards for Natural and Mechanical Venti- lation). This standard recommended ventilation rates for vari- ous residential and commercial spaces on an occupancy-density basis. In 1981, ASHRAE adopted Standard 62-81 (Ventilation for Acceptable Indoor Air Quality), which also recommended ventilation rates for various residential and commercial spaces on an occupancy-density basis but distinguished between smok- ing and nonsmoking occupancy. Modal ventilation rates in this standard equaled 36 cfm/occupant for smoking occupancy and 7 - cftn/occupant for nonsmoking occupancy. As noted in Equation 6-1, ventilation rates are incorporated as ach. ASHRAE 62-q3 recommended from 15 to 25 cfm/person for general office space with an estimated 10 pereone/1,000 ft2 density, while ASHRAE 62-81 recommended 20 cfrn/pereon when smoking is permitted and a 5-cfm/person'minimum for nonsmoking occu- pancy with an estimated 7 persons/1,000 [t2 density occupancy. Assuming full occupancy and an 8-ft ceiling, ASHRAE 62-73 ven- tilation rate ranges are 1.13 and 1.9 ach, while ASIIRAE 62-81 recommends a rate of 1.3 ash for smoking occupancy and 0.26 ach for nonsmoking occupancy. When considered on a epace-by-apace basis for commercial or residential environments as recommended by either ASHRAE 6273 or ASHRAE 62-81, the ach rates vary considerably, depending on the use of the space and whether smok- ing is permitted. In estimating ach's for inputs into Equation 6-1, to assess either current or past RSP concentrations in occupied space due to ETS, the following points should be keht in mind: 9VG98449 87 . There are no data to indicate the current or past distribu- tion of ach's currently are or have been in a variety of commercial spaces in which smoking is or has been permitted. • Air-exchange rates are calculated from cfm/pereon rates specified in the standards for full occupancy. To the extent that occupancy is less than or greater than the designed figure, the cfm/person could be significantly different. • Ventilation codes are equivalent to design standards. In ac- tual - tual practice, the heating, ventilation, and air-conditioning - (HVAC) system may not operate as designed. Interior alterations, - --- - - modifications in occupancy, maintenance, and repair of equipment, and operator practice can significantly affect the performance of the HVAC system. Air-infiltration values~ in housing are induced by differences in pressure across the structure envelope. Limited data exist to -- - indicate what the current or past distribution of air-exchange - rates in houses in the United States are or what the intra- or -- --- - - interseasonal variations are. One study of seasonal infiltration rates of 312 houses in North America (Grirnrsud et al., 1982; Figure 5-3) found a median value of 0.5 ach. This study was based --- - on new energy-efficient houses. Another study (Grot and Clark, 1979; Figure 5-3) of 266 low-income houses in North America -- -- - - - found the median seasonal air-exchange rate to be 0.9 ach. Air- infiltration rates for both studies were taken without occupants in - --- the houses. Normal activities of occupants would add an average 0.10 to 0.15 ach to the values reported in these two studies. Ventilation or infiltration rates in commercial and residential buildings can vary by an order of magnitude among and within buildings, season to season and within a season. Unfortunately, there are few data available that would allow for an accurate estimate of the distribution of air-exchange rates in commercial and residential spaces currently or over the past several years. A range of 0.4 to 1.6 seh would seem reasonable. Rensoval by Surfaces Next to ventilation, the major mechanism for removal of sus- pended particulate matter is surface deposition. Surface deposi- tion of particles indoors is a function of several variables, including particle size and composition,, temperature, humidity, type and quantity of surface material in a room, surface-to-volume rates,

88 d 0.5 1.0 1.5 2.0 2.5 3.0 INFILTRATION RATES h:h) 3.5 0.5 1.0 ' 1.5 2.0 2.5 3.0 3.6 INFILTRATION RATES (ach) FIaURE 6-S Histograms of Infiltration values In two different samples of North American housing. (a) Average seasonal inRlitratiou of 312 recently constructed houses; the median value 4 0.6 air changes per hour (adA). Reprinted with permission from Qrimsrud.t al. (1082). (b) Average seasonal Infiltration of 266 older lorr-income houses; the median is 0.9 adh.-Reprinted - with permiasion of Grot and Clark (1979). and turbulence. In laboratory studies under conditions of ideal mixing, surface-deposition rates (h'1) for ETS were found to vary from an equivalent 0.1 h°1 to 1.8 h'1 (Leaderer et al., 1986). The greater the degree of turbulence introduced into the chamber and the higher the surface-to-volume ratio, the higher the surface deposition. One recent chamber study evaluated the importance of ina- terials (rugs, wall paper, and painted wall board), surface area of 89 materials, temperature, humidity, and turbulence on the deposi- tion rate of RSP generated by tobacco combustion under condi- tions tions of ideal mixing (Leaderer et al., 1986). In these experiments, the deposition rate of RSP was determined by monitoring the de- cay of RSP and carbon dioxidq (injected into the chamber) under the experimental conditions examined. The difference between the RSP and carbon dioxide (nonreactive tracer gas) represented the RSP deposition rate. This study found the most pronounced impact on deposition in this chamber to be the air-recirculation rate (fresh-air-ventilation rate held constant) or turbulence. T%e type and quantity of material, temperature, and humidity were also found to impact particle deposition in a significant way. The results of this study indicate that a particle deposition rate of 0.2 to 0.8 h'1 for ideal mixing might typically be encountered in occupied spaces. Mizing Once released into an enclosed space, air contaminants move through it by dispersion. Dispersion determines where the high and low concentrations of the contaminant will occur. Dispersion - is controlled by diffusion, which is the movement from areas of high to low concentrations, and by mixing, which is the movement - of air in the space. When ideal mixing occurs in a space, i.e., m= 1 in Equation 5-1, no spatial gradient of an air contaminant like 1ZSP exists, and the full effectiveness of ventilation and sink rates in removing the contaminant is seen. In controlled laboratory studies, ideal mixing is easily obtairied through the use of mixing fans or the rapid recirculation of the air. In occupied spaces, however, ideal mixing is hardly ever obtained unless a great deal of turbulence is introduced to the space and the supply- and exhaust-air system is carefully designed. Less than ideal mixing - can result in a pronounced concentration gradient of a contaminant in the space. 'rhe ventilation rates and removal by surfaces under those conditions are not as effective in lowering air-contaminant levels. The mixing term is usually defined as the ratio of effective - ventilation to theoretical ventilation. In an occupied space, the value of the mixing factor is affected by the source and its use, room geometry, air supply and exhaust design, air-flow rates, obstacles in a room, and activity of the occupants. In addition, the mixing factor is specific for a precise 4bG9e44g

91 location. No data exist that would indicate the distribution of the - - values for the mixing factor in occupied spaces. A limited number of studies show a range of mixing factors from 0.1 to near 1.0 (Brief, 1960; Kasuda, 1976). Volume The volume of the space in which smoking occurs is highly variable. It can range from a few cubic meters in a car to several ,thousand cubic meters in large auditoriums or sports arenas. The highest RSP levels from ETS will tend to occur in smaller spaces with high smoking rates. Predicting Environmental Tobacco Smoke Exposures from Tobacco Combustion Utilizing Equation &1, expected RSP concentrations indoors from ETS can be estimated for a range of the input parameters that realistically can be expected under normal smoking occupancy. Figures 5-4 and 5-5 allow for the easy calculation of RSP levels • due to ETS as a function of emoking rate, ventilation, sink rate, mixing, and volume of the space (see the example outlined in the legends of these figures). The calculations treat the spaces of concern (e.g., multiroom home or a single room in a house) -- - as a single compartment with no air-cleaning devices. The total amount of RSP (in milligrams) from ETS can be determined in Figure 5-5 as a function of the smoking rate and effective removal rate (N). The removal rate is equal to the sum of ventilation plus removal by surfaces times the mixing factur.. `l'he total amount of RSP calculated from Figure 5-4 is then eutered into Figure fi-6 to determine the RSP concentrations (in micrograms per cubic - meter) expected for a given volume of space. The calculations used to generate these figures assume that: • an average total RSP emission rate is 26 mg/cigarette, • the emissions are nearly consistent and averaged over a 1-h - period, e near steady-state or equilibrium conditions are reeched quickly, • no air-cleaning devices are in use, • background levels of RSP in a space are zero, and 3Qoo ~oea sooo :'N~\ \',. \'.\',_ .i 11 t0o 7o 50 30 20 a ~ N 7 480 360 240 Ieo 120 90 60 48 is 24 Is 12 6 4 2 I • • 02 _ 03 05 0.r i 2 33 Tp N (air cAonqes per hour) FIQURE 6-4 Uiagram for calculating the RPS mass from ETS emitted into any occupied space as a function of the smoking rate and_ removal rate (N). The removal rate is equal to the sum of the ventilation or Infiltration rate (n,) and removal rate by surfaces (n.) times the mixing factor m. The calculated ETS RSP mass determined from this figure serves as an input to Figure 6-6 to determine the ETS respirable suspended particulate mass - concentration in any space in µg/ms. Smoking rates (diagonal lines) are given as cigarettes smoked per hour. Mixing is determined as a fraction and n. and n, are In air changes per hour (od). All three parameters have to be .estimated or measured. Calculations were made using the equiiibrium form of the mau-baiance equation __ (Equation 6-3) and assume a fixed emission rate of 26 mg/cigarette of RSP. Shaded area shows the range of RSP emissions that could be expected for a residence rrith one smoker smoking at a rate of either 1 or 2 cigarettes per hour for r the raege of mixing, ventilation, and removal rates occurring in residences under steady-state conditions. f RVe9sZ,c,R

92 0 ~ 3 5 10 20 50 Ifq 200 509 1000 30~9 I0U00 Totol RSP Emitted (mq) FIGURE 5-5 Diagram to calculate the ETS RSP concentration ta n apace as a function of the total mass of ETS RSP emitted (determined from Figure 6-4) and the volume of a space (diagonallines). The concentrations shown assume a background level In the space of sero. The particulate concentrations shown are estimates during smoking occupancy. The dashed horisontal lines (A, B, f), and D) refer to Atational Ambient Air Quality Standards (health-related) for total suspended particulates established by the U.S. Environmental Protection Agency. A ts the annual geometric mean. B is the 24-hour value not to be exceeded more than once a year. 0 Is the 24-hour air pollution emergency level. D is the 24-hour signiHcant harm level. Shaded area shows the range of concentrations expected (from Figure 5-4) for a range of typical volumes of U.S. residences and rooms Ia thae residences. 0 a one-chamber model is appropriate. In these f gures, the RSP-emission rate is assumed constant. lf, in fact, this rate is variable, then the predicted RSP level will also vary. As already discussed, the input parameters to Equation 5-1 are known to vary greatly under realistic occupancy conditione, with few or no data available on the distribution of the values of those input parameters. Figures 5-4 and 5-5 highlight the large effect that small variations in the input parameters can have on the predicted RSP concentration. 93 For example, for a range of conditions that can be expected to be encountered in private residences with one smoker (shaded areas in Figures &4 and 5-5), RSP levels in residences or public places during smoking occupancy can vary by more than two or- ders of magnitude from approximately 17 to 5,000 µg f ine. This example assumes one smoking resident smoking at a rate of either one or two cigarettes/hour. Relatively easy-to-obtain information on some of the input parameters, such as building volume or typical - - smoking densities, obtained through questionnaires or observation, can serve to significantly reduce the range of estimated exposures. It is also clear from Figures 5-4 and 5-5 that, for the vast majority of conditions, RSP levels due solely to ETS can be expected to - equal or exceed levels specified in National Ambient Air Quality Standards for the total suspended particulates (Code of Federal -- Regulations, 1985). These standards are health-based and reflect - --- different averaging times as well as levels of exposure. Direct com- parison of exposures with the standards requires consideration of particle size, concentration, and time. The physical and chemical nature of the particulate matter resulting from tobacco smoke is different from particulate matter observed outdoors in ambient air. These different particulate matters no doubt have different - biological effects. Therefore, direct comparisons of exposures to outdoor standards should be made with caution. Application of Respirable Suspended Particulates Model - The most extensive uee_ of the mass-balance equation for as- sessing RSP levels due to ETS in occupied spaces has been by Repace and Lowrey (1980). Drawing upon the best available data from several sources, including both measured and estimated pa- rametera, they proposed and applied in field observations a con- densed version of the mass-balance ass-.balance equation for estimating RSP exposures due to ETS TS in a variety of indoor micro_ e_ nvironme_ nts._ Their model is C',e = 866D. + n., (5-3) where G`,, is the equilibrium of concentration of RSP due to ETS - expressed. in micrograma per cubic meter, D, is the density of active smokers expressed as units of burning cigarettes observed in a space per 100 m3 over the sampling time frames, and n, is the 6VG98UI8

94 ventilation or infiltration rate in ach. The constant term (650) is calculated from a standard set of assumed conditions for smoking ,.. rates, RSP emission rates, mixing factors, ventilation rates, and sink rates. These standard sets of conditions are derived largely from experimental data and building standards. Although many of the input parameters were estimated from the literature, which is based on limited experimental data, Repace and Lowrey (1980, 1982) applied Equation 5-3, or similar equa- tions, to a variety of situations and found that they produced reasonably accurate estimates in a limited number of occupied spaces with smoking occupancy. Apparently, easy-to-obtain data on building volumes, design occupancy, smoking occupancy, type of ventilation systems, and building standards can improve the prediction of RSP concentrations. In using Equation 5-3, the ma- - _ -- jor assumptions deal with mixing, ventilation rates, and sink rates. Additional8eld testing of the Repace and Lowrey model, as well as a better understanding of the variability of the input parameters, either estimated or measured for use in Equation 5-3, is needed. SUMMARY AND RECOMMENDATIONS In investigating the adverse health and comfort impact of air - - contaminants, it is important to specify the exposure to a specific air contaminant or a class of air contaminants on the time scale - --- - - corresponding to the health or comfort effect being evaluated. Ac_- -- - curate data on exposure is essential to minimize inisclassiRcation of exposure in epidemiologic studies of air contaminants. In the absence of an indicator of the dose of the contaminant, target tis- sue exposures may be estimated by use of biological markers, by personal monitors, or by the air monitoring of rnicroenYironmente in which people spend time combined'with Lime activity patterns. ETS is comprised of several thousand chemicals in both the gas and particulate phases. While several individual constituents of ETS have_ been measured in a number of microenvironments as a proxy for ETS (nicotine, CO, acrolein, etc.), none have met all of the criteria necessary for a suitable proxy, nor has an individual - - contaminant been uniformly accepted or recognized as represent- ing ing ETS exposure. New methods of measuring nicotine in air hold promise for using nicotine as a suitable proxy for ETS, but consid- erable development and testing need to be done. The single largest component of ETS by weight is the RSP, which refers to particles 95 less than 2.6 pm and is highly variable in chemical composition. The RSP fraction of ETS is currently the best and most-utilized general category of air contaminants to represent ETS exposure. A limited number of studies that employed personal monitors to measure total RSP found that individuals who reported being exposed to ETS were exposed to RSP concentrations consistently greater than those whb reported no exposure. Furthermore, the distribution of RSP concentrations varied widely (from 10 Ng/m3 to more than 200 µg/ms). The limited number of samples, lack of data on the environments where the exposure took place, and lack of a specific proxy for ETS do not permit accurate estimation of the ETS exposure or extension of the data to a larger population. They do indicate, however, that individuals who report exposure to ETS will have greater RSP exposures than those who do not. Measurements of RSP levels in various indoor environments (residences, offices, restaurants, bars, bowling alleys, airplanes, arenas, etc.) have clearly shown that RSP levels will be consider- ably above background levels (outdoor levels or nonsmoking levels) - when smoking is reported in the space. Modeling of RSP concentrations due to ETS in any indoor environment usually utilizes a simplified form of the mass-balance equation. These models are typically single-chamber models that assume steady-state or equilibrium conditions to estimate RSP levels and require as input parameters an RSP-emission rate for tobacco combustion, number of cigarettes consumed, ventilation or infiltration rates, removal rates by surfaces, air mixing in the -- space, and volume of the space. Information on the current or past - distribution of these input parameters in the range of microenvi- ronments in which individuals spend the majority of their time - - (residences, offices, etc.) is not available. Tho variability of one or more of the input parameters can make a difference of as much as an order of magnitude in the estimated RSP concentration. A....l itional variability in the estimated RSP levels is introduced to the extent that the equilibrium assumptions do not hold_ (i.e., an intermittent rather than continuous source). Gathering data on easily measured input parameters such as smoking rates or volume can substantially reduce the variability of the estimated RSP levels. Limited field tests of the general -- equilibrium model, in which some of the input parameters were measured and others were estimated either from chamber studies or building codes, have predicted RSP levels reasonably well over a

96 wide range of values of input parameters. While the predicted level of RSP exposure due to ETS may be highly variable using models, it is clear from the models that, even using the most conservative estimates for the input parameters, RSP levels when smoking is allowed will result in substantial increases over nonsmoking occupancy RSP levels. This is consistent with the concentrations measured through personal monitoring or area monitors in various microenvironments. What Is Ifnown 1. Various individual chemical constituents of ETS have -- - been measured in indoor spaces as proxies for ETS, but their suit- ability as proxies for ETS exposures has not been well established. 2. The total RSP, measured by personal monitors, has been found to be elevated for individuals who reported being exposed to ETS as compared with those who reported no exposure. S. The distributions of RSP measured by personal monitors and by portable monitors vary widely. However, levels of RSP measured in various indoor environments have clearly shown that RSP levels will be considerably above background levels when smoking is.reported in the space. 4. Limited field tests of the mass-balance, general-equilib- rium model in which some of the input parameters are measured and others are estimated have predicted RSP levels reasonably well over a wide range of values of input parameters. What Scientific Information Is Missing 1. There is a lack of data on the environments where mear - surements have been taken. Consequently, an accurate estimate of the ETS exposure or extension of the data to a large population based upon present data may not be possible. 2. A suitable proxy or tracer air contaminant is not avail- able for total ETS exposure. Nicotine may be a good indicator for exposure to the vapor phase. However, the relative proportions of various constituents of ETS in the particulate and vapor phases need further study to determine the extent to which a tracer for one phase can be used to infer exposure to the other phase. 3. Information on current or past distributiens of the input parameters for the ntass-.balance models of RSP concentrations is 97 not available for a range of microenviro_ nments in whic_ h_ individuals spend the mqjority oUtheir time. - 4. When levels of various constituents of ETS are measured in field situations, data should be gathered on input parameters such ae smoking rates or volume so that_ a detailed field_ evaluation of the equilibrium model can be made. 5. ETS exposure in epidemiologic studies needs to be im- proved. Questionnaires must be validated. Personal and microen- vironmental monitoring studies should be conducted to determine the predictive value of various exposure assessment methodolo- gies. This might be achieved as parE of a nested design in a larger epidemiologic study. 6. `1'he variability of RSP emissions into the environment - and the relationship between vapor apor and particulate phases need -- to b i ti t d f e nves ga e or a variety of brands of tobacco. REFERENCES I AS_HRAE Standards 62-1973. Standards for Natural and Mechanical Venti- lation. Atlantas ASHRAE, 1973. ASHRAE Standard. 69-1981. Ventilation for Acceptable Indoor Air Quality. Atlanta: ASHRAE, 1981.19 pp. - Badre, lt., R. Cnillerme, N. Abram, M. Bourdin, and O. Dumas. Pollution atmospherique par la fum6e do tabac. Ann. Pharm. Fr. 3g:ee3.e6F, 1978. Brief, R.S. S_ imple way to determine air contaminants. Air Eng. Z:39-44, 1960. Bridge, D.P., and M. Corn. Contribution tion to the assessment of nonsmokers to air pollution from cigarette and cigar amoke In occupied spaces. Environ. ltes. 6:19Y-Z09, 1972. Brunekreef, B., and J.S.M. Boleij.. Long-term_ average suspended particulate concentrations in smokers' home.. Int. /1rch. Occup. Enxiron. Health 50:299-302, 1982. Olausen, 0., W.S. Cain, P.O. Fanger, and B.P. Leaderer. The in8uence of aging, particle filtration and humidity on tobacco smoke odor, pp. 3l6-360. In P.O. Fanger, Ed. OLIMA 2000. VVol. _1. 4. Indoor C_ limate._ Copenhagen: VVS Kongress-VVS Messs, 1986. Qode ofFederal Regulations (CFR). National primary and secondary ambient air quality standards. Code Fed. Regul. 40(PT60):600-673, 1986. Cuddeback, J..6., J.R. Donovan, and W.R. Burg. Oecupational_ aspects of passive smoking. Am. Ind. Hyg. Assoc. J. 37:383-287, 1976. - Elliot, L.PR, and D.R. Rowa. Air quality during public gatherings. J. Air Pellut. Control Assoc. 36e836-838, 1976. Esmen, N.A. Dharacterisation of contaminant concentrations in enclosed - spaces. Environ. Sci. Technol. 12:337-339, 1978. I-,691Rf.4e

99 Eudy, L.W., F.A. Thorne, D.L. Heavner, O.R. Clreen, and B.J. ingebrethseu. Studies on the vapo_ r-particulate phase distribution of environmentnl nicotine by selected trapping and detection methods. Presented at the - 39th Tobacco Chemists' Research Conference, Montreal, Canada, Oct. -- 2-5, 1986. First, M.W. Environmental tobacco smoke measurementt Restrospect and prospect. Eur. J. Respir. Di.. 6(Suppl.):9-18, 1084. Caluskinova, V. 3,4-Benspyrene determination in the smoky atmosphere of social meeting rooms and restaurants. A contribution to the problem of the noxiousness of so-called passive smoking. Neopiasma 11:486-488, -- _- 1984. Crimsrud, D.T., M.H. Sherman, and R.C. Sonderegger, Calculating infil- tration: Implications for a construction quality standard, pp. 422-452. Proceedings of the ASHRAE-DOE Conference on the Thermal Perfor- mance of the Exterior Envelope of Buildings 11, held in Las Vegas, Nlev., Dec. 1982. New York: ASHRAE. 442 pp. Grot, R.A., and R.E. Clark. Air leakage characteristics and weatherisation - techniques for low-income housing, pp. 178-194. Proceedings of the ASHRAE-DOE Conference en the Thermal Pa: formance of the Exterior Envelope of Buildings, held in Orlando, Fla., Dec. New York: ASHRJIE, 1979. Hammond, S.K., B.P. ).eaderer, and A. Roche. Collection and analysis of nicotine as a marker for environmental tobacco smoke in personal samples. Atmos. Environ., in press. Hawthorne, A.R., D. Gammage, f1.8. Dudney, B.E. Ilingerty, D.D. Schureeko, D.C. Par:yek, D.R. Womack, S.A. Morris, R.It» Westeley, D.A. White, and J.M_. Schrimscher. Air Indoor Quality Study of Forty East Tennessee Homes. ORNL 6965. Oak Ridge, Tennessee: Oak Ridge National Laboratory, 1984. 134'pp. Hoegg, U.R. Cigarette smoke in closed spaces. Environ. Health Perspoct. - - - 2e117-.128,1972. Iehisu, Y. General equation for the estimation of indoor pollution. Environ. _ - --- Sci. Technol. 14:1264-1267, 1980. Jaffe, J.H. Behavioral pharmacology of tobacco use. Life Sci. Res. Rep. -- - - 8:175-198, 1978. Just, J., M. Borkowska, and S. Masiarka.. Zaniecsyescsenie dymen tyto- - - - niowym powistrsa kawiarn Warasawakich. (Tobacco smoke in the air of Warsaw coffee house.) Rocs. Panstw. Zakl. Hig. 23s119-136, 1972. Kusada, T. Control of ventilation to conserve energy while maintaining acceptable indoor air quality. ASHRAE Trans. 82:1189, 1976. Leaderer, B.P., W.S. Cain, and R. Isserofi, and L.©. Berglund. Ventilation requirements in buildings. II. Particulate matter and carbon monoxide _ from cigarette smoking. Atmos. Environ. 18:99-108, 1984. Leaderer, B.P., S. Rones, P. Bluyssen, and H. Van De Loo. Chamber studies of NO2, SO2 and RSP deposition rates Indoors. 1'roceedings of the - lis, 79th Annual Meeting g of Air Pollution Control Association, Minneapolis, Minn., June 23-27, 1988. APCA No. 86-38.3, 1986. Moschandreas, D.J., D.J. Pelton, and J. Zabransky. Comparison of indoor and outdoor air quality. EPRi EA-1733. Electric Power Research Institute, 1981. ZgG913448 Moschandreas, D.J. H:xposure to pollutants and daily time budgets of people. Bull. N.Y. Acad. Med. 67:846-869, 1981. - - - Muramatsu, M., S. Umemura, T. Okada, and H. Tomita. Estimation of personal exposure to tobacco smoke with a newly developed nicotine personal monitor. Environ. Res. 36:118-227, 1984. Muramat.u, T., A. Weber, S. Muramatsu, and F. Ackermann. An experi- - - rpental study on irritation and' annoyance due to passive smoking. Int. Arch. Occup. Environ. Health. b1:306-317, 1983. Nitschke, I.A., W.A. Clarke, M E. Clarkin, G.W. Traynor, and J.B. Wadach. Indoor air quality, infiltration and ventilation In residential buildings. NYSERDA *86-10. Albany: New York State Energy Research and -- - - Development Authority, 1985. Parker, G.B., Ci.L. Wilfert, and G.W. Dennis. Indoor Air Quality and Infil- tration in Multifamily Naval Housing, pp. 1-14. Annual PN WIS/APCA - - Meeting, beld In Portland, Oreg., Nov: 12-14, 1984. Penkala, S.J: and 0. Do Oliveira. The simultaneous analysis of carbon - monoxide and suspended particulate matter produced by cigarette smok- ing. Environ. Res. 9:99-114, 1976. Perry, J. Fasten your seatbelts: No smoking. B.C. Mnd. J. 16:304-305, 1973. Repace, J.L., and A.H. Lowrey. Indoor air pollution, tobacco smoke, and public health. Science 208:484-472, 1980. Repace, J.L., and A.H. Lowrey. Tobacco smoke, ventilation, and indoor air quality. ASHRAE Tran... 88:894-914, 1982. Rickert, W.S., J.C. Robinson, and N. Oollishaw. Yields of tar, nicotine and carbon monoxide in the sidestream smoke from 15 brands of Candian cigarettes. Am. J_ Public Health 74:228-231, 1984. Schenker, M.B., T. Smith, A. Munos, S.' Woskie, and F.E. Speisea Diesel exposure and mortality among railway workers: Results of a pilot study. Br. J. lnd. Med. 41.3Z0-.327, 1984. - -- -- - - Sexton, K., J.D. Spengler, and R.D. Treitman. Personal exposure to res- pirable particulates: A case-study In Waterbury, Vermont. Atmos. Environ. 18:1886-1398, 1984. Shair, F.H., and Heitner, K.L. Theoretical model for relating indoor pollutant - --- -- concentrations to those outside. Environ. Sci. Tech. 8:444-451, 1974. Spengler, J.D., D.W. Dockery, W.A. Turner, J M. Wolfson, and B.J. Ferris, Jr. Long-term measurements of respirable sulphates and particles inside and outside homes. Atmos. Environ. 16:23-30, 1981. Spengler, J.D., R.D. Treitman, T.D. '1'bsteson, D.T. Mage and M.L. Socsek. Personal exposures to respirable particulates and Implications for air pollution epidemiology. Environ. Sci. Technol. 19:700-707, 1985. Sterling, T.D., and E.M. Sterling. Investigations on the effect of regulating smoking on levels of indoor pollution and on the perception of health on comfort of office workers. Eur. J. Respir. Dis. 85(Suppl 133):17-32, -- --- -- 1983. Ssalai, A, Ed. The Use of Time. 'Daily Activities of Urban and Suburban Populations. The Hague: Mouton, 1972. 868 pp. Tnrk, A. Measurements of odorous vapors in test chamber: Theoretical. ASHRAE U(5)s66-8, 1963.

100 U.S. Depsrtment of 'l~anportation aQC ~f . amokin ninntra[ plorttleircr~att' cat'wn, and Wel[ars. Health up _ 6 , Federal Av .v Wa.hington, D,O.: U.S. Department of Tcanportatioa_ , [+'ducation ~ad tion Administration, .nd U.S. Department - of Health, - Welfare, National lnstltute of for occapatlonal Safety and 1(ealth,1971, tlowrnment. Leadon 85 pp. and hnlth, p. 83• ln U.K. U.K. E3overnment. Smoking -_ ort 11179. Annual R~epoct, Laboratory of the Ginvernment t+hemLt Rep London: U.K. ©overnment, 1980. 197 pp. Wallsee, L.A., and W•H• Ott. Personal moniton: A aate-o[ tha-art snrvey, J, Air Pollut. Control Auoe. 32:801-810, 1982. Weber, A. Acute affects of .aviroamental tobacco smoke. Eur. J. R.esPir• Dis, gg($upp1. 133:98-108, 1984. at work. lnt. Arch. Occap• Weber, A„ and T. Fischer. PM1980,moking Environ. Health 47s209-3Z1, - ean. Irritating effects on man of alr Weber, A., a• Jermini, and E. arandj -- J. Public Health 88:67R-878, pollution due to sigarette amok4~ Nn' erimental 1976. e~, passive smoking in sxp Weber, A., T. Fischer, and i.. (~randl and field conditions. Environ. Resn P~~i~* ~~ok nga Irritating effects Waber, A., T. Fischer, and E. (iraadl of the total smoke and th. ga• Ph~s. lnt. Arch. UcEnP. Environ. Health 43:183-193, 1979b. ~wsaa~ and H: W. SthliPkoeter. Pat• Winneke, d., K. Plischke, A. terns and determinsnts of reaction to tobacco s j ~~dY~ll, e and i]. Saa- e:cposure setting, pp. 361-366, In B. Bergland, del1, Eds. Indoor Air, Vol. ~t~ ° Im_,Y $w dea_ ~ StockLolmt Sv+~bh and Housing Epidemiology. 1984. Council for Building Research, ins. yentilacioa reguirement~. Yaglou, C.E., E.t7. RiCsy, and D.I. G>gg ASHRAE Trans. 4g:133-182, 1938. M994L8 " 6 Assessing Exposures to Environmental Tobacco Smoke Using Questionnaires ' The active component(s) of environmental tobadco smoke (ETS) associated with various health effects may be different for acute and chronic outcomes. Also, the mechanisms of ac- tion differ. Furthermore, as discussed in Chapter 2, the relative - - concentrations of various components of ETS change over time, i.e., as the smoke ages. Therefore, the use of a single proxy pol- lutant, such as respirable particulates, or an indirect measure of ETS limits the ability to assess responses to E'1S exposure. For some investigations, indirect assessment is probably not adequate to evaluate health effects for at least two reasons. First, the to- bacco smoke components that affect the health outcome may not be related to the indirect assessment in a simple way, e.g., vapor- phase-component concentrations cannot be adequately measured by particulate-phase components. Secondly, a variety of host fac- tors affect the actual dose received so that assessment of exposure does not accurately (or completely) represent dose (see Chapter A variety of methods is used to estimate individual exposures associated with human health effects in industrial and nonindus- trial _ -- trial settings. These exposure indicators may be direct-such as the use of personal monitoring data or biochemical measures ob- tained tained by testing body fluids for the compound or its metabolites- or or indirec-t--suc:h as the use of data from m interview responses of - family members regarding activities of the subject and modeling based on environmental monitoring of the ambient or industrial setting. The resulting data from direct and indirect indicators of - - exposure can be expressed in quantitative or qualitative terms. 7). 101

102 The advantages and disadvantages of the various exposure men- aures used in industrial and nonindustriat settings are summarized in Table 6-1. The issues raised in this table are directly relevant to assessing ETS exposure. The use of surrogate measures derived - from questionnaire responses and the issues resulting from use of these measures are discussed in this chapter. EXPOSURE HISTORIES DERIYED FROM QUESTIONNAIRES Questionnaire responses of study subjects or family/household members are used for two purposes. First, questionnaires are ueed - - - to obtain data on the physical characteristu:s of each environment and the time-activity patterns of the individual in each envirun- ment. These data can be used with individual monitoring data -- - to estimate (usually by modeling) the air•-containinant levels in the microenvirontnent and to estimate time-weighted, integrated individual exposures. Second, questionnaire responses provide a basis for classification-of individuals into broad categories of ex- posure - - posure based on self (or proxy) reports of exposure to individuals who smoke. Questionnaires of the latter type have -provided the bases for associating ETS to the increased risk of nonmalignant and malignant disease. There are several major issues in el>idemiologic studies of -- health effects of exposure to ETS that rely on indirect measures _--- of exposure as derived from questionnaire data. First, the assessment of ETS exposures associated with acute health effects iequires a different approach than that for chronic health effects. Acute health effects, such as respiratory infections, are manifested shortly after exposure and are of short duration. By inference, these health outcomes depend only on exposures in the recent psst. In contrast, chronic health effects are conditions that are associated with long-term exposure to L:TS, that is, they are manifested after some prolonged period of time and are of long duration. In ®valuating the association of ETS with chronic dis- eases, knowledge about the duration of exposure and the duration of time from initial exposure to disease onset is more important than the duration of the disease. Second, quality of information obtained by interview or self- administered questionnaires may vary among studies anti inay vary for different disease outcomes. For example, the assekwment ?4 .1 V 103 ap ~'V ~- J.I ~ A ~ 8 ,g a a H ~ ,~ E ,s R c3 3 g" ~'~ y e~ rJ n x , i o t~ .r rti ~-i ~r vi ° o r a M -- Nrf g9

TABLE 6-1 Continued Ilndicator Advantages Disadrantages C. Employer on other reports of exposure to compound-quafrta- tisx D. 5elf,reports of exposure to com- pound-qtralitatene 2. INDIRECT A. !Bioloyical monitoring ('] ), with e'aromosomr, studies- quantitative or qYaGrari.r (2) by meaatring changes in bio- eb.miaalrespoms (r.S_ s1e+aLed ~ V.e0 ot NAioqyaw.q p.odre}{ee in -ponSei to cyiAldeIeipoiutQ)- qtanthative or qnalitatirr B. Aeea industrial'hygidne or ambient monitoriag-quatttiratire C• Employer work aroa assignment mcosds (rork histories')_qyalita- ei.,e (specific estimates may'be made using job-exposure linkage) orqrannitatirr (miy be derdloped by' using duation of,time spent in different en.ironments) 1.'pro.ides details of accidental releases 2. Can indicatrsafety prncedureslptoteative measures 4. ', Pro.iderdetailcof acddenW aeleases 2• Can indicate personaU'byYtene and'safety habits 3. Can obtain chronology of work experience with multiple agent exposures '1. Identified aan_as in the genetic material 2. IIndicates, systettiic exposure to a mutagen 1. 'Identifies alterations in normal constitu- ents of body fluids and changes ia, rrte of woe+erl' bioahmI ieat prOCa.rs 1. Doatments concentration of agenY in work ea.irvnment 2• Variety af inessun'ment techniques arail. able 3. Can be', petformed easily by'the employer II. ' Can peovid'r chtonologic work experience for duration of exposure 2. Can indicate exposure to multipk agents 3. ' May provide suppiementary linformation 1. Data may be incomplete (unreported) 2. Exposure quantified subjectively 3. Episodic measurement ofiunusuol oeettrtrnces rather than ••a-ge" workday exposure 1t poteatirl for recall bias 2. Empk>yea may' bei unaware tlf exposure 3. Potential for' falsification of etyposure for per- sondigain 4. Poteatial'for lost, to folkw.-up (missing inforttta- tion) in aetrosptxtive studies 1. Fspensire. due to need for soeeially trained persoonel and sophisticated equipmeat 2. Relationthip'between changes in mutation,ratet and reproductive outaomes' is unknown 3. Results may be confounded' by smoking and emFtotttnental factors!(e.g.. effect of smoking on sister ¢hsomatid exchanges in Jymphocytess radiation effeas) 4. Indi.idusll rariabilitv in baseline rates 5. Mast chromosomal aberrations ate'nortspecift'c 1. Does, not'quantiEy'botly'burden 2' ' Resuhs may, be confounded by drugs. nutrition. anddllCafC 3. Requieai understanding of oompound's metabo- lism in'body I. , May not correspond with results of personal sampling 2. 'Measurements~ haft mutEipk sourasof variation 3. : Does not indicate specific exposntx level for mdiridttal employees 4. No information abouti previous e:posnres S. Type of sample''tskea atu,y'be'inaPptoPute, for heaklt effects being studied 1. May be' ineomplate or may'be unavailable 2. Records not designed'forresearch,purpeses 3. Presumed expostne'by work assignment may be based tm'subjective aiteria 4. Record t+eriew is tittte-aonsuming Actirity diaries of study subjects, tecording time spent in different miaoemvrcoumeats D. Surrogate (next tlf kin) interriew, 1. Can obtain infotmatitm about c+otlfound- 1. Limited by know+kdge of employee's work responsestsgarding work history iogfactarts 2. Msy' produce overestimate or underestimate of and activity history of study sub- 2. Identifies major agents to which exposed , exposure ject-qr.alitative (specific estimates 3. May provide supplementatY:dentographic 3. •d"une-consuming to locate and' interview made using' al jtib-exposure link- information about employees 4. Lack of validation of, data age) or quantftaafve (estimates S. Differential quality of information by degrce' of developed using duration of time kinship spent',in environments)

106 of maternal smoking during the first year of life of a child may be a much more accurate measure of exposure to E'I'S reluted to res= -- -- piratory illness than a summary history of ETS exposure related to lung cancer. Data quality for ETS exposures can be affected in major ways by differential and nondifferential misclassification of exposure. In Chapter 12, the impact of ntisclassifying exposed subjects as nonsmokers, when they are in fact current smokers or exemokers, is discussed. Therefore, it is important to determine whether nonsmoking subjects are, in 'fact, never smokers or cur- rently rently nonsmokers, i.e., examokers. Another source of bias is the misclassification of exposure among nonsmokers. That is, non- smokers who say they have not been exposed may in.fact have had significant exposures. In both cases, detailed probes are needed. Third, the role of major confounding exposures needs to be assessed. For instance, occupational exposures to other air con-' taminants may cause pulmonary disorders. Fourth, the evaluation of ETS exposures should attempt to assess all such exposures rather than focus solely on exposures from smoking by family members (spouse, mother, or father) or focus solely on the home environment. An adequate assessment of total ETS exposure will necessitate a consideration of exposure - - levels in specific microenvironments-such as home, school, work, vehicle, and recreation-and the duration of time an individual is exposed in these environments. Developing such a measure is complex even for relatively acute health outcomes, such as acute cardiovascular, respiratory, or neurotoxic symptoms, for which it may be sufficient to estimate recent exposures. Developing a comprehensive measure to ETS exposures is far more complex for diseases with long induction times, such as cancer and chronic obstructive pulmonary disease. The data required for modeling a long-term integrated ETS exposure may be far more detailed than are available or can be reliably obtained. Further, when a surrogate informant is used, that person most likely will be able _ to report on exposures in only some of the iiiicroenvironmente. In this case, it may be impossible to develop a comprehensive index. 107 ENVIRONMENTAL TOBACCO SMOKE EXPOSURE DATA FOR STUDIES OF ACUTE AND CHRONIC HEALTH EFFECTS The acute health effects of ETS in children, such as respiratory -- illnesses, have been assessed in the National Health Interview Survey (NHIS) by determining smoking status of one or both parents or smoking status of adults in the household (Bonham and - Wilson, 1981). In this national probability sample of households, parental smoking histories and reports of respiratory illness among children were obtained at one point in time. By contrast, in the Harvard Air Pollution Respiratory Health Study (Six Cities Study), information on current smoking habits for parents and all household members who smoke regularly in the home is obtained annually to determine amount of cigarette smoking in the home environment to which the children aged 6-13 years are exposed (Ware A al., 1984). (In Chapter 11 the assessment of exposure to parental smoking in studies of respiratory illness in children is discussed in more detail.) In studies of chronic health effects in adults, such as cancer, exposure of nonsmokers to ETS has been largely determined by smoking status of the spouse. Most studies of lung cancer among - - ---- - nonsmoking women have relied solely or principally on information __ regarding smoking status of the spouse to assess ETS exposures, -- - - with little attempt to corroborate self--reports of exposure to ETS. ' The difficulties in assessing ETS exposure are similar to diffi- culties --- -- culties of assessing occupational exposures (Axelson, 1985). Both exposures are complex and variable. The problem of obtaining adequate information about ETS exposure might be overcome by obtaining data from multiple respondents and by using corrobo- rating procedures. However, the conceptual difficulty concerning the determination of exposure is unresolved or unaddressed in most studies. Exposure to a substance involves a varying inten- sity over some period of time prior to the development of disease. These factors may influence the absorption and distribution of an agent in the body as well as the biotransformation and ex- cretion of the agent. Therefore, these factors probably influence the risk of the health outcome of interest. For exposures extend- ing over.long periods of time, a simple "cumulative dose" usually is calculated by a time integration of the intensity. The estimate of SS6994,4g

108 exposure over long periods of time is sxpreHSed as average nuun- ber of cigarettes per day or the calculation of "pack years" This type of measure does not provide for an independent considera- tion of latency, does not consider variability in exposure over the time period, and represents two components of exposure, one of which may be more precisely measured (duration) than the other (intensity) (Doll and Peto, 1978). Axelson (1985) describes some ~ sophisticated adjustments that have been proposed for weighting time periods of exposure to estimate cumulative-dose measures. These proposed methods have not been widely adopted, prob- ably due to both the complexity of the method- as well as the recognized limitations of exposure data typically available. The more common, simplified procedure is to apply an appropriate induction/latency period in the analysis of studies of cancer or other chronic diseases. This practice suggests, however, that more attention be given to identifying the separate effects of late (re- cent) exposures versus early (remote) exptoures on development of various diseases.. These effects may also lie mediated by the age at which the exposures occurred. The proposal described by Johnson ajkd Letzel (1984) advo- cates a method of assessing exposures to CTS experienced over an entire lifetime. The major limitation of this approach is that it has not been validated. Johnson and Letzel (1984) argue that since no objective criteria for lifetime exposure to ETS exists, a direct validation of an instrument to assess lifelong ETS exposure cannot be obtained. They propose that the instrument be vali- dated on a recent time frame, such as 24-hour data. From these data the investigators argue by analogy that the method, when expanded to a longer time frame, can be regarded as valid. While this approach may seem less than ideal, the constraints due to data availability and quality emphasize the importance of the type of methods development and corroboration illustrated by the work of Johnson and Letzel (1984). DATA QUALITY Misclassification of individual ETS exposure may be differ- ential ential (biased) or nondifferential (random). DilFerential misclas- sification would result in a distortion of the estimate of risk in either direction, depending on the direction of the misclassifica- tion. Nondifferential misclassification would result in a reduction 4,SG98448 109 of power in a study, thus making it more difficult to detect a true association of exposure with risk of disease. - One form of differential misclassification that is a major con- cern - cern in studies of ETS exposures is the active smoking status of study subjects. This misclassification rnay be considered differen- tial because spouses and children of smokers are more likely to be smokers (or have smoked) themselves, even though they are reported as "nonsmokers." The effects of this differential mis- classification are discussed in Chapters 11 and 12.. One way to minimize this problem is to have multiple questions that probe for previous cigarette usage, even if the subject has defined himself or 'herself as a nonsmoker. Another form of differential misclassification is that resulting from the biased reporting of exposure to ETS by individuals with existing respiratory diseases, such as asthma or chronic bronchitis. - Ono might conjecture that individuals with existing respiratory •diseases may be more or may be less likely to report exposure to ETS.than individuals without such existing conditions. In studies of ETS exposures, inforination about the smoking - habits of the subject, family, and household_ members is obtained by interviews with the study subject when available, or by inter- view with a family member when the study subject is deceased or unavailable. '1'hat is, surrogate respondents may be used to collect information regarding personal exposures of the study subject. The validity of surrogate information in most studies is un- certain, certain, and the direction of any potential bias is rarely known (Gordis,1982). The feasibility of this approach for a variety of ex-s posures and habits has been examined (Pickle et al., 1983). Also, several studies have assessed the reliability and validity of surro- gate respondents for various kinds of exposures (Rogot and Reid, 1975; Kolonel et al., 1977; Marshall et al., 1980; Baumgarten et al., 1983; Humble et al., 1984; Greenberg rg et al., 1985; Herrmann, 1985; Lerchen and Samet, 1986). In all of these studies, agreement between -self and surrogate responses improves when the mnount.of detail required for the response is decreased. This observation was first reported by Rogot and Reid (1975) and subsequently observed in studies comparing self versus spouse/surrogate responses. Lerchen and Samet (1986 reported perfect agreement of - cigarette-smoking status (ever _never) as reported by lung can- cer cases and their wives, but only 6E3 (86%) of the 77 wives - married to smokers were able to supply complete details about

110 their husbands' cigarette-smoking habits. In this study, agree- ment (expressed as correlation coefficients) wss <luite good for all smoking-related variables, such as age at which the subject started to smoke (0.48), total years of smoking (0.91), and average number of cigarettes smoked per day (0.44). The mean values reported_ by cases and their wives were not significantly different for any variable. Overall, the agreement observed for self- and surrogate-reported smoking reiated information was better than the agreement fot education, occupation, and dietary information. Pershagen and ' Axelson (1982) also reported perfect agree- ment for smoking status information obtained by interview with a close relative (parent, wife, or child) for 141ung cancer cases when information was compared with that obtained previously by the plnnt physician. Their inquiry was limited to smokor/nonsmoker status. Damber (1986) and Pershagen (1984) reported 99% agree- ment between reports of close relatives and hospital records for ever/never smoking studies in a sample of 86 patients admitted for respiratory disease. The agreement for number of years smoked (±5 years) was 74%. Other studies have noted additional features of the responses from surrogates. The report by Pickle et at. (1983) indicates that respondents other than spouse and direct next-o-L-kin (siblings, parents, and children) are more likely to not know relevant in- formation. Marshall et al. (1980) demonstrated the increase in sensitivity obtained by combining information from two or more surrogate respondents, and Herrmann (1985) showed that hus- bands reported data for wives as reliably as wives reported-expo- sures of husbands. Recent data from the NCHS Epidemiologic Followup Study (NHEFS) in 1982-1984 of participants in the National Health and Nutrition Examination Survey (NHANES I) in 1971-1975 provides a strong confirmation of these earlier reports (S.R. Machlin, J.C. Kleinman, J.H. Madans, National Center for Health Statistics, personal communication). This analysis is based on a subsample of 5,669 individuals with data regarding baseline smoking status available from both NHANES I and NHLFS. Agreement *ratea between NHANES I and NIIEFS for the 5,029 subject responses versus the 640 proxy responses at follow-up are compared ('lh- bles 6-2 and 6-3). When smoking status is broadly defined as ever/never, the 91% agreement rate for proxy responses complLres quite favorably with the 95% agreement rate of subject responses SSG9N4419 111 (Table 6-2). Similar high agreement is observed for proxy and sub- ject responses at follow-up when smoking status is considered as current/not current (lkble 6-3). Additional analyses of these data to assess the factors associated with agreement between baseline and follow-up responses considered age, race, gender of subject, type of respondent at follow-up, and smoking status at baseline. Estunates of the relative odds of disagreement indicated that only the effect of race did not interact with any of the other variables - included in the multiple logistic model. Significant two-way inter- actions were observed for type of informant and age of subject, - - - baseline smoking status and gender of subject, and baseline smok-. -- - ing status and age of subject. These,results suggested that proxy respondents were more than twice as likely to misclassify smoking - ---- - - status for subjects less than 65 years of age, but not for subjects age 65 years and over. When amount smoked (current amount at -- ---- -- baseline versus usual amount at follow-up) is compared for smok- ers ok- ers only, the agreement rates are substantially affected by type of respondent; 55% agreement for subject responses versus 35% agreement for proxy responses. When this comparison is made with nonsmokers included, a much higher rate of agreement for both subject (80%) and proxy (74%) responses is observed. This comparison is strongly influenced by the substantial proportion of nonsmokers (over 60%). Of concern, however, is the he high propor- tion - --- tion of self-reported current and former smokers at baseline who are reported as never smokers at follow-up; 5.6% by self respon- dents and d 12.9% by proxy respondents. These results are discussed later in the section concerned with confounding. Another large cohort study in England and Wales provides information regarding the proportion of people who say that they - have never smoked but, in fact, have done so ln the past (N. Britten, University of Bristol, England, personal communication). A large longitudinal study of children born in 1 week in England and Wales in 1946 has included several follow-up visits, the most recent of which was done in 1982 when the subjects were 36 years of age. Table 6-4 presents some results. A portion (4.9%) of the subjects said they had never smoked as much as one cigarette a day in 1982 when in fact they had previously reported that they smoked. These subjects had reported smoking at a rate of about half the current smokers. Nearly all of exsmokers (93%) had smoked 10 or inore years earlier.

112 TABLt3 6-2 Percent Distribution of Smoking Status at Baseline Exam - - (NHANES 1, 1971-75), According to Smoking Status at Follow-up (NHEFS, 1982-84) by Type of Respondent at Follow-up w Smoking Status Reported at Follow-up - b ~ Baseline Smoking Ever Never Status (NHANES O No. Percent No. Percent Total No. G E ~ ~ 7~pe of Follow-up Re.rpondent: Self ~ Ever 2,675 95.6 125 5.6 2,800 Never 122 4.6 2.107 94.4 2.229 ~ Total 2.797 100.0 2,232 100.0 5,029 Typs of Follow-up Retpondentz ft:oxy Ever 329 95.1 38 12.9 367 V Never 17 4.9 256 87.1 273 Totat 346 100.0 294 100.0 640 ~Q. SOURCE: Information obtained from National Center for Health Statistics (S. tt. Machlin, 1. C. Kleinman, J. H. Madan:, personal communications). P4 ~ .~ TABLE 6-.3 Percent Distribution of Smoking Status at Baseline Exam - - - (NHANES 1. 1971-75), According to Smoking Status at Nollow-up (NHEFS, 1982-54) by Type of Respondent at Follow-up o Smoking Status Reported at Folbw-up ~ Baseline Smoking Current Not Current Tota( .X Status (NHANES 1) No. Percent No. Percent No. ~ ~ 2ype qf Follow-up Reipondent: Self Current 1,722 89.5 124 4.0 1,846 0 Not Current - Total 202 1.924 10.5 100.0 2.981 3.105 96.0 1W.0 3,183 5,029 ~ 0 7)Ve af Follow-rp Respondent: kroxy Current 186 83.4 24 S.8 210 Not Current - 37 16.6 393 94.2. 430 x Total 223 100.0 417 100.0 640 SOURCE: Information obtained from National Center (or Health Stalixtia (S. K. Machlin, 1. C. Kkinman, and J. H. Madans, personal communication). GSO8449 Io~•vVS°° 1. r ry q~ [`j Vl ! h 113 !Or, Y..., g C4 ...4 . e, ~. ..,

114 Therefore, both longitudinal studies indicate that about 5% of self-reported lifelong nonsmokers may, in fact, have smoked. Rogot and Reid (1975) observed that there was a tendency of surrogate informants to report a higher tobacco consumption than previously reported by the study subjects. However, Lerehen and Samet (1986) observed no such differential in the reporting by wives of amount smoked by their husbands as compared with that reported by husbands. The body of evidence on surrogate responses to questions about smoking status suggests that the validity of such data may be limited and that spouses and, perhaps, other close family members can provide an accurate, but simple, smoking history (ever/never, smoker/nonsmoker). However, detailed information about amount and number of years smoked may be inaccurate and may result in substantial misclassification of study subjects by ex- posure posure status. These findings, although from a limited number of studies, have direct implications for the studies of ETS expo- sures -- sures where ETS exposure information is derived from surrogate - -- - reports. It should be noted that in the special instance where the spouse surrogate is reporting on his pera-nal smoking history, the information regarding ETS exposure of the nonsmoking study subject may be more accurate with regard to lionie exposures than the report by the study subject. Cotinine, the major metabolite of nicotine, can be detected in blood, urine, and saliva of active cigarette nmokers and of thone passively exposed to ETS. Coultas et al. (1986) demonstrated that nonsmokers exposed to cigarette smoke in their homes have de- tectable levels of salivary cotinine that increase as the number of smokers in the home increases from 1 to 2 or more (Table 8-6). - Biochemical corroboration is not as promising for remote expo- - - sures to ETS. Corroboration of historical exposures, therefore, must rely on other methods, such as review of historical records. _ - Results of recent biochemical measures may be used to corroborate self-reports of recent exposures for individuals for whom reports of both recent and remote exposures are available. The quality of historical data for an individual can be inferred from data using results from biochemical corroboration: -'1'his approach has been -- - proposed by Johnson and Letsel (1984). The true validity of retrospective ETS exposures is impossible to establish. Wherever possible, other methods to corroborate exposure estimates should be used to assess and confirm the quality 09Fi9@44g 115 Tasl.e 6-S Salivary Cotinine Concentration (ng/ml) in Nonsmokers by Age and Number of Active Smokers in Household Number of Household Smokers Atie, yr None One Two or More Younger Ih.n 6 11;1.7;68• 3.8;A.1;41 5.4;5.6;21 6-17 0,1.3;200 1.8;2.4;96 5.3;5.6;2S Older than 17 0,1.5;316 l).65;2.8;60 0;3.7;12 'Median; mean; number of subjects. Coultas el al. (1986). of self- and proxy reports of ETS exposure as well as active smoking -- - status of study subjects. Other methods currently available for comparison with questionnaire and interview responses include biochemical measures, environmental modeling, review of existing records, and reports of additional respondents. OTHER VAR.IABLES Confounding factors that should be considered in the design, collection, and use of questionnaire ire data are other risk factors associated with the disease that may or may not be correlated with exposures to ETS. In the case of lung cancer, such risk factors include, but are not limited to: . • occupation and industry of employment, • exposure to specific respiratory carcinogens, such as as- bestos, arsenic, radon, etc., in occupational or nonoccupational settings, • dietary factors, • fantily history of cancer (Ooi et al., 1986), . residential history, . housing characteristics, . years of education, and a socioeconomic status. Confounding factors relevant to the assessment of pulmonary func- tion and respiratory illness are listed ted in Table 11-1. In addition, examokers and current smokers have been (or are) exposed to active smoking for some period of time. Therefore, these individ- uals may have been exposed to higher concentrations and longer

116 duration of ETS, due to their own smokuig patterns. Thus, an --- - evaluation of the increased risk associated with exposure to ETS for any disease that is strongly associated with active smoking will need to control for smoking status of the individual study subjects. The confounding effects of active smoking were not adequately controlled in several investigations of lung cancer (discussed in Chapter 12). This concern is particularly relevant in studies of acute respiratory illness in children and adolescents where the study subjects may be disinclined to report their smoking behav- ior accurately or the parents may be unaware of their child's active smoking (described in Chapter 11). A history of exposure to all other known or suspected con- founding factors should be obtained in a comparable manner for cases and comparison subjects by interview and corroborated whenever possible by comparison with existing records or self- reports obtained before development of the disease. The exposure data collected should strive to be as detailed as possible with re- spect to intensity, duration, and calendar time for all exposures, including ETS exposures. However, one should be cognizant of the limitations imposed on data quality, especially when the inves- tigation relies on surrogate responses. Such quantification at best provides an approximation of exposure, whether the information is obtained from the individual himself or from a surrogate. SUMMARY AND RECOMMENDATIONS There are problems with self- and proxy reports of ETS ex- posure inferred from questionnaire responses that limit the utility of these data. The best method by which to estimate individual ETS exposures is not known, and this lack of information ham- pers all efforts at assessing data quality, including data validity. At present all methods used and proposed are indirect, although some provide quantitative measures and some qualitative mea- sures (smoker/nonsmoker). However, information on exposure from monitoring and detailed environmental-modeling studies of RSP indicate that only 30-40% of the variation in exposure can be explained using this approach (see Chapter 5). Further, biochem- ical methods to assess E'1'S exposure are extremely limited in the assessment of historical exposures that are most important with regards to chronic health effects. Therefore, exposure data derived . 117 from questionnaire responses have an extremely important role in existing and future studies of ETS exposures. What Is Known 1. Surrogate responses from spouses or close family members can provide data as accurate as self-reports for simple ever/never smoker status and current amount smoked. However, with such simple classifications, an error rate of about 5% is observed where- by ever smokers are misclassified as lifelong nonsmokers. This error is present for self-respondents as well as proxies. What Scientiflc- Information Is Missing 1. Differences in exposure levels between home and work envi- ronnients have not been described in existing studies. In addition ddition to the amount of time that an individual may spend in a work setting, the actual exposure may vary within the setting due to physical characteristics of the work environment as well as the number of active smokers present. 2. Future investigations should be concerned with detailed - -- -- - characterization of ETS that would provide a more precise estima- tion of individual exposures and include additional considerations of physical characteristics of the environment, activity patterns of the study subject, and ages at which exposures occurred. These data could be entered into a model, from which exposure estimates - - can be made. 3. Because of the importance of tnisclassification of active - smoking statue, repeated and complementary efforts to determine and corroborate smoking status should be made in the collection - of exposure data. Specific probes regarding former smoking status - might be included in the questionnaire, even if the study subject - has defined himself or herself as a nonsmoker. 4. Confounding factors should be considered in the design, collection, and use of questionnaire data. These will vary with the health effect being assessed. The evaluation of ETS expo- sures should attempt to assess all such exposures, including both the home and work environment rather than focus solely on the smoking status of one family member, e.g., spouse. 5. The comparability of questionnaires used to assess_ ETS has not been established, and this would be desirable. 19G9Q44B

118 REFERENeE9 Axelson, 0. Dealing with the exposure variable in occupational and environ- mental epidemiology. Seand. J. Soc. Med. 13:147-15R, 1986. Baumgarten, M., J. Siemiatycki, and a.W. Gibbs. Validity of work histories obtained by Interview for epidemiologic purposes. Am. J. Epidemiol. 118:5_83-691,1983. Bonhsm, a.8., and R.W. Wilson. Children's health In famiiies with cigarette -- smokers. Am. J. Public Health 71:290-293, 1981. - Goultas, D.B., J.M. Samet, O.A. Howard, Q.T. Peuke, and B.J. Sk_ipper. Salivary cotinine levels and passive tobacco smoke exposure in the home. - Am. Rev. Respir. Dis. 133:A167, 1988. Damber, L. Lung cancer in males: An epidemiological study in northern Swe- den with_ special regard to smoking and occupation. UmeG University Medical Dissertations, Umei, Sweden, 1986. 135 pp. Doll, R., and R. Peto. Cigarette smoking and brochial carcinoma: Dose and time relationships among regular smokers and lifelong non-smokers. J. - _- Epidemiol. Comm. Health 32:303-313, 1978. Qordis, L. Should dead cases be matched to dead controls? Am. J. Epidemiol. -- -- 116:1-6, 1982. Greenberg, E.R., B. Roeaer, O.H. llennekens, R. Rinsky, and T. Colton. An investigation of bias in a study of nuclear shipyard workers. Am._ J. Epidemioi. 121:301-308, 1985. Herrmann, N. Retrospective in[ormatlon from questionnaires. 1. Gompara- _ bility of primary respondents and their ne*t-ot kin. Am. J. Epidemiol. 121:937-947, 1986. Humble, O.(i., J.M. Samet, and B.E. Skipper. Comparison of self= and surrogate-reported dietary information. Am. J. Epidemiol. 119:88-98, 1984. Johnson, L.Q., and H.W Letsel. Measuring passive smoking: Methods, - - - - problems and perspectives. Prev. Med. 13:706-716, 1984. Kolonel, L.N., T. Hirohata, and A.M.Y. Nomura. Adequancy of survey data collected from substitute respondents. Am. J. Epidemiol. 106:476-484, 1977. - Lerchen, M.L, and J.M. Samet. An assessment of the validity of questionnaire - - responses provided by a surviving spouse. Am. J. Epideaniof. 1R3(3):481- 489, 1986. Marshall, J., R. Priore, B. Haughey, T. Rsepka, and S. Graham. Spouse- subject subject Interviews and the reliability of diet studies. Am. J. Epidemiol. - 114e876-883, 1980. 41 4oi, W.L., R.a. Elston, V.W. ehen, J,E. Bailoy-Wilson, and H. Rothschild. Increased familial risk for lung cancer. J. Natl. Cancer Inst. 72:217-222, - - _ 1986. Pershagon, f3. Validity of questionnaires data on smoking and other ex- posures, with special reference to environmental tobacco smoke. The -- - Respir. Dis. 133(Suppl.):78-80, 1984. Pershagen, 0., and 0. Axelson. A validation of questionnaire information on - occupational exposure and smoking. Scand. J. Work Environ. Health 8:24-48, 1982. 119 Pickle, L.W., L.M. Brown, and W.J. Blot. Information available from sur- rogate respondents in ea.e-control interriew studies. Am. J. Epidemiol. 118:99-108, 1983. Rogot, E., and D.D. Reid. The validity of data from next-of-kin in stu_d_ ie_. - of mortality among Immigrants. Int. J. Epidemiol. 4:61-.64, 1976. Ware, J.H., D.W. Dockery, A: Spiro lil, F.E. Speiser, and B.G. Ferris, Jr. Passive smoking, gu cooking, knd repiratory health of children living in six cities. Am. Rev. Respir. Dis. 129O(i0-374, 1984. Z9E9B4.49

121 7 Exposure---I)ose Relationships for Environmental Tobacco Smoke ESTIMATING DOSE When considering the risks of exposure to.environmental to- bacco-smoke (ETS) by nonsmokers, it is not enough to evaluate -- - - exposure and response. The actual dose received should be con- sidered. Typically, for smokers, the exposure is given in terms of number of cigarettes smoked per day or cuinulative pack-years. For nonsmokers, the exposure is usually characterized in terms of particle or gas concentration in micrograms per cubic meter. But what is known about the total integrated done to the respiratory tract resulting from exposure to ETS by nonsmokers? What frac- tion - tion of inspired particles and gases 'is deposited and fails to exit • with the expired air? Moreover, what is the fate of the deposited --- - smoke? Although highly variable in concentration, ETS includes many of the same constituents as the smoke entering the activo smoker's - -- - - lungs. Both particulate and gaseous phases are present, as de- scribed scribed in Chapter 2. In principle, the retained dose for either - inhaled particles or gases can be approximated in a straightfor- ward - ward manner: Dose = V x(GJ x CE. (7-1) The deposited dose, in micrograms per hour, equals the ventilation rate in cubic meters per hour (V) times the concentration of par- ticle or gas in thee inspired air in milligrams per cubic meter ((E1), times the collection efficiency (CE). CE has no dimensions; it is the fraction of the inhaled particle or gas that deposits and thus 120 C3G9Q44B fails to exit with the expired air. Thus, the dose is directly pro- portional --- portional to three variables: ventilation, pollutant concentration, and the fraction deposited. First, connider ventilation (Y). The standard 70-kg adult at rest breathes about 7,6 L/ntin (International Commission on Radiological Protection, 1975). However, a value of 20 L/min - would be more appropriate for adults in indoor environments who periodically stand, walk, type, or perform other modest tasks. During heavy exercise, ventilation can increase by a factor of as much as 10, to exceed 100 L/min (International Commission on Radiological Protection, 1975). The concentration of various constituents in ETS ([C]) that might be encountered in various situations has been discussed in - - ---- - Chapters 2 and 6. PARTICLE SIZE For particles, collection efficiency (OE) is determined primar- ily by two factors: particle size and breathing pattern. If the geometric size, shape, and density of the individual particles or droplets are known, then the distribution of particle diameters can be described. Because it ie a better predictor of the behavior of the particle in the respiratory tract, aerodynamic diameter rather than optical measurement is used to express the range of particle sizes. Aerodynamic diameter is defined as the diameter of a sphere of unit density that has the same settling velocity as the particle being measured. It may be expressed as the count median aerody- namic diameter (CMAD) or mass median aerodynamic diameter - (MMAD). These are, respectively, the diameters for which half of the number (or mass) of the particles are less than that diameter -- and for which half exceed it. The particles in mainstream cigarette smoke have ' been mea- sured by several investigators using a variety of analytical devices. Because of the different apparatus and methods of smoke genera- tion and dilution, results vary. However, to an order of magnitude, the findings are reasonably consistent. McCuaker et al. (1983) used a device called the single particle aerodynamic relaxation time (SPART) analyzer to size mainstream particles from several brands of cigarettes, with and without filtera. The MMAD for all brands averaged approximately 0.46 mm and was not markedly

122 different when the filters were removed. Particulate concentra- tions tions per milliliter ranged froin 0.3 x 100 to 3.3 x 10°, depending on whether the cigarettes were rated ultralow, low, or medium in tar content. Hinds (1978) compared the particulate size distribution iii cigarette smoke using an aerosol centrifuge and a cascade impactor. Although these devices are based on different physical principles, the MMAD values were comparable to those measured by Mc- Gusker et al. (1983), ranging from 0.37 to 0.52 µm. Variations depend primarily on the dilution of the smoke. Keith and Derrick (1960) used a specially modified centrifuge, termed a conifuge, to analyze cigarette smoke and reported MMAI) and concentration - -- values similar to Hinds (1978) -and McCusker et al. (1983). Partic- ulate - - -- ulate analysis by a light-scattering photometer yielded a MMAll of 0.29 µm and particulate concentrations of 3 x 101e/m1. - - - - Time and concentration can modify tobimxo smoke. Cigarette smoke aerosols contain volatile components, and evaporation grad- ually reduces particle diameters. It is also true that when the par- ticle concentrations are extremely high, like those encountered in mainstream smoke, the aerosol can agglomerate rapidly because nearby particles collide with each other and coalesce. If smokd is cooled (reducing the vapor pressure of volatile components) and diluted in room air (reducing the probability uf particle collisions), the size of the particles will become more stable. Particle size may also change within the human respiratory tract. After air contain- ing smoke is drawn into the mouth and upper respiratory tract, it becomes humidified. Smoke particles can grow in size because of their affinity for water, termed hygroscopicity (Hiller, 1982a). BREATHING PATTEItN Particle size is a critical factor in determining the collection efficiency, but breathing pattern is also important For example, large slow tidal volumes will favor alveolar deposition, while high inspiratory flows will promote deposition at bifurcations in the air- ways. Breath-holding is also important. Thu greater the elapsed time before the next expiration, the higher the fraction of inspired - particles deposited, since there is more time for particles to sedi- ment -- ment or diffuse. Individual anatomic differences may influence the amount and distribution of deposited partich:s. 'Phe cross section of airways will influence the linear velocity of the inapired air. 123 Increasing alveolar size decreases alveolar deposition. Preexisting disease can also modify the deposition of smoke. For environmen- tal tal tobacco smoke (diameters of particles ranging from 0.1 µm to 1 µm) the sedimentation and diffusion mechanisms will be the primary mechanisms of deposition. Changes in the rate and pattern of breathing associated with exercise can also affect the total dose of cigarette particulates deposited in the lungs.. Bennett et al. (1985) reported that exer- cise eise increased the percent deposition of experimentailly generated aerosols (MMAD of 2.6 pm) in humain subjects. The reason for this observation was that during exercise, breathing patterns change so that flow rates are increased. Increasing the flow rates also increases the inertial impaction. Also, exercise is frequently asso- ciated with a shift from nose to mouth breathing. Consequently, the filtration of large particles that takes place in the upper respi • ratory tract no longer occurs. Increased deposition was also mea- sured in exercising hamsters that inhaled a radiolabelled aerosol (activity median diameter of 3 pm) (Harbison and Brain, 1983). These results are relevant to those who breathe air containing ETS when their minute ventilation tion is_ increased while working or during periods of exercise. DEPOSITION OF CIGAItETTE SMOKE PARTICLES The factors discussed in the previous sections indicate that ex- - __ perimental measurements of the concentration of smoke aerosols in indoor environments, i.e., exposure concentrations, are insuf- ficient for predictions of smoke deposition. ETJJ smoke is con- stantly changing, thereby complicating the collection of accurate and reproducible data regarding its particulate size. In addition, alterations in respiratory structure and respiratory rate can affect deposition of harticulates. These complexities stress the impor- tance of actual measurement of regional deposition of cigarette - smoke particles in human lungs. However, little is published on this important area, despite the prevalence of passive smoking and concerns about its impact on human health. The majority of the available information on deposition of particles present in cigarette smoke is based on theoretical or physical models of the lungs and measurements of differences between the concentrations of tobacco smoke aerosol or model aerosols in inhaled and exhaled air. V%99448

124 A model to predict the percent of deposition of particles based on MMAD was developed by the Task Group on Lung Dynamics (1966) of the International Commission on Radiological Protoc- - - tion. The respiratory tract was divided intn three main regions: nasopharynx, trachea and bronchi, and the alveolar, ln conjunc- tion with estimates of particle clearance, deposition calculations were made for these regions at thre-e different inhalation volumes. This model suggests that about 30% of the particles within the size range present in cigarette smoke will deposit in the alveolar region and Fs-10% in the tracheobronchial region. This model also emphasizes the impact of particle solubility on the total integrated dose with time. Brain and Valberg (1974) developed convenient nomograms and a computer program to calculate how particle solubility and particle size signi8cantly affect the net amount of particulates retained in the lungs. Although the basic outline of the model is generally correct, more recent measurements suggest that values for alveolar deposition of particles 0.1-1.0 µm are too high by a factor of at least 2 (Heyder, 1982). The extent to which ETS particles are hygroscopic and increase in size within the reepi- ratory tract is an important and unresolved issue that adds further uncertainty. Aerosol deposition has also been studied in airway casts. Phys- ical models of the upper airways of human lungs have been made by a double-casting technique to study particulate deposition at several airway generations (Schlesinger and Lippman 1972). llif- ferent flow rates and particle sizes were used to study deposition patterns. Schlesinger and Lippman (1978) reported a correlation between the deposition sites of test aerosols in their lung casts and the most common sites of origin of bronchogenic carcinoma in smoking humans. Both occurred preferentially at bifurcatiuns. Martonen et al. (1983) added an oropharyngeal compartment and a replica cast of the larynx to the tracheobronchial casts in or- der to better simulate airflow patterns in the upper respiratory tract. They used these models to evaluate the amount of cigarette smoke condensate deposited in the airwayn at different flow rates. More condensate was present in areas where airways branched and especially at the bifurcation points, indicating increased lev- els of impaction. Aerosol was also deposited preferentially along posterior airway walls of the branching regions. Hiller et al. (1982a) measured the collection efficiency in adults of an aerosol containing three different sizes of polystyrene latex 125 spheres In nonsmoking humans. They measured a 10% deposition for 0.6-µm (MMAD) spheres, which is similar to the results of Davies et al. (1972) and Muir and Davies (1967) using 0 b-µm aerosols and Heyder (1982) using aerosols that were 0.2 to 1.0 µm in size. The size ranges of these aerosols are comparable to those experimentally measured in cigarette smoke, as previously discussed. In contrast to passive smoking, the estimates of the collection efficiency of smoke particles during active smoking are substan- tially higher (about 70%) for at least two reasons (Hiller et al., 1982b). First, the much higher particulate concentrations in main- stream smoke may give rise to more agglomeration and greater hy- - droscopic growth in the respiratory tract. Both processes produce larger particles with higher collection efficiencies. Second, and more important, the breathing pattern used by the active smoker _ is markedly different than normal breathing. It is characterized by a slow deep inspiration followed by breath-holding. This in- creases the average residence time of the smoke particles and thus _ increases the fraction of inhaled particles that deposit in the lung. To compare the amount of smoke deposited in the lungs of an active smoker with an individual exposed to ETS, first consider a pack-a-day smoker (about 20 cigarettes during an 8-hour period). The average tar rating in mainstream smoke (MS) over the past couple of decades has been about 14 mg/cigarette. Therefore, the total amount of tar inspired is 280 mg/8 h. Assuming a collection efficiency of 70%, the amount of tar deposited is 196 mg/8 h. As pointed out in Chapter 6, smoke particles can range from -- 50 to 500 µg/ms in public places where smoking occurs and from 20 to 1b0 µg/ms in homes with smokers. Consider a nonsmoker who - -- - breathes at 10 L/min, or 4,800 L/8 h. With modest exercise, this could increase to 20 L/min, or 9,600 L/8 h. Based on estimates by Hiller et al. (1982a,b), the collection efficiency of particles in - ETS is about 10 .:. Therefore, the total amount of smoke particles deposited in a nonsmoker in these environments for 8 h could range from approximately 0.0096 mg/8 h = 20 µg/ms x 4.8 m'/8 h x 0.1 to an extreme of 0.5 mg/8 h = b00 µg/me x 10 ms/8 h x 0.1. This would be approximately 0.005% to 0.26% of that amount of tar deposited in the active smoker's lungs after smoking 20 cigarettes. The active smoker, of course, also breathes the ETS, so that the total dose received by the active smoker is the mainstream smoke plus a passive smoking dose equivalent to that received by the 5908L48

126 nonsmoker exposed to ETS. However, since the closo received due to breathing ETS-contaminated air is so sinall, this additioni.l contribution to the total dose is negligible. , Benzo[alpyrene (BaP) is one of the primary constituents (if particles in mainstream smoke. From Table 2-10 one can estimate that a nonsmoker exposed to ETS receives a higher relative dose of BaP than of RSP. However, the ambient measurements, which are used to estimate the dose for the nonsmoker, may be clevated in view of the high outdoor concentrations that are reported in these studies. More data on the fate of BaP in ETS and on ambient concentrations are needed before estimates of the relative doses - -- - can be made meaningfully. Although the amount of smoke deposited in the lungs of non- smokers during exposure to ETS is small compared with that . encountered by tho active smoker regarding mainstream smoke, it may differ in composition and toxicity. For example, as discussed in Chapter 2, certain constituents are present in much higher con- centrations in sidestream smoke as compared with mainstream smoke (Weiss et al., 1983). These possible differences in composi- tion must be explored. PARTICLE itETENTION IN THE LUNGS The amount of p- articles present at different sites in the lungs is not only dependent on deposition. Retention of smoke depends on the_ balance between the amount of each constituent that deposits in the respiratory tract and the efficiency of the lung clearance mechanisms in the airways and alveoli. Clearance mechanisms are a dynamic component of normal lung function and operate to keep the lung clean and sterile. Particles depositing in the airways are entrained in the mucus layer that lines the airway. This layer is swept toward the mouth by the action of ciliated cells and eventually swallowed. Mucus transport is approximately 1-Z cm per minute in the trachea, but is slower in smaller airways. In addition, macrophages present in the airways may phagocytose deposited particulates and be carried towards the mouth by the mucociliary transport system. Particulates reaching the alveolar region-those that are usually less than several micrometers-e are soon engulfed by alveolar macrophages. Some of these cells gradually migrate towards the airways and exit the lung via the 127 mucocillary escalator. Dissolution of particles is an additional important clearance mechanism. Lung disease and cigarette smoking itself can affect particle clearance and retention in smokers' lungs. Previous studies have shown that smokers have differgnt aerosol deposition patterns and - slower clearance rates than nonsmokers (Albert et al., 1969; San- - - -- chis et al., 1071; Cohen et al., 1979). These alterations in clear- -_ ance are, in part, caused by components within cigarette smoke that affect the quantity and rheological properties of the mucous. Components of cigarette smoke, also, can impair phagocytosis by alveolar macrophages (Ferin et al., 1965). Clearance mechanisms in smokers may be further compromised by lung diseases, such an - emphysema and fibrosis, and by exposure to other air pollutants. Measurements of the long-term retention of compounds as- - sociated with cigarette particulates in the lungs are di[ficult to. - estimate from data obtained with airway casts or from differences between inhaled and exhaled aerosol concentration, since these methods do not take into account clearance mechanisms. Un- fortunately, few data are available regarding the actual retention and sites of deposition of cigarette smoke particles in either non- - smoking humans or animals exposed to ETS. The most accurate - method that could be used is quantification of particulate deposits in individual pieces of tissue dissected from the lung. Impossible in living animals, this is a tedious procedure in animal lungs or human material obtained at surgery or autopsy and is especially difficult for large lungs. One can also attempt to quantify dose by examining saliva, serum, or urine. These possibilities are discussed in Chapter 8. GASES IN ENVIRONMENTAE TOBACCO SMOKE In addition to the particulate phase, we must also consider -- - - exposure-doso relationships for gases in L'TS. As before, breathing pattern influences gas uptake. Of particular importance is the difference between oral and nasal breathing. Breathing by mouth increases the exposure of the airways, while breathing by nose (as would be true for nonsmokers exposed to ETS most of the time) e) o(fers some protection for the lower respiratory tract. The most important variable determining the amount and site - - - of uptake is the water solubility of the gas in question. Gases that are highly soluble in water, such as formaldehyde or acrolein, will 99E96446

128 be almost completely removed by the upper respiratory tract, es- pecially during nasal breathing. The concentration of other garies, such as the oxides of nitrogen, which have an intermediate solu- bility, will decrease as the inspired bolus henetrates deeper and deeper into the lungs. There will be uptake of gas in the upper airways, but signilicant amounts will also penetrate to respiratory bronchioles and alveoli. Finally, there are gases of low solubility, such as carbon monoxide. No significant uptake of CO occurs - - in the upper airways, and it is only slowly absorbed across the air-blood barrier. In the absence of heavy exercise and very high ' ventilation rates, many hours are required to establish an equilib- rium between inspired CO and carboxyheinoglobin in the blood. As was true for particles, we can estimate the gas uptake for active smokers and for passive smokers. As reviewed in Chaltter 2, CO from ETS can range from less than i to 8 ppm. If the background air has little or no CO, even the upper estimates of 8 ppm will have a negligible effect on carboxyhemoglobin levels. Almost 2 hours would be required to reach 1 % carboxyhemoglobin (Peterson and Stewart, 1975). This is approximately the same as background levels of carboxyhemoglobin, which are associated with endogenous production of carbon monoxide. Even after 15 hours, when the equilibrium value of 1.7% OOllb is fiually reached, the effect should be insignificant. However, if air pollution from mobile and stationary sources produces higher background levels of CO, then an incremental exposure of 1 to 8 ppm could produce some added burden of carboxyhemoglobin. Reactive or highly soluble gases such as formaldehyde, acro= lein, or oxides of nitrogen present a different situation. Acrolein has a very high water solubility (40 g/100 ml). Because of thie high solubility in the airway lining fluids, one would anticipate a collection efficiency approaching 100%. Moreover, this would occur rapidly, so that acrolein is classiRed as an upper respiratory tract irritant. According to Table 2-10, there are between 60 and 100 µg of acrolein generated per cigarette. Thus, from 20 cigarettes, 1.2 to 2.0 mg of acrolein would be deposited in the respiratory tract of the active smoker. - Chapter 2 suggests that levels of acrolein in public places where smoking is permitted could range from 10 to 50 pg/m'. Using similar wtsumptions to that made for particles, we estimate that the nonsmoker would inhale 4.8 to 10 rn3 of air per 8 hourt. Assuming a collection efficiency of 100%, the total amount of 129 acrolein deposited in the passive smoker would be approximately 0.048 to 0.1i mg. We select 1.6 mg/8 hours as the mid-range dose for the active smokers, which assumes 20 cigarettes smoked per 8 hours with 80 pg acrolein per cigarette. Using this value, the nonsmoker exposed to ETS for 8 hours would then receive approximatuly 3 to 31% of that received by the active smoker. When the contribution of ETS is included for the active smoker, the nonsmoker exposed to ETS for 8 hours would receive between 5+utd 24% of that of an active smoker. The relatively high dose of acrolein received by the nonsmoker reflects the high collection efltciency for this hydrophilic component and the persistence of vapor-phase components in the air even when filtration is used. Table 2-10 gives comparisons of the amount of other materials inspired for both active smokers and individuals exposed to ETS over shorter periods of time. SUMMARY AND RECOMMENDATIONS A number of studies have measured the levels of specific con- stituents of ETS under natural conditions (reviewed in Chapters 2 and 5). The extrapolation from relative exposures to relative doses received is difficult. Variation in the percent of time individuals --- - spend inn particular environments such as home, workplace, and so forth, and the variations in uptake and clearance, discussed in this chapter, will affect the actual dose received. - Using a simple, first-approximation model for exposure and retention, the relative daily dose received for a nonsmoker exposed to ETS can be compared with the dose received by an active smoker. - For RSP, the estimates were up to 0.26%. For acrolein, a hydrophillic, vapor-phase constituent, the relative dose is es- timated to be much higher, 3 to 31%, whether or not the ETS exposure of the active smoker is considered. Nicotine, another constituent that appears primarily in the vapor phase of ETS, has an estimated relative dose of up to 1% (see Chapter 8). The extent to which these are Indicative of the relative ex- posures -- posures to specific constituents that are important for particular health effects in active smokers or in nonsmokers exposed to ETS cannot be determined for any of the health elfects reviewed later - in this report. Nevertheless, the estimated relative exposures give L969g4L8

130 some idea of the potential range of relative exposures, for conatitu- tents that are found both in the vapor phase and in the particulate - phase. Because of the range of estimated relative doses, it would be ideal_ to make estimates of the relative dose based on the spe- cific constituent(s) that are most relevant to the health effect being assessed. However, many of these specific constituents, for instance the carcinogenic constituents such as benzo(nlpyrene, N-. - nitrosodimethylamine, and N-nitrosodiethylamine, are difficult to measure; therefore, there are not enough data available to make meaningful estimates of the relative doses of these constituents. Also, biological markers might be potentially informative indica- tors of the relative doses. However, as reviewed in Chapter 8, to date only carbon monoxide, nicotine, and cotinino have been measured extensively in humana. What Is Known 1. Particle size and breathing pattern are critical factors in the deposition of ETS in humans. 2. Theoretical models predict that 30 to 40% of the particles with the size range present in cigarette smoke will deposit in the alveolar region and 5 to 10% in the traeheobronchial region. 3. The collection efficiency of smoke particles during active - smoking has been measured to be about 70%. On the other hand, the collection efficiency is estimated to be ouly 10% for nonsmokers - exposed to ETS. What Scientific Information Is Missing 1. Actual measurement of regional deposition of cigarette - smoke particulates in human lungs is not available. 2. There are little data regarding the actual retention and sites of deposition of ETS particulates in either humans or animals. 3. The concentrations of various components in vapor and _- particulate phases of MS and L`PS differ. Consequently, research is needed, particularly for vapor-phase components, to see how these differences affect dose. 9JU9t3f.4MF3 I 131 REFERENCES Albert, R.E., M. Lippmann, and W. Briscoe. The characteristics of bronchial clearance in humans and the effect of cigarette smoking. Arch. Environ. Health 18:738-766, 1969. Bennett, nett, W.D., M.S. Messina, and a.C. Smaldous. Effect of exercise on deposition and subsequent retention of inhaled particles. J. Appl. Physiol. 69:1048-1064, 1986. Brain, J.D., and P.A. Yalberg. Models of lung retention based on ICRP Task Group report. Arch. Environ. Health 28:1-11, 1974. Cohen, D., 8.F. Aral, and J.D_ Brain. Smoking impairs long-term dust clearance from the lung. Science 204:614-617, 1979. - Davies, Q.N., J. Heyder, and M.fl. Subba Ram.. The breathing of half- micron .ero.ols. 1. Experimental. J. Appl. Physiul. 34:691-800, 1972. iPeriu, J., 0. Urbankova, and A. Ylokova. Influence of tobacco smoke on the' - elimination of particles from the lungs.. Nature 206:616-618, 1966. Harhison, M.L., and J.D. Brain. Effects on exercise of particle depositio_ n in Syrian goklen b_ amsters. Am. Rev. Respir. Dis. 19-8t904-908, 1983. - Heyder, J. Partiele transport onto human airway surfaces. Eur. J. Respir. Dis. 63(Suppl. 119):29-60, 1982. Hiller, F.(Q., M.K. Masnmder, J.D. Wilson, 1-'.Q. McLeod, and R.C. Bone. Human respiratory tract deposition nsing multimodal aerosols. J. - -- - Aerosol. Scl. 13:337-343, 198Z.. Hilier, F.Q., K.T. ivtcCusk.r, M.K. Masumder, J.D. Wilson, and R.C. Bone. Deposition of sidestream cigarette smoke in the human respiratory tract. Am. Rev. Itespir. DL. 125:408-408, 198Zb. Hinds, W.O. Sise characteristics of cigarette smoke. Am. lnd. Hyg. Assoc. -- -- -- J. 39,48-64, 1978. International Commis.ion on Radiological Protection (IC_RP), Task Group -- of Committee 2 of the International f3ommission on Radiological Pro- - - - tection. Physiological data for reference man, pp 348-347. In ICRP. Report of the Task Group on Reference Man (ICRP 23). New York: Pergamon, 1976. Keith, C.H., and J.C. Derrick. Measurement of the particle size distribution -- - -- - and concentration of cigarette smoke by the `conifuge! J. Colloid Sci. - 16:340-358, 1960. - Martonen T.B., and _J.E. Lowe. Cigarette smoke pattern in a human res- piratory tract model. Proc. Ann. Conf. Eng. Med. Biol. 25:171, 1983 (abstract). McCusker K., P'.A. Hiller, J.D. Wilson, M.K. Masumder, and R. Bone. Aero- dynamic sising of tobacco smoke particulate from commercial cigarettes. Arch. Rnviron. Health 38:316-318, 1983. Muir D.O.F., and O.N. Davies. The deposition of 0.6 µm diameter a_ eros_ o_ b_ in the lungs of man. Ann. Occup. Hyg. 10:181-174, 1987. Petorson, J.E., and R.D. Stewart. Predicting the earboxyhemogiobin levels - resulting from carbon monoxide exposures. J. Appl. Physiol. 39:633-638, 1976. Sanchis J., M. Dolo.kh, R. Chalmers, and M.T. Newhouse. Regional distribution and lung clearance mechaninus in smokers and non-smokers, - -- - --- pp. 183-191. In E.H. Walton, Ed. Inhaled Particles, Part 111. Surrey, E-nglLndi Unwin Brothers Ltd., 1971.

132 Schlesinger R.B., and M. i.ippmann. Particle depo.io~E i ~E3s 2a7 Z51~ 1972. upper tracheobronchiai tree. Am. ind. Hyg. Schlesinger R.B., and M. Lippmsnn. Selective particle deposition and bron- chogenic carcinoma. Environ. Res. 16:424-431, 1978. Task Group on Lung Dynamics. Deposition and retention models for internall dosimetry of the human respiratory tract. Health Phys. 11c17g-207, - -- 1966. Weiss S T., I.B. Tager, M. Schenker, and F.E. Speiser. The health effects of involuntary smoking. Am. Rev. Respir. Du. 1Z8:933-941, 1983. G9(;9gZ4i9 __- O Assessing Exposures to Environmental Tobacco Smoke Using:Biological Markers Previous chapters have dealt with the formation and compo- sition of tobacco sidestream smoke, its contribution to environ- mental mental tobacco smoke (ETS), and the conditions that govern the physicochemistry and toxicity of E`1'S. Personal monitoring of ex- poeure and analysis of the respiratory environment enable us to estimate the level of toxic agents for individuals exposed to ETS. Studies on the uptake of smoke constituents by individuals and on the metabolic fate of such constituents can provide informa- tion relative to epidemiologic observations and the actual exposure ~ levels of different populations. Exposure to ETS may depend on several factors, including . : the number of smokers in an enclosed area, the size and nature of the area, and the degree of ventilation. '1'hus, optimal assessment of exposure should be done by analysis of the physiological fluids of exposed persons rather than by analysis of respiratory environ- - menL. The development of new biochemical methods enables us to obtain measurements of exposure to ETS by determining the uptake of speciBc agents in body fluids and calculating the risk relative to that of the exposure of active smokers. The uptake of individual agents from ETS can be determined by biochemical -' measures that have been developed for assessment of active smok- ; ing behavior, as long as these measures are sensitive and specific : enough for quantitating exposure to such ageits by nonsmokers. 133

134 BIOLOGICAL MARKERS IN PIIY9IOLOGICAL FLUIDS Tblocyanate The hydrogen cyanide (HCN) absorbed from tobacco smoke is detoxiHed in the liver, yielding thiocyanate (SGN'). However, SCN- in serum and other. biological fluids does not exclusively originate from inhaled tobacco smoke. Thiocyanate also can be derived from the diet (IIaley et al., 1983; Jai-via, 1985). Before 1975, primarily two celorimetric methods were used for the manual determination of thiocyanate in biological flu- ids (Aldridge, 1944; Bowler, 1944). Subsequently, the automatic method by Butte et al. (1974) has found wide application in coin- paring physiological fluids from smokers and nonsmokers. It un- tails determination of thiocyanate by its reaction with ferric ions, which yield a color complex with maximal absorbance at 460 nin, the intensity of which can be measured in an autoanalyzer. In sera of nonsmokers, Butts et al. (1974) determined up to 95 µmol/L of SCN-. The critical value in differentiating between smoknre and nonsmokers was 85 pmol/L of SCN-. ln other investigations, 100 µmol/L of SCN- was found to be the critical level for serum (Junge et al., 1978) and for saliva (Luepker et al., 1981). This fact and the low concentrations of HCN in ETS (Holfmann et a1.,19kl4) explain why some investigators were unable to distinguish between nonsmokers exposed to ETS and those without any exposure to tobacco smoke (Hoffmann et al., 1984; Jarvis, 1985). Similarly, the mean serum level of SCN- in healthy pregnant women at term who were exposed to 1•:TS (35.9 pmol/L) was not distinctly different from that in those without E'rS exhosure (32.3 µmol/L), nor was there a measureable difference in SCN' levels in the umbilical cords of the neonates (2E3 versus 23 µmol/L) (Hauth et al., 1984). In one study, it appeared that there was a trend toward higher thiocyanate levels in the saliva of nonsmoking children residing with smokers compared to the SCN' levels in saliva of children without ETS exposure, yet this trend was insignificant (Gillies et al., 1982). In a stud- y of six volunteer nonsmokers exposed to a smoke-filled room for 4 hours, there was a significant iucrease in salivary SCN-. However, the SCN- values of the nonsmokers 136 exposed to 1i:TS were not distinguishable from those nonsmokers kers free of tobacco smoke exposure (Pekkanen et al., 1976). In another study, mean serum thiocyanate levels were reported to be significantly higher (p < 0.002) for children and adolescents with exposure to cigarette smoke at home (n = 14; SCN- = 97,3.f 45.4 ptnol/L) than for those not exposed (n = 10; SCN- = 54.2 ± 11.3 µmol f L). The authors of tho_ latter study also reported a weak correlation between thiocyanate concentration and number mber of cigarettes smoked per family (Poulton et al., 1984). This study was criticized because some of the determined thiocyanate levels were within the range reported for heavy cigarette smokers. It is likely that there was deceptive reporting of adolescent smoking status (Jarvis, 1985). Based on the observations to date, the level of thiocyanate in saliva, serum, and/or urine is not o_t useful as an indicator for the uptake of ETS by a nonsmoker. CarLon Monoxide and CarUoxyhomoglobiu Carbon monoxide (CO) in the body originates from endoge- nous processes as well as environmental sources. The endogenous production of CO is primarily a consequence of the breakdown of hemoglobin and of other heme-containing pigments. Healthy adults produce about 0.4 ml of CO per hour (0.5 mg/h; Coburn et al., 1964). This provides the major portion of CO that is found as carhoxyhemoglobin (COHb) in nonsmkers. Iu nonsmokers with- out occupational exposure to CO, COHb ranges from 0.5 to 1.5% (National Research Council, 1981; Wald et al., 1981). The inhalation of CO from the environment is followed by an increase of the CO concentration in the alveolar gas and by diffusion froni the gas phase through the pulmonary membrane into the blood. CO is complexed with blood to form COHb and, - as such, is transported throughout the body. Complexing it with hemoglobin occurs with a strong coordination bond with the iron of home, a bond that is about 200 times_ stronger than that with molecular oxygen. CO is only slowly released from the blood in the process of exhaling. In the case of nonsmokers who_ have been exposed to elevated levels of CO in the air for a few hours, the half-life of COHb lasts 2-4 hours (National Research Council, - 1981). Monitoring of absorbed CO in the blood is done primarily by the analysis of CO in alveolar gas and by the analysis of COHb OLFi9Q44g

136 in blood. The most widely used technique in the clinical labo- ratory is the determination of COHb with automated differential spectrophotometry (National Research Council, 1977). The deter- mination of CO in exhaled air by standardized gas analyzers has been used less frequently. However, the portable 'Ecolyzer' and other similar instruments have proved to be reliable instruments - for the recent validations of the reported smoking habits among populations in field studies (Vogt et al., 1979). The data from both measurements, amount of CO in the alveolar gas and the concentration of COllb in blood, are well correlated. Theoreti- cally, the slope of the graph relating the percent of concentration of COHb to alveolar CO should be about 0.1511 at CO concentra- tions tions of 0-50 ppm. Most laboratory studies have confirmed this - correlation experimentally (National Research Council, 1981). In the case of cigarette smokers who have inhaled puffs of smoke con- taining 20,000-50,000 ppm of CO, the correlation between exhaled CO and COHb is also in good agreement (r = 0.97; Heinemann et al., 1984). The COHb levels are of value for comparing degrees of smoke inhalation. In a study of inen aged 34-64 years, cigarette smok- -- ers had on the average 4.7% of COHb; cigar smokers, 2.9%; pipe smokers, 2.2%; and nonsmokers, 0.9% (Wald et al., 1981, 1984). However, measurements of exhaled CO or COHb are not valid in- dicators of chronic exposure to ETS. A study of 100 se1[ reported nonsmokers who weere divided into four groups-without exposure to ETS, with little, with some, and with a lot-revealeci no sig- nificant differences in measurements of expired CO (5.0-5.7 ppm; mean, 5.61 f 2.70 ppm) or COHb (0.80-0.SM°~6; mean, 0.87 :L 0.67%) (Jarvis and Russell, 1984). This observation is also sup- ported by a study of six nonsmoking flight attendants who served in the smokers' section of a trans-Pacific aircraft. Preflight COHu levels_ were 1.0 t 0.2% and postflight levels (after serving round- trip) were 0.7 ± 0.2% (Foliart et al., 1983). Heavily smoke-polluted environments can lead to elevated ab- sorption of CO. Thie was shown for seven nonsmokers exposed for 2 hours in a pub, whose exhaled air revealed an average of 5.9 ppm of CO, a level that corresponds to the alveolar gas of a smoker after smoking one cigarette (Jarvis et al., 1983). Another study showed that twelve nonsmokers, eliming the nonaircondi- tioned environment of a room with four smokers who smoked four cigarettes each within 30 minutes, had an t;OIlb increase -of the T1.G981.118 I 137 same magnitude as that measured in a smoker after consuming one cigarette (Huch et al., 1980). Even.though tobacco smoke is a major source for indoor air pollution, additional sources rces may contribute to increased CO con- centrations ui air and, consequently, to higher COHb levels in exposed subjects. Such sources include gas stoves, faulty furnaces, and space heaters (National Research Council, 1981). For exam- ple, kerosene heaters can be a major source for indoor pollution. Depending on the model and flame setting, kerosene spaco heaters generate up to 6.5 mg of CO per minute of operation (Leaderer, 1982). In summary, CO in alveolar air and as COHb in nonsmokers ---- originates from endogenous processes as well an from environmen- tal - --- tal sources. ETS is an important pollutant of indoor environments; _- however, except for highly polluted settings, CO levels in exhaled air and COHb levels in the blood are not statisically significantl elevated following exposure to ETS, although acute short-term exposures from 3-4 hours may be detected if blood or expired air is sampled within 30 minutes of the end of exposure. In sum, how- ever, measurements of exhaled CO and of COHb. are not useful - indicators of exposure to ambient ETS except in acute exposure studies in the laboratory. CO measures are a markeF of gas-phase exposure to ETS. Nicotine and Cotlnine Disregarding nicotine-containing chewing gum and nicotine aerosol rods an aids for smoking cessation, the presence of nicotine _ and that of its major metabolite, cotinine, in biological fluids is entirelY due to the. ex - posure to tobacco, tobacco smoke, or environmental tobacco emoke. The determination of nicotine and cotiuine in saliva, blood, or urine of active and passive smokers is done primarily by gas chromatography (GG) with a nitrogen- sensitive detector and by radioimmmnoassay (RIA). The GC method requires great precaution in order to avoid contamination by traces of nicotine from the environment or from solvents and/ar equipment. This is of major importance for sam- ples containing nicotine at levels <20 ng/ml of fluid, as is the case in nonsmokers exposed to ETS (Feyeraband and Russell, 1980). The GC method can be used to measure concentrations of nicotine as low as 1 ng/ml and concentrations of cotinino as low as 5 ng/ml

138 in samples of physiological fluids (Jacob et a1., 1981). An expori- enced chemist can analyze up to 25 samples per day for nicotine and cotinine. The radioimmunoaseays for nicotine and cotinine represent probably the most direct technique available. These assays have only low cross-reactivities with other naturally occurring metabo- lites of nicotine. The sensitivity of these assays is about 0.6 ng/ml for both nicotine and cotinine and has inter- und intra-assay vari- ations of ±5% (Langone et al., 1973; Hill at al., 1083). An experi- - enced biochemist with automated equipment can analyze up to 8l/ samples (plus 20 control samples) per day. So far; the RIA method has been used by a limited number of laboratories because it re- quires the synthesis of specifc nicotine and Eotinine derivatives for the generation of serum albumin conjugates and the raising uf antibodies to these conjugates (Langone et a1.,1973). In addition, - the RIA method also requires careful drawing and handling of samples to avoid contamination. Table 8-1 presents results from the major studies on the up- take of nicotine by nonsmokers under acute exposure conditions. These data show that exposure to high levels of LT-S in labora- tories can lead to a significant uptake of nicotine. This uptake is clearly reflected in the concentrations of nicotine in plasma (up to 0.9 µg/mI for nonsmokers compared with a mean value of 14.8 µg/ml for smokers, an increase oE 15-[old) and in urine (84 ng/Inl for nonsmokers, compared with 1,750 ng/ml, a increase of 20-fald) (Russell and Feyeraband, 1975; Hofl'cnann et al., 1984). The sig- nificantli higher values for nicotine in the plasma compared to urine ma; be explained by the short initial half-life in smokers of ard ra1`+-ct?7 s!:c=: Lerm',--al ha!f-life in s=kers of 2 eader dzlj liie conario-.:... Wh"a -t er_;l?=bf Matsukurs at al. (1984), the data demonsLrate that the insoluntart expasure of the passive smoker amounts to a few, percent or lew :f ':2it 3, _`^cr_at eE ztatiL-e that is inhaled by a cigarette smaLK; as &tetDOimA --- ~ 6- \' Z4fi9Bzz8 I,. 139 C [q~V Q _ U ^~Wi~ C n'C zlo~-S~?i d~ri W 1? ~ r ' e. a ~ ° y a e -H +i .\ ~~ic~~'^a6 ~-H~~ L~ s ~ G ~ g I I g ~ ~ w sA 4 A~AF~AAESAA~oii'~q~z z° N ~ i R

140 s i i ~8g.~$' .~ g ` 1:e~o~i`oa~ ~ ~ 0 I $ ~39~~pC~i C p p 1 _d F ~' ~° a e F e e~i r, m-±m ., :eiF3 ~ 6r E 141 to lese than 1% of the mean values observed in physiological fluids of active smokers, even though some nicotine measurements in plasma give a higher reading (Jarvis et al., 1984). In a large-scale study of 839 nonsmokers (identified by their questionnairo response and also having a cotinine concentration of <20 ng/ml of saliva), cotinine levels increased with the number of smokers in the home for each of three age groups examined inde- pendently (<5, 6-17, and >18 years). The cotinine levels in saliva were found to be significantly associated with increasing number of smokers per household within each age group. The median salivary cotinine levels in adult smokers was 287 ng cotinine/ml (Coultas et a1.,1986). Matsukura et al. (1984) report that cotinine in the urine of ETS-exposed nonsmokers reaches an average of 1.56 f 0.57 µg/mg of creatinine when 40 or more cigarettes per day have been smoked in the home of the exposed subjects. In the case of cigarette smok- ers, they found cotinine levels of 8.57 f 0.39 ug/mg of creatinine in urine. This study has been questioned because its findings of cotinine in urine of both active and passive smokers indicate levels substantially higher than those reported in other studies (Adlkofer - et al., 1985; Pittenger, 1985) (see Ghapter 12). Nicotine uptake by infants of cigarette-snioking mothers ap- pears to be higher than is generally observed for the adult non- smoker. The amount of cotinine excreted in the infant's urine ---- - has been found to be correlated with the nuiryber of cigarettes smoked by the mother in the 24 hours preceding the measurement (Greenberg et al., 1984). The analysis of nicotine and cotinine in physiologic fluids can be misleading if made on very light smokers or nonsmokers who either sniff tobacco or are tobacco chewers or snutF dippers. In the case of the very light smker, nicotine and cotinine values may be similar to those of nonsmokers who had exposure to high levels of ETS (Russell and Feyerabend, 1975; Wald et al., 1984). In the case of individuals who use tobacco nasally, or orally, on a regular basis, the nicotine and cotinine values may approach those of heavy cigarette smokers (Russell et al., ]980; Russell et al., 1981; Palladino et al., in press). In both groulm, the analysis of COHb will reveal that these subjects are light smokers or nonsmokers, respectively. However, nicotine and cotinine levels for such persons are clearly not valid for the detennination of their exposure to ETS.

~ z,tnl;e'8-2 Nicotine Uptake by'Nonsmokers 'F_x,posed to ETS Under Daily Life'Conditions° ~ No. of ~ Group of Nonsmokers ~ Authors Nonsmokas Examined Results ~ ~ Russell and' Feyerabend. 1975 Hospital eeqployees It aiter' lunch) i 't oll ri ng/ml on , ect ne c (u p' I (a) Grou 14 Nicotine in,urine• 12.4 tilbl9 ' , i(b) Group 2 13 Nicotine in urine: 9.1 8.9 t attentt d'out l ng/mi - Fa7erabend tst al.. 1982 p oyees an liospitYl emp (a) nonexposed to; ETS during the marning 30 Nicotinel in the uiitte: ' (self report) Nicotiae'in salica: 5.9 _ 4.4 poxdi to ETS during the awraing (b) ex Nicotinr in urine: 21.6 ± 28.9 ' , (seJt tepott) Nicotine in srliwa: 9.7 10A ± d 6 Nicotine in nerum: ng/ml Foliart et al.. 1983 Flight atten ants (San 1 Francisoo-To$lro-'San Francisco) (a) before flight L.6 ,(b) alterlnight 3.2 ± 1.0 atients d ff ng/enl Wald aial.,'1984 outp an Hospital sta (a) ', nonexposed to E7LS 22 Cotinine in urine: TOW~) (b) exposed tolE'f5 (seN'tcpott) 1199 Cotinine in urine: &0 (1.4-22i0) ng/ml 1984 Wrld and Ritehie (a) husbands of nonsmokars 101 Cotinine in urine: 8:5 t 1.3 . (b) husbands dt smokers 20 Cotinine in urine: 253't 14.8 ' pie sa Ef i 7 ng lad Before After 1983 Jan.is at al.. ice, n: an o m Employees ' 1 11 (Il') collection at 111:30a.m.'(1)'and 7:45'p.m. (timc bctweeni collections' inclttd'mg 2-h stay Niootine in plasma: 0.76 2.49 in smoking "pob") Nicotine in sali.a: 1.90 43:63 Nicotiae in urine: 10'S11 92.63 Cotinine in plasma: 2.07 7.33 Cotinin+e', in saliva: 1.50 8:04 Cadains' in otiac 4.80 12:9a • (Diffaences'betwxn'I and II ate stitisticaDy . hi8hh' s±gnificanr',P values tange'ftom Jarvis et a1..,1985 Nonsmoking school children (11-16-yr-old) <0.01 to <0.001) ng/mll I. Npitlta parent smoked 269 Cotinine in saliva: -' . 0.44 :t II. pAly father smoked 96 Cotininr in saliva: - 1.31 t 1.21 III. ' Only motheo smoked ' 76 Cotinwei in saliva: 1.95 t 1 71 Matsukura et al.. 11984 IV. Bothipatsaa, sawl:ed 472 nonsmokers Urine collection in: the mornieg 128 Cotiniae in saliva: . 3.38 * 2.45 (a) smokers mbome 272 µg cot./mg ctat. 10.79 t 0 11 (b) nonstttokees'in home Cigatetta smoked per day in home of nonsmokers 200 ~ . !0.51 t:0:09 1-9 25 µg cot./mg crest. 0 31 y- 0 08 10-19 57 . . ' 0.42 ± 0.10 20-29 99 0 87 ± 0;1'9 30.39 38 . 1 03t0.25 >40 28 . 1.56 t 0.37 Greenberg et'al.,'1984 Unspecified Infants under 10 months of age 1 (not breastfed) 25 0-56=0.16 (a) not, exposed to En I8 Urine ng nic:/tng creat. 0(0=59) Urine ng nic./mg atat. 4(0-045) Saliva ng nic./mg crest. 10 (0-3) (b) exposed to ETS 28 Urine ng nic./mg xteat. 53'(0-370) Urine ngicot./mg creat. 351 (41-1,885) Saliva ng nic./mg ereat. 12.T,(0-166) Saliva ng cot./mg ctpt. 9'(0-25) 'Albbrevfations: cot., cotinine; ctr.at„ creatinine; nic., nicotine.

144 TAULa 8-3 Approximate Relations of Nicotine as a Paramster Between Nonsmokers, Passive Smokers, and Active Smok4rs" Nonsmoken wilhout Nonsmokers with E'fS Exposure ETS Exposure Active Smokers 94 No. _ 46 No. = 54 No. - St% of Active % of Active, ' Nicotine/Cotinine Mean Value Smokers' Value Mean Vaiue Smokers Value Mean Value Nicotine (ng/m1I in plasma 1.0 7 0.8 5.5 14.8 in saliva 3.8 0.6 5.5 0.8 673 in urine 3.9 0.2 1z.1• 0.7 1.750 C i i ( /mU otnne_ng in plasma 0.8 0.3 2.0* 0.7 275 in saiiva 0.7 0.2 2.3'0 0.8 310 in urine 1.6 0.1 7.700 0.6 1,390 'Differences between nonsmokers exposed to ETS compared with nonsmokers without ex- posure:'p < 0.01; a'p< 0.001.' SOURCE: lervis ci al., 1984. Cotinine elimination in the plasma of nonsmokers exposed to ETS was reported to be slower than cotinine elimination in the plasma of active smokers. Cotinine elimination from urine was also significantly slower. In a study of 10 chronic smokers and 4 nonsmokers experimentally exposed to ETS, the half-life of elimination of cotinine from plasma was 49.7 hours in nonsmokers and 18.5 hours in smokers (Sepkovic et al., 1986). Disappearance of cotinine from urine was also significantly slower in nonsmokers than in chronic smokers (32.7 hours vernus 21.9 hours). These preliminary data need to be considered when using cotiliino to quantify the dose in nonsmokers exposed to ETS. In summary, the determination of nicotine and, especially, of cotinine in saliva, blood, and/or urine of nonsmokers -exposed to ETS represents at present the most appropriate assay for estitnat- ing long-term (average daily) exposure. However, venipuncture needed to get serum samples is often impractical, if not irnpos- sible. The use of saliva for nicotine and cotinine assays, doNpite some advantages, also has certain inherent weaknesses, such as uncharacteriatically high readings immediately after heavy E'.l'S exposure and the need to wait several hours after exposure fiir the -q4698449 145 cotinine concentration to stabilize (Hoffinann et al., 1984). Saliva is a particularly erratic source on which to make nicotine measures. - Urinalysis for cotinine is the preferred method for assessment of long-term ETS exposure, because the sampling is noninvasive, the excretion rate of cotinine is only slightly dependent on the pH of urine, and assessment of the average daily exposure on the b$ sis of cotinino levels is independent of the restrictions posed by variations of the half-life of nicotine in smokers and nonsmokers - (Beckett et al., 1971; Klein and Gorrod, 1978). Creatinine-Reference Compound for Urine Analysis Urine sampling does have some associated problems. Often it is impractical to collect 24-hour urine samples for the analysis of - biological markers of direct exposure to tobacco smoke or to ETS unless undertaken under strict medical supervision, such as in a metabolic ward. In this case, the ratio of biological markers to creatinine is often used to allow for variations in fluid intake (and - excretion (see Table 8-1). Creatinine excretion varies from person to person, but the daily output for each individual is almost constant from day .to - day. Urinary creatinine bears a direct relation to the muscle mass of the individual.. The milligram amount of creatinine excreted during 24 hours per kilogram of body weight is often expressed as the creatinine coefficient. The coefficient varies from 18 to 32 in - men (total excretion 1.1-3.2 g/day) and from 10 to.25 in wome_n_ (total excretion 0.9-2.5 g/day). The coelfcient is lar el - - - 8 y indepen- dent of variations in diet, since creatinine in healthy persons is of - --- en ogenous origin. In older people, the daily output of creatinine may decrease to 0.5 g/day. In cigarette smokers, urinary output of ereatinine in men appears to decrease with greater number of cigarettes smoked per day (Adlkofer et al., 1984). However, this finding needs to be confirmed. Based on the variations in daily croatinine excretions in the urine, one has to be aware of the limitation of the factor "arnount of - biological marker per milligram of creatinine " In a study with 15 adult male cigarette smokers, the daily creatinine excretion varied between 1.0 and 2.5 g and the cotinine excretion between 1.3 and 13.1 mg (Hoffinann and Brunnemann, 1983). liowever, in certain cases, such as with healthy infants, the daily variations in urinary excretion are rather small. Thus, the measured nanograms of

146 cotinine per milligram of creatinine in urine reflect the inhalation of environmental nicotine from ETS rather woll (Greenberg et al., 1984). For the determination in urine, creatinine is complexed with picric acid and the resulting red color is measured epectrophoto- metrically, a task now predominately done with an autoanaly$or' (Faulkner et al., 1976). - Although the determination of cotinine in urine without refer- ence to creatinine ht+s resulted in meaningful slata in some studiex, the standardized cotinine levels per unit of creatinine may give a more stable measure of ETS exposure-particularly when limited urine samples must be used. Hydroxyproline Inhalation of nitrogen dioxide causes degradation of lung col- lagen and elastin (Koemidar et a1.,1972; HatLon et a1.,1977). This degradation results in elevated urinary excretion of hydroxyproline (Lewis, 1980). It is thus possible that the NOz in tobacco smoke, and even NOz in ETS, has the same lung-damaging effect as pure NO2. Kasuga et al. (1981) reported two studies in which healthy cigarette smokers excreted significantly more hydroxyproline thiun healthy nonsmokers and exsmokers. In the case of 6- to 11-year-uld children of smoking parents, Kausga et al. (1981) found elevated hydroxyproline levels in the urine. Because of the relatively low concentration of NOz in ETS (see Chapter 2), this finding was unexpected. Adikofer et al. (1984) were unable to confirm this finding in a study of 23 nonsmokers exposed to ETS. At present, the question of quantitative aspects of urinary hydroxyproline excretion in nonsmokers exposed to ETS is not settled. It will require additional studies before this compound ~ and its ratio to croatinine can be used as indicators for the degree of ETS exposure. N-Nitrosoproline N-nitrosoproline (NPRO) in urine reflects endogenous forma-n tion of nitrosamines, many of which are known animal carcinogens - (Preussmann and Steward, 1984; Yainio et al., 198b). NPRO ap- pears neither to undergo metabolism in mamnials nor to alkylate 147 cellular macromolecules.. NPRO is considered to be nonmuta- genic and noncarcinogenic and is excreted nearly quantitatively in urine.. It has been shown that endogenous forniation of N PRO is - - - - - significantly uicreased in cigarette smokers (Hoffmann and Brun- nemann, 1983; Ladd et al., 1984; Scherer and Adlkofer, in press). The increase is probably due to the high concentrations of nitrogen oxides in tobacco smoke that serve as nitrosating agents and the elevated concentration of thiocyanate in smokers that catalytically --- --- enhance the endogenous formation of nitrosamines such as NPRO. - - -- -- These effects are sbsent in nonsmokers without ETS_ exposure. . In one 6-day study, four rnale uonsnlokers with controlled di- ets were exposed to known degrees of ETS for three periods of 80 nunutes each on day 3 and day 4. Their 24-hour urine voids were analyzed for NPRO and for cotinine. While the cotinine lev- els In the urine of these nonsmokers increased from 5-7 ng/ml to 215-360 ng/ml, the NPRO excretion did not significantly change (Brunnemann et al., 1984). In another controlled study with 10 nonsmokers exposed to ETS containing 45 ppb of NOz, 400 ppb of NO, and 22 ppm of CO, urinary output of NPRO was also not elevated while COHb had increased significantly (Scherer and Adlkofer, in press). Although these two studies require con(irma tion and should include analytical assessment of nitrosothioproline (NTPRO) (Tsuda et al., 1986), another endogenously formed ni- trosamine, at present neither NPRO nor NTPRO measurement in urine can be used to indicate exposure to ETS. Ar-omatic Amines During the burning of cigarettes, 20-30 tunes more aromatic - - amines are released into the eidestrearin smoke than are present - in the mainstream smoke (see Chapter 2). Although at this time there is a lack of analytical data, it may be assumed that indoor - - - - - environments that are strongly polluted with E'M contain ineasur- - - ably higher aniounts of arometic aminea than ambient air without -- - -- - tobacco smoke pollution. Preliminary data indicate that t free aniline and o-toluidine, serving as surrogates for aromatic amines, are increased, although not signi6cantly, in the 24-hour urine voids of cigarette smokers (3.1 ± 2.6 pg and 6.3 ± 3.7 pg) compared with nonsmokers (2.8 - ± 2.6 pg and 4.1 ± 3.2 µg) (El-Bayounty et al., in press). The g4afi9BL4+S

148 next step requires the assay of the metabolites of aniline and o- toluidine in the urine of both smokers and nonsmokers. A study of the urinary excretion of aromatic amines in passive smokers would be indicated only if the total amounts of individual amines and their metabolites in smokers' urine are found to be significantly increased. GENUTO7CICITY OF THE U1tINE The evaluation of the genotoxicity of urine in nonsmokers Viith ETS exposure must consider the possibility of confounding effects, because DNA modifiers may be present in urine as a consequenca of dietary intake or as a secondary result of the activity of infectious agents in the urine of the host. NeverthetosEi, urinary. constituents may be DNA modifiers, because the inhaled agonta are known or suePected mutagens or because the inhaled agents lead to the formation of such biologically active compounds. Since 1975, the most widely used assay for genotoxicity, of human urine is the deterunination of mutagenicity in bacterial- tested strains with and without activation by enzyme-induced liver homogenate. In 1977, Yamaeaki and Ames reportod the preaence of inu- tagens in the urine of cigarette smokers, thus suggesting a cor- relation between mutagens in smokers' urine and increased risk for bladder cancer. Since publication of these data, other studies have reported an association of urinary mutagens that are active. in bacterial tester strains with cigarette smoking (International Agency for Research on Cancer, 1986), but not all results f'roin these studies have been consistent. One reason for the divergent findings could be the in8uence of dietary factors on the mutageni in the urine of smokers (Saason at al., 1985) and, perhaps also, nonsmokers exposed to.ETS. Three studies have attempted to explain the possible niuta- genic activity of the urine of nonsmokers exposed to ETS. In one study, fractions and subfractions were isolated by high-pressure liquid chromatography (1iPLC) from the urine of five passive smokers. Upon metabolic activation by S9 liver homogenates from rats pretreated with 3-mothylcholanthrene, these materials were mutagenic in TA-bacterial tester strains (Putzrath at al., 1981). It appeared that these mutagens are a complex mixture of urinary 149 components in the polar lipophilic subfractions. Due to a lack of diet control, these results are ambiguous. In a second assay of urine for bacterial mutagenicity, 8 male nonAmokers ('lb and 35 years of age) were placed in a poorly ven- tilated room (10 m') with 10 smokers for an 8-hour period (Boa et al., 1983). The 12-hour urine samples of the nonsmokers were collected before, during, and after exposure to ETS. Metabol- ically activated concentrates of the urine samples were analyzed for niutagenic activity in the toster strain, TA 1538. Urine samples collected directly after exposure to ET$ were significantly more mutagenic (relative activity: 3.9 f 1.0) than urine samples of the same nonsmokers prior to (3.1 ± 0.7) or long after ETS exposure (2,6 f 0.5). In the third study, six women who were medical students were exposed to E'.t's in a 10-m' exposure chamber on 2 consecutive days for one 3-hour session in the mornings and a 2-hour session in the afternoons. During these sessions, three of the women smoked a total of 30 cigarettes per day of a low-yield filter-tipped brand (5.4 mg tar, 0.4 nicotine, 4.6 mg CO); the other three women did not smoke. After 3 days without exposure and without cigarette smoking by any of the women, the exposure was repeated with reversal of the roles, so that those who had proviously been non- smokers now were smokers, and vice versa. The CO concentration in the chamber averaged 3.0 ± 0.9 ppm. The uptake of smoke was assessed by determination of COHb, cotinine, and thiocyanate in the plasma. Urine samples were collected at the end of the daily smoking periods. Urine was concentrated according to Yamasaki and Ames (1977) and tested for mutagenicity with tested strain TB98 using rat liver homogenate for metabolic activation (Sorsa et al., 1985). As is evident from the data in Table 8-4, COHb values for nonsmokers and passive smokers were indistinguishable, while there was a trend for higher plasma eotinine values in the passive smokers. The authors observed an increase in the mutagenicity of the urine of passive smokers during the period of study. The differences observed were not significant. On the basis of presently available data, it is likely that the exposure of nonsmokers to heavy ETS increases the potential for metabolically activated -i - - - genotoxc activity of their urine above and beyond the mutagenic activity that is observed in urine of the same nonsmokers before and_ long after exposure to ETS. However, before validating the Ames bacterial assay for inutagenicity as an Z4F3JQF•u5aC3

150 ~ ~ ~ a 4A]. no00 ~ +1 ., o ~ 1 +1 +1 ~ s ~ ~I C :~ ~ W ~ . ~ ~ ~ ~ ~ ~ ZVI V 1*4 N 0 ~ .~ H o•wa ~ a al~ +1 +1 _1i ~ .om~ 0 ~ Rf-- JR V ~5 ~ i ~~ tt a ~ N -N ~ `l Ir " A ~ ~ +1 ~ a ~ c ~ w F ~ .g i C! N C! rF 0 151 appropriate inethod for estimating the genotoxic effect of the urine of BTS-expoHed nonsmokers, the method itself and the diet of test subjects have to be standardized. Research in this area is needed, as are studies on the isolation and identification of the active agents in the urine of ETS-exposed nonsmokers. Adducts Formed in Passive Smokers -- upon Exposure to ETS Since about 1975, highly sensitive methods have been devel- oped for the determination of protein- or DNA adducts of environ- -- mental carcuwgens and toxic agents in circulating blood. Meth- ods probing these reactions for the toxic agents known to occur in tobacco smoke and ETS include determination of hemoglobin adducts of nitrosodimethylamine, methyl chloride, vinyl chloride, and benzene (National Institute of Environmental Health Sciences, 1984), as well as 4-aminobiphenyl (Green et al., 1984). DNA adducts with the smoke carcinogen, benzo[a)pyrene (BaP), have been described (Santella et al., 1985), and the tobacco-specific 4- (N-inethyl-N-nitrosamino)-1-(3-pyridyl)-1-butanone (NNK) leads to O°-methylguanine in DNA (Hotfmann and I[echt,1985). RIA's have been developed for quantitative determination of both the BaP-DNA adduct and O°-methylguanine (Porera et al., 1982; Foilt:s et a1.,1085). So far, the method for the determination of the DNA adducts has been applied to the analysis of benzo(alpyrene in smokers (Shainsuddin et al., 1985). In addition, the hemoglobin-4- aminobiphenyl assay has been used. for the analysis of the blood of - --- - smokers (Tannenbaum et al., in press). In both cases, only a lim- ited number of samples have been analyzod for these adducts. Nev- ertheless, the data appear encouraging. Another sensitive method for quantifying- DNA adducts is the P32-poetlabelling technique, which has been applied to human tissues (Gupta et al., 1982; . Everson et al., 1986). Validation and quantitative determination of the uptake of tobacco smoke carcinogens is urgently needed. Assays of adducts of BaP, aromatic amines, and tobacco-specific nitrosamines with protein or DNA in the circulating blood are the most promising teste of exposure to tobacco smoke. Once such asgays have been advanced to yield reproducible, informative methods in smokers, they may be subsequently refined to such sensitivities that they gLG9B44B

152 will also furnish reliable data on such adducts in the blood of passive smokers. . . FUTURE NEEDS At present, the best method for quantifying human exposure to ETS is the assay of nicotine and cetinine in urine and possi- bly saliva. Nicotine and cotininp can also be determined in seruni samples, but these samples require invasive techniques. ln smoke- polluted environments, nicotine is present in the vapor phase as a free base, thus its uptake by the passive smoker may not be rep- resentative of the uptake of acidic and neutral smoko components from the vapor phase nor of any component in the particulate phase. Thus, future studies should be concerned with develop- ing techniques to measure the uptake by the nonsmoker of various other types of tobacco-specific ETS components. This may include assays for the vapor-phase 9-vinylpyridine or flavor components - that are indigenous to tobacco. Particulato-phase agents to be - traced could include solanesol, tobacco-apecilic nitrosainines, aud polyphenols such as chlorogenic acid or rutin. These components are likely to be found only in trace amounts in E'1'S, and, thus, only minute quantities would be found In the circulating blood of passive smokers, making the development of assays difficult. The development of new trace methods for quantifying the levels of some tobacco-specific materials.in nonsmokers may require the identification of adducts formed between the ETS components and the proteins in blood. This approach would require the develolr ment of highly sensitive methods such as immunoassayp (e.g., RIA, ELISA) or postlabelling with radioinotopea or other inarkers.. The epidemiological studies on the effects of exposure to E'1S by nonsmokers have to consider a number of non-ETS-related factors. This fact underlines the urgent need for the development of highly sensitive dosimetric methods for ETS-specific carcinogens that can be applied in field studies. SUMMARY AND RECOMMENDATIONS Passive smokers are exposed to trace amounts of toxic agents including tumor initiators, tumor promoters, carcinogens, and organ-specific carcinogens when inhaling ETS. The determination of thiocyanate, nicotine, and cotinine in body fluids such as saliva, • 153 serum, and urine, as well as quantitication of CO in alveolar air and GOlib in blood, has been useful for the assessment of the habits of tns ividuals mid groups of smokers of cigarettes, cigars, and pipes. Currently, for measuring the exposure to E'1`S by nonsmokers, nicotine and cotinine appear useful. In acute exposure studies, COlIb can be a useful marker. Nicotine and cotinine, however, may not be directly related to the carcinogenic potential of the smoke. Indicators that are related to the carcinogenic risk are needed. 'tb assess the risks - involved in the exposure to carcinogenic agents from ETS, sensitive dosimetry methods for tobacco-specific compounds are urgently needed. During the last decade, immunoassays and postlabelling methods have been developed for tracing toxic and carcinogenic agents in circulating blood. These methodologies should be used for the development of dosimetry studies in nonsmokers exposed to ETS. Protein and DNA adducts may provide exposure measures that could be effectively used in epidemiologie studies. What Is Known 1. Deterniinations of thiocyanate, nicotine, and cotinine in saliva, serum, and urine, as well as quantification of CO in alveolar air and carboxyhemoglobin in blood, have been shown to be useful parameters for the assessment of the habits of individuals and groups of active smokers of cigarettes, cigars, and pipes. However, Jin goneral, only nicotine and its metabolite cotinine have proven 'useful for measuring the exposure to ETS of nonsmokers. 2. Assessment of average daily exposure on the basis of cotinine - levels in saliva and urine is independent of the restrictions posed by variations of the half-life of nicotine in smokers and nonsmokers. 3. The determination in urine of the amount of cotinine per milligram of creatinine should provide a more stable measure of recent environmental exposure to nicotine from ETS than cotinine without reference to ereatinine, particularly when limited volumes of urine are available. '.. It is likely that the exposure of nonsmokers to ETS increases the mutagenic activity of their urine over the activity observed_ in_ urine of the same nonsmokers when not exposed to-ETS. 6Z.698448

154 WLat Scientific Information Is Missing 1. The ,question of urinary hydroxyproline excretion in non- smokers exposed to ETS is not eettled. 2. A study on the urinary excretion of aromatic amines in nonsmokers exposed to ETS is needed in order to correlnte the total amounts of individual amines and their metabolites in the urine of nonsmokers exposed to ETS. 3. Where oxposure.histories can be specified clearly, validation of the use of adduct assays to determine and quantify uptake of tobacco smoke carcinogens is needed. 4. Information is needed on ceFtain tobacco-specific cunetitu-s ents and their fate in the ETS-expoeed nonsmoker, inr,luditig solanesol, tobacco-specific nitrosamines, and polyphenols such ita chlorogenic acid or rutin. 5. Knowledge of the levels of nitrOSothioproline following ex-e posure to ETS as well as nitrosoproline is needed. 6. Knowledge of the effects of diet is needed when interpreting results of the Ames bacterial assay for mutagenicity of the uriue of ETS-exposed nonsmokers. 7. Identification of the mutagenic agents in the urine of ETg- exposed nonsmokers needs to be made. 8. Future studies should be concerned with methodologies that enable us to assay the uptake by the nonsmoker of various other types of ETS components that are tobacco-specific. 9. New trace methods will have to be developed for dosimetry studies of carcinogens involving adducts (DNA and protein) and the development of highly sensitive methQds such as immunoassays or postlabelling for other products. 10. The epidemiological studies on the effects of ETS exposure in nonsmokers should consider a number of non-fi.TS-related fac- tors. This fact underlines the urgent need for the development (if highly sensitive dosimetrie methods for E`1'S-specific carcinogens that can be applied in field studies.. REFERENCES Adlkofer, F., Gt. Scherer, and W.D. Holler. Ilydroxyproline excretion In uriae of smokers and passive smokers. Prev. Med. 13A170-679, 1984. Adlkofei, F., 0. Scherer, and U. von Hees. Passive smoking. N. Rngl. J. Med. 312:719-720, 1986. 155 Aldrige, W.N. A new method for the estimation of micro quantities of cyanide and thiocyanate. Analyst 89t242-286, 1944. Bec_k_ett, A.H., J.W. Gorrod, and P. Jenner. The analysis of nicotine_- 1'-N-oxide In urine, in the presence of nicotine and cotinine, and its application to the study of in vivo nicotine metabolism in man. J. Pharm. PharmacoL 23:668-818, 1071. Benowits, N.L., P. Kuyt, and P. Jacob Ifi. Circadian blood nicotine concen- trations - tratinns during cigarette smoking. Olin. Pharmacol. Z'her. 32:768-764, 1982. Bos, R.P.0 .1.L.G. Thenws, and P.Th. Henderson. Excretion of mutagens in human urine after passive smoking. Cancer Lett. 19;86-90, 1983. Bowler, R.G. The determination of thiocyanato in blood serum. Biochem. J. 38:385-388, 1944. Brunnemann, K.D., J.O, Scott, N.J. Haley, and D. Hoffmann. Endogenous formation of N-nitrosoproline upon cigarette smoke inhalation. IARC -- Sci. I'ubi. 67:819-828, 1984. B_utts, W,C., M. Knehne_man, and G.M. Widdowson. Automated method for determining serum thiocyanate to distinguish smokers from nonsmokers. - - Olin. Chem. 20:1344-1348, 1974. Cano, J.-1'., J. Catalin, R, Badre, 0. Dumas, A. Viala, and R. Guillerme. Determination de Ia nicotine par chromatographie en phase gaseuse. Ii. Applications. Ann. Pharm. Fr. 28:633-840, 1970. - - Coburn, H.F., W.J. Williams, and R.E. Forst.r. Effects of erythrocyte destruction on carbon monoxide production in man. J. Olin. Invest. 43: - - 1098-1103, 1964. Conltas, D.B., J.M. Samet, C.A. Howard, G.1`. Peake, and B.J. Skipper. Salivary cotinine levels and passive tobacco smoke exposure in the home. Am.. Itev. Respir. Dl.. 133:A157, 1988. El-Bayoumy, K., J.M. Donahue, 8.8. Hecht, and U. Hoffmann. Identification and quantitative determination of aniline and toluidines in human urine. - Cancer Res., in press. E_rerson, R.B., E. Randerath, RJN. Santella, R.O. Cefalo, T.A. Avitts, and K.. Randerath. Detection of smoking-related covalent DNA adducts ducts In human placenta. Science 231p64-67, 1986. Faulkner, W.R„ and J. W. King. Renal function, pp. 981, 997.999. In N.W. Tiets, S. Berger, and W.T. Caraway, Eds. Fundamentals of Clinical - - - Chemistry. Philadelphia: Saunders, 1976. Feyerabend, C., T. Higgenbottam, and M.A.H. _Ruuell. Nicotine conce_n- arations in urine and saliva of smokers and nonsmokers. Br. Med. .1. 284:1002-1004, 1982. - Nbyerabend, Q., and M.A.H. Russell. Assay of nicotine in biological materials: -- - Sources of contamination and their elimination. J. Pharm. Pharmacol. - -- 32:178-181, 1980. bblles, P.G., N. lkushin, and A. Castonguay. Measurement of 0e-mothyl- deoxyguanosine In DNA methylated by the tobacco-specific carcinogen 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone usingg biotin-avidin en- syme-linked immuno.orbent assay. Csrcinogenesis 8;989-99$, 1986. Foliart, D., N.L. Benowits, and C.E. Becker. Passive absorption of nicotine e in airline flight attendants. N. Engl. J. Med. 30®:1106, 1983. oaGse4.Ls i

156 Gillies, P.A., B. Wilcox, ©. Coates, F. Kristmundsd"ottir, and D..1, Reid. Uns of objective menurement In the validation of self-reported smoking In children aged 10 and 11 years: Saliva thiocyanate. J. Epidemiol. f3omut. Health 36:205-208, 6.206-208, 1982: Green, L.C., P.L. Skipper, R.J. Tdresky, M.S. Bryant, and S.R. Tannenban:u. In vivo dosimetry of 4-aminobiphenyl in rats via a cysteine adduct In hemoglobin. Cancer Res. 44:4264-4269, 1984. Greenberg, R.A., N.J. Haley, R.A. Etsel, and F.A. Loda. Measuring the exposure of infants to tobacco smAke: Nicotine and cotinine in urine ---- and_ saliva. N. Engl. J. Med. 310 1076-10780 1984. f3upta, R.C., M.V. Reddy, and K. Randerath. 33P-postlabeling analysis of non-radioactive aromatic carcinogen-DNA adducts. Carcinogenesis 9:1081-1092, 1982. Haley, N.J., C.M. Axelrad, and K.A. Tilton. Validation of seif-reported - -- - smoking behavior: Biochemical analyses of cotinine and thiocyanate. Am. J. Public Health 93:130l-1307, 1983. Hatton, D.V., C.S. Leach, A.E. Nicogosslan, and N.N. Ferrante. Collagen breakdown and nitrogen dioxide inhalation. Arch. Environ. Health - 32:33-36, 1977. Harke, H; P. Zum Problem des Passiv-Rauchens. Milnch. Med. Wochenschr. 112:2328:2334, 1970. Hauth, J.C., J. Hauth, R.B. Drawbaugh, L.O. Qilstrap 111. and W.P. Pierson. Passive smoking and thiocyanate concentrations in pregnant women and newborns. Obstet. (lynecol. 63:519-522, 1984. Heinemann, (i., H. Schievelbein, and F. Richter. Die analytische und diag- nostische Vsliditnt der Bestimmnng von Darboxyh"amoglobin im Blut - - - - -- --- ---- - - - und Kohlenmonoxid in der Atemluft von Rauchern und Nichtrauchern. . J. Clin. Chem. Olin. Biochem. Z4:'249-436, 1984. Hill, P., N.J Haley, and E.L Wynder. Cigarette smoking: Carboxyhe- moglobin, plasma nicotine, cotinine and thiocyanate levels vs. self- reported data and cardiovascular disease. J. Chron. Dis. 3Q:439-449, 1983. Hoffmann, D., and K.D. Brunnemann. Endogenous formation•of N-nitroso- proline in cigarette smokers. Cancer Ree. 43:6670-6674, 1983. Hoffmann, D., N.J. Ilaley, J.D. Adams, and K.D. Brunnemann. Tobacco sidestream smoke: Uptake by nonsmokers. Prev. Med. 13:8t)8-817, 1984. Hoffmann, D., and S.S. Hecht. Nicotine-derived N-nitrosamines and tobacco- related cancer: Current status and future directions. Cancer Res. 45:936-944, 1986. - Huch, R., J. Danko, L. Spatling, and R. Huch. Itisks the passive smoker runs. Lancet 2:1376, 1980. International Agency for Research on Cancer (IARQ) Monograph. Evaluatiua of the Carcinogenic Risk of Chemicals to Humans, Vol. 38, pp. 1d3-314.. Tobacco Smoking. Lyons: IARf3, 1986. 421 pp. Jacob, P., Ill, M. Wilson, and N.L. Benowits. Improved gas chromatographic method for the determination of nicotine and cotinine in biologic 9uitls. J. Chromatogr. 224;81-70, 1981. Jarvis, M.J. Serum thiocyanate in passive smoking. Lancet 1:109, 1986. Ja_rvis, M.J., and M.A.H. Russell. Measurement and estimation of smoke dosage to non-smokers from environmental tobacco smoke. Eur. J. Respir. Dis. 65(Suppl. 133):68-76, 1984. 157 Jarvi., M.J.; M.A.H. Russell, and C. Feyerabend. Absorption of nicotine and carbon monoxide from passive smoking under natural conditions of exposure. Thorax 38:829-833, 1983. Jar:ris, M., 11. 11:nstail-Pedoe, C. Feyerabend, 0. Vesey, and Y. Saloojes. Biochemical markers of smoke absorption and self reported exposure to passive smoking. J. Epidemiol. Comm. Health 38:336-339, 1984. Jarvis, M.J., M.A.H. Russell, C. Feyerabend, J lt. Eiser, M. Morgan, P. Qammage, and E.M. Gray. Passive exposure to tobacco smoke; Saliva cotinine concentrations In a representative population sample of non- smoking schoolchildren. Br. Med. J. 291:927-929, 1986. Junge, B., U. Borges, M.-B. Berkhols, W. Thefeld, and H. Hoffmeister. Thiocyanat Im Serum als Indikator f8r die Schadstoffbelastung durch a ranc . Arbeitsmed. Sosiaimed. Praeventivmed. 13:13-18, 1978. Kasuga, H., H. Matsuki, and F. Osaka. Smoking effects and urinary by- --- droxyproline. Presented at the 9th International Scientific Meeting of International Epidemiological Association, Edinburgh, Scotland, 1981. 14 pp. Klein, A.B., and J.W. Gorrod. Metabolism of nicotine in cigarette smokers - during pregnancy. Eur. J. Drug Metab. Pharmocokinet. 3:87-93, 1978. Kosmidar, S., K. Ludyga, A. Misiewica, M. Drosda, and J. Sagan. Experi- mental and clinical investigations on the emphysema forming action of nitrogen oxides. Zentralbl. Arbeitsmed. Arbeit.schuts 22:328-388, 1972. ------ Lndd, K.F., Ii.L. Newmark, and M.O. Archer. N-Nitrosation of proline in smokers and nonsmokers. J. NatI. Cancer Inst. 73:83-87, 1984. Langone, J.J., H. ajika, and H. VanVunakis. Nicotine and its metabolites. Radioimmunoassays for nicotine and cotinine. Biochemistry 12:502b- -- 6030. 1973. Leaderer, B.P. Air pollutant emissions from kerosene space heaters. Science 218:1113-1116,1984. Lewis, T.R. Criteria relevant to an occupational health standard for nitrogen - - dioxide, pp. 381-376. In S.D. Lee, Ed. Nitrogen Oxides and Their Effects on Health. Ann Arbor: Ann Arbor Science Publications, 1980. Luepkeq R.V., T.F. Pechacek, D.M. Murray, C.A. Johnson, F. Hund, and D.R. Jacobs. Saliva thiocyanate: A chemical indicator of cigarette - - - smoking in adolescents. Am. J. Public Health 71:1320-1324, 1981. Matsukura, S., T. Taminato, N. Kitano, Y. Seino, H. Hamada, M. Uchihashi, H. Nakajima, and Y. Hirata. Effects of environmental tobacco smoke on urinary cotinine excretion in nonsmokers. N. Engl. J. Med. 311:838-833, 1984. National. Research Council, flommittee Medical and Biologic Effects of En- vironmental Pollutant._ Carbon Monoxide. Washington, D.C.: National Academy Press, 1981. 239 pp. National Institute of Environmental Health Sciences (NIEHS). Biochemical and cellular markers of chemical exposure and preclinic.l indicators of disease, pp. 197-Z67. In Human Health (n the Environment. Some Research Needs: Report of the Third 1tsk Force for Research rch Planning - in Environmental Health Science. Bethesda, Maryland: National In- stitutes -- stitutes of Health, 1984. 407 pp. (Available from the U.S. Government Printing Office as NIH Publ. No. 86-1277.) iesse44e

168 Palladino, a., J.D. Adams, K.D. Brunnemann, N.J. Haley, and D. Hoffmann. Snuff-dipping in college students: A clinical pro8le. Military M*d., In press. Pekkanen, T.J., 0. Elo, and M.L. Hanninen. Changes In non-smokers' saliva thiocyanate levels after being In a tobacco smoke-Hlled room. World Smoking Health 1:37-39, 1976. Perera, F.P., M.C. Poirier, S.H. Yuspa, J. Nakayama, A. Jaretski, M.M. Curnen, D.M. Knowles, and LB. Weinstein. A pilot project In molecnlar cancer epidemiology: Determination of benso(a)pyrene-DNA adducts In animal and human tissues by immunoa.says. Carcinogenesis 3t1!105- 1410,1982.. Pittenger, D.J. Passive smoking. N. Engl. J. Med. 312:730, 1985. Poulton, J., ©.W. Aylsnce, A.W.J. Taylor, and E3. Edwards. Serum thio- cyanate levels as Indicator of passive smoking In children. Lancet 2:1405-1406, 1984. Preussmann, R., and B.W. Stewart. N-nitroso carcinogens, pp. 843-828. In 0. Stearle, Ed. Carcinogens. American Chemical Society Monograph 182. Washington, D.as American Chemical Society, 1984. Putsrsth, R.M., D. Langley, and E. Eisenstadt. Analysis of mutagenie activity In cigarette smokers' urine by high performance liquid chro- matography. Mutat. Ros. 86:97-108, 1981. Russell,- M.A.H., and E3. Feyersbead. Blood and urinary nicotine in_ non- smokers. Laneel 1:179-181, 1976. Russell, M.A.H., M. Raw, and M.J. Jarvis. Clinical use of nicotine chewing- -- gum. Br. Mad. J. 280:1699-1602, 1980. Russell, M.A.H., M.J. Jarvis, d. Devitt, and C. Feyerabend. Nicotine intak.. by snuff users. Br. Med. J. 283;814-817, 1981. Sa_ ntella, R.M., L.-L. Hsieh, C,-D. Lin, S. Viet, end I.B. Weinstein. Qeaa- titation of exposure to benso(a)pyrene with monocbnal antibodies. Environ. Health Perspect. 62:96-99, 1986, Sasson, I.M., D.T. Coleman, E.J. LaVoie, D. Hoffmann, and L+I.L. Wynder. Mutagens in human urine; Effects of cigarette smoking and diet. Mutat. Res. 168:149-167, 1985. - Scherer, (3., and F. Adlkofer. Endogenous formation of N-nitrosopre ins in smokers and nonsmoker., In D. Hoffinann and (7. Harri., Eds. Mechanisms In Tobacco Carcinogenesb (llanbury Report 23). Cold Spring Harbor, I*lew York: Cold Spring Harbor Laboratories, In pres.. Sepkovic, D.W., N.J. Haley, and D. Hoffmann. I:limination from the body of tobacco products by smokers and passive smokers. J. Am. Med. Assoc. 266:863, 1986 (letter). Shamsuddin, A.K.M., N.T. Sinopoli, K. Hemminki, R.R. Boesch, and CA. - Harris. Detection of benso(a)pyrenes DNA adducts In human whits blood ceils. Cancer Res. 46:68-88, 1986. Sores, M., P. Einisto, K. Husgafvel-Pursianinen, H. JErventaus, H.. Klvistii, Y. Peltonen, T. Tuomi, and S. Valkonen. Passive and active expoiure to cigarette emoka In ssmoking experiment. J. Toxicol. Environ. IlealtA 16:523-534, 1986. 159 Tannenbaum, B.R., M.S, Bryant, P.L. Skipper, and M. Maclure. Hemoglobin adducts of tobacco-related aromatic amines= Application to molecular epidemiology. In D. Hof[mann and 0. Harris, Fds, Mechanisms In Tobacco Carcinogenesis (Banbury Report 23). Cold Spring Harbor, New York: Cold Spring Harbor Laboratories, In press, Tsuas, M., J. Niitu.uma, 8. Sato, T, Hirayam., T. Kakisoe, and T. Sugimura. Increase iu the levels of N-nitrosoproline, N-nitrothioproline, snd'N- nltro-3-methylthioproline In human urine by cigarette smoking. Cancer Lett. 90:177-124, 1986. Vainio, H., K. Ilemminki, and J. Wilbourn. Data on the carcinogenicity of chemicals In the IARC monographs programme. Carcinogenesis 6:1863- 1666, 1986. Vogt, T.M., S. Selvin, and J.H. Billings. Smoking cessation program: Base- line carbon monoxide and serum thiocyanate levels as predictors of outcome. Am. J, Publie Health 69:1166-1169, 1979. Wsld, N.J., and 0. Ritehie. Validation of f studies on lung cancer In non- smokers married to smokers. Lancet 1:1607, 1984. Wald, N.J., M. Idle, J. Boreham, A. Bailey, and H. Van Vunakis. Serum cotinine levels In pipe smokers: Evidence against nicotine as cause of coronary heart disease. Lancet 5:775-777, 1981. Wald, N,J., J.Boreham, A. Bailey, Q. Ritchie, J.E. Haddow, and (3. Knight. Urinary cotinine as marker for breathing other peopk's tobacco smoke. Lancet 1:390-Z31, 1984. Yamasaki, E., and B.N. Ames. Concentration of mutagens from urine by adsorption with the nonpolar resin XAD-2: Cigarette smokers have -- mutagenic urine. Proc. Natl, Aced. Sci, USA 74:3356-3559, 1977. I ZgG9844R

III HEALTH EFFECTS POSSIBLY ASSOCIATED WITH EXPOSURE TO ENVIRONMENTAL TOBACCO SMOKE BY NONSMOKERS £8r,9gG4g

9 Introduction Epidemiologic and experimental studies seek to determine if a relationship exists between a particular exposure and particu- lar health elfects. When the exposure is via the air, as is the - - case with environmental tobacco smoke (ETS) exposure to non- smokers, tho organs that are directly exposed include the eyes, nose, throat, and lungs. Clinical, epidemiologic, and animal stud- ies have shown, generally speaking, that air pollutants can have major health effects on the respiratory system (National Research Council, 1985). Experimental research using animals (Chapter 3) and research with biological markers in hunians (Chapter 8) in- dicate that various constituents of the smoke are absorbed into the blood and, therefore, are transported to organs and tissues of the body. Consequently, the range of possible health effects of exl>osure to liTS may be very broad and vary enormously in their effect on the individual. Effects may be reversible or irreversible, discomforting, or life-threatening. In the following chapters, several possible health effects that have received substantial attention are reviewed. Many of the health effects associated with active smoking have been evalu- ated in -studies of nonsmokers exposed to ETS. These include: acute, noxious sensory irritation; nonmalignant respiratory symp-. toms and disease; decrease in pulmonary function; lung and other caucers; cardiovascular disease; relative growth, ear infections in children; and low birthweight of children of nonsmoking women. Nonsmokers commonly complain of the perception of tobacco smoke and its irritating, noxious, or annoying qualities. However, in most such spontaneous instances, these complaints are voiced VSG9844e 1 163

164 because the subjects can sse another person actively smoking in their vicinity. Chapter 10 •reviews experimental studies that evaluate these acute comfort aspects under controlled conditions. Chapters 11 and 12 assess and evaluate possible nonneoplastic and neoplastic pulmonary effects of exposure to ETS by nonsmok- ers. Over the past 15 years, a number of studies in children and in adults have assessed various possible acute and chronic pulmonary effects subsequent to long-term exposure to ETS. Individuals who have chronic lung diseases, such as patients with asthma, alpha=L- antitrypsin deficiency, or cystic 6brosis, are potentially hypersen- sitive to the effects of ETS exposures. Chapter 13 reviews and evaluates reports of cancers other than lung that may be associated with exposure to ETS in nonsmokers. Chapter 14 discusses the possible association of exposure to ETS with chronic and acute cardiovascular responsoe and cardio- vascular diseases in nonsmokers. Individuals with chronic disease that compromise the cardiovascular system, such as patients with a history of angina pectoris, are at a high risk for developing - - - abnormal cardiovascular responses following exposure. Chapter 15 considers evidence that a number of other health - - effects are linked to ETS exposure in children of smokers, includ- ing lower relative growth, frequency of ear infections, and low birthweight (with nonsmoking pregnant mothers). The studies reviewed here are epidemiologic and experimeu- tal. Epidemiologic studies lnclude case-control studies, in which - subjects are selected according to whether or not they have the health outcome being studied, and cohort (or prospective) stud'ies, in which subjects are classified according to whether or not they have been exposed to ETS. eross-sectioual studies are thase in which an assessment is made of a population at one point in time. - Longitudinal studies follow a group of persuns over time. In ex- perimental studies, subjects are exposed to ETS uuder controlled conditions often using chainber studies. Most studies of ETS have been cross-sectional rather than longitudinal. To be informative, a study must evaluate a sufficient number of people to provide a precise estimate of the effect; obtain valid iuformation regarding the history of exposure and health status of the individuals; and, of course, the statistical analyses must be appropriate to the study design. The appropriate design and use of these epidemiologic- methods for the study of air pollution and i,ossiblo health effects 165 are discussed in general terms in the monograph "Epidemiology and Air Pollul,ion" (National Research Council, 1985). REFERENCE National Research Council, Committee on the Epidemiology of Air Pollu- tants. Epidemiology and Air Pollution. Washington, D.G.: National Academy Pres., 1986. 224 pp. SeG9ef,48

167 10 Sensory Reactions to and Irritation Effects of- Envir-onmental Tobacco Sniolce R 4 AIR SPACE PER PERSON (m3) 5 10 /5 20 In this chapter, the acute sensory reactions from exposure to ETS are discussed. These reactions include perception of odor and irritation of eyes and upper airways. Methods for evalunt- ing these psychoscnsory phenomena include controlled chamber studies, where ventilation and smoking raten are manipulated and evaluated in terms of reported perception by a s)nall number of eubjects. ODOR The perception of odor is often the earliest indicator of ex- posure to many airborne contaminants, but not for all. For sotne individuals, odor tnay merely be a nuisance. For others, odor is an early indicator of a complex reaction to exposure to ETS involving allergic and other physiologic responses. Considerations of sensory reactions have a central role in the development of guidelines for ventilation requirements for occupied spaces. The amount of ventilation, or number of air exchanges, needed to eliminate unacceptable odors and irritation commonly exceeds that required to meet any other needs, such as control of _ - carbon dioxide. For a number of years, quite apart from concerns - - about possible adverse health from exposure to ETS, ventilation engineers have viewed ET8 as the most problematic common in- door d_ oor contaminant (Leonardos and Kendall, 1971). Efforts to derive functional relationshii/s between the atnount - of a contaminant generated in a space and the arnoui/t of outdoor air, i.e., ventilation, necessary to control its odor began in the ~ 20 ~ ~ lo C4, fUNCT1ON DERIVED FROM DATA OF YAGLOU. RILEY. AND C066INS11938) 1t3 lo 5 . ~ . . . 0 200 400 60o coo AIR SPACE PER PERSON (CU. FT.) FIGURE 10-1 Relationships between ventilation rate and air space per person In an environmentai chamber according to three criteria: (A) main- tenance of oxygen concentration; ntrationi (B) control of carbon dioxide to a level of 0.6% (9.6 cubic feet per minute)i and (C) control of body odor at a moderate te level under sedentary conditions of occupancy, no smoking. 19309 (Yaglou et al., 1936). Function C in Figure 10-1 is derived front experiments of Yaglou et al. (1936), where judges assessed the odor generated by occupants sitting quietly in an environmental chamber. The function depicts the combination of air space per person and ventilation rate of the air space (outdoor air) per person necessary to maintain odor at a moderate, acceptable level under steady-state conditions. Theoretical functions A and B, which fall below C, implying less need for ventilation, represent the outdoor air needed to maintain oxygen at a minimum of 20% and the air necessary to hold carbon dioxide at a maximum 0.6%, respectively. The decrease in curve C at low occupancy density (large air space per person) resulted most likely from the instability of oc- cupancy body odor in Yaglou's chamber. That is, occupancy odor decays relatively rapidly on its own (Yaglou and Witheridge, 1937; Clausen et al., 1984). Tobacco smoke odor, on the other hand, exhibits relative stability. When smoking has ceased in an unventilated room, the odor will remain at the about same level over many hours (Yaglou and Witheridge, 1937; Clausen et al., 1985). In a diagram such as Figure 10-1, a function for tobacco 3969Q4.49 166

168 smoke odor, like functions A and B, would lie indehendent of the size of the space or of air epace per occupitnt. In this respect, tobacco smoke odor behaves es a simple contaminant and ventila-- tion requirements for reducing tobacco smoke odor should depend strictly on rate of smoking. Twenty years after his study on occupancy odor, Yaglou - (1955) reported a st nall experiment on tobacco smoke odor. $tudy- ing the very high smoking rate of 24 cigarettes per hour. generated by six of nine occupants in his 1,410-cubicfoot chamber, he re- ported the need for 40 cfm (cubic feet per minute) per smoki:r, or 600 cubic feet per cigarette, in order to achieve moderate, ac- ceptable odor. At about the same time, Kerka and llumphreys (1956), using similar psychophysical techniques, estimated the re- quirement at 2,250 cubic feet.per cigarette, or 300 cfm per smoker smoking 8 cigarettes per hour. At a smoking rate of 2 cigarettes - per hour, this would be 75 cfm per smoker. Recent results have estimated ventilation needs closer to those of Kerka and Humphreys (1956) than those of Yaglou (1955), but have also uncovered limitations on ventilation as a solution to the odor problems produced by ETS. Figure 10-2 shows how tobacco smoke odor varied over time for three smoking rates and various ventilation rates (Cain et a1.,1983). The line connecting the open - - - squares in the left panel depicts the level of odor generated by nonsmoking occupancy with low ventilation. It shows that even In the presence of higher ventilation rates, smoking generated more odor than simple occupancy. The psychophysical judges in the experiment, a mixed group of smokers and nonsmokers, assessed acceptability in addition to per- ceived intensity. Figure 10-3 shows the percent of dissatisfaction as a function of ventilation rate per cigarette. The ventilation rate that would lead to 20% of judges dissatisfied is 4,240 cubic feet per cigarette (shown by the vertical dashed line). Twenty percent dis- satis6ed is the maximum level allowed by recommendation of the American Society of lIeating, Refrigeration and Air-Conditioning Engineers (ASHRAE,1981). On the realistic assumption that the percentage of people actually smoking in a space at any given time will equal about 10%, ventilation rate per person (smokers and nonsmokers) would need to be 63 cfm (see Figure 10-3) to reduce odors rs to a level that would satisfy 80% of the judges. Despite ASHRAE's goal of satisfying at least 80% of visitors to a space, none of its recommendations for ventilation are as high . . ~ ~~..... =:A8i0 II~~III ~~~•Q7c8 ~ 169 1 I 1ana ) Ouj.V°ursoVw Irhawoy r-- l I I $ $ LRG99449

170 Ventilation Rate per Occupant (/-sec 's) 1 2 3 5 7 10 20 30 30 Ventilation Rate per Cigarette (ins) 10 20 30 SO 100 300 ' 1 ' SbOb 1000 _2000 200 500 Ventilation Rate per Cigarette hts) , 10 2a '50_r_r__M& Ventilation Rate per Occupant (cfm) F1CiURE 10-3 Percent of Judgments of unacesptaNe odor quality of air versus ventilation per cigarette and ventilation per occupant, assuming that 10% - of occupants in a.pace will be smoking at any time. Data from Cain at ai. -- _ - (19t;S).- as 53 cfm per occupant. For offices where smoking is allowed, the ASHRAE recommendation is 20 cfm per occupant. For many other smoking areas, however, the ASHRAI:E recommendation is - 35 cfm per occupant. Such recommendations did not result from experiment, but rather from a consensus procedure of expert heat- ing and refrigerating engineers that weighed available information. ing The bulk of the data on the acceptability of odor and irritation . _ from ETS was not available at the time AS)IRAE prepared its - - standard in 1981. The standard was, however, the first to specify - the need for 4 to 5 times greater ventilation rates during smoking occupancy as compared with nonsmoking occupancy. The most common rate specified for smoking occupancy is 35 c-fnt per oc- cupant, whereas 7 cfm per occupant is the most common rate specified for nonsmoking occupancy. This means that in a space where smoking is allowed, the pollution generated by smoking creates the greatest need for ventilation. According to the data of Cain et al. (1983) (h'igure 10-3), ASHRAE's proposed ventilation rate of 35 c-fnt per occupant dur- ing smoking will lead to 25% of visitors being dissatisfied (76af6 171 satisfied) with the odor. Data suggest that the difference in sat- --- n isfaction between smoking and nonsmoking occupancy, and hence the difference in recommended ventilation rates, arises largely be- cause cause of the intensity of the odors (Figure 10-4), rather than the quality of tho odors. At equal.odor intensity, the occupancy odor and tobacco smoke odor are disliked about equally. -- An additional factor affecting annoyance with odor is that nonsmokers are much more likely thau smokers to object to to- bacco smoke odor. Figure 10-5 depicts relative dissatisfaction with tobacco smoke odor at various intensities, expressed in terms of equivalent levels cif butanol. At 32 ppm (butanol level 2), 1% of smokers found the odor unacceptable, while 20% of nonsmokers found it so. The odor had to rise to 256 ppm (level 5) before as 2 3 4 5 6r8910 Odor Intensity (cm) FIGIURE 10-4 Percent of Judgments of unaeeeplaNe odor quality of air versus odor inten.ity assessed by means of a graphic rating procedure during various conditions of smoking and nonsmoking occupancy. Each p~int represents the outcome from a particular combination of contaminant generation (number of occupants or rate of smoking) and ventilation rate. Data from Cain at ai. - (1g6a). BB~~B~GQ

d 172 FIGURE 10-6 Percent of judgments of unecceptal/i odor qualitlr of air derived from tobacco smoke odor related to equivalent level of bntanol (parts per million at top; 1092 at bottom). Judgments accumulated across all conditions of smoking (4,367 judgments). Lefts Data from all visitors. Right: Data from smokers and nonsmokers plotted separately. Data from - Cain et al. (1983). many as 20% of smokere found the odor unacceptable. In terms of practical solutions to the odor problem caused by tobacco smoke, the difference between smokers and nonsmokers may prove insur- mountable. Under usual levels of smoking, no realistic level of ventilation will drive tobacco smoke odor as low as the equivalent of 32 ppm butanol (butanol level 2). IRILITATI4N Ventilating and air-conditioning enginer+rs have typically con- cerned themselves with the reactions of visitors to enclosed spar.es - - on the assumption that visitors will exhibit more sensitivity than occupants. As society has become more concerned with the health risks of smoking in the recent past, research on consequences of ETS exposure has focused on the occupant. Included within this concern have been the sensory reactions of occupants. Figure 10-6 illustrates changes in tobacco smoke odor and irri- - tation over time for occupants. Whereas perceived odor magnitude may fade due to olfactory adaptation, irritation may increase. Also apparent in this figure is a relationship between relative humidity and odor or irritation perception. In the low relative humidity conditions, both odor and irritation were exacerbated. 173 Receptors for irritation exist throughout the nasal, pharyn- geal, and laryngeal areas and on the surface of the eyes. The receptors comprise free nerve endings of the fifth, ninth, and tenth cranial nerves and form the mediating elements of what is known as the commots chemical sense. Although particularly sensitive to corrosive stimuli, the common chemical sense responds to almost any airborne organic material at high concentration (Cain, 1981). The comnion chemical sense (or irritation perception) is char- acterized by a tendency to respond more vigorously over time (Coanetto-Munis and Cain, 1984). A person in an environment -- - with a low-level irritant may even fail to notice any irritation at --- - first. Once irritation has begun, however, it may persist even after removal of the stimulus. I 2 3 TIME. min FIGURE 10-6 Changes In odor and irritation during continuous, short-term exposure to cigarette smoke generated in a chamber. Veatilation equalled 14 - ' cfm Jrer cigarette and ambient temperature equalled 2b°-f3. Relative humidity (RH) was 30% In ene condition and 65% In the other. Adapted from Kerka and l f umphreys (1956). (NG9RU8'

•y! kr0albn Ind.x 3-1 2-1 1 /f< ,< i ,. --- 174 ~-_- 10 ppra a pprn 2.8 ppm / e.. ~ .. 1.3 ppm qonltol 20 40 e0 oapo.r, wln " wopp'" ! 'appM <* < E'' 2.5 ppst 0 20 t 40 ~ oo ..pe.r. Wn FIGURE 10-7 Eye irritation related to duration of exposure and concentra- tion (parts per million carbon monoxide) of ETS. Left: Eye Irritation Index. Right: Eye blink rate. From Weber (1984). Charnber Studies A number of chamber studies have examined irritation and odor from tobacco smoke (Weber et al., 1976a,b; Weber et al., 1979a,b; Weber and Fischer, 1980; Muramatsu et a1.,1983; Weber, 1984). The major findings include:. . Irritation from ETS varies with both concentration (mea- sured --- sured as an increment in carbon monoxide as a surrogate for ETS) --- and time over long durations, as shown in Figure 10-7. e The eyes are the most readily affected site for irritation, with the nose second. e Rate of eye blinking correlates well with estimates of eye and nose irritation when the level of ETS is high (i.e., level of ETS such that the carbon monoxide concentration is at least 5 ppm), though eye blinking seems a less sensitive index than psychophysical judgments (Figure 10-7). . Degree of annoyance (a compoeito index of impressions as defined by Weber) reaches a steady state much more rapidly than irritation, presumably because odor contributes to annoyance. . Degree of annoyance depend almost entirelyy on the gas phase of ETS. Filtration of the particles is followed by only a small, though relatively constant, reduction in annoyance. OGG9gl,4g . 175 • Eye irritation and increased eye blink rate depend almost entirely on the particulate phase of ETS. Particle filtration dimin- ishes ishes the sense of irritation greatly. e Weber (1984) suggested that ETS corresponding to CO concentrations of 1.5 to 2.0 ppm should form the maximum permis- sible level of exposure in envirimmentally realistic circumstances. At about 2 phm of CO, almost 20% of occupants report "strong" or "very strong" eye irritation. Gain et al. (1983) and Clausen et al. (1985) found that incremental CO concentrations of I to 2 ppm led to 20% of visitors becoming dissatisfied with the air. One of the more important issues with respect to control of ETS is whether filtration of the particles will reduce discomfort. As indicated above, Weber and colleagues found only a small reduction of annoyance when they filtered the particles with Cam- bridge pads. Since their criteria for annoyance largely assessed - odor, their data largely agree with those of Clausen et al. (1985), - -- - who found that electrostatic precipitation of the particles caused no significant reduction in odor perceived by visitors to a cham- ber. Nevertheless, Weber and associates did find that filtration reduced reported eye irritation considerably. This led them to -- - draw the conclusion that eye irritation derived largely from the particulate phase of ETS. Cain et al.. (in press), on the other hand, ' - found only a slight reduction of irritation following electrostatic precipitation of the particles. This disparity suggests the need for a more direct comparison of the sensory effecte of the two filtration - methods and for chemical analysis in order to determine whether Cambridge pads remove a vapor-phase constitutent of ET_ S_ that is left airborne by electrostatic precipitation. Field Studies Winnecke et al. (1984) argued that when people engage in so- cial activities (e.g., playing cards or games) they become somewhat less critical of the environment and will tolerate a level correspond- ing to more than a.5-ppm increment in CO. It is suggested that when undistracted, occupants of chambers in experimental studies might complain about circumstances that would go unnoticed in life situations. On the other hand, irritation may prove relatively resistant to distraction. Restaurants would seem to offer a realistic proving ground for the interpretation of the chamber studies.

176 Weber ot al. (1979a) found that more than 20% of occupants in restaurants included in a Zurich study reported eye irritation when CO, used as a surrogate for tobacco smoke levels, increimed by 2 ppm above background. There was a direct relationHhip between reported irritation and CO concentration in four restau- rants surveyed. In a study of more than 40 workrooms, Weber and Fischer (1980) also found a similar association between the -- - - concentration of CO and reported eye irritation. Judgments of dissatisfaction, whether taken inside a chamber or in the field, may well vary as the social acceptance of cer- tain odors, including ETS, changes with time. For this reaHon, judgments of some attribute, such as eye irritation, or judgmonts of odor intensity, particularly those that entail a reference such as butanol, should form the information of interest for long-term - considerations. Dissatisfaction measures may be more variable. IiYPEItSENSITIVE I NDiVIDUALS Individuals with chronic lung diseases, such as asthma and vasomotor rhinitis, may be more sQnsitire to the acute irritating effects of exposure to ETS (see Chapter 11). In addition, many people without active diseases report allergic or allergic-like syanp- toms as a result of exposure to ETS (e.g., Speer, 1968; Zussuian, 1974). Reported symptoms include eye irritation, nasal symptoms, headache, cough, wheezing, sore threat, and nausea. The percent of people who report these responses varies with the nature of the exposure. These reports have led to the belief that a tobacco smoke allergy may exist. Several investigators have studied immediate cutaneous hy- persensitivity to extracts of tobacco leaves. Zuseman (1974) found that 16% of 200 atopic patients reported that they were clinically sensitive to ETS exposure. All of them did develop erythema dur- ing the intradermal tests. Becker et aL (1976) found that one-third (11 out of 31) of human volunteers, including smokers, exhibit hy- persensitivity to a glycoprotein purilied from cured tobacco leaves (TGP-L) and from cigarette smoke condensate (TGP-CSC). Re- ports of immediate skin reactivity suggest an bnmunological basis for clinical sensitivity to tobacco smoke. Tobacco smoke has been shown to contain immunogens that can stimulate immune responses to tobacco leaf extract in experi- mental animals (Lehrer et al., 1978; lfecker et al., 1979; (;leich and 177 Welsh,1979). However, the extracts differ and there is controversy concerning the purity of tobacco glycoprotein isolates (Becker et al., 1981; Bick et al., 1981). In a recent study of Lehrer et al. (1984), skin prick tests of 93 subjects were done, including 60 of whom claimed clini- cal sensitivity to tobacco smoke. The group included atopic and nonatopic individuals. Approximately 50% of the atopic subjects had positive skin tests to leaf extracts or cigarette smoke con- densate (CSC). Fewer than 5% of nonatopic individuals had a positive reaction, independent of whether they claimed to be_ sen- sitive to ETS exposure. Radioallergoeorbent tests (RAST) were also conducted. Forty-five percent of atopic individuals and 6% of nonatopic individuals were positive for leaf extracts. There were no significant•diR'erences in specific serum IgE antibodies among smokers, examokers, or nonsmokers. Fewer than 6% of ei- ther ther group responded to CSC. Because there was no relationship between subjective tobacco smoke sensitivity and reaction to the various tests, the authors concluded that the reported subjective sensitivity is probably not related to hypersensitivity to tobacco leaf or smoke antigens. In summary, experimental and clinical studies have indicated that there are immunogens in ETS and that a portion of the pop- ulation is sensitive as shown by dermatological tests. However, the specific agent responsible for this reactivity has not been con- clusively identified. Furthermore, there is some question as to whether reactions to skin tests are correlated with subjective com- plaints. It is clear, however, that a substantial number of atopic individuals will have positive skin tests to tobacco smoke or to- bacco leaf extracts. More research needs to be done to characterize the imiuunogens and explain the relationship n_ ship between subjective symptoms and skin tests. SUMMARY AND ItEGOMMENDATIONS There are a number of acute, noxious effects of exposure to ETS by nonsmokers that may occur. These include de annoyance with odor, eye irritation, throat irritation, and immunological - - responses. The specific constituents that olicit these responses are -- - no _nown. TGG98~48

178 What Is Known Odor 1. ETS arouses odor responses. The objectionable odor gener- ated by ETS greatly exceeds that generated by simple occupancy under comparable conditions of occupancy, density, temperature, and relative humidity, and is more persistent. 2. Tobacco stnoke odor is stable over time. Ventilation re- quirements for tobacco smoke odor will therefore vary in strict proportion to the number of cigarettes smoked. - - 3. Rooms (and other spaces) where there is smoking require much more ventilation than.spaces with nonsmoking occupancy. During smoking, ventilation requirements that satisfy at least 80% of visitors to a room exceed 50 cfm per occupant. - - 4. Nonsmokers and visitors to rooms appear to set a more - stringent criterion than smokers for acceptability of tobacco smoke odor. Current ventilation guidelines for smoking occupancy will _ apparently fail to satisfy a criterion level of 80% of visitors (mixed group). It is not clear that any practical ventilation rate could sat- isfy 80% or more nonsmokers under typical conditions of smoking occupancy. Irritation • 5. Low humidity may exacerbate odor and irritation responses to ETS.. 6. Whereas odor will govern the reactions of visitors to a eniok- ing space, irritation will largely govern the reactions of occupants. Over time, eye irritation grows to become the most important - negative response of the occupant. Dissatisfaction observed in chamber studies is commensurate with that found in field studies., 7. Eye blink offers a reasonable correlate of sensory irritation at high levels of smoke (i.e., levels of ETS such that the -concen- -- - tration of CO is at least 5 ppm), but not at low levels. 8. Filtration of particles from ETS via an electrostatic pre- cipitator causes no decline in odor to visitors and no meaningful decline in odor or irritation to occupants. This suggests that - irritation and odor derive primarily from gas- or vapor-phase con- stituents. - 9. Filtration of particles via a Cambridge pad reduced irrk tation, but not odor, to occupants. Perhaps the Cambridge pad zsESS~,~,~ 179 removes some critical vapors from the smoke along with the par- ticles. 10. A substantial portion of atopic individuals are sensitive to tobacco leaf or tobacco smoke extracts as shown by skin tests. However, cutaneous sensitivity appears not to correlate with sub- -- - jective symptoms. What Scientific Information ls Miesing 1. The outcomes obtained in chambers regarding dissatisfac- - -_ tion created by the odor and irritation of ETS should be fur- ther verified in Reld.situations. The chamber studies imply that there must be considerably more than 20% dissatisfaction in places where smoking occurs even when current ventilation standards are met. 2. Prospects for abatement of discomfort through filtration of the vapor or particulate phases of ETS should receive attention. 3. Objective physiological or biochemical indices should be sought to validate reports of chronic irritation of the eyes, nose, and throat. 4. Research is needed to determine specific constituents that are the irritants in ETS. 5. Information is needed on the prevalence and severity of al- lergic and hypersensitive responses to tobacco smoke in the general p= ula - op_ tion and in atopic individuals. 6. Further research needs to be done to determine the specific elements that are immunogenic in extracts of tobacco smoke and to relate immune response on skin tests to subjective complaints of eensitivity to tobacco smoke. 7. Research is needed to evaluate the - medical importance of positive reactions to RAST tests of tobacco leaf products for -- s atopics. ItEFE- - ~'NCE3 ASIIRAE Standard 62-81. Ventilation for Acceptable Indoor Air Quality. Atlanta: ASHRAE, 1981. 19 pp. Becker, Q.Ci., T. Dubin, and H.P. Wiedemann. Hypnr.eneitivity to tobacco sntigen. Proc. Natl. Acad. Sei. USA 73:1712-1716, 1978. Becker, E7.f3., lt» Levi, and J. Zavees. Indnction of 1g1C antibodies to antigen isolated from tobacco leaves and from cigarette smoke condensate. Am. A P 1 th l . a o U8:2 .9-iE64, 1979.

180 Becker, C.Q., N. Van Ilamont, and M. Wagner. lbbacco, cocoa, coffee, and ragweed: Cross-reacting allergens tha activate factor-XX1-dependent pathways. Blood 58:861-867, 1981. Bick, R.L., R.L. Stedman, P.L. Kronick, E. Hillman, and J. Fareed. Stud- les related to tobacco glycoproteins A claimed activator of coagula- tion flbrolysis, complement, kinin and a clahsmd allergen. Thromh. - Haemostasis 46:231, 1981. Cain, W.S. Olfaction and the common chemical sense: Similarities, dif- ferences, - - ferences, and interactions, pp. 109-121. In 11.R. Moskowits and 11. Warren, Eds. Order Quality and Intensity as a Function of Chemical Structure. American Chemical Society Symposium 148. Washington, D.C.: American Chemical Society, 1981. Cain, W.S., B.P. Leaderer, R. Isseroff, L.(3. Berglund, R.J. Huey, E.D, -- - Lipsitt, and D. Perlman. Ventilation requirements In buildings. 1. Control of occupancy odor and tobacco smoke odor. Atmos. Environ. 17:1183-1197,1983.. Cain, W.S., T. Tosun, L.a. See, and B.P. Leaderer. Environmental tob_a_eco_ smoke: Sensory reactions of occupants. Atmoa. Environ., iu press. Clausen, (l. P.O. Fanger, W.S. Cain, and B.P. Leaderer. Stability of tobacco smoke odor In enclosed spaces, pp. 437-441. In 11. Berglund, T. Lindvall, and J. Sundell, Lds. Indoor Air, Vol. 3. Sensory and Hyperreactivity Reactions to Sick Buildings. Stockholm, Sweden: Swedish Council for - Building Research, 1984. Clausen, (i., P.O. Fanger, W.S. Cain, and B.P. Leaderer. The in8uenu of aging, particle filtration and humidity on tobacco smoke odor, pp. --- -- 246-350. In P.O. 8'anger, Ed. Glima 2000, Vol. 4. Indoor Climate. Copenhagen, Denmark: VVS Kongress-VVS Messe, 1986. Cometto-Munis, J.E., and W.S. Cain. Temporal integration of f pungency. Chem. Senses 8:316-347, 1984. E3leich, Q.J., and P.W. Welsh.. Immunochemicai and physicochemicai prop- erties of tobacco extract. Am. Rev. Respir. Die. 120:996-1001, 1979. Kerka, W.F., and C.M. Humphreys. Temperature and humidity effect on odor perception. ASHRAE Trans. 81:631-668, 1956. Lawther, P.J., and 1f.T. Commins. Cigarette smoking and exposure to carbon monoxide. Ann. N.Y. Acad. Sci. 174:136-147, 1970. • Leonardos, C3., and D.A. Kendall. Questionnaire study on odor problems of enclosed space. ASHRAE Uans. 77, 101-112, 1971.. Lehrer, S.B., M.R. Wilson, and J.E. Salvaggio. Immunogenic properties of. tobacco smoke. J. Allergy Olin. Immunol. 62:368-370, 1978. Lehrer, S.B., F. Barbandi, J.P. Taylor, and J.E. Salvaggio. Tlobacco smoke 'sensitivity"-Is there an immunological basis? J. Allergy Olin. Im- _ - : munol. 73.440,R46, 1984. Muramatsu, T., A. Weber, S. Muramatsu, and F. Altermann. An experi- ---- mental study on Irritation and annoyance due to passive smoking. lnt. Arch. Occup. Environ. Health 61:306-317, 1983. Speer, F. Tobacco and the nonsmoker., Arch. Environ. Health 1¢:443=446, 1968. Viessman, W. Ventilation control of odor. Ann. N.Y. Acad. Sei. 116:630-637, --- - - - 1964. Weber, A. Acute effncts of environmental tobacco smoke. Eur. J. Respir. Die. 68(Suppl. 133):98-108, 1984, 181 Weber, A., and T. Fischer. Passive smoking at work. Int. Arch. Occup. l0nviron. Health 47:409-941, 1980. Webor, A., T. Fischer, and E. arandjean. Passive smoking In experimental and field conditions. Environ. Res. 30:306-216, 1979a. - Weber, A., T. Fischer, and E. flrandjean. Passive smoking: Irritating effects of the total smoke and the gas phase. Int. Arch. Occup. Environ. Health - 43:183-193,1979b. Weber, A., 0. Jermini, and E. drandjean. Irritating eftects on man of air pollution due to cigarette smoke. Am. J. Public Health 66:672-676, 1976a. Weber, A., T. Fischer, and E. C3randjean. Objektive und subjektive phys- iologische Wirkungen des Passivrauchens. lnt. Arch. Occup. Environ. Health 37:277-288, 1976b. Winnecke, ©., K. Plischke, A. Roscovann, and H.W. Schlipkoeter. Patterns and determinants of reaction to tobacco smoke In an experimental exposure setting, pp. 351-356. In B. Berglund, T. Lindvall, and J. Sundell, Eds. Indoor Air, Vol. 2. Radon, if'assive Smoking, Particulates, and Housing Epidemiology. Stockholm, Sweden: Swedish Council ncil for - --- Building Research, 1984. - Yaglou, C.P. Ventilation requirements for cigarette smoke. ASHRAE Trans. 01:46-32, 1965. Yaglou, C.P., and W.N. Witheridge. Ventilation requirements, Part 2. ASHRAE Trans. 43:4Z3-436, 1937. Yaglou, O.P., I..Q. Riley, and D.I. Coggins. Ventilation requirements. ASHRAE Trans. 49:133-183, 1936. -- Zussinan, B.M. Tobacco sensitivity In the allergic population. J. Asthma Res. 11x169-167, i974. CGt;9el4Q

183 11 Effects of Exposure to Environmental Tobacco Smoke on Lung Function and Respiratory Sympt oms This chapter discusses epidemiologic studies of nonsmokers ex- posed to tobacco product smoke that have evaluated lung function or respiratory symptoms, most of which have evaluated children. The effects of active cigarette smoking are briefly reviewed to re- count the reasons why certain aspects of lung function have been studied in nonsmokers. The plausibility of finding sinlilar effects in nonsmokers exposed to ETS is discussed and the studies found in the literature are assessed. LUNG FUNCTION AND SYMPTOMS IN ACTIVE SMOKE1tS Cross-sectional studies of smokers have demonstrated that smokers, compared with nonsmokers, have (1) att increased prevar lence of chronic cough, chronic sputum production, and wheezing and (2) decreased lung function (see U.S. I'ublic Health Service, 1984, for an extensive review). The effects of suwking on 1xith respiratory symptoms and lung function may be seen within a few years of the onset of regular etnokinr (U.S. Public Health Service, 1979, 1984; Woolcock et al., 1984). Longitudinal studies have demonstrated that the mean rate of ducline with age of the 1- -- second forced expiratory volume (FEVt) is greater in smokers tllan in nonsmokers. In some amokers, the rate of decline of FI':YI is rapid, leading to clinically important chronic airflow obstruction. 1`he_ structural changes associated with active cigarette snlok- ing are seen in both the conducting airways and the pulmonary parenchyma (for a more detailed description, see U.S. Pu1,6c QeWst oeN mN.plalrl M'pMrophy ; hypsrplssla, mucous plsnd 1 N1 Increesed .uscePllblNty to vlrel low.r respiratory - haot InUcllons coush & sputum FIGURE 11-1 Known and suspected meehanisms for effects of tobacco smoke on airways. Solid lines = known mecheqi.ms; dashed line. = au.gected mecLanisms. - Health Service, 1984). In the large airways there is hypertrophy - - and hyperplasia of the mucous glands. These changes are followed by an increase in mucus production that leads to lncreased cough and aputum l/roduction. Structural changes in smaller airways range from relatively mild inflammation to narrowing and closure -- of airways duo to inAammation, goblet cell hyperplasia, and in= traluminal mucus. Changes in the parenchyma include increased nunlbers of inflammatory cells and ultimately destruction of the -- -- - alveolar walls, most commonly in the central part of the lobule, i.e., the development of centrilobular emphysema (see Figure 11-1). -- The link between airway disease and parenchymal disease is poorly unders_ tood. Smokers with severe functional impairment usually have nn appreciable amount of emphysema (U.S. Public Health Service, 1984). Cessation of smoking leads to a rapid decrease in respiratory ---- - symptoms, an improvement in lung function, and a shift towards the nonsmoker's rate of decline of FEVI (U.S. Public Health Ser- vice, ---- vice, 1979, 1984). These improvements are usually seen regardless of the functional level at which cessation occurs. PATHWAYS FOR EFFECT OF A6'i1VE 8MO_ KIN4_ ON AIRWAYS AND PARENCHYMA ~nr~wnrnuoon ~ ~ ~ hl l ~ ranc a ~ hypsrrsspon.Msness nsuUOpldlln/lux Incissssd ; permsablNty 'j .y to atlerycns ~ ~ sisNass / X 1 f burden ~ 1 4E VGE'i9944e 182

184 Population-based studies of. adults have generally shown a strong dose-response relationship between FE:VI with dose iiiea- sured either in terms of years smoked, the number of cigarettes per - day, or the integrated dose, i.e., pack-years (U.S. Public Health Service, 1984). It is worthwhile noting, however, that in two ma- jor studies (Burrows et al., 1977a; Beck ot al., 1981) the active smoking dose accounted for only about 15% of the variation of FEVI even after age and height adjustment. Most of the variance could be attributed to the naturally occuu ring large variability in pulmonary function. Another reason_ the active smoking dose did not explain much of the variance is that the number of cigarettes an individual smokes cannot readily be translated into the doie of - - smoke that is delivered into the airways and parenchyma. Many - - factors, such as puff volume and lung volume at which_ inhala- tion starts, clearance rates, and airway geometry of the lungs of exposed individuals, will influence the dose and the distribution of the smoke within the lungs. • Variabiiity in individual suscepti- bility to the effects of cheinicals deposited in the lung has been demonstrated in studies of animals (Evans et al. 1971, i975,1lf78). PLAUSIBILITY FOR AN EFFECT DUE TO I'AS3iVE SMOKING The dose of cigarette smoke delivered to the lungs of non- smokers exposed to ETS is both qualitatively and quantitatively different from mainstream smoke, being a small fraction of that delivered to the lungs of an active smoker (see discussions in Chap- - ter 7). Exposure to constituents of tobacco sinoke may begin in utero and continue throughout childhood -through ETS expoHUre. During these periods, the lung is undergoing both growth and remodeling. Therefore, the lung of the fetus and young child may be particularly susceptible to environmental insults. Despite qualitative differences between mainstream smoke, sidestream smoke, and ETS, it has been customary,to assume - - that exposure to ETS approximates a low-dose exposure to to- bacco smoke. The ability to measure responses to low doses de- pends on the shape of the dose-response curves, the sensitivity and specificity of the measurement tools available, and whether there - is a threshold of exposure below which there is no response in any individual. 185 The assumed shape of the dose-response curve determines what kinsls of effects would be expected and the estimates of the - - probability of detecting them. If the dose-response curve were linear with a shallow slope, or a slope concave to the dose axis, the response at low doses might be so small that it would be difficult lt to detect. In such a situation, only the very susceptible portion of the population might have detectable effects. It is likely that there - is a distribution of susceptibility to the effects of ETS within the population, such that there will be some persons who will respond at low doses ses and some persons for whonk many years of heavy - exposure may be needed to cause the sanie symptonis or change in lung function (Cockcroft et al., 1U83). If individuals who are most susceptible to the irritating effects of cigarette smoke on the lower respiratory tract do not start to smoke or, having started, soon quit as smokers, then a population - of nonsmokers would be more likely to include the uiost suscep- tible individuals than a population of smokers. The existence of different subpopulations introduces an additional complication to the extrapolation from high-dose exposure in active smokers to the low-dose exposures of nonsmokers. In addition, it is likely that the development of respiratory - disease or symptoms, lung function level, and rate of decline reflect the cumulative burden of many environmental exposures and other - insults, such as respiratory infections (Purvis and Ehrlich, 1963) to the luug. Furthermore, it might be hypothesized that the cumulative burden may interact with the individual's genetically -- determined susceptibility. METHODOLOGIC CONSIDERATIONS FOR EPIDEMIOLOGIC STUDIES A recent report of the National Research Council (1985) is devoted to methodologic issues of epidemiology and air pollution. - In this section, many of the problems are reviewed briefly. Study Design and Analysis Chronic pulmonary effects of ETS have been the subject of - - several recent reviews (Lee, 1982; Weiss et al., 1983; Surgeon - - - General, 1984; Guyatt and Newhouse, 1985; Taylor et al., 1986) and symposium or workshop reports (U.S. Public Health Service, S6G98448

186 1983; Gammage and Kaye, 1984; Rylander, 1984). Many of thu studies reported in these reviows had not been originally designed - to study chronic pulmonary effects of ETS exposure. Instead, these data sets were reanalyzed to address the question of the pulmonary effects of ETS. This use of these studies suggests thu need for caution when interpreting their results. Several analytic approaches were used iii the reported stud• ies. Independent risk factors, such as age attd sex, usually neod to be taken into account, but this was not always done. Several statistical approaches, such as stratification or regression analysin, are used to take into account the effects of tootentially confound- ing variables. For >tnost of the potentially confounding variabiett, researchers do not -agree on the nature of the roles of the variables - - a» confounders and, hence, on the appropriate ways to introduce these variables into the data analyses. Assessing Exposure Interpretation of epidem'iological studies is hampered by the existence of factors that interact with and modify the response to -- - exposure and by confounding factors that are associated with tlte same symptom complex as exposure to_ ETS, such as coughing, production of sputum, and wheezing (see Table 11-1). Thede variables must be assessed and accounted for in the statistical analyses where possible. Unreported active smoking could lead to a large bias. Underre- porting of smoking is likely in studies of older children, particularly when parents answer questionnaires for their children. Children. - who have parents who smoke are themselves more likely to smoke. Therefore, because active smoking is likely to have a considerably greater impact on respiratory symptoms and lung function than - exposure to ETS, miselassil'ication of the children who smoke will tend to overestimate the effect of exposure to ETS. - For blue collar males, occupational exposure can also be itn- portant and may interact with both direct cigarette smoke and - - ETS. Many pulmonary toxicants can exist in the workplace. Fur- thermore, ETS exposure can occur in the workplace. Similarly, comparison of inner-city-dwelling persons with less urban, or sub- urban, controls can lead to biases. 187 TASt,e 1 1-1 Potentially Confounding and Effect Modifying Factors In Epidemiologic Studies of Exposure to Environmeutal Tobacco Smoke Unreported nreporied active smoking Tobacco products Marijuana . Clove cigarettes f_kvebpmental factors Maternal smoking during pregnancy Factors related to outdoor environment -- Outshrar temperature, humidity Respirable and nonrespirable particulates, e.g., fugitive dust Pollens and other allergens Factors related to indoor environment Crowding Number.nd age of siblings Total number of peopkfanimals in dwelling unit Total number of smokers in dwelling unit Houschold conditions Frequency of air exchanges - - Temperature and humidity Use and condition of air conditioning units Conditions of child care facilities Unvenled combustion products from heatbiglcouking stoves Respirable and nonrespirable puticplates, e.g., worid smokes Pollens, molds, mites Allergens and infectiou: organisms Formaldehyde Factors related to work/hobbics Work/hobby-related exposure to gases, fumes, particulates Misceilaueous factors Annoyance response to tobacco smoking Reporiing bi.ses Assessing Respiratory Variables Methods commonly used to assess the elfect of passive smoking - on the respiratory system, such as respiratory symptom question- naires and measurement of lung function, may lead to some error. The problems associated with the respiratory symptom ques- tionnaires include: • Different questionnaires are used in studies. Differences in how the questions are asked can sometimes lead to large differences in answers. For instance, asking `Are you a smoker?" may ay elicit a "No" response from an examoker whereas th_ equestion "Ifave you ever smoked?" would be answered "Yes". 9GG984.48

........ -.~;:;;,;;; 188 . Some studies use a self-administered questionnaire, where- as other studies use a trained interviewer. `Aaincd interviewers can determine whether the subject understands the questiounairo - and can follow a prescribed set of probing questions that may help -- - to resolve the specific nature of not-well-deacribed symptoms. -- . Some studies have parents complete the questionnaires for the children, whereas other studies have the child answer tho questionnaire. For older children, parents may not be aware of active smoking by the child and exposures to ETS in environments outside the home. . Questionnaires necessarily involve some subjective ele- ments that are prone to recall bias. For example, a smoker who is symptomatic may be more likely to report the same symptom in his/her child (Schenker at al., 1983; Ferris at al., 1985). Many tests are prone to measurement error, which tends to obscure dilferenc-es between groups of subjects. For example, it may be necessary to repeat lung function measurements for a given individual and to average results to get a reliable es4imato. Lung function tests are often not sensitive to the structural and functional changes associated with lung disease (Drill and Thomas, 1980). CROSS-SECTIONAL STUDIES In the following sections, selected cross-sectional studies of respiratory symptoms, lung function, and respiratory infections and longitudinal studies of lung functions are reviewed. The stud- ies reviewed here are larger studies in which attempts have been made to standardize assessments and many of the datargathering techniques, including interviews. Studies of Respiratory Symptorns In Children Almost all of the cross-sectional studies that have compared children of parents who smoke with the children of parents who do not smoke have reported increased prevalcnccs of respiratory symptoms, uHually cough, sputum, or wheezing, in the children of smoking parents. Some studies, including some that have not found a statistically significant increme in the prevalence of res- piratory symptoms in ETa exposed cbildron, have demonstrated 189 an increase in respiratory symptom prevalence with an increasing number of parents or other adults who smoke in the home (see below). Three problems are especially important for studies of respi- ratory symptoms in children, i.e., underreported active smoking on the - part of children, recall bias leading to overreporting of symptoms by parents, and the confounding variables of infections in parents. All three may lead to overestimation of symptom prevalences among children of smokers. Recall bias would occur if parents who have respiratory symptoms are more likely to re- port those symptoms in the children. (The possiblity also exists that parents with these symptoms would look upon them as so commonplace as not to be worthy of mention). Parents who are smokers are also more likely to have more respiratory symptoms and respiratory infections. Respiratory infections (and, as a conse- quence, symptoms) among children of smokers may be the result of direct transmission of infectious agents from the parent or may be caused by inflammation and nd irritation of lung tissues due to ETS exposure and consequent increase in susceptibility to infection, It has been observed that parents, especially mothers, who have a - history of severe respiratory illness report higher rates of respira- tory tory symptoms in their children (Schenker at al., 1983; Ferris et al„ 1985). Various ways of dealing with these potential sources of bias have been proposed.. Restricting the study or analysis to children below age 8 is likely to eliminate bias due to underreporting of children who currently smoke. It is more difficult to handle the overreporting of symptoms in children when h_ en the parents have respiratory symptoms. An additional problem for interpretation of parental reports of respiratory symptoms was noted by Schenker et al. (1983). In their study, children whose questionnaires were completed by fathers had signiAcantly fewer symptoms reported than children with mother-completed questionnaires. There was no comparison of questionnaires completed separately by both mother and father - for the sanne child. Because the rates for symptoms as reported by the mother were similar to what was found in other studies and because the fathers reported significantly fewer symptoms, the investigators suggested that fathers underreported symptoms in their children. r 4f 695448

190 Table 11-2 reviews several selected cross-.sectional studies of respiratory symptoms in children and adults. Lebowits and Bur- rows (1976), reporting on children in the 7licson Epidemiologic Study of Obstructive Lung Disoase, emphasized the need for con- trolling for parental symptoms. They reported that children had a higher prevalence of respiratory symptoms if they lived in house- holds with adults with the same symptom, regardless of the family - -- - smoking habits. When the presence of symptoms in the adults was taken into account by partitioning households based on presence - or absence of adult symptoms(s), the odds ratio that remained was greater than unity but was no longer statistically significant (Mantel-Haenszel odds ratio for all respiratory eymptorns calcu- lated _ lated from data presented is 1.15 (95% confidence limits of 0.91 to 1.98)]. Most symptoms were reported more frequently for children in currently smoking families. Ferris et al. (1985) have argued that correcting for parental. symptoms represents an overcorrection for respiratory symptoms in children since it also corrects for the parents' smoking habits. In the Harvard Air Pollution Respiratory Health Studies (Six- Cities Study) of 10,106 white children aged 6-9 years, the variable indicating whether the parent had a history of bronchitis, emphy- sema, or asthma was found to be a highly significant independent risk factor for cough and wheeze and a history of respiratory ill-s ness among children (Figure 11-2). Children whose parents had a positive history had 72-155°f6 higher symptom and illness rates than children whose parents had no history of these illnesses. Ad- - - - justment for parental respiratory history reduced the size of the estimated effects of maternal smoking on respiratory symptoms and illnesses by 20 to 30%, but the associations remained statisti- cally significant for most of the outcome symptom and respiratory illness variables (odds ratios of 1.23 and 1.28, respectively). In both the Lel,owits and Ferris studies, adjustment for par- ental symptoms or respiratory illness decreased the strength of the apparent association between exposure to ETS and reepira- tory symptoms, but did not eliminate it. 'rhis finding leads to the reasonable conclusion that the exposures typical of ETS are sufficient to cause respiratory symptoms in some children. The ui- creases in frequency of cough were 20 to 50%, and as high as 90%, when there were smoking parents. The increases in frequency of wheezing were more variable, which may indicate the difficulty in a ~ ~ ~ 0 ~ ~ 9 u b ~ L% ~ ~ ! ~ ~ ~ N ~ h 191 .11 1101-:91 R ~ I N f3GG9B44B

~ ~ ~ e v E 9 ~ ~ y 0 ~ ~ ~ ~ ~ i *. I 192 ° ~ Y sti~ 3 - 192 CURRENT MATERNAL SMOKING (CIG/DAIf) FIGURE 11-2 Relative odds of respiratory illness or symptom. versus average daily cigarette smoking by the child's mtlur. Reference value is - zero cigarettes per day.. From Ferris et si. (1985). assessing this symptom. Furthermore, there appears to be a dose-e response relationship between exposure and the likelihood of the - child's developing respiratory symptoms or a respiratory illness. In the Harvard Study, a significant dose-response relationship was reported; the more mothers who smoked, the greater the risk of reepiratory symptoms and illnesses among their children. Studies of Lung Function In Children A more quantitative measure of the impact of ETS on the lung is obtained by measures of lung function. Many of the studies that have examined the relationship between passive smoking and lung function have been cross-sectional. 6669B44B

194 Most studies have examined the effect of exposure to parental smoking rather than ETS exposure outside the immediate family. It is assumed that children are less likely than adults to be exposed to occupational irritants. The cumulative burden of respiratory insults is, therefore, likely to be smaller in children than in adults. It is often difficult (but not impossible) to measure lung function in young children and also hard to dissect out the relative con- tribution of ETS and that of natural variation and the effect ffect of respiratory infections to pulmonary damage. A majority of the studies (reviewed in Table 11-3) has shown a small decrease (up to 0.5% FEVI per year) in rate of increase in lung function associated with normal growth in children living with one or more parents who smoke compared with those living with nonsmoking parents (Table 11-3 and Figure 11-3). These differences have usually been statistically significant. Although the mean effect is small, there are individuals in each study who have large decrements in growth of lung function, Some studies have - - - found a dose-response relationship with the number of smokers in the home or the amount smoked (Hassclblad et al., 1981). Ware (1984) shows (see Figure 11-4) a highly significant negative association between maternal smoking level and FEVI at both the baseline and follow-up examinations. For a child of a mother who smoked one pack of cigarettes per day compared with a child of a nonsmoking mother, the FEVi was 0.7 f 0.2% lower at the baseline examination and 0.8 ± 0.2% lower at the follow-up examination 1 year later. This amounts to a 10- to 20-m1 difference for a child with an FEVI between 1.5 and 2.5 L. In most studies, only the maternal effect was statistically significant. This may be because mothers usually spend more time with their young children than _ fathers. A study carried out in Shanghai in the People's Republic of China reported a clear paternal effect. Chen and Li (1986), in a cross-sectional study of 303 boys and 268 girls aged 8-16, found that the number of cigarettes smoked by fathers was linearly re- lated to a decrease in FEYI and FEFs5_rb%, the average forced expiratory flow during the middle half of the period of expira- tion. None of the mothers in this study were smokers; therefore, there was no maternal effect in that population. Differences in - - father's smoking status accounted for 0.5% of the variation among individuals in FEVI and 1.2% of the variation in FEF23_76%.. 444 vn 195 I 0004B44B

I i4 196 8 ~15 S. it 0.!^ Y Wk ';5g u3 ~~~ L it I jila I 197 FEVi FVC Vmax g0 p s,003 pK.02 NS MALES 110 a Both p+renls never srnokers ~ Both parents cwient snakers FEV1 FVC Vmaxg0 NS NS pc.02 FEMALES FI(QURR 11-3 Me.a pere.nt lung funetbn, by pawntnl smoking, of non- smoking malee and femala, ages 10-19, 1983-1fl86, from Tecumseh, Michigan. surcha.l .t ai. (1986). The most important contributors to variation in lung function among children are size-related factors such as sex, age, and height. These accouut for about 50-60% of the variation (Comroe et al., 19s2).. It is not Ilossible to determine whether C't'S is directly caus- ing ing the decreased lung function observed in children of smoking ;' parents or if an increased infection rate in these children (see be- k low) ia responsible for the decrease. The annual small decrease f in FEVI, which is related to exposure to ETl3, is unlikely to be 'F clinically significant. However, the effect may lie important in two $ respects.. First, the existence of statistically significant dilferences related to parental smoking leads to the conclusion that there are pathophysiologic effects of exposure to E'lS in the lungs of.the growing child. It may be an in utero effect, an effect on the grow- ing and remoaleling lung, or both. Second, it raises the question of whether the child who is adversely affected by parental smoking 100 80 1oo44e44e

i98 +.02 +.01 0.00 3 -.02 -.03 } 1 ~ I I t t t 0 20 40 60 MOTHER'S DAILY CIG FIGURE 11-4 Mean of pulmoaary function residuel (f I SD) by mothers' reported daily cigarette smoking, compared with chiklreu whose mothers have -never smoked. Squares represent the first exaruination (n - 7,112) aud triangles represent the second examination (tt =®,278). From Ware et at (1984). may be at an increased risk for the development of chronic airflow obstruction in adult life. An accelerated decline in lung functiun - could increase the risk of chronic pulmonary disease (Sarnet ot nl., 1983). Studies of Lung Function bt Adults White and Froeb (1980) studied 800 nonsmoking, tniddle-aped subjects, out of a total population size of 2,100, and found a sniall statistically signilicant decrease (8%) in FXFza_76% in both nien and women who were nonsmokers exposed to ETS. The reported reduction in FEFz6_ra96 for ETS exposed uonsmokers was alnioat identical to that of the smokers of 1-10 cigarettes per day. This raises questions about their findings. This study may suffer from problems of selection bias in the allocation (if subjects to categories and the absence of any examokers (Adlkofor, ct al., 1980; Avindo, 1980; Huber, 1980; Lee, 1982). 199 A cross-sectional study from France (Ii;aulfmann et al., 1985) supports the conclusions that exposure to ETS may have an etfect - on lung function in nonsmoking king adults. The French Cooperative Study surveyed more than 7,800 adult residents of seven cities in - France in 1975 and< found 1,675 were true nonsmokers. In men - -- ~- - and women over 40, nonsmokers of either sex who had a spouse --- - who smoked lrad a significantly lower FEFzb_ra% than those living with a nonsmoker. These differences were not explained by social - --- -- clasH, educational level, air pollution, or family size. Among the wonten, there was also a significant difference in FEVI and a dose effect was seen with the amount smoked by their husbands. These differences, only apparent in persous over 40, were small and were uncovered only following detailed examination of the data after the hopulatiou had been stratified by age. Two other cross-sectional studies involving adult women have found an effect of exposure to ETS on lung function. In a study of 220 married women aged 25 to 69 years from five U.S. cities, Ifauffmann and coworkers (1986) reported that standardized resid- uals for FEVI and FEVt/FVe* for the group identified as pas- sive smokers were intermediate between the results of nonsmokers and current smokers. In a study of 163 nonsmoking women liv- ing in a rural area of the Netherlands, llrunekreef and coworkers (1985) found that those exposed to ETS tended to have slightly lower mean values for all of the lung function variables measured. These differences reached statistical significance for peak Ilow and FE1096_g6% in the 40- to 60-year-olds. The numbers in each of their groups were small. No information was given on possible chiklhood exposures to cigarette srnoke of the wnmen studied. Kentner and coworkers (1984), in a study or 1,351 white collar workers (941 tnen and 410 women) in northern Bavaria, and Com- stock et al. (1981), a study that included 1,724 adults residents of Washington County, Maryland, examined the potential effects of ETS. In these studies, information was collected from subjects using questionnaires and the subjects were then classified as never smoked, exsmekers, and current smokers. The Kentner et al. study evaluated honte and workplace exposures, whereas the Comstock et al. study evaluated only home exposures. In the IEentner et al. (1984) study, an additional classification was made for other smok- ers, representing those who were cigar and pipe smokers. These ~ LrVO is the forced vital capacity. ZQD4Q449

200 inveetigators found no significant reductions In lung function with - ETS exposure. In view of the large number of factors that affect lung function, - it is not surprising that it is difficult to document the extent to which a single type of exposure affects lung function. The lungs of adults have been subjected to many environmental exposures and potential insults over a lifetime, making it unlikely that a specific effect could be isolated. The variability in lung function due to differences of the other factors tends to obscure effects of a single variable. In addition, results in adults should be evaluated for possible misclassification of exsmokers or occasional smokers as nonsmokers, as well as possible confounding by occupational exposures to other pollutants or to ETS. LONGITUDINAL STUDIES OF LU NG FUN CTION IN CHILDitEN AND AD ULT3 An important unanswered question is whether exposure to ETS affects the way the lungs grow and develop during childhood. Respiratory symptoms, by themselves, may have little clinical significance but would be irriportant if assuciated with a change iri the rate of lung growth and development or the development of pulmonary pathology at older ages. There is evidence from two cohort studies (Table 11-4) that parental smoking may affect the rate of lung growth during child- hood. Tager and coworkers (1983), who have followed 1,166 el- ementary school children in East Boston, Massachusetts, over a 7-year period, reported that maternal smoking was associated with a reduced rate of annual increase in FEVr and Fl,',F2s_-rs%. There was a reported 3-5% decrease in expected lung growth over the 7-year period. IlurchHel and coworkers (1988) examined pulmonary function in 3,482 children in Tecumseh, Miclugan Children 0 to 19 yoare old were followecl for 15 years, during which tinie questionnnire information was collected from both paients. FEYI and I+VC values were significantly lower by 5% in mde nonsmokers 10 to 19 years of age w-ho9e parents were current smokers. The Harvard Air Pollution Respiratory Health Studies (Forris - et al., 1985; Berkey ot al„ 1986) (Figure 11-5) show' a relatively smaller effect than that reported by Tager aiid coworkers (14183). The Harvard Wedy included 7,834 children between the ages of ! 1 I ~ I h 201 $ €e ~ Is Jilt1 ~ 9~ V V 8-a ~51 '0 11 .. o a .. -•~ ,,~ .~, ~~ AI ~ &4 w EOOZ.9444

202 6 and 10 years who were followed over a 6-year period. Children whose mothers smoked one pack of cigarettes per day had FEVI levels, at age 8, that were 0.81% lower than children with nonsmok - - ing mothers. Growth rates for FEVt were approximately 0.17% per year lower. For a child aged 8 years with an 1rEVl of 1.82 L, this corresponds to a deficit in rate of growth of F1.V1 of approxi- mately 3 ml per annum and a deficit of 13 ml by age 8. In contrast to the lower FEVI seen in children whose mothers smoked, higher levels for FVC were observed in children with smoking mothors compared with children whose mothers did not amoko. For exam- ple, average FVC at age 8 for a child whose lnother smoked one pack per day, was 0.33% higher than a child with a nonsmoking mother. On the other hand, the growth rate I'or FVC was 0.17% lower for a child with a smoking mother. This would be equivalent to a 2.8 percent decrease in pulmonary development throughout childhood and implies a decrease in the development of pulmonary function in children of smoking parents. In view of the effects that climatic conditions can have on - housing characteristics, and subsequent ventilation rates, it would be advantageous to conduct longitudinal studies in regions of the - - United States other than the Northeast. In any future studies, great care should be taken, as it was in the two cohort studies, to. account for potential confounding variables in the analyses, sudi as socioeconomic status and gas cooking. Another aspect that --- - - deserves more attention in future studies is the effect on children's pulmonary function when parents stop smoking. THE EFFECT OF PASSIVE SMOtC1NG ON ItESPIR.ATORY INFECTIONS There is now strong evidence that bronchitis, pneunionia, and other iower-respiratory-tract illnesses occur more frequently (at least during the first year of life) in children who have one or more parents who smoke (ace Table 11-5). Evidence that this increased frequency of acute respiratory infections continues into later childhood is-less convincing, although the evidence from both - cross-sectional studies and cohort studies shows such a trend. -- - Harlap and Davies (1974) followed a cohort of 10,672 infants born in Israel between 1965 and 1968. Admissions to the hospital - during the first year of life were recorded. Information about ma• ternal smoking was obtained during the pregnancy only. Infants VO0/'sh4S as0 a2o o.to J Q ~ O N W at i I 7 8 203 i I 9 10 SLOPE a -.04339 f Yf1 LEVEL - -0993 8 9 -0.20 - 6 7 J 10 1.70 1.60 1.50 1.40 < 1.30 120 1.10 1.00 v 0 A ~ .95 m N 0 m .90 E3 ~ 4 ~ .85 m 0 ~ 0 .80 AGE N'1GIUltE 11-6 Calculation of growth rate and level of In(FEVt) for an Individual child. ti'he ruiduale In the upper panel, i.e., the difference between observed and predicted In(FEVI), were regressed on age In the lower panel. - - From Ilericey it al. (1986). with major congenital malformations and those dying before their first birthday were excluded from the study. For the total popu- lation studied, there were 25.4 admissions per 100 babies under - - - - 1 year of age. The infants of mothers who smoked had a 27.5% - greater hospital admission rate for pneumonia and bronchitis than - children of nonsmoking mothers. A dose-response relationship was also found between the amount of ruaternal smoking and admis- sions to hospital for pneumonia and bronchitis. Colley (1974; Leader et al., 1976) carried out a similar study in London. The study involved a birth cohort of 2,205 infants born between 1963 and 1965. In this group of children, the incidence of pneuntonia and bronchitis in the first year of life was associated with the parents' smoking habits. This was true whether or not the parent has respiratory symptoms. The incidence was lowest for children of nonsmoking pkreots, highest in families where both

~ O ~ rABt.E'i1-5 ~Ch'ildhood Respira'tory'Lract IIlness and Passive'Smoking ~ Facpusuro Assessment ~j Studv Source of'Subjeets Subjects and,Health Information Findings Comments Harlav and Bicth cohotr'WVest YWl itlfants in achort df Ilntrnatall interview of 1. 'Significantly more Davies. 7erusakm: i19t5+1969 10.672 adteittcd'to mothers admissions forbnon- 197+4 hospiital in 7ernsalem ehitir or pneumonia. especiaNy' in .rintet., in infants whose mothers smoke 2. Dose-tssponse'for nutnberof cigarettes by mother and, excess of bronchitis and pneumonia Colley. Birth: cohort; Harrow. 2.ZOS infants Annual'foilo.v-qp'by 1. lneitlence of ,ptrcumo- 1974; !UK:;1963-1965 health visitors' for 5 nia and bronchitis in l.esder yr qeestiotuui" fmst year: aasociatr:d etal.. adminirteted'by with parental smokinp,: 1976 trainedihealth viditor ii6dence lowest with 0 Ramamallio. Birth cdDon: Ydorthern 1978 Fnland; 1966 LL.lu" U exposed. 1.821 uneqposed; ager0-5 Smoking determined in interview during pregnancy Said i Cohosti Franaq; 3,920 ehi/deem; ages 5elf-administeted by el sll„ 11975-1976 10-20 children . 11978 i Fergusson Birth cohort !Chrsst- 1?6-5 infants Folb+-up by stxnetturd et al.. ehut ¢h. New Zealand; intet.iewa with 1981 1977 mother at'bitt`h, +4 mo. 1. 2, and 3 yr, diaries icept by tooth- ers on thild's history of inediea'1 csee; check .rith hospital records both nonsmokers. highest with i both smokea, intermediate .rith one smoker 2. Associations'itaconsis- tmt after l;yr 3. 'Ia, first .ear of life. EI'S qqposure tloubbd risk for paeaaroeia/ 'b.eeehitis Significant increase in hospitalization for >espiratoey iliMas Increase in tonsillectomy and7oradetwidertomy 1. Lower respcator.. iUnn% signdinMig telated to mother's stnoking' in fitst year of life, egui.ocsf in se¢ond, and abseat in third 2. No effect of paternal smoking 3. ' Linear dosnre;poase between materelkl smoking and'incidence of',kwwer respiratoty infections Pedteira. Birth eohott from 1.144 infants followKd Interview with mother 1. ?raaheitis and bron- 19'85 practice of four for l,yr after birfh at fust;..ell baby chitm signifuwl;ly Zediatric++ns in exam carrie id out by related to materaal suburb df Washing- doetoc; al4sdbsequttnt smoking ton.' D:C.:' 1976-',198I1 office risits in first , 2. No dose-response year of life'for lower relationdhip respiratory tract 3. Bronchiolitis not infection related to;parental smoking lufotsnation about moth- et!s smoking obtained peenatally. aot concur- raa't with child?t admis- sion; no'infoimation about'father's smoking obtained Mfost lintportant determi- nant of respiratory inness .aa bevndiitis or pneumonia' in sibling: anaksis not controlled for number of siblings !Qnly.n1ate71al smoking evaluated; categories based on smoking during pregnancy smoking by,parems may not': hanceoincided or p M.Ap operations Analysis eoatiolled'for maternal age, eflua- tiott. faaoiiy.size, family li.ing aonditions NoadJ'usttnent made:for potentially confounding ,.-ariabler zessivelr affluent area and' low maternal smoking rate (19%)

9Q04B44B ~ w .i: ~ ~ 206 ~~ 1 ' . ~ 11 .1 a I x .§ ~ .11 1 ~ ~~ u ° I H ~ ~a A ; ~ ~ ~~.. O " ae' GI~i I y_ U O ~~, ~ N 207 parents smoked, and intermediate where one parent smoked. This effect was not seen consistently ovor age 1. A third birth-cohort study, involving 1,265 children in New Zealand, was reported by Fergusson et al (1081) They studied !he children from birth to age 3 years and found an increase in both brouchitis/pneumonia and lower respiratory i1Ltess during the first year in children whose mothers smoked. During the second year, the relationship between maternal smoking and lower respiratory illness was equivocal. The relationship disappeared by the third year. There was no effect observed of paternal smoking on the incidence of lower respiratory illness. Using logistic regression, they found that the rates of lower respiratory illness were related to maternal smoking. For each /ive cigarettes smoked per day by the mother, there was an increase of 2.6-3.fi lower respiratory "events" per 100 children at risk. Adjustment for maternal age, education, family size, arid fainily living conditions did not change the relationship. Rantakallio (1978) studied the effect of maternal smoking dur- ing pregnancy on morbidity and mortality of children to age 5 based on 12,t)(i3 births. Smoking status on the mother was only available from antenatal interview. Perinatal mortality was not higher among children of smokers, however, 1>ostneonatal mor- tality (between 28 days and 5 years) was significantly increased. Children of suiokers were hospitalized for respiratory illness sig- nil'ienntly more often than children of nonsmokers and the average duration of hospitalisation was longer among children of smokers. '1`wo case-eontiol studies evaluated smaller groups of children hospitalized for respiratory infection and nonhospitalized controls. Pullan and Hay (1982) studied 130 children who were hospitalized with a documented respiratory syncytial virus (RSV) infection in infancy and 111 controls. They found that children hospitalized with documented RSV infections tions were more likely to have mothers who smoked and that the children had an excess of wheeze and asthma and lower levels of pulmonary function, which persisted to age 10. Sims ot al. (1978) also suggested that cigarette smoking by parents during a baby's first year of life is associated with a__n increased risk of RS V infections. Speizer et al. (1980) studied approximately 8,000 children, aged (3.10 years, from six communities in the United States as part of a prospective study of the health effects of air pollution (Harvard 11it Pollution Respiratory Health Studies). Parental smoking and

208 sex of the child was associated with respiratory disease before age 2, after other variables had beon taken into account. Children from households with gas cooking also gave a history of more frequent respiratory illness before age 2 than children front households wi1.h electric cooking'. Dutau et al. (1981) studied 892 children under age 6 in the south of France who were seen by a pediatrician or hospitalized for various reasons. They found a significant correlation between tlte, annual incidence of pulmonary infections and the total number of cigarettes smoked inside the house. Pedreira et al. (1985) followed all newborns (1,144 infants) seen by a group of pediatricians for a first well-baby examination between 1976 and 1981. They found that tracheitis and bronchitis occurred significantly more frequently (89% and 44%, respectively) in infants whose parents smoked and that maternal smoking iin- posed greater risks upon the infants than paternal smoking. One study looked at the frequency of tonsillectomies and/or adenoidectornies in children (Said et al., 1978). They found the frequency was significantly increased among children with smoking parents. However, the smoking status reported for the parent may have been current smoking status, even though the operations had occurred 5 to 15 years previously. All the studies that have examined the incidence of respira- tory - - tory illnesses in children under the age of 1 year have shown a positive association between such illnesses and exposure to );l'S. The association is very unlikely to have arisen by chance. It may --- - represent a direct association between ETS exposure and disease (a causal explanation) and/or an indirect one (noncausal) arising because children living in homes of smokers are at risk of such diseases for other reasons. Some of the studies have examined the possibility that the association is indirect by allowing for confou_ nd- ing factors-such as social class, parental respiratory illnesses and birthweight-and have concluded that such factors do not explain the results. This argues, therefore, in favor of the causal expla- nation. Such au explanation is supported by the evidence of a dose-response relationship specific for rospiratory disease (Tables 11-6 and 11-7). Also, the mother's smoking is more likely to af- fect the infant than the father's smoking, since the proximity of - - mother and child is closer during the child's first year wheu the e(lect is more marked and consistent than later in childhood (see - Fergusson et al., 1981). This also supports a causal, rather than 4004,94449 209 an indirect, explanation. Therefore, the evidence indicates that smoking in the home does increase the incidence of respiratory illness in infauts. The meehaniszn for this increase is less certain. It could rep- resent a direct effect of ETS on the respiratory tract of the infant or it could be due to such infants' being exposed to more parental respiratory infections as a result of their parents' smoking. Ei- . ther way, smoking in, the home appears to increase the rate of respiratory illness in young children. W7IEN DO PULMONARY EFFECTS OF PASSIVE SMOKING OCCUR? The weight of evidence is that there are clearly observable effects of ETS on the respiratory system. These effects include an increase in the incidence of acute respiratory infections in early infaucy; increased prevalance of cough, sputum production, and wheezing; and a decrease both in lung function measured at an instant in time and in the growth of lung function. The finding of differences in symptom prevalence, respiratory infection rates, and lung function among children exposed and not exposed to ETS is often interpreted 'as evidence of a chronic effect of ETS on the airways. This is probably true, and it is unlikely that ETS is not an upper- and lower-respiratory-tract irritant in children. The possibility that there is an effect of maternal smoking in utero as well must be considered. Evidence of an in utero effect in pregnant rats exposed to whole tobacco smoke has been reported by Collins et al. (1985). These investigators reported that pregnant rats exposed to smoke daily from day 5 to day 20 of gestation, when compared with control rats, showed reduced lung volume at term and saccules that were reduced in number and increased in size. The internal surface area of the lung was decreased. The relevance of this study to maternal smoking during preguancy in humans is not yet clear and deserves further investigation. Other factors that may alter the time when ETS effects dur- ing childhood include the relative immaturity of the immunologic system an the growth and remodeling that are occurring in the immature lung. The infant lung differs in a number of impor- tant ways frouw the adult lung: (1) 'f _lymphocyte and macrophage functiori are iiot fully developed at birth, (2) there is increased susceptibility to infection as $ result of comparatively immature

210 211 'rABLB 11-7 Admission Rates in the First Year of Life for Bronchitis and Pneumonia per 100 Infants by Maternal Smoking and Number of Cigarettes Smoked Daily (Number of Infants in 1'arentheses) Nonsmokera Never Smoked Former Smokers $mokers (tigarellct per day) Total (8.900) (786) 1-10 11-20 31-F (10,672) (747) (179) (60) 9.6 7.8 10.8 16.2 31.7 9.8 NOTI?: Differences among three calegorks of smokerp < 0.001. SOUItCE: Harlap and Davies (1974). lung defenses, (3) the internal diameter of the_ small airways is ex- treinely small and vulnerable to obstruction, and (4) the newborn child has its full complement of airways at birth but only a small proportion of the alveoli. During childhood the.airwaye grow in - internal dianleter, and the alveoli both multiply and increase in size. The question of the timing of the effect of ETS on the grow- ing and developing lung remains to be elucidated If the effect is in utero, the question of how this carries over into infancy and childhood muet be addressed. Likewise, the carryover effects of - - - - increased incidence of respiratory infections in infancy must be - - determined. In this regard, there is already some information relating early childhood respiratory illness to subsequent respi-y ratnry symptoms and impaired lung function later in childhood (Woolcock at al., 1984; McConnochie, 1985). Evidence is also accumulating that respiratory infections in early childhood are re- lated to an accelerated decline of h'EV1 in adult life (Burrows et al., 1977b; Lebowitz and Burrow, 1976). If this is so, and if expo-e sure to ETS increases susceptibility to acute respiratory infections in infancy, ETS may have a carryover effect into adult life. From the evidence to date, it appears that the effects of ex- posure to ETS may start in utero by altering Lhe growth pattern of the fetal lung. In infancy, exposure to ETS may increase sus- ceptilibity to viral respiratory infections that in turn may have a OOQ4/Vfa(.Iv

212 carryover effect into later childhood and adult life. Direct effects (of ETS as an airway irritant are also likely, although the dose by itself may be insufficient oxcept for the most susceptible individuals tu cause symptoms and/or functional impairment. it is unlikely thlA exposure to ETS can cause much emphysema. As one of the many pulmonary insults, however, ETS may_ add to the total burden of environmental factors that become sufficient to cause chronic airway or parenchylnal disease. STUDIES OF ACUTE PULMONARY ErFECTS Several studies have examined acute responses to ETS. Be- cause asthmatics may be hypersensitive to exposures to noxious agents, a number of studies have also searched for acute effects of exposure to ETS among asthmatic populations. Other studies have been conducted on normal healthy adults. Normal Subjects Pimm et al. (1978) compared various physiologic responses of nonsmokers to either room air or room air plus machine-generated cigarette smoke. Each smoke exposure consisted of combustion of four cigarettes to produce an extremely polluted room with high levels of -carbon monoxide (24 ppm) and particles (greater than 4 mg/m3). Pulmonary function tests, nitrogen washout curves, blood carboxyhenioglobin levels, and heart rates were measured before, during, and after a 2-hour exposure. A few statistically significant differences between smoke and ambient air exposure - days were found. The differences were small and were considered by the investigators to be of questionable importance. Subjec- tive tive complaints were common in this and other acute cigarette smoke exposure studies, particularly eye irritation and cough. CO and suspended particles are thought to be less important than the phenols, aldehydes, and organic acids in producing this sytnp• tomatology (Hinda and First, 197b). Shephard et id., (1979b) utilized a protocol silnilar to Pimm et - al. (1978) but under conditions of intermittent moderate exercise (increasing the respiratory volume per minute 2.5 titnes). Moder- ate and heavy E'1'S exposures were considered, associated with CO - concentrations of 20 and 31 ppm, respectively. Neither exercise SOQ~+B~~B 213 TASIS 11-8 Pneumonia and Bronehilis by Parents Smoking In First Year of Follow-up, Annual Incidence per 100 Children (Number of Infants in Parentheses) Both or One Fssmokers or Both Smoking Habits ' One Both Nunsanokers Changed Snaker Snakers AII 7.8 9.2 11.4 17.6 11.5 (372) (675) SOURCE: Colkp et.l. (1974). (552) (478) (2.077) nor exposure level significantly influenced symptomatology. Small decrements (:l-4%) in FVC, FE1Vi, V,„„so%, and V,,,as26% (the volumes of air expired during the first half of the period of forced expiration or first quarter of the period, respectively) were noted in response to smoke exposures; however, static lung volumes were unaffected. Eye irritation and odor complaints were very common. One subject complained of wheezing and chest tightness, although his laulmonary function was not signi6cantly impaired. Subjective symptom scores were higher overall for the higher smoke exposure (13.8 versus 10.3 points/subject at the lower exposure). A few subjects reported cough, nasal discharge, or stuffiness and throat irritation. Asthmatic Subjects A number of studies have examined acute pulmonary re- sponses of asthmatic patients to exposure to ETS (Table 11-8). - However, the mechanisms for bronchoconstriction among asthmat- ics differ. Therefore, the comparison between study populations and between individuals within studies is difficult. Shephard et al. (1979a) examined asthmatic persons to de- terntine whether their response to ETS exceeded that of normal subjects in a previous study. The subjects (9 men and 5 women; av- erage -- erage age, 97 years) were exposed for ~ hours to machine-gcnerated smoke (CO, 24 ppm). None of the patienta had current respiratory infections, but_ some may have had associated chronic bronchitis or pulmonary emphysema. No significant alterations in dynamic lung volumes (FEVI, V„,,,&0sc: and Vm,,,2a%) were detected when the asthmatics' responses to ambient air and cigarette smoke were

214 compared. A small, but significant, decrease in total lung capacity ('1'LC) was noted, although preexposure TLC wae slightly higher than that on the same exposure day (96.5',"6 and 103.6% relative to ambient.air TLC, respectively). The lack of measurable chauge was interesting in light of a reported history of exacerbation with exposure to ETS by four subjects. Acute symptomatic responses during the experimental study were similar to those seen in the investigators' previous study of normal individuals; however, more complaints of tightness in the chest (43% of subjects) and whee$-. ing (36%) were inade by asthmatic subjects. It was concluded that asthmatics did not have unusual measurable responsiveness to ETS exposure in this study. The findings of Dahms et al. (1981) contrast with those of Shephard et al. (1979a). The exposure in this study was less In- tense, i.e., 1 hour at CO levels of 15-20 lshm. The patients were 16 to 39 years old, had mild impairment, and were on medication, except for the restriction that no bronchoolilators might be used within 4 hours previous to the test. Five of the patients reported specific complaints when exposed to ETS. When compared with control subjects, asthmatics showed signilicant pulmonary func- tion changes following 1 hour of smoke exposure. FVC decreaeed 20% and FEVi declined 21.4% in the asthmatic subjects. These decreases are very large compared with the other studies. Baeied on a 0.40% increase in blood earboxyherueglobin, the environ- mental CO concentration was calculated to be between 15 and 20 ppm-compared with approximately 24 ppui in the Shephnrd et al. (1979a) studies. Reasons for the discrepancy . between the Dahms and Shepliard studies results are neit clear, nor do Dahme et al. (1981) cite or discuss the earlier Shephard et al. (1979a) findings. Knight and Breslin (1985) evaluated six nonsmoking patients. The details of the subject population and exposure conditions were not specified. They measured a mean fall in FEVi of I t% following exposure to ETS. Using a histamine inhalation test, they found that the provocative concentration (or dose) that produced a 20% fall in FEVI (PCzeF1±.V1 or PDzoFEVI) decreased following exposure to ETS. This indicates an increased bronchial reactivity to histamine. The authors hypothesized that the airways may be primed to react more vigorously to other triggers. Wiedemann et al. (1986) evaluated nine astlimatic individuala (aged 19 to 30 years) with normal or nearly normal lung function 215 for both lung function and airway reactivity following exposure to ETS. Six patients reported a history of reaction to ETS. These subjects, all of whom were off ined'tcation, were exposed for 1 hour (CO between 40 and 50 ppm). Their carboxyhemoglobin levels increased an_ average of 0.86% (p < 0.001), FVC decreased 2% (p < 0.01), and FEVi declined 1% (not statistically significant). Airway reactivity was assessed using a niethylcholine challenge test. The 1'Dz 0FEV1 increased from 0.25 f 0.22 on the day before exposure to 0.79 f 1.13 postexposure (p < 0.05), indicating a decrease in airway reactivity following exposure. The magnitude (if this decrease_ was small, and the clinical meaning of the change is uncertain. There are a number of possible reasons for the apparent in- consistency among these studies, not the least of which is small sample sizes. The subjects have not been characterized fully. As noted by the authors, the stability of patients and mechanisms of bronchoconstriction differ among subjects. For instance, patients were included in several of these studies, regardless of whether - they were hypersensitive on the methylcholine challenge test. Fur- ther, some studies were performed on medicated patients. None of the studies could be performed blind to the presence of ETS. 'Yherefore, the authors could not exclude the possibility that pul- monary fusiction changes could be emotionally related to cigarette smoke exposure, especially in those patients who reported previous histories of adverse response to ETS exposure. There are several issues that are unresolved by these studies. --- - For instance, what proportion of a clearly defined population of asthmatics do react to ETS? If the patients are selected according to methylcholine or histamine responsiveness, criteria should be given for the extent of responsiveness, since it is a continuum. To address the issue of degrees of sensitivity, the appropriate case- control or cross'-over studies, with carefully selected populations, ~ need to be done. Mechartlsrns of ltesporise The mechanisms responsible for eye irritation and rhinitis, as well as possible changes in airway size, are almost entirely un- known. They could represent irritant effects from gases such as oxides of nitrogen, acrolein, animonia, and other reactive con- stituents. Lundberg et al. (198;1) reported that throat irritation OTQ48448

216 and local edema may be due to vapor-phase components that stim- ulate substance P release frctm local capsaicin-sensitive afferent neurons in the airway mucosa. It is also possible that an allergic- mechanism could be involved.. Several authors have described nl- lergic reactions to cigarette smoke (see, for example, Zussmann, 1970). Cutaneous hypersensitivity to tobacco antigens has been described in clinical settings (Bec-ker et al., 11)76). Constituents of tobacco smoke have also been shown to Le iuimunogenic in laboratory animals (Becker at al., 1979; Gleich and We1sL,1979). During the last 10 years, 13ecker and colleagues (1979, 1981; Becker and Dubin, 1977) have isolated a tobacco glycoprotein both from cured d tobacco leaves as_ well as from cigarette smoke con- densate. Animals that were previously sensitiZed to this antigen had both pulmonary and cardiovasculer changes when challenged (Levi et al., 1982). However, the role, if any, of this antigen, as well as other antigens that may be present in tobacco smoke, in the pathogenesis of cardiopulmonary diaease in active smokers, let alone nonsmokers exposed to ETS, remains controversial. SUMMARY AND ItECOMMENI)A'rIUN3 There have been many studies of respiratory effects of expo- sure to ETS to children. In view of the weight of the scientific evidence that ETS exposure in children increases the frequency of pulmonary symptoms and respiratory infection, it is prudent to eliminate smoking and resultant ETS from the environments of small children. What Is Known 1. Children of parents who smoke compared with the children of parents who do not smoke show increased prevalences of respi- ratory symptoms, usually cough, sputum, and wheesing.'rhe odds ratios from the larger studies, adjusted for the presence of parental symptoms, were 1.2 to 1.8, depending on the symptoms. These fin4ings imply that ETS exposures cause respiratory symptoms in some children. 2. Estimates of the magnitude of the effect of parental smoking on FEVI function of children range from zero to approximately 0.5% decrease per year. This small effect is unlikely by itself to be clinically significant. However, it may reflect pathophysioiogic 217 effoets of exposure to ETS in the lungs of the growing child and, as such, may be a factor in the development of chronic airflow - obstruction in later life. 3. Bronchitis, pneumonia, and other lower-respiratory-tract illnesses occur up to twice as•ofton during the first year of life in children who have one or more parents who smoke than in children of nonsmokers. , What Scientific Information Is Missing 1. ETS exposure during childhood may influence the devel- opinent of airway hyperresponsiveness in adult life. Research is needed to address this issue. Zb evaluate the timing of physiologic changee during development may require anuual studies. 2. Future cross-sectional studies of ETS exposure and lung function in adults need to be designed to control for other factors that may affect lung function. ion. 3. Little information is available from long-term longitudinal studies of the effect of exposure to ETS by nonsmokers on lung function in either children or adults. Studies need to be carried out in areas with different climates and characteristics of housing over long enough periods of time to assees the effects of changing smoking patterns. Animal studies may also be required to address these longitudinal questions. Intervention studies, in which par- ents stop smoking in the presence of children, should be_ done to assess the reversibility of these effects. - 4. The pathophysiologic mechanism of increased susceptibility to viral infections in very young children exposed to ETS has not been clarified. 6. The extent to which normal and asthruatic adults are af- fectod by short-term exposures to ETS needs to be studied further. 8. The few studies of the effect of short-term ETS exposure of - -- - asthmatic patients and of nonasthmatics are not consistent. This - -- - may be because they have not been conducted under adequate control and have examined persons with considerable variability in the severity of asthmatic disease and airway responsiveness. Future studies should carefully define the populations when ad- dressing issues of frequency of reaction to ETS and should be done separately on hyperresponsive and nenhyperresponsive patients when addressing issues of severity of reaction to ETS. LLO4e449

218 7. Studies of other patients with obstructivo lung disorders, ~ such as cystic fibrotic and alplra-1-antitrypsin patients, ueed to bo done. Future studies need to identify susceptible eubpopulatione, if they exist, who are unusually vulnerable to tho acute effects c:f ETS exposure. 8. There is no consensus on how to deal with data on parental respiratory symptoms. Investigations shoulil report on rates of childhood illness/symptoms using analyses that are both adjusted and unadjusted for parental symptoms. 9. There is need for information on cLanges in pulmonary function between the end of the peak growth period and adult life' to assess the possible reversibility of effects. REFERENCE4 Adlkofer, F., a. Scherer, and H. Weimann. Small-airways dy.functinn In passive smokers. N. Engt. J. Med, 303:392, 1980. Aviado, D.M. Small-airway dysfnpctioa in passive nnoker.. N. Engl. J. Med. 303:393,1980, Beck, C3.J., Q.A. Doyle, and E.N. Schachter. Smoking and lung function. - Am. Rev. Respir. Dis. 123:149-166, 1981. Bec_ ker, Q.C3., T. Dubin, and H.P. Wiedemann. Hypersensitivity to tobacco - - antigen. Proc. Natl. Aced. Sci: USA 73:1712-1716, 1976. Becker, 0.0,. and T. Dubin. Activation of factor XII by tobacco giycoprotein. J. Exp. Med. 148:467-487, 1977. - -- Becker, O.a., R. Levy, and J. Zavecs. Induction of IgE antibodies to antigen isolated from tobacco leaves and from cigarette smoke condensate. Am. J. Pathol 98:249-464, 1979. Becker, Q.Ei, N. Van Hamont, and M. Wagner. 21ibacco, cocoa, coffee, aud ragweed: Cross-reacting allergens that activate factor-Xll-depeadaat pathways. Blood 68:801-867, 1981. Berkey, O.S., J.H. Ware, D.W. Dockery, B.(3. Ferris Jr., F.N'« Speiger. Indoor air pollution and pulmonary function growth In preadolescent childran. Am. J. Epidemiol. 123:260-260, 1986. Bland, M., B.R. Bewley, V. Pollard, and Mai. Banks. Effect of childnn's and parents' smoking on respiratory symptoms. Arch. Dis. Child. 63s100-. - _ --- 105, 1978. Brunekreef, B., P. Fischer, B. Remijn, R. Van der Lsnde, J. Schouten and P. Quanjer. Indoor air pollution and Its effect on pulmonary functbn of adult non-smoking women. 1l1. Passive smoking and pulmonary function. Int. J. Epidemiol. 14:227-430, 1986. Burchfiel, O.M., M.W. Higgins, J.B. Keller, W.J. Butler, W.F. Howatt, and I.T.T. Higgins. Passive smoking, respiratory .ymldoms aud pulmonary function: A lougitudinal study In children. Am. R.ev. Respir. Uis. 133:A167, 1988.. 219 Burrows, B., R.J. Knudson, M.fl. Dline, and M.D. l..bowits. Quantitiative relations between cigarette smoking and ventilato ry function. Am. Rev. Respir. Dis. 115:196-205, 1977a. Burrow., B., R.J. Knudson, and M.D. L.bowit.. The relationship of child- hood respiratory Illness to adult obstructive airway disease. Am. Rev. - -- Respir. Dis. 116:761-700, 1977b. Chen, Y., and W.X. Li. TLe effect of passive nnoking on children's pulmonary onary function in Shanghai. Am. J. Public Health 78:616-618, 1986. Ohen, Y., W. 1a, and-S. Yu. Influence of passive smoking on admissions for respiratory illne.s in early childhood. Br. Med. J. 203:303-306, 1986. E3ockcroit, D.W., B.A. Berscheid, and K.Y. Murdock. Unimodal distribution - - - of bronchial responsiveness to Inhaled histamine In a random human population. Chest 83:761-764, 1983. Colley, J.R.T. Respiratory symptoms in children and parental smoking and phlegm production. Br. Med. J. 2:101-204, 1974. Qoiloy, J.R.T., W.W. Holland, and ILT. Gorkhill. InBuence of passive smoking au parental phlegm on pneumonia aud bronchitis in early childhood. Lancet 2:1031-1034, 1974. Collins, M.H., A.0. Moesinger, J. Kleinerman, J. B_assi, P.. Rosso, A.M. Collins, L.e. James, and W.A. Blanc. Fetal lung hypeplasia associated with material smoking: Morphometrk analysis. Pediar. Res. 19:408- 412, 1985. - -- Comroe, J.H., lt.E.. Forster II, A.B, Dubois, W.A. Briscoe, and E._ Carlsen. The Lungs Clinical Physiology and Pulmonary Function Tasts. Chicago: Year Book Medical Publ., Inc., 1962. pp. 323-364. Coautock, (3.W., M.B. Meyer, K.J. Hebing, and M.S. Tockman. Respiratory effects on household exposures to tobacco smoke and gas cooking. Am. Rev. Respir. Dis. 124:143-148, 1981. Dahm., T.E., J.F. Bolin, and R.a. Slavin. Pas.ive smoking: Effects on bronchial asthma. Chest 80:630-634, 1981. Dodge, R. The elfects of indoor pollution on Arizona children. Arch. Environ. Health 37:161-166, 1982. DDrill, S., and 1t. Thomas. Evaluation of Short-llern: Bioassays to_ Predict Functional Impairment. Virginia: The MITRE Corp., 1980. DutNU, f3., O. Enjaume, M. Petru., P. Darcos, P Demeurisse, and P. Rochiceloli. Enquete epidemlologne sur la tabagteme passif des enfante do 0 a 6 ans. Arch. Fr. Pediatr. 38:721-.726, 1981. Evaus, MJ., W. Mayr, R.F. Bils, and O.Q. Lsoeli. ERects of ozone on cell renewal In pulmonary alveoli of aging mice. Arch. Environ. Med. 22:460-463, 1971. Eva:u, M.J., L.J. Cabral, R.J, Stephens, and 0. Freemaa, 1976. Transforma- tion of alveolar type 2 cells to type 1 cells following exposure to NO2. Exp, Mol. 1lathol. 22:143-160, 1976. Evans, M.J., N.P: Dekker, L.J. GsbraLAnderson, and a. Freeman. Qu_an- titation of damage to the alveolar epithelium b means of t _ y ype 2 cell proliferation. Am. Rev. Respia. Dis. 118,787-790, 1978. Fergiuson, D.M., L.J. Horwood, F.T. Shannon, and B. Taybr.. Parental smoking and lower respiratory Illness In the first three years of life. J. Epidemiol.. Comm. Health 35:180-184, 1981. zio4e44s

220 Ferris, B.d., Jr., J.H. Ware, O.S. Berkey, D.W. Do<kery, A. Spiro 111, and F.E. Speiser. Effects of passive smoking oa health of childreu. Environ. Health Perspect. 82:289-296, 1986. Gammage, R.B. and S.V. Kaye. Indoor air and hnmau health, pp. 196- 200. Proceedings of the Seventh Life Sciences 8ymposium, Knoxville, Tennessee, Oct. 20-31, 1984. 430 pp. (3leich, G.J., and P.W. Welsh. Immunochemlcai and physicochemical prolrs srties of tobacco extracts. Am. Rev. Respir. Din. 120:996-1001, 1979. Guyatt, C3.H., and M.T. Newhouse. Are active and passive smoking harmfuF7 Determining causation. Chest 88:446-451, 1986. - Harlap, S., and A.M. Davies. Infant admissions to hospital and maternal smoking. Lancet 1:629-63'l, 1974. Hasselbiad, V., C.fi. Humble, M.GI. Graham, and H.S. Anderson. Indoor l environmental determinants of lung function In children. Am. Rev. Respir. Dis. 123:479-486, 1981. Hinds, W.O., and M.W. First. Concentrations of nicotine and tobacco smoke in public places. N. Engl. J. Med. 492:844-846, 1976. Huber, E3.L. Small-airways dysfunction in passive satokers. N. Eug1. J. Meil. 303:392, 1980. Kauffmann, F. Selection bias of P1MZ subjects. Am. Rev. Respir. Dis. 131:800-801, 1985. Kauffmann, F., D.W. 1)ockery, F.E.. Speiser and B.Ct. Ferris, Jr. Respiratory _ symptoms and lung function in women with passive and active smoking. ' Am. Rev. Respir. Dis.,133:A167, 1988. Kentner, M., Ci. Tr_iebig, and D, Wsltie. The influence of passive smokiug on pulmonary function-A study of 1,351 offlce workers. Prey. Med. 13:068-869, 1984. Knight, A., and A. B. Breslin. Passive cigarette sn1oking and patients with . asthma. Med. J. Aust. 4:194-196, 1986. Lebowits, M.D., and 11. Burrows. Respiratory symptoms related to smoking habits of family adults. Chest 89:48-60, 1976. Lebowits, M.D., D.B. Armet, and R. Knudson. The effect of passive smokb:g on pulmonary function In children. Environ. Int 8:371-373, 1983. Lee, P.N. Passive smoking. Food Cham. Toxlcol. 40h323-330, 1082. Leeder, S.R., R. Oorkhill, L.M. Irwis, W.W. Holland and J.R.T. Collay. Influence of family factors on the incidence of lower respiratory• Niness during the first year of life. Br. J. Prey. Soc. Med. 30:203-212, 1976. Levi, R., J:H. Zavecs, J.A. Burks, and C.f4. Becker, Cardiac and pulmonary - anaphylaxis in gninea pigs and rabbits induced by glycoprotein iwlated from tobacco leavas and cigarette moke condensate. Arn. J. Pathol. --- - 108:318-326,1982. Lundberg, J.M., f7.R. Martling, A. Saria, K. Folkers, and S. Rosell. Cigarette smoke-induced airway.oedeina due to activation of capsaicin-snuiUve vagal afferents and substance P release. Neuroscience 10:1361-1368, 1983. McConnochie, K.M., and K.J. fioghmann. Predieting clinically tigni8caat - lower respiratory tract iliness in childhood following sniid bronchiolilit. A:n. J. Dis. Child. 139:846-831, 1986. McConnochie, K.M., and K.J. Roghmann. Parental smoking, presence of older sibling, and family history of asthma increase riek of bro_nckiolitis. Am. J. Dis. Child. 140:808-812, 1986. ` a-O4aC3M4e I 221 Nationai Research Council, Committee on the Epidamiology of Air Pollu- tants. Epidemiology and Air Pollution. Washington, D.C.: National Academy Press, 1985. 224 pp. Pedreira, F.A., V.L. (Juandoio; E.J. Feroli, G.W. Mella, and I.P. Weiss. Involuntary smoking and incidence of respiratory Illness during the first year of life. Pediatrics 76:694-597, 1986. - Pimtn P.E., F. Silverman, and R.J: Shephard. Physiological effects of acute passive exposure to cigarette smoke. Arch. Environ. Health 33:201-213, 1978. Purvis, M.R., and R. Ehrlich. Effects of atmospheric pollutants on suscepti- bility to respiratory Infection. Ii. Effect of nitrogen dioxide. J. Infect. DIs. 113:72-76, 1963. Pullnn, C.R., aud E.N. Hey. Wheesing, asthma, and pulmonary dysfunction 10 years after Infection with respiratory syncytial virus in infancy. Dr. Med. J. 284:1665-1669, 1982. Rantakallio, P. Relationship of maternal smoking to morbidity and mortality of the child up to the age of five. Acta Paedfatr. Scand. 67:621-631, 1978. Rylander, R. Environmental tobacco smoke and lung cancer. Eur. J. Respir. Dis. 133(Sappl.):197-133, 1984. Said, f]., J. Zalokar, J. Lellouch, and I.. Patois. Parental smoking related to .demoidectomy and tonsillectomy in children. J. Epidemiol. Comm. Health 32.07-101, 1978. 8_cheuker, M D., J.M. Samet, and F.E. Speiser. Risk factors for childhood respiratory disease. The effect of host factors and home _me environmental exposures. Am. Respir. Dis. 128:1038-1043, 1983. Schiliing, R.S.F., A.D. L.tai, S.L. Hui, ©.J. Beck, J.11. Schoenberg, and A. Bouhuys. Lung function, respiratory disease, and smoking In families. Am. J. Epidemiol. 106:274-283, 1977. Sheplurd, R.F., R. Collins, and F. Sihrerman. •Passive' exposure of asth- matic subjects to cigarette smoke. I:nviren. R__es. 20s392-40Y, 1979a. Shepi:ard, R.J., R. Collins, .nd_ F. Silverman. Responses of exercising subjects to acute 'passive' cigarette smoke exposure. Environ. Res. 19:279-291, 1979b. Sims, D.a.; M.A. Downham, P.S.. Gardner, J.lf. Weblr, and D. Weightman. Study of 8-year old children with a history of respiratory syncytial virus bronchiolitis in infancy. Br. J. Mad. 1:11-14, 1978. Speiser, F.E., B. Ferris, Y.M. Bishop, and J. Spengler. Respiratory disease rates and pulmonary function in children associated with NO2 exposure. Am. Rev. ltespir. Dis. 121:3-10, 1980. Tager, l.B., S.T. Weiss, A. Munos, B. Ro:ner, and F.E. Spdiser. Longitudinal • study of tlie effects of. maternal smoking on pulmonary function in children. N. Engl. J. Med. 300:899-703, 1983. Tager, Lti., S.T. Weiss, B.. Rosner, and F.E. Speiser. Effect of parental cigarette srnoking on the pulmonary function of children. Am. J. l;pidemiol. 110,16-28, 1979. Tashkin, D.P., V.A. Clark, M. Simmons, C. Reams, A.H. Coulson, L.B. Ilourque, J.W. Sayre, R. Detels, and S. Rokaw. 7'he UCLA population atudies of chronic obstructive respiratory disease. Am. Rev. Respir. Dis. 129:891497, 1984.

222 Taylor, L.D., S.D. Greenberg, and P.A. Bu81er. Health effects of Indoor passive smoking. Tex. Med. 81:36-41, 1086. U.S. Public Health Service. Smoking and Health. A Report of the Surgeon - General. DHEW (Pils) flubl. No. 79-60088. Itockville, Maryland: U.S. Department of Health and Human Services, Public Health Service, Office on Smoking and Health, 1979. U.S. Public Health Service. The Health Consequences of Smoking: @ardio- vascular Disease. A Report of the Surgeon General. DIIHS(PHS) Publ. No. 84-50204. RockvUk, Maryland: U.S. Department of Health and Human Services, Public Health Service, Oflice on Smoking and Health, 1983. 384 pp. U.S. Public Health Service. The Health Consequences of Smokings Chronic Obstructive Lung Disease. A Report of the Surgeon General. DHIIS -- - (PHS) Pubi. No. 84-60206. Rockville, Maryland: U.S. Department of Health and Human Services, Public Health Service, O1Bce on Smoking and Health, 1984. 646 pp. Ware, J.H., D.W. Dockery, A. Spiro III, F.F.. Speiser, and B.(3. Ferris, Jr. Passive smoking, gas cooking, and respirato ry health of children living In six cities. Am. Rev. Respir. M 139s988-S74, 1984. Weiss, S.T., I.B. 'lkgnr, F.Lr. Speiser, and B. IRa.ner. I'ereistent wheese: Its relation to respiratory illness, cigarette smoking, and level of pulmonary function In a population sample of children. Aso. ltev. Respir. Du. 122:897-707,1980. Weiss, S.-T-., 1.B. Tsger, M. Schenker, and F.E. Sl:eiser. 'I'he health elfects of involuntary smoking. Am. Rev. Respir. Dis. 128:933-942, 1983. White, J.R., and H.F. Froeb Small-airways dysfunction In nonsmokers chran- icaily exposed to tobacco smoke. N. Engl. J. Med. 902s7S0-743, 1981). --- - - Wiedemann, H.P., D.A. Mahler, J. Loke, J.A. ifirpnlto, P. Snyder, and H.A. Matthay. Acute effects of psiisive smoking on lung fuuction and airway reactivity In asthma subjects. Chest 89:180-1H6, 1986. Woolcock, A.]., J.K. Pest, S.R. Leeder, and O.I1.B. Blackburn. The deveNop- ment of lung function In Sydney childrens Blfects of respiratory illnew and smoking. A ten-year study. Eur. J. Itespir. Dfe. 46(Suppl.132):1- 137, 1984. Zussman, B.M. Tobacco sensitivity In the allergic patient. Ann. Allergy - 28:371-377, 1970. VTQ4Bh4B 12 - Exposure to Environmental Tobacco Smoke and Lung Cancer The risk of lung cancer in cigarette smokers is directly related -- -- - -- - to the number of cigarettes smoked. At low-to-average levels of - smoking, this relationship is approximately linear and with no apparent threshold, although there are good theoretical reasons - to believe that the true dose-response curve should be curvilinear - - and probably quadratic (Doll and Peto, 1978; Usart and Schneider- - -- mau, 1979). Among smokers, an increase in exposure leads to an increase in risk, as long as the additional tobacco smoke, whether through active or passive smoking, reaches the bronchial epithe- liulu. Passive smoking would, therefore, be expected to cause some increase in risk of lung cancer in active smokers, as well as in any - - -- other persons in whom the appropriate tissues are exposed. The studies reviewed in this chapter have attempted to ad- dress -- - - dress the questions of whether an increase in risk of lung g cancer --- - - does occur in nonsmokers exposed to ETS and whether the dose- response reslionse relationship is similar to that in smokers. In part, this -- -- - -- - depends on whether there is a threshold dose of cigarette smoke exposure below which there is no increase in risk. Biological theory -- and current- evidence on low-dose exposure to carcinogens do not - -- - - - -- -- - - provide evidence for such a threshold, and it is generally thought that one Is unlikely (Office of Science and 1lechnology Policy, 1985). If there is no threshold, it follows that exposure to tobacco smoke -- - - - at low coneentrations, such as that experienced by nonsmokers - exposed to ETS, will cause an increased risk of lung cancer. The risk, of course, will be expected to be very much smaller than that associutod with active smoking because of the much lower exposure of the bronchial epithelium to tobacco smoke. 223

224 -TABLB IZ-t Urinary Cotinine (ng/nd) in Nonsmokers According to Number of Reported Hours of Lxywsure to Other People's Tobacco Smoke Within the Past 7 Days (Inciuding Day Urine Sample Was Coilccted) Duration o/ E.posure Urinary Colinine Quinlile Limits (h) No. , mean ± SD" _ 1st 0.0-1.5 43 2.8 t 3.0 2nd 1.5-4.5 47 3.4 t 2.7 3rd 4.5-8.6 43 5.3 t 4.3 4th 8.6-20.0 43 14.7 t.19.5 S1h 20.0-80.0 4S 29.6 ± 73.7 All 0.0-80.0 221 11.2 ± 35.6 'Trend with Increasing exposure was significant (p < 0.001). SOURCE: Wald el al. (t984). USING HIOLO GIGAL iV1ARKLRS TO ESTIMATE RISK Cotinine, a metabolite of nicotine, while of itself not con-. sidered a carcinogen, is a useful marker of exposure to tobacco smoke, whether through active or passive smoking. Table 12-1 ehows that the mean unnary cotinine concentration increases with the estimated exposure to other people's tobacco smoke over the past 7 days. Much of these data, collected in the United Iting- dom (Wald et al., 1984), showed that nonsmokers had, on average, about 0.4% of the concentration of urinary cotinine found in active • smokers. Similar work done in Japan suggested that nonsmokers had relatively high cotinine levels, about one-sevonth the levels in average Japanese smokers (Matsukara et a1.,1Q84). 'J'he reason for low- this difference is not known and it needs to be investigated. llow- - ever, in both countries studies showed increasing uriuary cotinine levels in proportion to the estimated increasing I.TS exposure. In most of the epidemiologic studies that assessed the relation- ship of lung cancer to ETS-exposed nonsmokers, the measure of exposure used was "living with a smoking spouso* '-Phe observed risks of lung cancer for nonsmokers were cAampared for those living with a smoking spouse and those living with ntinnmokers. While it is reasonable to believe that people living with smokers would be more heavily exposed to ETS than people livillg with nonsmokers, 225 this would seem to be a relatively insensitive measure of expo- - -- sure. Many people who are exposed to other peoples' smoke may not always be married to smokers. Even if they are married to smokers, they are likely to be exposed to their spouses' smoke for only a relatively small pruportion of the day. The possibility exists that they may be exposed to other people's smoke, for instance, at t work, or while hile in other public places. A study using urinary cotinilie levels as -a measure of expo- sure, however, showed that "marriage to a smoker" may identify - individuals who are more exposed to tobacco smoke in general, not simply from their spouses (Wald and Ritehie, 1984). Table 12-2 shows that the exposure to other people's smoke was greater for men married to smokers than for men married to nonsmokers (me- - - dian hours of reported exposure of 21.1 and 6.5 hours per week, respectively). Of particular relevance for epiderniologic studies is the fact that exposure is greater outaide the holne as well as within the home.. A reasonable interpretation of this fact is that men married to smokers might be more tolerant of other people's smoke than men married to nonamijkers and are less likely to seek - - out smoke-free environments outside the home. Silnilar results, based on questionnaire information, have been reported by others (Friedman et a1.,1983). These resulte corroborate the use of n spouse's smoking his- tory - tory as a method of classifying nonemokers into groups that have different exposure levels to tobacco smoke. Using data from the TABLB 12-2 Urinary Cotiniue Concentration and Number of Reported -- - Hours_ of Cxposure to Other People's Tobacco Smoke Within the Past 7 Days in Nonsmoking Married Men_ According tl) Smoking Habits of Their Wives Urinary Cotinine Concentration Exposure to Other Peopic _s Smoke In Preceding Weck. h Smoking Calegury No. nf , ngfml Total Outside Home of Wlre Mon Mean (SE) Median Mean (SE) 6fedian Mean (SE) Median - - - Nonsnake_r 101 9.S(t.3r 5.0 U.Q(1.2)k h.S 1Q.0(1.2)° 6.0 Smokcr 20 25.2(14.8) 9.0 23.2(4.1) 21.1 16.4(J.3) 10.7 NOTE: Differences (nonsmoking wife v.nus smoking wife); •p < 0.05; kp < - r 0.001; p < 0.06 (Wilyo:in rank sum lest). SOURCE: Wald and Ritchie (1984). STOLB449

226 British study (Table 12-2), the relative urinary cotiniue levels in three groups-nonsmoking men married to nonsmoking women, nonsmoking men married to smoking women, and men who were themselves active smokers--were in the ratio of 1:3:215 (actual mean values were 8.5, 25.2, and 1,826 ng/rnl, respectively; Wald and Ritchie, 1984, and personal communication). Assuming a sim- ilar half life of cotinino in smokers and nonsmokers, this suggests - that exposure to ETS among nonsmokers who are exposed is about 1% (i.e., 25.2/1826) of that of active smokers. Similar results were reported by another United Kingdom study (Jarvis et al., 1984) and one from the United States (Haley at al., 1986).. However, the halE life of cotinine in nonsmokers may be roughly 50% longer --- - than in active smokers (Kyerematen at al., 19112; Sepkovic at al., 1986), thereby changing the estimate of relative exposure by up to 50%. Assuming a usage of 20 cigarettes (onn pack) per day by active (male) smokers and assuming a linear rolatiunship between number of cigarettes smoked per day and urinary cotinine level, this represents exposure to smoke equivalentti of roughly 0.1 to 0.2 cigarettes per day. Others have estimated rigarette equivalent exposures of 0.2 to 1 cigarettes per day (Klosterk6tter and Gono, 1976; Hugod et al., 1078; Vutuc, 1984). Urinary cotinine is at present the best marker of tobacco smoke intake for passive smoking dosimetry because it is highly sen$itive and specific for tobacco smoke. Because it can be measured di- rectly in nonsmokers as well as active smokers, it makes it possible to estimate the relative exposures of the two groups (see Chap- ter 8). With other markers or with other substances in tobacco smoke, this is not currently possible. Estimates nwst be made or the extent to which these substances are inhaled in mainstream smoke, on the one hand, and released into room air, diluted, and then inhaled by nonsmokers, on the other (Chapter 7). Both of these estimates involve more assumptions in estiniating the actual intake. Whether a urinary cotinine measurement can provide a rea- sonable basis for computing a first estimate of the risk of lung cancer arising from LTS exposure depends in part on whether the intake of the relevant carcinogens in active and passive smokers is directly proportional to the relevant intake uf nicotine, frow which cotinine is derived. Our lack of knowledge of which specific - - smoke components are responsible for causing lung cancer and our 3LO~,AGGB 1 227 present inability_ to measure their intake directly creates uncer- tainty. But, as a first approximation, it is reasonable to assume propor tionality. Based on the above dosimetric eonsiderationn, the risk of lung cancer from ETS exposure among nonsmokers in the United King- dom aud United States would be small. Assuming linearity in the dose-response relationships, the risk would be about ut 1% of the excess risk in active smokers. This is equivalent to a relative risk pf 1.14 in males, given that the relative risk in average male active smokers is 10 to 15 times greater than in nonsmokers (Haminond, L966; Doll and 1'eto, 1978). For ET$-exposed women, the average relative risk may be less. If the cotinine data suggesting greater ETS exposures in Japan are correct, the excess risk in Japan would . be greater. ASSESSING THE RISK FROM EPIDEMIOLOGIC STUDIES OF LUNG CANCER - AND EXPOSURE TO ETS ' Some of the epidemiologic studies on the possible relationship between ETS exposure of nonsinokers and lung cancer have been discussed elsewhere (Rylander, 1984; Samet, 1985; IARC, 1986). The majority of studies of lung cancer in nonsmokers and ETS exposure classify subjects on the basis of whether the nonsmoker - lives with a smoker. Eighteen such studies were identified, aiid the analysis presented below is based on 13 studies listed in Tables 12-3 and 12-4. '1'he other 5 studies were excluded for the following ceasons: • Knoth et al. (1983), no reference population was given; • Miller (1984), study reported all cancers but did not report on lung cancers separately; • Sandler ot al. (1985), included very few lung cancer cases; • Koo at al. (1984), a more recent analysis of the population was presented in Koo at al. (1986); and - - -- • Wu et al. (1985), raw data wore not presented. Otherwise our analysis used data from all the studies, thereby reducing the possibility of bias arising out of selecting only some of the studies that met minimal standards. T+.b1e 12-3 gives the characteristics of the 13 studies included in the analysis. The relative risk estimate, together with its 95%

rrABu12-3 'EpidGm'iologic Studies of Lung Cancer and Exposure to Environmental Tobacco Smoke: Methodological Description of Studies Induded', in Analysis F-:yposnts Assestment' Typo df ptoeq Not Environments Stuqy Subjects Interview Informants? ~Wlanrieil EcsmokerS& Assessed Gomments Chan and lllong' Kong, t3!9: 84 Isten.iew. Fung. cases (out: of 189); not blind 1982 139 orthopedic eonnws Ztiehopoulos. ' Gteecev 62, cases (out lnter.ie.r, et al.. of 1102): 'N9D otllho- not blind 1983 pedic eontrols (out d(251) Correa L,ouisiana: 30' (22F, lntet.iew, et al.. 8M) cases'(out oY blinded 1983 35)313 (133F, 18aM) hospinl wntrols 1(dbeases nanla[ad to .....kl..a) No criteria No criteria No criteria given tnea B^K" No "L'xzposr3" Fsdude if smoked within ,ptiot 2o yr; "'nonaooka"' il no smoking in 20,yt; "es- s-ookc" if stocpeH 5-ZO.r before Honte': and Little' information on wotkplaa methods or selection of conttnk: tw xdinstmcnts of odds tatio: high cancer sate'forSouth China Spouse Ecduded': adenocanei- (curteat nomas and tertuinal and btotuliial: ari;inal former) sam,pk sitnilar age and SES. no tnatch on final sample Yes (249% o[ Esauded, Eourinnoe: used Spouse, cases,' 119% include pack yr of parents of controls) "e.et ;husbanfl rs.-ried" tio a43istmmt for age. ttue. or'hospital admission: tepotted odds ratio for older than 40; esciudcd bsondhioaivcolar pnmr Kabat and Multicenter USAs 78 lntavie+.. No 24 nses and Orily dati for 'l yr Workplace, 1 r,ses cmttvk W,yoder. (2SM, 53F) casa: na' blind ZS oontrols home ch 1954 78 (2SM. 53F) !had no mat ed, for' age. sez. tace, ltospital controls I(twn- spouse, . and date interviewed tobacco cancets) Buffkr Texas: 4'1 cues (out of Interview Yes ~ MIo a~iteria No critetia given S ouse qti i sl ul i et rl., 460): 192',popula- , tivan p g n ;pop at on matdbll age raee 1984 tion-based controls , , rital status. count r• (out of 492) - no match on final ' umple Garfinkel M. !Ghio: '134 cases Interview. Used data on Exclude, es,posed Home, octt- No dose-response ct al., age 40+; 402 aobn blinded telative, side botne 1 effeet: oonecqed for 11965 ntwer othernise age. SES. date ..unexposed» dia~ perehagen Sweden: 67 registry Ma1kd Yes "[3naaposed" : Exclude Spouse Previous intrr.ietz et al., casc;: 347: imontzdls , in nreps puants, 4961,11963 .rith worlcplatt foUow•up 1984: N possibk' interaction eD Akiba ]apan:' 113I(9+4F, Interview, et al., 19M) eaxs (out of not blind 1986 164); 380 (27DF, 11aMt) controls (match age, city, vital statuy) Yes, (90% of Excluded casat, 88% of controls) Exclude; spouse - nonsmoker"'if no smoking' in pciar 'SD yr with tadoq:.adjusted for occupation. radon, urban; matched foz' age, vital status S,pouse. Selaeted from atomic parents bomb survivors: a.etagrage mom than 70; no adjust- ment for radiation do.e

230 231 confidence interval, is shown in Figure 12-1 for each of the studies. Also shown in Figure 12-1 is the summary estimate based on the studies combined. The relative risk estimetes of lung cancer in nonsmokers in association with ETS exposure, together with the data used for calculating them, are given in Table 12-4. The data given in this table permit readers to combine any subset of the 13 studies which they may wish to consider. A summary estimate - of the relative risk for the selected studies can then be calculated using the general method described in Appendix B. The method weights each study by its statistical precision and avoids making inappropriate comparisons across different studies. In the course of examining the data, several such subset analyses were conducted and the results are presented below. The overall summary relative risk of lung cancer among non- amukers in association with ETS exposure was 1.34 (95% confi- dence - dence limits 1.18-1.53). For all women the relative risk was 1.32 (1.16-1.51); for men it was 1.62 (0.99-2.64). The wide confidence limits for meu reflect the fact that most of the data were based on nonsmoking women rather than nonsmoking men. For studies conducted in the United States, the relative risk was 1.14 (0.92- 1.40). Considering only the largest studies (those with expected number of lung cancer deaths of 20 or more), the relative risk estimate was 1.32 (1.15-1.52). The confidence limits on each of these estimates all include the overall summary estimate of 1.34. I 0 A A CORRECTIONS TO ESTIMATES FOR SYST EMATIC ERRORS Two alternative explanations can be given for the finding of an increased risk in the epidemiologic studies. The finding may represent a direct and causal eEE'ect of ETS exposure on lung cancer in nonsmokers; or it could be due in whole or in part to bias, either in the form of systematic errors in the reporting of information or a confounding factor that is associated both with lung cancer and the fact•of living with a spouse who smokes. An important question to answer is 'What true risk, modified by a reasonable set of bias-pr-0ducing factors, could lead to the average _ risk indicated by the epidemiologic studies?" In the following sections two computations are given that estiniate how much the true relative risks might be modified as a result of these possible kinds of misclassification. M494L8

zAel.e 1t2-4 Summary of 'Epidemiologic 'Studies of 'Irislc'Based on Faposura Assessed by'Spouse Smoking Hab'its, When J4,vailable, or Smokimg'by the Household Cohabitants Luny, Caacest' in °E:pased- Group 95% S d ' S l Var of Cooifwlenoe tu y No. tu ly Authors Location 5e: Obs. :E:p. fJ-'E . (O-E) Risk" Limits CasNCwarrdl Srrdies 1 Chan and' Fun j, Hong Kong F 34 37.7 -3.7 13:01 0.75 0.44 1.30 1982 2 T'richopoulos Graeae F 38 29:3 8.7 11.70 2.13 1.118 3.78 at al.. 11993 3 Comrr:a et il., U.S.A. F 14 10:6 3.4 4.75 2.03 0:83 5.03 1'983 M 2 1.2 0.8 0198 2.29. 0.31 16:50 4 [Qabat and U.S.A. F 13 13,7 -0.7 3L06 0.79 0.2b 2.43 HYyatles, '1984 M 3 5.0 0.0 1SI 1.00 0.20 4:90 5 i Butlkr et tWl,. ' U.S.1A. F 33 34.1 -1:1 4.78 0.80 0.32 • 1 ~.99 1988/ M 5 6.6 -1.6 2.37 0.50 0:14 1:79 6 Gasa..at~ _ aL. C'3_'4_ F a'2 893 2.5 22.33 1.12 0.74 1.69 1985 7 P.esAagen .K al.. I Sweden F 33 29.6 3.4 13:88 R.28 p:73 2.116 b ¢sa a .r1u'ba et at,. Jap,.n F 73 67:4 5.6 14.19 1.48 !0.88 2.50 1986 M 3 1:8 11.2 1.38 2.45 10.46 11106 9 Koo et ai.. ,Hong'Koaq mPress 110 I l.ee a ai.. '.1'986 England F F 51 22 45.3 21:9 5.' 0.1 13.19 4.71 1.5•t 1.03 0.90 10:41 2.64 2.47 1"1 8 79 7 10 2 56 1 30 . . . 1038 4.42 D"urcl- f- CtarConaaCSradiff 426 40'1.0 25:0 114.40 1.24 L04 1.50 CoGort. Prarpserrs Strdiet 11 GarFinkei, i1981 U.S.A. 12 F 88 8118 6.2 30.82 1.18# "' 10.90 1.5+/ GiIlis iet a1., 1984 Scotland F 6 6.0 10:0 1.58 1.00b 0.20 4:91 13 ' M 4 2.3 1.7 1.40 3.256 0.60 17.65 Hirayama. 1984 Japan F 146 129.5 16.5 34.83 11.63 1.25 2.11 M 7 3.3 3 7 3.02 . 2.25 1.04 4.85 l7vr,aU f- Pirotpssri.e Srydiet 251 222.9 28.1 71.65 1.44 1.20 11:72 Orao2 jor All S:rdier 692 637.7 53.1 186.0 1.34 1:1B IM 'Risk is 8i.en aa calculated odds ratios for case-control studies, (see Appendisi E for eak.tilatioos) and publishatl n:latire Adt for abhoM, prospective stndies. bRs :o of age stan'dasdized moaalit7 raus. N w w

234 ElS E%PoSURE 9Y N6MSMOKERS AMD LINIO CANCER 10 ! s 3 a 1 o.A as 0.3 ol A • N tluAlos s_ N sWdIft wMh r NrM se.op.clod o..N C • e<ry w.lvie..a" D . diy nrb eM~ E•NU.S.OWN&W a a 1 0 a i , ~ ~ 01 ' a-' f 10 11 1! t! A s G 0 E STUDY NUMBER 1! 3 4 S e 7 _ .~.:~. C/1SE-COM1NO1. S1U91ES iROSpECl1YE STl1aE! SUMMARY OoOs R/1iI9S 'i+'IEiURE 12-1 Passi.e smoking and lung cancer. The relative risk (point estimate and 96ei6 confidence Interval) of lang cancer In nonsmokers whose spouses smoke compared with nonsmskers whose spouses do not smoke for each of the studiea given In Table 19-4 and the summary estimate based on all the studies combined. The figures for females are shown first for studies based on male and fesnsle .nbjects. STUDIESs 1. Chan and FLng, 1982; 2. 1lrichopoulos at al., 1983; 8. Correa at a1., 1983; 4. Kabat and Wynder, 1984; 6. Bulfler et al., 198416. Garfinkel et al., 1985; 7. Pershagen at al., In press; 8. Akiba et al„ 198619. Koo et aL, In press; 10. Lee at al., 1986; 11. 6arfiakel, 19811 12. aillis at al., 1984; 1S. Hirayam., 1984. Misciassi$ed Examokers and tlte 'rendency for- Spouses to Have Similar Smoking Habits One source of petential bias that would inlluenEe the estimatis - of relative risk is that some people who occasionally smoke or who have smoked in the past may report that they have never smoked. Having smoked, these people are somewhat more likely to develop lung cancer than would true lifelong nonsmokers. Because smokers tend to marry smokers, they are +.lio more likely to have a spouse who smokes or did smoke in the pnet.. Table 12-8 shows 235 that the biaA produced by this cnisrePorting could be serious. The arguments presented in this table are an extension ofidea_s discussed by Lee (Lehnert et al., 1984). From -the table it is apparent that for this misclaseificatiou to fully account for the e observed excess risk, it would be necessary that 8°f% or more of smokers and oxsmokers report themselves as nonsmokers and that their smoking habits and history be identical with those of the - - self-reported nmokers: The proportion of people who say that they are nonsmokers, but who in fact do smoke, can be estimated using biochemical markers of tobacco smoke absorption. They appear to constitute - about 0.5-3%, depending on the population studied and the ques- tionnaire used (Wald et al., 1981; Saloojee et al., 1982). The proportion of people who smoke or have done so in the past but who say they have never smoked has also been n estimated in two cohort studies (see Chapter 6). In one of these studies (N. Britten, England, personal communication University of Bristol, England; see Table 6-4), information on smoking was obtained in detail in a longitudinal study. A proportion (4.9%) of the subjects said they had never smoked as much as one cigarette a day in 1982, when in fact they had previously smoked and reported so in pre= vious interviews. These subjects, however, had smoked at a rate of about half that of the current smokers and nearly all of them (93%) had stopped smoking 10 or more years earlier. Similar, or alightly higher, misreporting has been noted for older persons (see Chapter 6). llowever, older persons are likely to have smoked less and to have quit longer ago. Zhble 12-5 is based on the assumption that people who fail to report that they have been smokers have the same risk of lung cancer as the average current smoker. Ae indicated in Table & 4, "misclassified smokers" are tnore likely to have been exsmokere who failed to record the fact that they had smoked at some time in the past or, if they were current (or recent) smokers, they smoked fewer cigarettes per day than the average smoker (Table 6-4). In either event, their spouses' risk of lung cancer would be lower than for the spouse of a current smoker. The American Cancer Society's study of smoking (Ilammond, - - ,1960) reported that women who smoked 20-30 cigarettes a day had a 4.9-fofd 'increased risk of lung cancer compared with reported nonRmokere. The British Physicians' Study (Doll et al., 1980) yielded an estimate of 6.4. Both studies were conducted a number OZd4e449

236 TAat6 12--S Illustration of a Bias Likely to Affect Pussive Stnoking Studies ASSUME_: (1) proportion of smokers among women = 35% (li) proportion of smokers among men - 50% (ifi) aggregation ut smokers with smokers and nonsmokers w{Ih nonsmokers ~ 3s True Situation 35% Smokers 3S,o00 S 1 8% Misclasxitied as non(smokers - Odds of having a spouse who smokes Spouses Assume true RR of lung cancer associated with smoking (8.0) and spouse smoking o) . (I Rate%10,000/10 lcan Number of lung cancers 2,g0U S ~ 2tt 1 : `2. 1.9l7.1 S(/ 1) 2 NS (b) 25.6b0.tt S(c) 39,339.2 NS (d) 8.0 8.0 1.0 1.0 40.0 40.0 5.0 5.0 7.79 (e) J.41(o 12.g:! lg) 19.67 !b) ' _ Observed SitratJon o d Observ 67,E00 (65,000 NS + 2,gOl) smokers miselassl- . n e Observed no. in spouse groups fled as nonsmokers) 27,607.9 S (a+E) 40,192.1 NS (b+d) of lung eancen Observed no 20.62 (e+g) 23.08 (t+h) . Observed rate/10,t100/!0 years 7.47 s:» Observed RR of lung cancer associated with spouse smoking 1.30 CONCLUSION: Misclaaitkation error would increase the trua relative risk of 1.0 to 1.311. •Ratio of eross•products in a 2 X 2 table of smoking status (11es or No) by spouse smoking status (Yes or No). ABBREYIATIONS: S= smoker; MS = nonsmoker; RR !' relative risk. 1.3:2 257 of years ago. With the increased duration of smoking in women in recent years, these relative risks also should have increased. The relative risk estimate may be as• high as 8.0. To the extent that this may be an overestimate, it will tend to exaggerate the effects of misclassification. The lung cancer relative risk for persons misclassified a:+ nonsmokers is, for the reasons given, probably less than half of that for correctly classified active smokers (relative risk of 4) and probably. closer to one-quarter (relative risk of 2). - Applying the same argument illustrated in Table 12-5, the -- - misclassification effect on the relative risk is given in Table 12-6 - (N. Wald and K. Nanchahal, personal communication), assuming that the risk of lung cancer of lnisclassiHed nonstnokers is half that of current smokers (relative risk = 4.0) or one-quarter (relative risk = 2.0). 'lhble 12-6 shows the possible effect of a noitrandom marriage (aggregation) pattern. In this table the extent of nonrandom association is described by an "aggregation" factor. The degree of aggregation is estimated by the ratio of the cross-products in a 2 x 2 table of sntoking status of study subjects by spouse smoking status. For the computations in 'Pable 12-8, three aggregation factors are assumed, 2.5, 3.5 and 4.5. The slnoker aggregation factor (from epidemiologic studies) appears to be about 3 to 4(eee Table 12-7; Wald et al., personal communication). '1`he' overall effect on an assumed true association between passive smoking and lung cancer, i.e., the atruo' relative risk, is shown in Table 12-6 for relative risks ranging from 1.0 (i.e., no as- sociation) to 1.25 (i.e., 25% increase in lung cancer risk associated --- - with passive smoking). It is assumed that 35% of women smoke - ---- - and 60% of men smoke.. Also, the effects of the misclassification -- - -- - ---- - of between 2% and 10% of smokers as nonsmokers is shown. The most plausible assumptions are a relative risk of 2.0 to 4.0, an ag- - - - - gtegation factor of 3 to 4, and a misclassification rate of 2 to 7%. To use Table 12-6, locate the rows and columns that correspond - - - the the above most plausible assumptions. The entries in the body ---- of the table that are approximately 1.84, i.e., the observed overall relative risk, correspond to the set of parametric values that, with plausible assumptions of the bias, would inflate a true relative risk to the observed values. Inspection of tlke data within the body of Z3tble 12-6 shows that an observed relative risk of 1.34, given the range of atfsumpti4ns specified in the table, could come about - - if there were a true relative risk of no less tLan 1.15. That is, TZ04S4lae

238 239 TABLE 12-6 Estimates o f the Obse rved Reiat ive Risk of Lung Cance r TAI/L612-6 Continued from Studies of Married - Nunsmok ers; Assu A ming 35 % of Women an kers S d True Relative Risk M i 50% of Men In the Gen eral Popula re tion Current mo arr age 1'roportken of Misclassitied Smokers Passive Misclassitkd Aggregation --- True Itelative Risk Marriage proportbn of Misclassified Smokers Snwkers Smokcre Fadors 2% 4% 6% 8% 10% 8 Passive MisclassUied Aggregation 10% .0 2.5 1.26 1.30 1.35 1.3$ 1.42 Snwkers Smokers• Factork 2% 4% 6% 8% 3.5 1,28 1.35 1.42 1.47 1.52 00 2.0 1 2.5 1.01 1.02 1.03 1.04 1.04 4.5 1.30 1.39 /.47 1.54 1.61 1.25 2.0 ` 2.5 1.26 1.27 1.27 1 28 1 29 . 5 3 1.01 1.03 1.04 1.05 1.06 . . . 4.5 1.02 1.03 1.04 1.06 1.07 3.5 1.26 1.27 1.28 1.29 1.30 4.5 1.26 1.28 1.29 1 3/1 1 31 4.0 2.5 1.03 1.05 1.118 1.10 1.12 17 . . 4.0 2.5 1.27 1.30 1.31 .1 1 35 3.5 1m 1.08 1.11 1.14 1. 0 . 3.5 1.29 1.32 1.35 1.37 1 40 5 4 1.05 1.09 1.13 1.17 1.2 . 8.0 . 2.5 1.06 1.12 1.17 1.21 1.25 36 4.5 1.29 1.33 1.37 1.40 1.44 8.0 2.5 1.30 1.35 1.39 1 43 1 46 5 3 1.09 1.17 1.24 1.30 1. . . S . 5 4 1.11 1.20 1.29 1.37 1.43 3. 1.33 1.40 1.46 1.52 1.57 . 4.5 1.35 1.44 1.52 1.59 1 65 0 2 2 5 1.06 1.07 1.08 1.08 1.09 . . 1.05 . 3.5 1.06 1.08 1.09 1.10 1.11 'Subjects who have smoked either lu ihe past or currently. but claim to be lifelong non- 4.5 1.07 1.08 1.09 1.11 1.12 smokers. 4.0 2.5 1.08 1.10 1.13 1.15 1.17 eMarriage aggregation factor defined as ratio of cross-produMs of 2X 2lable of smoking 3.S 1.09 1.12 1.16 1.19 1.21 status of study wbject by smoking status of spouse. 4 5 1.10 • 1.14 1.18 1.22 1.25 OT Tfi 8 0 . 2 5 1.11 1.17 1.21 1.26 1.29 N B; e.alues inside the boxes indicate those situations that are most plausibk, based n th f d . . 3.5 1.14 1.21 1.28 1.34 1.411 o o er sources fi ata for parameters, and yield observed relative risks of about 1.34. 4.5 1.16 1.25 1.33 1.41 1.411 1.10 2.0 2.5 1.11 1.12 1.13 1.13 15 14 1 1 1.14 14 1 3.5 1.11 1.12 . . . 4.5 1.12 1.13 1.14 1.16 1.17 4.0 2.5 1.13 1.15 1.17 1.19 1.21 3 5 14 1 1.17 1.20 1.23 1.26 1 . 4.5 . 1.15 1.19 1.23 1.26 1.311 TABLE 2-7 Number of Smokers and Nonsmokers According to the ' 0 8 2.5 1.16 1.21 1.26 1.30 1.3.1 $muking Nabils of fheir Spouses and the Odds Ralio Indicnting the . 3 5 19 1 1.26 1.33 1.39 1.44 Extent of Such Marriage Aggregation' . . 5 1 51 - , 4.5 1.20 1.30 1.30 1.4 . 1S 2.0 1 2.5 1:16 1.17 1.17 1.18 1.19 Females Males . 5 3 1.16 1.17 1.18 1.20 1.20 . 4.5 1.17 1.15 1.19 1.21 1.22 Spouse Smoker Nonsmoker Total Smoker Nonsmoker Total 4.0 2.5 1.18 1.20 22 1.22 1.24 28 25 1 1 1.26 1 31 Smoker 53 13 70 20 11 31 3.5 1.19 1. . . 1 31 . ~n Nonsmoker 47 83 130 S3 Li0 133 4.5 1.19 1.24 1.27 . 8.0 2.5 1.21 1.26 1.30 1.34 t:3tt AQ 1110 100 200 73 91 164 3.5 1.23 1.31 1.37 1.43 1.48 4.S 5 2 1.25 21 1 1.34 1.2? 1.43 1.50 1.22 1.23 1.56 1.24 Odds ratio 3.1 2.3 1.20 2.0 . 3.5 . 1.21 1.22 1.23 1.24 1.25 '@nsed on interviewing 200 women and 164 men atlending a health screening center In 4.5 1.21 1.2.1 1.24 1.25 1.27 I.ondon or working in the Civil Service In Newcastle In 1985. 4.0 2.5 3.5 1.23 1.24 1.2R 1.27 1.27 1.311 1.29 .33 1,.10 1: Y~ SOURCL: W.kl ct al., personal comnwnicatiun. 4.5 1.24 1.211 1.32 1.36 L;19 ZZD4.R448

240 a true relative risk of 1.15 or more could, by a reasonable set of misclassification biases, be elevated to 1.30 in an epidemiologic study. Stated differently, this implies that reasonable misclassili- cation does not account for the total increased risks reported by the epidemiologic studies, leaving the conclusion that the risk of lung cancer followiiig exposure to other people's smoke, as judged by whether a nonsmoker has a smoking spouse, would be increased by a minimum of 15%, and most probably increased by 25% (i.e., 1.25). (If the percentage of women smokers were as high as 500A, it would be 1.20.) The study by Garfinkel at al. (1985) provides data relevant to the misclassification of exemokers and the tendency for spouses to have similar smoking habits. In this study, subjects were in- terviewed if the hospital record indicated nonsmoker or made no mention of smoking status. From interviews by the investigators, it was determined that 40% of the women had actually smoked. Among these women who smoked, 81% had husbands who smoked, but only 68% of the women who were in fact nonsmokers had hus- bands who smoked, yielding an aggregation factor of 2.0. Effects of Incorrectly ClaHSifying Persons as Unexposed to E'rS In the studies that classify nonsmoker exposure based on whether or not the spouse smokes, some of the "unexposed" non- smokers, i.e., married to nonsmokers, are likely to be exposed in other settings. For instance, some nonsniokers married to non- smokers may be exposed to ETS in the workplace. Therefore, some individuals in the baseline, unexpused," group for these studies must have been exposed, and hence have risks greater than unity if there is an ETS effect. In the studies, which do ask about exposure to ETS in all environmeids, there still tends to be misclassification of some nonsmokers ati "unexposed," because there may be a tendency to overlook episodes of exposure. The data from urinary cotinine studies and the observed rel- ative risks can be used to estimate this offect. The only known source of cotinine in the body is from nicotine, which is virtnallr exclusively derived from tobacco, with Lite exception of nicutine chewing gum and nicotine aerosol rods. 'l=hercfore, if people who actively use tobacco or nicotine-containing aisls to help stop suwk- ing are excluded, cotinine can be used as an objective measure of 241 I recent) exposure to tobacco smoke in nonsmokers. For the follow- n ng argument, the cotinine in body fluids is compared for the two groups of nonsmokers, those who reported exposure to ETS and those who reported no exposure. Since both groups are nonamok- ers, the concern of whether or not the clearance rates for nicotine or nicotine metabolites differ between smokers and nonsmokers is not germane to these estimates. In the study by Wald and Ritchie (1984), the urinary coti- nine levels aniong nonsmokers exposed to smoking spouses were 3 times those of nonsmokers married to nonsmokers. Using a linear model of risk and assuming that the 3:1 ratio represents a lifetime difference, the implied relative risk of these two groups would be equal to: Rn a 1+ 9ftdN ~ risk to "exposed' nonsmokers 1-h Pdw risk to "unexposed" nonernokere' (12-1) where dH is the dose received by nonsmokers who are self-declared "unexposed" and p is the increase in risk per unit dose received (for details, see Appendix C). This eciuation assumes that the lifetime carcinogenic dose received by nonsmokers who say that they are "exposed" is 3 times that of truly unexposed nonsmokers, assuming cotinine levels to be a proxy for carcinogenic constituents of ETS. When Equation 12.1 is set equal to the relative risk, one can solve for QdN. In the previous section, it was noted that the true relative risk in likely to be 1.25 and, as argued above, probably lies between 1.15 and 1.35. Consequently, relative risk values of 1.25, 1.15, and 1.35 will be considered. Using these values, pdN will be 0.14 ("ranging" 0.08 to 0.21). Therefore, the relative risk of solf-identilied "unexposedp nonsmokers compared with truly unexposed nonsmokers is: 1+fidN 1 ' (12-2) which would be 1.14 ("ranguig" 1.08 to 1.21). The relative risk of "expoeed" nonsmokers compared with a truly unexposed non- smoker is: •1 + 9ddN 1

242 which would be 1.42 ("ranging" 1.24 to 1.61). That is, the in- creased risk of lung cancer ss a result of chronic exposure to l,'TS, corrected for the effect of not identifying /t truly unoxposed refer- ence _ - ence group of nonsmokers, is likely to be at least as large an the observed risk. We can say, therefore, that while the epidemiologic studies show a consistent and, in total, a highly significant association between lung cancer and ETS exposure of nonsmokers, the excess might, in principle, possibly be explained by bias. However, de- tailed consideration of the nature and extent of the bias shows that given some reasonable assumptions the bias would be insuAicient to explain the whole effect. In fact, thero are some types of bias that lead to underestimates of the effect. It must be conc-luded, therefore, that some, if not all, of the eifect reported in spouse studies is causal. UTIiEIt C©N$IDE1t.ATI()N3 Some of the spouse-slnoking studies show a dose•response e[-& fect with rates increasing with increasing exposure as measured by increasing levels of cigarette consumption by the smoking spouse (see Tables 12-8 and 12-9). A dose-reeponse relationship also sug- gests a causal explanation, although biases could also operate to affect this estimation. It is possible that a person misclassitied as a nonsmoker married to a smoker will have a cigarette consumption that is correlated with that of his or her spouse. A misclassified nonsmoker married to a heavy smoker would, therefore, have a higher risk of lung cancer independent of spouse is smoking than a misclassified nonsmoker married to a light smoker, thus giving the appearance of a dose-response relationship between ETS exposure and lung cancer. This possible pseudo-dose-response effect arises only as a re- sult of misclase+ifyiug smokers as nonsntokers. It is of interest, - therefore, that one study has reported an effect of passive smoking - - in smokers as well as nonsmokers (Akiba et al., 1986). However, it does not appear that adjustment has been made for amount smoked. To khe extent that smokers married to smokers may smoke more than the smokers married to nonsmokers, this would bias the results. 243 ,rAsl.n 12-8 Risk of Lung Cancer in Nonsmokers According to Cigarette Consumption of Spouse Au/hors l:ass-Eaetrul .1Yrdiss '1'rkhnpouku el al., 1983 Uarnnkel et al., 1985 , Akiba el d., iu press l'aAovf Studies ilirayama, 19114 t,arfinkel, 19H1 Findings Exsmoken 1.0 1-20 cig. per day 2.4 21 + cig. per day 3.4 1-191o1al cig. per day 0.84 20-39 total sig. per day 1.08 40+ tolal cig. per day 1.99 elgarfpipe 1.13 1-19 cig. per day 1.3 20-29 cig. per d.y 1.5 30+ cig.. per day 2.1 1-19 cig. per day 1.45 20+ cig. per day 1.91 1-19 cig. per day 1.27" 20+ cig. per dar 1.10' 'Morlality ratios, not relative risks. NO'rH: Relative risk for selt-reported unexposed Is assuuwai to be 1.0. Most of the studies considered the histological type of lung cancer. in general they showed a higher proportion of adenocarci noma in ETS-exposed nonsmokers than would be expected among active smokers. This is to be expected in view of the fact that - the proportion of adenocarcinomas is, in general, higher among nonsmokers. Adding some nonadenocarcinoma-type disease, pos- sibly as a result of ETS exposure, would reduce this proportion. It would nonetheless leave the proportion of adenocarcinomas higher than would be found among lung cancer cases among active anlok- ers. If there were a high relative risk of adenocarcinoma associated with ETS exposure of nonsmokers, it would suggest a real e8oct, but the published data are insufficient or_ not ot presented in a way - to allow assessing this issue at'this time. Two studies have examined the risk of lung cancer associated with passive smoking using parental smoking as a measure of exposure instead of spouse smoking (Correa et al., 1983; Sandler et al., 1985b). The first found an association with maternal smoking -- (1lR = 1.66, p < 0.05) but not with paternal entoking (RR = Vzo48449

244 0.83). The second found no significant association with smoking of either parent. The bias discussed in connection with spouse-smoking studius is likely to apply also to parental-emoking studies. In addition, these two studies included active smokers ae well as recorded nou- amokers, and it is likely that the children of smokers start smoking at a younger age and possibly smoke more than do smoking chil- dren of nonsmoking parents (U.S. Public Ilealth Service, 1984). This would also be a source of bias. TABLE 12-9 Risk of Lung Cancer In Nonsmokers According to Duration of Smoking of Spouse or Other Measures of Exposure Not Shown in Table 12-7 _ Authors Euse-CoNrof Studies los ci al.. 198.1 o i h T Findinits Total no. of ell. (in Ihuusanils): u op c r 1-99 1.3 100-299 2.5 300+ 3.0 Correa el al.. 1983 Total pack-year:: Males FEm!lej 16 1 1-40 . 3.52 41+ 2 07 AII 2.0 . Koo et at., 1964 Total hnurss 1-3.499 13d 3,SU0+ 1.02 Any 1.24 Cartinkel et al., 198S No. of h/day: 1-2 t.ust S yr 1.59 twf 25 yr 0.77 3-6 1.39 1.34 >6 0.94 1.14 k reyhur; l I EMGurst 11 Pe ha en et al. In ptcss rs g •_- I fd or Less lhan 15 c g- ay So 6 iobacco/wk for kns than 30 yr 1.6 0.8 M ote than 15 ci6./dait or _ 50 6 tobacco/wk for nwte than 30 yr '.0 .4 Akibo et a1., in press p.ck-daYs within last 10 yr; < 5.00 1.0 5.000-9.999 2.6 11/.000+ 1.8 --- NOTE: Relative risk for sslt-rep"rted unexposed (s assomed to he 1.0. 245 Pershagen et al.. (in press) reported that the relative risk for lung cancer in women married to smokers and living in a home that had measurable radon levels was increased relative to the effects of living with a smoker or living in a home with radon. They suggested that this might represent an interaction between exposure to ETS and' radon. ltesearc_ h needs to be done that explores this association further iu light of recent reports of high radon concentrations in homes (Code of Federal Regulations, 1985). SUMMARY AND It,E C OMMENDATI ONS The weight of evidence derived from epidemiologic studies shows an association between ETS exposure of nonsmokers and lung cancer that, taken as a whole, is unlikely to be due to chance or systematic bias. The observed eatimate of increased risk is 34%, largely for spouses of smokers compared with spouses of nonsmokers. One must consider the alternative explanations that this excess either reflects bias inherent in most of the atudies or that it represents a causal efCect. Misclassification can have con- tributed to the result to some extent. Computations of the effect of two sources of misclassification were presented. Computations taking into account the possible effects of misclassiRed examokers and the tendency for spouses to have similar smoking habits placed the beat estintate of increased risk of lung cancer at about 25% - in persons exposed to ETS at a level typical of that experienced - _ by nonsmokers married to smokers compared with those married to nonsmokers. Another computation using information from co- tinine tinine levels observed in nonsmokers and taking into account the effect of making comparisons with a reference population that is - truly unexposed leads to an estimated increased risk of about one- third when exposed spouses were compared with a truly unexpoaed population. The finding of such an increased risk is biologically plausible, because nonsmokers inhale other people's smoke and, as a result, absorb smoke components containing carcinogens. What Is Known 1. A sumnlary estimate from epidemiologic studies platcea the increased risk of lung cancer in nonsmokers ntarried to smokers compared with nonsmokers married to nonsmokers at about 34%. szo4e44e I

24¢ Assuming linearity at low-to-average doses and a constant pro- portionality of nicotine and carcinogens in mainstream smoke and ETS, extrapolation from studies of active smokers using relative urinary cotinine places the risk at about 10%. 2. To some extent, misclassification (bias) may have con- tributed to the results reported in the epidemiologic literature. However, bias Is not likely to account for all of the increased risk. The best estimate, allowing for reasonable trtisclassification, is that the adjusted risk of lung cancer is increased about 25% (i.e., I1R - - = 1.25) in nonsmokers married to smokers compared with non- smokers married to nonsmokers. When one allows for exposure to nonsmokers who report themselves as unexposed, the adjusted increased risk is at least 24%. The adjusted increased risk to a group of nonsmokers married to nonsmokers is at least 8% (i.e., - - RR = 1.08) compared with truly unexposed subjects. This excess risk may come about from expoaures in the workplace or other public places. What Scientific Information Is Missing 1. It would be useful to quantify the dose-response relation- ship - ship between ETS exposure and lung cancer more precisely using biological markers of exposure. Studies should be done that incor- porate these biological markers. 2. Laboratory studies. would be important in determining the carcinogenic constituents of ETS and their concentrationH in typical daily environments and in facilitating understanding of possible dose-response relationships. 3. The interaction between ETS and radon exposure, which - can increase risk of lung cancer, is worth es amining further. REFERENCES Akiba, S., W.J. Blot, and H. Kato. Passive smoking and lnng cancer aniong Japanese women. Fourth World Conference on Lung Caacer, Toronto, Canada, Aug. 46-30, 1986. Akiba, S., H. Kato, and W.J. Blot. Passive smokLig and lung cancer among Japanese women. Cancer Res. 46:4804-4807, 1986. BufSer, P.A., L.W. Pickle, T.J. Mason, and C. Qontaut. The causes of lung cancer In Texas, pp. 83-99. In M. Misell aail P. Correa, Nds. Lung Cancer: Causes and Prevention. New York: lfarlag-Chemie International, Inc., 1984. 247 Chan, W.O., and S.O. Fung. Lung cancer in non-smokers in Hong Kong, pp. 199-202. In E. flrnnd:nann, Ed. Cancer Campaign, Vol. 6. Cancer Epidemiology. Stuttgart: Gustav Fischer Verlag, 1982. Code of Federal Regulations (OFR). National primary and secondary ambient air quality standards. Code Fed. Regul. 40(Pt. 60):600-673, 1985. Correa, P., L.W. Pickle, B. Fontha:n, Y. Lin, and W. Haenssel. Passive smoking and lung cancer. Lancet 4:696-697, 1983. Doll, R.D., and R. Peto. Cigarette smoking and bronchial carcinoma: Dose and time ralationships among regular smokers and lifelong non-smokers. J. Epidemioi. Comm. Health 3z.303-913, 1978. Doll, R., R. Uray, B. Hafner, and R.. Peto. Mortality In relation to smoking: 22 years' observations on female British Doctors. Br. Med. J. 967-971, 1980. Friedman, G.D., D. Pettiti, and R.D. Bawol. Prevalence and correlates of -- pauive smoking. Am. J. Public Health 78:401-406, 1983. QarBnkel, L. `Pime ttends In lung cancer mortality among nonsmokers and a note on passive smoking. J. Natl. Cancer Inst. 6e:1081-1066, 1981. Garfinkel, L., O. Auerbach, and L. Joubert. Involuntary smoking and lung cancer: A case-eontrol study. J. Natl. Cancer last. 76s463-489, 1986. dart, J.J., and M.S. Schneidermn. eLow.risk` cigarettes: igarettes: The debate continues. Science 204:688-690, 1979. Siillis, C.R., D.J. Hole, V.M. Hawthorne, and P. Boyle. The effect of environmental tobacco smoke In two urban communities In the west of Scotiand. Eur. J. Respir. Dis. 133(Suppl.):1s1-198, 1984. Haley, N.J. and D. Hoffmann. Analysis for nicotine and cotinine In hair In determination of cigarette smoker status. Olin. Chem. 31:1698-1600, 1985. Haley, N.J., 1). Hoffmann, and E.L. Wynder: Uptake of tobacco smoke aomponents. In D. Hoffmann and 0. Harris, Eds. Mechanisms In Tobacco Oarcinogenesis (Banbury Report 23). Cold Spring Harbor, New York: Cold Spring Harbor Laboratories, 23:3-19, 1986. Hammond, B.C. Smoking In relation to the death rates of one million men and women. Natl. Cancer inst. Monogr. 19:127-204, 1966. Hirayama, T. Cancer mortality In nonsmoking women with smoking husband_s_ on a large-scale cohort study In Japan. Prev. M.d. 1'J:680-690, 1984. Hirayama, T. Non-smoking wives of heavy smokers have a higher risk of lung cancer: A study from Japan. Br. Med. J. 282t183-186, 1981. Hugod, O., L.II. Hawkins, and P. Astrup. llarpo.ure of passire smokers to tobacco smoke constituents. Int. Arch. Occup. Environ. Health 42;21-29, .1978. Jarvi., M.J.., li. Tunstall-Pedoe, O. Feyeraband, O. Yes.y, and Y. Saloojee. Biochemical markea of smoke absorption and self reported exposure to passire smoking. J. Epidemiol. Comm. Health J8:366-339, 1984. Kabat, Q.O., and E.L. Wynder. Lung cancer In nensmokers. Cancer 63:1214- 1231,1984. Klosterkotter, W., and E. tdono. Zum Problem das Passivrauchens. Zentrabl. Bakteriol. Hyg., I. Abt. 1: Orig. B 162:51-69, 1976. Knoth, A., H. Ilohn, and F. Schmidt. Pauive smoking as a causal factor of bronchial carcinoma In female nonsmokers. Medisinisch Klin. 78:86-69, 1983. 9ZO4B44B

248 Koo, L.(7., J.H-O. Ho, uid N. Lee. An analysis of some risk factors for lung • cancer In Hong Kong. Wt. J. Cancer 35:149-166, 1986. Koo, L.Ct., J.H-©. Ho, J.F. Ihaumeni, W.J. Blot, J. Lubin, and B.J. Stone. - Measurements of passive smoking and estimates of risk for lung cancer among non-smoking Qhineu females. Fourth World Con(erence on Lung - - Cancer, Toronto, Canada, Aug. 26-30, In press. Kyerematen, G.A., M.D. Damiano, B.H. Dvorchik, aml E.S. Vesell. Smoking- _ _ induced changes (n nicotine disposition: Application of a new HPLfl assay for nicotine and its metabolites. Olin: Pharmacol. Ther. 82:789- 780, 1983, Lee, P.N., J. Chamberlain, and M.R. Alderson. lt.elationship of passive - smoking to risk of lung cancer and other smoking-associated g-associated diseases. Br. J. Cancer 64:97-106, 1986. Lehnert, a., L. Garfinkel, T. Kirayama, D. Schmilh, K. f)berla, E.L. Wynder, and P. Lee. Rountable discussion. Pr.v. Med. 13:730-746, 1984. - Matsukura, S., T. Taminato, N. Kitano, Y. Seino, H. Ilamada, M. Uchihuhi, -- - - H. Nakajims, and Y. Hirat.. Effects of environmental tobacco smoke on urinary cotinine excretion In nonsmokers. N. Engl. J. Med. 311:828-832, - 1984. Milier, (3.H. Cancer, passive smoking and nonemployed and employed wives. West. J. Med. 140:882-886, 1984. -- Office of Science and Technology Policy.. Chemical carcinogens: A review of the science and its associated principles, Feb. 1986, pp. 10371-10443.- - Fed. Regist. 60:60(14 Mar. 1985). Pershagen, (1., Z. Hrubec, and C. SYensson. Passive smoking and lung cancer in Swedish women. Am. J. Epidemiol., In preu, 1988.- - Rylander, R. Environmental tobacco smoke and lung cancer. Eur. J.. Respir. Dis. 133(Suppl.):127-188. - Samet, J.M. Relationship between passive exposure to cigarette smoke and cancer, pp. 227-940. In R.B. Ciammage, S.V. Ifaye, and V.A. Jacobe. Indoor Air and Human Health. Chelsea, MichiKan: Lewis Pub., Inc. -- Sandler, D.P., R.B. Everson, and A.J. Wikox. Passive smoking In adulthood _ and cancer risk. Am. J. Epidemiol. 121:37-48, 1985a. - Sandier, ndler, D.P., A.J. Wilcox, and R.B. Everson. Oumulative effects of lifetin:e passive smoking and cancer risk. Lancet 1:812-316, 1986b. Saioojee, Y., C.J. Vesey, P.V. Cole, and M.A.H. Rnsuell. C7arboxyhemoglobia and plasma thiocyanate: Complementary indicators of smoking behan- - _ _- iort Thorax 87:641-646, 1982. Sepkovie, D.W., W.J. Haley, and D. Hoffmann. Eliiaination from the hody of tobacco products by smokers and passive pnwkers. JAMA 268:80:1, 1988 (letter). Trichopoulos, D., A. Kalandidi, and L. Sparros.. l.ang cancer and passive smoking: Conclusion of Greek study. Lancet 20077-078, 1983. U.S. Public Health Service. The Ilealth-Oonsequences of Smoking: Chronic - Obstructive Lung Disease. A Report of the Surgeon General. DHIIS (PHS) Publ. No. 84-60205. Rockville, Maryland: U.S. Department of Health and Human Services, Public Health Service, Office sn Smoking - _ and Health,,1984. 646 pp. - - - - Vutuc, C. Quantitative aspects of passive smoking and lung cancer. Pu+v. Med. 13:698-704, 1984. 249 Wald, N.J. Validation of studies on lung cancer In nonsmokers married to _- smoka. Lancet 1:1067, 1984 (letter). Wald, N.J., and 0. R3tchie. Validation of studies on lung cancer In non- smokers married to smokers. Lancet 1:1007, 1984. Wald, N.J., M. Idle, J. Boreham, and A. Bailey. Carbon monoxide In breath In relation to smoktng.and carboxyhemogbbin levels. Thorax 88:986-.889, 1981. Wald, N,J., J. t3or.Lam, A. Bailey, O. Ritchie, J.E. Haddow, and 0. Knight. Urinary cotinine as marker of breathing other people's tobacco smoke. Lancet 1:230-231, 1984. Wu, A.H., B.E. Henderson, M.O. Pike, and M.O. Yu. Smoking and other risk factors for lung cancer In women. J. Natl. Cancer Iast. 94:747-761, 1985. 42.O4844R

13 Cancers Qther ~rhail Lung Cancer The association of lung cancer with exposure to ETS has yielded relative risks of 2 or less for nonsmokers. As cancer of the lung is the cancer most strongly associated with active smok- ing, weaker effects would be expected for c,ancers that are less closely related to smoking. The first emphasis in this chapter is on smoking-related cancers, because these might be more plausibly associated with exposure to ETS. However, exposure to ETS oc-- curs at earlier ages than active smoking; thus, there may be effects of ETS exposure on risk for other cancers. 9MOKING-RELATEi) CANCERS Active tobacco smoking is an important cause not only oLlung cancer, but also of bladder cancer, cancers of the pancreas aud re- nal pelvis, and probably of the nasal sinus and kidneys. Oral, oropharyngeal, hypopharyngeal, laryngeal, and oesophageal can- cers are also strongly associated with active smoking, especially in conjunctiorl with the use of alcohol. Primary cigar and pipe smok- ers face a somewhat lower risk for cancer of the lung than cigarette smokers, but theie riek for cancer of the larynx, pharynx, oral cav- ity, and esophagus is similar if not greater than that of cigarette smokers (U.S. Department of'Health and Human Services, 1982). Also, lip cancer is associated with tobacco smoking, as well as pan- creatic cancer t.nd, perhaps, renal adenocarcinoma. An increased risk of cervical cancer has been observed in tobacco smokers, but 251 TAeLZ 13-1 Studies of Passive Smoking and Cancers Other Than Luug - Cancer wilh Significantly Increased Risks Author Study Design ' Size (Cases and Polwlalbn 4~c Controls) Tumor Outcome Studied Odds Ralios= Hirayama, 191N Cohort 34/91,5•10 Brain 3.0;6.3;4.3 28l91,540 Nasal sinus I.7;2.0;2.6 Miller, 1984 Case-Control 123/537 All sites 1.40 (3illis, 1984 Cohort Male: 8/827 Sites 0.5 (hl) Female: 43/1917 other than 1.26 (F) Sandler et al., Case-Control 518/518 lung All sites 1.6 19952 (Iotal)in- Breast 1.8 (adulthood cludes smokers Cervix 1.8 exposure) Endocrine 3.2 Sandkr et el., Case-control 869/409 glands All sites 1.4;2.3;2.6 1985b (subset) in- Breast 2 0;2.4;3.3 (lifetime cludes smokers Cervix 1.6;3.6;3.4 exposure) Leukemia and 2.5,5.1;6.8 Sandler et al., Case-control 438/470 lymphoma All sites 1.5 1985¢ (subset) in- Cerviz 1.7 (early life ciudes smokers okea Hematopo etic 2.4 exposure) _ /issue 'Given witb increasin8 dose, it available. the causal relationship is unclear (International Agency for Re- search on Cancer, 1986). The risk for these cancers to nonsmokers, exposed to ETS has been the subject of a few studies. Hirsyalna (1984; see Chapter 12 and 'lhble 13-1) examined cancers of the mouth, pharynx, oesophagus, bladder, pancreas, and cervix. The relative risks were not given, but they were reported to be insignificant. However, a rela4ionship between ETS exposure in nonsmokers and nasal sinus cancer was noted, with rate ratios for the aforementioned exposure categories of 1.7, 2.0, and_ 2.6, respectively (see Table 13-1). Sandler et al. (1985a; described iu more detail below and in Table 13-1) also did not find_ a significant odds ratio for any of the smoking-related cancers (including lung cancer), except for cervical cancer (p < 0.05). The odds ratios given for these cancers included snv,kers as well as nonstnokery. Therefore, since the odds ratios were not significant for the combined group, they would 8Z048414•S 250

262 not be expected to be significant for the nonsmokers analyzed separately. CANCERS NOT RELATED TO SMOKING Hirayama's (1984) study, based on a cohort of 91,450 non- smoking Japanese women, suggested an increased mortality fro[n brain tumors among women whose husbands smoked. The rate ra- tios were 3.0, 6.3, and 4.3 for exposure to husbands smoking 1-14, 15-19, or 20 or more cigar6ttes per day, as compared with non- smoking wives of nonsmoking husbands as the reference group.. A trend was noted for all cancer sites, but the risk elevation became insignificant when lung, nasal sinus, brain, and breast cancers were excluded. No significant associations were found for cancers of the - stomach, colon, rectum, liver, peritoneum, ovary, skin, or bone, or -- - - for malignant lymphoma or leukemia Sandier et al. (1986a,b,c), reporting on a case-control study from North Carolina, suggested an association of exposure to ETS at different periods during a lifetime with various types of cancer. People with cancer at any site, except basal cell cancer of the skin, were included in this study. The cases were drawn from a hospital-based tumor registry, irrespective of personal histories of smoking. Mailed questionnaires were used for collecting data on exposure, preceded by a telephone call for the control subjects, but not for the cases. Many of the odds ratios reported in these articles are for t e combined group, as briefly reported in 7ible 13-1. However, some results were reported separately for nonsmoking cases (No. = 231) and controls (No. = 235). The results discussed below are based on the latter group and thus reflect only 31% of the total eligible patient group. The overall crude cancer risk among individuals who were ever - married to smokers was 2.1 times that of those never married to -- - smokera Significantly elevated risks (p < 0.05) were seen ala.o for cancer of the cervix (odds ratio 2.1) and endocrine glands (odds ratio 4.4) (Sandler et al., 1985a). A nonsigniticant odds ratio of 2.0 was obtained for cancer of the breast. -- A subset of this study involved subjects who had lived witli --- - - both natural parents for most of the first l0 years of life and had information on the smoking habits of both parents and epoases. 253 Overall cancer risks were found to increase steadily and signifl- eantly with each additional household member who smoked (San- dier et al., 198bb). The overall risk was significant only for adult- hood exposure, either alone or in addition to childhood exposure (Sandler et al., 198bd) This trend appeared both for cancers tra- ditionally associated with smokiug and for other sites, with the -- - - etrungeat trend for the smoking-related cancers. The traneplacental and childhood exposures to ETS were spe- cifically cifically studied In another subset of the same study; however, the data were not adjusted for prenatal exposure (Sandier et al., 1986c). There were no significantly increased risks indicated for all sites or for specific cancers. liematopoietic-tissue-cancer risk had an odds ratio of 2.3 when maternal smoking was considered and 2.4 when paternal smoking was considered (significance not given). , Cancers of hematopoietic tissues have been reported as in- creased in children whose mothers smoked during and after preg- nwrcy. Neutol and Buck (1971) studied 65 cancer deaths among 89,302 children and found that the rate of leukemia among children of smokers was about twofold that of nonsmokers, but without a dose-reaponse trend. The total number of leukemia cases in this study was 22. Manning and Carroll (1957) studied 187 cases of leukemia, 42 cases of lymphoma, and 93 other cancers among chil-, dren, but found no effect of mothers' smoking habits. Neither of - - - - - these studies separated the effects of in utero exposure from the --- - - exposure to 14TS after birth. Two studies have evaluated all sites of cancer as a group. Miller (1984) questioned relatives of women who died between 1972 and 1976 in Erie County, Pennsylvania. lie found a non- significant increased risk (1.40) of any cancer among women whose husbands smoked. In another study (Gillis et al., 1984), a popula-. - -- tion in Scotland was followed up 10 years after an initial screening survey for cardiovascular disease. The Weat of Scotland Cancer Registry was screened for subsequent incidence of cancer. Among the nonsmoking males, there were 8 cases of cancer other than - lung cancer. The standardized mortality rates were actually de- creased among men whose wives smoked (ratio = 0.50). Among the nonsmoking women, there were 43 cases of cancer other than lung cancer, and the ratio of standardized mortality rates was - --- - - nonsignific-antly increased (1.26). 6Z048448

254 In the study by Preston-Martin at al. (1U82) of childhood brain tumors, caso and control mothers were similar in use of cigarettes during pregnancy. This is in contrast to the finding that significantly more case mothers than control mothers lived in a household with a smoker. The lack of an association of risk with maternal smoking, the association of smoking behavior with other lifestyle-related exposures, and the lack of apparent adjustment for smoking status of the mother make these uncorroborated results - difficult to interpret. INTERPRETATION Interpretation of these observations regarding a possible asso- ciation of ETS exposure and cancers with or without previously found associations with smoking is difGcult. Only a few studies have been reported, on cancer and ETS exposure other than lung cancer. The Sandler et al. (1985a,b,c) reports have been criticised for not maintaining matching (Higgins, 1985), for recruiting controls through two separate mechanisms, for conducting interviews in two different ways (telephone interviews or mailed questionnaires), and for insufficient information on other variables known to be risk factors for the various cancers that have been studied in these reports (Burch, 1985; Friedman, 1986; Mantel, 1986). '1'he criticism of the lack of information on other known risk factors - is of special concern and is also pertinent to the Ilirayama study. For instance, alcohol use, reproductive and sexual histories, and occupational exposures are important risk factors for. several of the cancers studied. Considering increased risks for hematological malignancies, leukemia has not been thought to be smoking-related even though there have been reports of higher leukemia risks among emokuES. However, there is a possibility that inhaled lead-210, originat- ing from the tobacco or from radon daughters attaching to E'L'S, could end up in the skeleton, especially in young individuals who are building up their skeletons, and would result in irradiation of the bone marrow. Such an explanation is presently highly spocu- lative, but increased concentrations of lea-4L-210 have been found in the skeletons of adult smokers (Iloltsman and Ilcewicz, 11166; Blanchard, 1967). Adults would be less semnitive to radiation than children. Austin and Cole (1986) suggest I,hat, in addition to the possible influence of radioactive elements, benaene, urcthane, and nitrosandnes enay be contributing factors. All of these chemicals are found in cigarettes and E'rS, aud they havo been shown to be le_ ukomogenic in experimental animals, or in humans. The findings of increased brain cancer associated with ETS ex- posure in the llirayarna (1984) study; and possibly in the Preston- Martin ot al. (1982) study, are of note. N-nitruso compounds are - - potent nervous system carcinogens in animals (Magee et al., 1976; Preuseman, 1084, 1986). , SUMMARY AND RECOMMENI?ATIONS These recent observations on a possible connection between ETS and various forms of cancer have created ntuch discussion and some confusion. The lack of consistency with other data ott tumors among children of smoking mothers and the appearance of tumors that are not clearly smoking-related call for further epidemiologic- research. Any new studies in this area will, hopefully, have a very careful, rigorous design, so that more definitive evaluation of this possible health hazard from ETS exposure is possible. What Is Known 1. There is no consistent evidence at this time of any increased risk of BTS exposure for cancers other than lung cancer. What Scientific Informatlon Is Missing 1. Smoking-related cancers other than lung cancer need to be studied with adequate numbers and good exposure data and with consideration of the potential confounding effects from other -- - known risk factors for these cancers. 2. Some cancers not related to active smoking, especially lyniphohematopoibtic neoplasms, should be studied in relation to ETS exposure, particularly in childhood. Then the possibility of a etiologic role of inhaled decay products of radon (like bone-seeking lead-210) should be considered. REFEItENCES Austin, H., enml P. Cole. Cigarette smoking aad kukemia. J. Chronic Dis. 59:417-421, 1988. OCQ4BLL9

256 Blanchard, R.L. Concentration of 21ePb and 21nPo h: human soft tissues. Health Phys. 11:836-833, 1967. Burch, P.R.J. LifetBne passive smoking and canuor risk. Lancet 1:860, 1985 (letter). Friedman, ©.D. Passive smoking In adulthood and cancer risk. Am. J. Epidemlol. 123:367, 1986 (letter). Q_ illi., O.R., D.J. llole, V.M. Hawthorne, and P.. Boyle. The sifect of environmental tobacco smoke In two urban communities In the west of Scotland. Bur. J. Respir. Dis. 33:8121-S1t8, 1984. Higgins, l. Lifetime pusive smoking and cancer riek. Lancet 2:867, 1986. Hlraya:na, T. Cancer mortality In nonsmoking women with smoking husbands based on a large seale cohort study (n Japan. Prev. Med. 13:880-0911, --- - - -- 1984. Holtsman, R.B., and F.H. Ikewics. Lead-210 and poloniu:n-410 In tissues of cigarette smokers. Science 163:1269-1280, 1906. IInternational Agency for Research on Cancer (IARO) Monograph. Evaluation of Carcinogenic Risk of Chemicals to Humans, Vol. 38, pp. 103-;114. Tobacco Smoking. Lyon: IARC, 1986. 421 pp. Magee, P.N., R. Montesano, and R. Preussman. N-Nitroso compounds and related carcinogens. ln: C.E. Seark, GL Chemical Carcinogens. - - Washington, D.C.: ACS Monogr. 173:491-846, 1976. Manning, M.D., and B.E. Carroll. Some epidemiological aspects of leukemia in children. J. Natl. Cancer lnst. 19:1087-10U4, 1067. Mantel, N. Passive smoking In adulthood and cancer risk. Am. J. Epidemiol. - - - 143:387-388, 1986 (letter). Miller, Ci.H. Cancer, passive smoking and nonemployed sud employed wives. West. J. Med. 140:832-836, 1984. - Neutel, f3.I , and O Buck. Effect on smoking duting pregnancy on the risk of cancer In children. J. Natl. Cancer Ina. 47:69-83, 1971. Preston-Martin, S., M.C. Yu, B. Benton, and 1I.E. Ilenderson. N-nitroso - - - - compounds and childhood brain tumors: A case-control study. Cancer Res. 42:6340-5246, 1983. Preussman, R. Carcinogenic N-nitroso compounds and their environmental ----- - significance. Natuewiesenchdten 71:26-30, 11,84. Proussman, R., and B.W. Stewart. Carcinogenic N-nitroso compounds and - - related carcinogens. In C.E. Searle, Ed. Chemical Carcinogens. 2nd Ed. Washington, D.C.: American Chemical Itocioty 182:043-838, 11184. S_a_ndler, D.P., R.B. Everson, and A.J. Wilcox. Psssive smoking In adulthood and cancer risk. Am. J Lpidemiol 121:37-48, 198Ga. Sandier, D.P., A.J. Wilcox, and R.B. Everson. Cumulative effects of lifelime pauive smoking on cancer risk. Lancet 1:314-316, 1985b. . Sandier, D.P., R.B. Everson, A..1. Wilcox, and J.1'. Browder. Cancer risk In adulthood from early life exposure to parents' smoking. ASu. J. Public Health 76:_487-494, 1986c. Sandler, D P., R.B. Everson, and A.J. Wilcox. The authors reply. Am. 1. Epidemiol. 123:369-370, 1988 (letter). U.S. Department of Health and Human Services. The Consequencn of Smoking. Cancer. A Report of the Surgeon General. DHSS (1'f1S) Publ. No. 82-6u179. Rockville, Maryland: U.S. I)epa_rtment of Health and Human Services, Public Health Servica, Ollice on Smoking and - - Health, 1982. 302 pp. LEO~,e~~e I I 14 Cardiovascular System The elfects of active smoking on exercise tolerance, blood pressure, and the risk of developing cardiovascular disease have boen reviewed elsehwere (U.S. Public Health Service, 1983). This chapter diNcusses studies of E'i'S exposure to nonsmokers and subaequent possible cardiovascular effects. 'rhe constituents that are thought to have the greatest offect on the cardiovascular system are carbon monoxide (CO) and nicotine. 'Fhe possibility. .exists that the mechanisms, as well as the magnitude of the effects, for acute and chronic cardiovascular effects may be different for _ exposure to whole smoke and to ETS. ACUTE CARDIOVASCULAR N FFECT3 OF fi.:NVIRONMENTAL TOBACCO SMOKE EXPOSUItE Administration of nicotine at level sin:ilar to those induced - ----- - by active cigarette smoking is shortly followed by increases in heart rate and blood pressure (U.S. Public Health Service, 1983). - --- - - Platelet aggregation has been shown to be increased in in vitro --- sLudies. CO rapidly combines with hemoglobin in the blood to --- -- fi,rm carboxyhemoglobin (COHh), thereby leading to some degree, of tissue hypoxia. CO combines with muscle myoglobin, which is fi,llowed by some muscle hypoxia. The level of exposure of the nonsmoker to_ these cigarel.te sntoke constituents, however, is less titan that of the active smoker, and the effects are expected to be less. Table 14-1 reviews some of the increanes in C©f1b levels as seen in both experimental and observational studies. The levels of 257

258 TABLa 14-f Carbon Monoxide and C_arboxyhcmoglobiu L.tpvels in Nonsmoking Individuals Experimental Studies /Gootrofled CAom6ers/ Carbu:yhemeglobin No of No. of CO, Study Cigarelles/h/!0 ma Subjecls ppn:' Coadrol C'hange Anderson and Dalhamm, 3.1 4.5 0.3 0 1973 +0 4 1981 Dahms et al. - t0 IS-20 0.6 . , 1970 Harke 3.9 7 30 0 0.9 + 1.2 , Huch el al., 1980 2.3 12 - 1.3 +0.5 Hugod et al., 1978 2.5 !0 20 0.7 +0.9 1978 Pimn et a1. 2.4 10 24 0.5 +0.3 , 2.4 10 24 0.7 +0.2 1977 Polak 6.7 15 23 2.0 +0.3 , Russell et al., 1973 . 15.1 12 38 1.6 + 1.0 Seppenen and Uusltalo, . 3.8 28 Ib 1.6 +0.4 1977 Srch, 1967 so 2 +3 O6servatiueal Studies Nnnexpusedc Expused Study SubjeLis/Bxlwwre No. of Carbo:y- Subjeels hemogklbin, % CO EX- pired, PPm Foliart et al., 1982 Flight attendantsf8 h 6 1.1t:0.7 Jarvis el al., 1983 Normal/publie house tor2h 7 4.7:10.6 Liglufool, 1972 Normal/submarine -:1.0 Wald et al., 1981 Participants in health screening program 6,641 Jarvis et al., 1984 Normal/self report 10 0.9:0.8 S.7:S.S SeppYnen and Uusitalo, 1977 Restaurant for 5 h (Ct1:2.3-IS ppm) 47. 2• 1:2• 1 Office for 8 h IS 2.3:2.3 (CU:2.5 ppm) _ 'Carbou monoxide (CO) measured as a pro:y to Indicate tlw: concentration of ETS ia lhe chamber. COHb commonly observed in active smokers are higher, rangius between 4 to 6 percent, rarely greater than 12 percent (Schievul-n bein and Richter, 1984). Because exposure of the nonemoker is qualitatively different than exposure to emokerp, a sintple scalin6 down of etfects observed in active smokers does not ai,pear to ite fully appropriate. Therefore, the effects of exposure to nicotine, 269 TASLV.14-2 Rcsting Acute Cardiovascular EFfects in Nondiscased - - - Huntans of Exposure to Environmcntal Tobacco Sntuke Results Aulhurs Study Population Conditions Measured Variable Before After Luguelle el al., 1971/ 40 children Room: 9 m3 Heart rate 89 97 ~ No. cig.: 6 Blood pressure 116/67 120/72 Harke and Blekhet 1, 10 Time: 1_5 min Room: n.g. Heart rate 72 t 8 74 ± 12 1972 No. cig.: 150 Bkrod pres.ure 123/84 121 /84 Time: 20 udn Skin lemparature (-°C/ndn) 0 0.0273 Rumuel at al., 1975 56 Room: 30 m3 Heart rate 72 t 10 71 ± 11 No. cig.: 6-8 Bhood pressure 117/71 117/71 Hunbman el al., 1978 8 Time: 20 ndn Room: n.g. Heart rate 73 79 No. cig.: 2-6 Blood pressure 107/67 114/68 Pimnr et al.. 1978 10 nnks Time: 10 min Room: 14.6 m3 Heart rate 84(F) 80(F) 10 females No cig.: 7 77(M) 70(M) Age = 22.3 Time; 2 h CO, or ETS noed_ to be separately studied. hr addition, consid- eration needs to be given to persons of different sensitivity or vulnerability. Heaithy Sub3ec ts 'Able 14-2 lists studies that report on the consequences of exposure of nondiseased individuals to ETS for periods up to 2 hours under experimental, resting conditions. There were no signiiicant changes noted in heart rate or blood pressurc in school- aged children or in adult men and women. ---- 'l~vo studies evaluated the physiologic responses to exercise with and without exposure to E'1.'S.. In the first, kimm et al. (1978) (see also Table 14-2) had subjects perform a 7-minute pro- gressive exercise test on an electronic bicycle ergometer. During dxercise, the women had higher heart rates after exposure to ETS when compared with control conditions (differences of 6.3 beats per luinute at 2 minutes and 4.5 beats ireh minute at 7 minutes, p< l).01). The recovery heart rates were not eigniRcautly differ- ent. The men, however, showed little diiference between test and zC04844s

260 control conditions (differences of -0.1 beats per minute at 2 min- utes ---- - - - utes and 1.5 beats per minute at 7 minutes). ln the second study, -- - - - Sheppard and colleagues (1979b) tested 11 males and 12 females - -- - - - --- , at two different levels of ETS (i.e., 7 cigarettes over 2 hours, CO =. --- 20 ppm, or 9 cigarettes over 2 hours, CO = 31 ppln). Under both . exposure conditions, contrary to expectations, both the increment in heart rate and average heart rate were less with ETS exposure. . In summary, for normal young adult males and females, no significant acute effects of ETS exposure on heart rate or blood - - pressure •llave been reported, either under resting or aerobic con- ditions. ditions. There have been several studies of exposure of normal sub-s jects under resting and aerobic conditions to low levels of CO but higher than those found with ETS exposure (reviewed in Envi- ronmental Protection Agency, 1984). No,significant effects were found in healthy, exercising subjects during short-term exposure (e.g., Drinkwater et al., 1974; Raven et al., 1S174a,bi DeLucia et al., 1983). Angina Patients Angina pectoris is a symptom complex Livolving feelings of pressure and pain in the chest, which is produced by Inild exercise or excitement, presumably because of insufficient oxygen supply to the heart muscle, Under conditions of ETS exposure, the CO levels are increased, thus possibly placing individuals with angina at an increased risk of recurrent epiaodee. Anderson et al. (1973) and Aronow and hia colleagues, in a series of experiments (1973, 1974; 1978, 1981) (Table 14-3), stud- ied angina patients under aerobic conditions with exposures to low levels of CO and to ETS. `1'en patients with diagnosed angina pec- toris, of whom two were smokers and eight exelnokers, were tested (Aronow et al., 1978). Significant increases in tiystolic blood pres- sure and heart rate, and decreases in time to onset of angina, were noted when the subjects were exposed to smoke in either venti- - lated or unventilated roonts (the actual levels of CO under these conditions were not noted). There were some subjective elements in the evaluation of these patients, and the physician conducting these tests was aware of the test conditions, i.e., smoking or not and ventilated or not. Consequently, the fintlutgs (if this study, in 261 TAst.e 14-3 Acute Cardiovascular Effects of Exposure to CO or Environmental'fobacco Smoke by Nonsmoking Angina Paticnts Study Design No. Conditions ' Results Anderson ol .1., Doubk-blind, IU" CO: 50 ppm Mean duration before onset 1973 Cross-over or 100 ppns of pain shorteoed (SO ppm Time: 4 h r S da f and 100 ppm); duration of pain Ionger (UK) ppm ys o only) Aronow aud Isbe_ Il._ Double blind_, IU° CU: Se ppm `I•imes until onset decreased; 1973 Cross-over Time: 2 h decrease in 1111 and heart Arooow, 1978 Not blinded 11}° No. cig.: IS rate at angina Ifarlier onset of angina; in- Time: 21t R+ams 311.28 m3 creased systolic BP and heart rate at angina et a1 A 1979 Double-blind 20 COHbp 4% Impairment in visualization rouow ., , ---- . Crosseover test Aronow, 1981 DDouble-blind, IS CO: SO ppm Time until onset decreased; Cross-over I'ime: I It decreased systolic BP and COHb: 2% heart rate al angina 'Includes five suiokeri and five nonsmokers. 6Not current suwkers. __ -- "Includes eight easmokers and two current smokers. the absence of a true double-blind approach, require verification by other research workers. The effects of rapid angina onset would be expected to be due to increased COHb levels. Anderson et al. (1973) and Aronow et al. (1973, 1981) exposed angina patients to low levels of CO. In these studies, angina pain appeared when COllb levels of patients were measurod at 2 and 4%. These studies have been reviewed ex- _-_- teusively as part of the Environmental Protection Agency's (1984) activity in establishing air quality criteria for carbon monoxide. The review group found that the results were suggestive for ef-s fects at COIIb levels above 3%, based on animal and theoretical models. There is concern that elevated levels of CO exposure rftay affect the electrical stability of the heart in previously compro- mitied heart muscle, thus possibly leading to sudden death. The levels reviewed in Table 14-1 are close to the 3% level. 'Phis sug- gents that there is reason to be concerned with possible effects of exposure. However, a finn qunntitative estimate of the risk to nonsmoking persons, under conditions of E'.1'S exposure, cannot be made from the literature at this tinre. CCo4e44e

....~..,.~..~.~~..~. 262 CARDIOVASCULAR DIS]:ASE M()RBIDI'i'Y AND MUxi.TAL1T'!€ Possible pathophysiologic mechanisms for the atherogenic in- fluence of cigarette emoking were reviewed in the 1983 Report of the Surgeon General. Experimental studies of subcutaneous or intravenous s administration of nicotine in rabbits (Schievelbein et al., 1970; Schievelbein and Richter, 1984) and monkeys (Liu et al., 1979) have demonstrated that long-term exposure leads to ar-c teriosclerotic lesions. Exposure to carbon monoxide also leads to - atherosclerosis in rabbits, pigeons, and other animals (Astrup and - Kjeldsen, 1979). Studies of whole tobacco smoke indicate that to-l tal serum cholesterol concentrations are increased and the ratios of the various lipoprotein fractions are changed (McGill, 1979). The contribution of whole tobacco smoke to modifying the lipoprotein fractions is not conclusive. However, there have not been experi- mental studies of the effects of ETS exposure or administration of ETS extracts. ' Smoking and Cardiovascular Disease The effects of active smoking on human health are sumnia- rised in the Surgeon General's report The Health Conaeguences of Smoking: Cardiovascular Disease (U.S. Public Health Service, 1983). The principal conclusions are that cigarette smokers ex- perience a 70% greater coronary heart disease (CHD) death rate than do nonsmokers and that smokers of more than two packs per day have 2 to 3 times greater CHD death rates than nonsmokers. The incidence of CHD ln smokers is twice that of nonsmokers. Heavy smokers (more than two packs per day) have an almost fourfold increase. The relative risk in smokers for sudden ileath is greater than that for all deaths from CHD. The relative risk in young smokers is greater than that in older stuokers. The rela- tive risk for young women smokers, especially those who usu oral contraceptives, is greater than 5. The excess relative risk associated with smoking decliues rapidly upon cessation of smoking, in some studies as much as 50% in 1 year. For exsmokers who previously smoked more than one pack per day, the residual excess risk also deElines, but never completely disappears. The decline in risk on cessation of smoking cannot be explauned by differences in known cardiac risk factors VE04844e 263 between Individuals who continue smoking and individuals who have qu-it. Smokers who have used only pipes or cigars did not appear to experience a substantially greater CHD risk than non- : smo ers. The rapid decline in risk associated with smoking cessation and the greater relative risk for sudden death suggest that active smoking can precipitate cardiac events in individuals with preex- isting coronary artery disease. Autopsy evidence of increased arte- rioscler-osis in smokers, coupled with the fact that risk of exsmok- - --- ors never returns to the levels found in nonsmokers, suggests that cigarette smoking is also implicated in the development of arte- riosclerotic cardiovascular disease (ASCVD) The mechanism by which cigarette smoke may lead to the development of chronic ASCVD, sudden death, or acute myocardial infarction is unknown. There appears, however, to be no threshold in the number of cigarettes smoked below which there is no increase in risk. -- - Data on uptake of cotinino by nonsmokers exposed to ETS indicate that the exposure in nonsmokers chronically exposed to ETS is approximately 1% that of an active smoker (who smokes one pack per day) (see Chapters 8 and 12). If the excess relative risk for CIID mortality or morbidity is a linear, nonthreshold function of dose and, further, if the excess risk of CHD in a one- pack-a-day smoker is twofold, then the relative risk from CHD in nonsmokers exposed to ETS (compared to true nonsmokers) would be approximately 1.02. Such relative risks would be difficult to detect or estimate reliably jn nonexperimental studies. Such small increases in relative risk are of the same order of magnitude as what might arise from expected residual confounding due to unmeasured covariates. Nonetheless, because of the large number of cardiovascular deaths each year, these possibilities deserve close attention and further study that could lead to firmer estimates of _- excess risk. Studies of Environmental Tobacco Smoke Exposure and . --- Mortality froni Cardiovascular Disease Garland et al. (1985) have reported that, in a prospective study of the effect of passive smoking, the age-adjusted rates of cardiac disease deaths in nonsmoking woonen whose husbands were former or current smokers were significantly elevated. It is

264 not certain, however, that the report is correct, because of a possi- ble.miscalculation or misuse of the Mantel-Haenasel statistic and some other methodologic problems. Data for the wives of former smokers were grouped with wives of current smokers. If this group, ing were made after examining the data, whirh indicated that tho risk was greater among the women whose husbnuds were former smokers, then this combination would be suspect. Tho p valuos based on the MantA-l-Haenazel test may be inappropriate in view of the small sample sizes. The authors employ the Cox Propor- tional Hazard analysis to control for other factors associated with cardiovascular risk, such as age, blood pressure, cholesterol, obe- sity, years of marriage, etc. '1'-hey report a relative risk for women married to current or former smokers compared with women mac . ried to never-smokers of 2.7 (Garland, 1985, corrected front an earlier report). The p value (< 0.10) associated with this esti-e mate is based on the asymptotic assumptions that are implicit in likelihood-based inference from the Cox model. These aesumh- - tiona may not hold for small sample.sizese ln summary, because of the small sample sizes, the significance calculations arising from • this study must be looked upon as approximations. Gillis et al. (1984) reported the results of a follow-up study of residents of two urban communities in'Scotland. Nonsmokers exposed to cigarette smoke in their homes had a slightly higher rate of myocardial infarction than those unexposed. The sample size was small, so that few of the results were statistically significant, and other risk factors for myocardial infarction were not controlled for. Hirayama (1984) rreported the results of a 15-year prospective study of nonsmoking Japanese women classified at start of follow- up by the smoking status of their husbands. A relative risk from ischemic heart disease of 1.3 was found for nonsmoking women - - whose husbands smoked niore than 19 cigarettes per day con~ pared with nonsrnoking women whose husbands did not smoke. A Mantel-Haenszel test for a linear trend was significant at the N< 0.01 level. It is unlikely that Hirayama's results can be explained by chance. The potential biases inherent in this study (see Chapter 12) limit the weight that can be placed on these results. The observed relative risk of 1.3 is at the upper limit of the expec- tations derived from extrapolations front active smokers, unless the uptake of the active component of cigarette smoke to which 2(f5 patisivo smokers are exposed is of the order of 10% of that of active aniokers. Mntsukura et al. (1984) have suggested that auch high levels of uptake in passive smokers may be seen in Japan. If there were independent evidence that nonsmokers exposed to other peo- plo's cigarette smoke do not differ on known risk factors for CHD -- --- - - - from unexposed nonsmokers, more reliance could be placed on Hirayama'e results. Svendsen et al. (1985) reported on the effect of cigarette smoke exposure to smoking wives among men participating in the Multi- ple Risk Factor Intervention Trial (MRFIT). MRFIT, which began in the mid-1970s, was a randomized primary prevention trial de- signed to test the effect of a multifactur intervention program on mortality from coronary heart disease in men with previous car- diac episodes. The men were chosen for participation if they had at least two of three risk factors for heart disease, including smoking, high cholesterol levels, or high blood pressure. The results re- ported by Svondsen et al. (1985), based on the group of men who never smoked but whose wives may or may not have been smokers, indicate no difference between exposed (i.e., smoking wives) and nouexposed (i.e., nonsmoking wives) of nonsmoking men for blood pressure or serum cholesterol. The MItFIT study demonstrates a roughly twofold increase in the risk of CHD mortality and morbid- ity among nonsmokers exposed to ETS. The sample size was small, and the results were not statistically signiRcant. Adjustment for other risk factors for CHD did not change the estimates of effect. SUMMARY AND ItECOMMEND.ATIONS What Is Known 1. No statistlcally significant effects of ETS exposure on heart rato or blood pressure were found in healthy men, women, and - school-aged children during resting conditions. During exercise there is no difference in the cardiovascular changes for men and women between conditions of exposure to ETS and control condi- - - tions. 2. With respect to chronic cardiovascular morbidity and mor- tality, although biologically plausible, there is no evidence of sta- tistically significant effects due to ETS exposure, apart from the study by Hirnyama in Japan. rIf,'0LRZ4f3

266 What Scientific Information In Missing 1. Experimental studies with animal models need to be per- formed with ETS to determine whether the cardiovascular changes - - seen following exposure to whole smoke also occur following expor sure to ETS. 2. Existing atudies have not provided evidence of serious harm in people with heart disease. With regard to angina Qns:+L, the findings are uncertain'and need to be repeated. REFERENCES Anderson, (1., and T. Dalhamn, Health risks dun to passive smoking (in Swedish). Li:karNdningen 70:2833-2838, 1973. Anderson, E.W., R.J. Andeiman, J.M. Strauch, N.J. Fortuin, and J.11. Knelson. Effect of low-level cstbon monoxide exposure on onset and duration of angina pectoris. Ann. Intern. Med. 79:48-60, 1973. Aronow, W.S. Effect of passive smoking on angina pectoris. N. I:ngl. J. Med- 299:21-24, 1978. Aronow, W.S. Aggravation of angina pectoris by two percent carboxyhe- moglobin. Am. lleart. J. 101:164-1b7, 1981. Aronow, W.S., and M.W. lsbell. Carbon monoxide effect on exercise-induced - angina pectoris. Ann. Intern. Med 79:392-396, 1973. Aronow, W.S., J. Cassidy, J.8. Vangrow, 11. March, J.C. Kern, J.R. ©old- smith, M. Khemka, J. Pagano and M. Vawter. Effect of cigarette smosking and breathing carbon monoxide on cardiovascular hemody- namics on anginal patlents..Circulation 50:3411-347, 1974. Aronow, W.S., R. Charter, and a. Seacat. Effect of 4% carboxyhemoglobia on human performance In cardiac patients. Prev.. Med. 8:682-588, 1079. Astrup, P., and K. Kjeldsen. Model studies linking carbon monoxide and/or nicotine to arteriosclerosis and cardiovascular disease. Pre.. Med. 8:296- 30Z, 1979. Bridge, D.P., and M. Corn. Contribution to the assessment of nennnokers to air pollution from cigarette and cigar smoke In occ_ npied spaca. Environ. Res. 6:19Z-209, 1972. Dahms, T.E. J.F'. Ilolin, and it.4. Slavin, Passive ,moking: Effects cN bronchial aathms. Chest 80:530-634, 1981. DeLucia, A.J., J.H. Whitaker, and L.R. Byrant. Effects of combined ex- posure to ozone and carbon monoxide (CQ) In humans, pp. 146-169. In S.D. Lee, M.d. Mustafa, and M.A. Mehlman, Eds. Advances in Modern Environmental Toxicology, Vol. V. International tiympoaium or the Biomedical Effects of Osone and Related Phatochemical Oxidu:ts. Princeton, New Jerseyt Princeton Scientific Publishers, 1983. Drinkwater, B.L., P.13. Raven, S.M.. Horvath, J.A. Gliner, R.O. Ruhling, asd N.W. Bolduan, and S. Taguchi. Air pollution, exercise and heat stress. Arch. Environ. Health 28:277-289, 1974. 267 Environmental Protection Agency. Revised Evahution of Health Effects - Associated with Carbon Monoxide Exposure: An Addendum to the - 1979 EPA Air Quality Criteria Document for t3arbon Monoxide. Publ. No. EPA-800/8-83-033F. Washington, ll.C.: U.S. Ciovernmunt Printing Office, 1084. FoI_iart, D., N.L. Benowita, and <).E, Decker. Passive absorption of nicotine In airline flight attendants. N. Engl. J. Med. 308:1106, 1982. Garland, C., E. Barrett-Connor, L. Snares, M. Oriqui, and D. Wingard. Effects of pas.ive smoking on ischemic heart disease mortality of aon- smokers. Am. J. Epidemiol. 121:046-860, 1986. Gillis, (_7.R., D.J. Hol., V.M. Hawthorne, and_ P. Boyle. The effect of environmental tobacco smoke In two urban cou:munitiea in the west of Scotland. Eur. J, Respir. Dis. 86(8133):121-1g0, 1984. Harke, H: P. Zum Problem des •Passiv-Rauchens.' MHnch Med Wochenschr. 51:3328-.5334, 1970. Harke, H: }?, and A. Bleichert. Zum Problem des Passivrauchens. Int. Arch. Arbeitsmed. 29:314-SZ2, 1972. Hirnyama, T. Lung cancer in Japan: Effects of nutrition and passive smoking, pp. 176-1U6. In M. Mise11 and P. Correa Eds. Lung Cancer: Causes and Prevention. New York: Verlag Chmuie, International, hic., 1984. Huch, IL, J. 1)anko, L. Spatling, and R. Huch. Risks the passive smoker -- runs. Lancet 3:1378, 1980. Hugod, O., L.11. Hawkins, and P. Astrup. Exposure of passive smokers to tobacE_o_ anoke wnstitnents, lnt. Arch. Occpp. Environ. Health 42:21-29, 1978. Hurshman, 1..(i., B.S. Brown, and 1t.O. C3uyton. The implications of sidestrsam cigarette smoke for r cardiovascular health. J. Environ. Health 41:146-149, 1978. Jarvis, M.J., M.A.H. Rusull, and 0. Feyernbend. Absorption of nicotine and carbon monoxide from passive smoking under natural conditions of exposure. Thorax 38:839-833, 1983. L.wther, P.J., and B.T. Oommins. Cigarette smoking and exposure to carbon monoxide. Ann. N.Y. Aced. Sci. 174:1a6-147, 1970. - Lightfoot, N.F. Chronic carbon monoxide exposure. Proc. R. Soc. Med. 85s798-799,1972. Lin, L.B., C.B. Taylor, S.K. Peng, and B. Mikkelson. Experimental srte- riosclerosis In Rhesus monkeys induced by multiple risk factors: Choles- terol, .itatnin D and nicotine. Arterial Wall 6:26-38, 1979. Luquette, A.J., O,W, Landess, and D.J. Merki. Some immediate effects of a smoking environment on children of elemsntuy school age. J. Scl:. Health 40:633-636, 1970. Mataukqra, S.,'1'. Taminato, N. Kitano, Y. Seino, II. ilamada, M. Uchihashi, lf. Nakajima, and Y. Hirata. Eifects of envirnameutal tobacco smoke on urina ry cotinine excretion In no:umuk.rs: Evidence for passive smoking. N. Eagl. J. Med. 311:898-831, 1984. McGill, H.f1. Jr. Potential mechsnisms for the augmentation of atheroscle- rosis and aetherosclerotic disease by cigarette smoking. Prev. Med. 8:390-403, 1979. Pimn:, P.E., F. Silverman, and R.J. Shephard. Physiological e(fTects of asute passive exposure to cigarette smoke. Arch. Enviroe. Health 33:201-313, 1978. `3EQLB449

268 Polak, E. Le papier A ctgarette. Son rQ1e dans la pnllution des lieux: habits. Tabagisme passif: Notion nouveiiee precise. Brux. Med. 67:336-.hA0, 1977. ,Raven, P.B., B.L. Drinkwater, R.O. Ruhling, N.W. Bolduaa, S. ltiguchl, J. Gliner, and S.M. Horvath. Effect of carbon nwnoxide and peroxyacetyl - - nitrate on man's maximal aerobic capacity. J. Appl. Physiol. 38.288-493, 1974a. Raven, P.B., B.L. Drinkwater, S.M. Horvath, R.O. Ituhling, J.A. aliner, J.C. Sutton, and N.W. Bolduan. Age, smoking habits, host stress, and - their interactive effects with carbon monoxide and peroxyacetylnitrate on man's aerobic power. Int. J. Brometeor. 18:222-232, 1974b. Rummel, R.M., M. Crawford, and P. Bruce. The physiological ellects +f. inhaling exhaled cigarette nnoke in relation to attitude of the nonsmoker. J. Seh. Heslth A6•6ZA-629, 1976. Russell, M.A.H., P.V. Cole, and E. Drown. Absorption by nonsmokers of carbon monoxide from room air polluted Ly tobacco smoke. Laucet 1:678-679,1973. Schievelbein, H., and F. Richter. The in6uence of passive smoking on the cardiovascular system. Prev. Med. 13:818-84A, 1984. Schievelbein, H., V. Londong, W. Londong, H. tfrumhach, V. Remplik, A. Schauer, and 11. Immich. Nicotine and arterioscherosis. An experimnntal contribution to the influeuce of nicotine on fat metabolism. '/.. _Iflin. Chem. Iflin. Biochem. 8:190-196, 1970. Sepp"anen, A., and A.J. Uusitalo. CaFbaxyhemoglobin saturation in r.lntiim to smoking and various occupational conditions. Ann. Clin. Res. 9:2B1- 288,1977. Shephard, R.J., R. Collins, and F. Silverman. "Passive" exposure of aeth- - matic subjects to_ cigarette smoke. Environ. Res. 20:392402, 1979a. Shephard, R.J., R. Collins, and F. Silverman. Responses of exen:ising subjects to acute 'panive' cigarette smoke exposure. Environ. Ras. 19:279-291, 1919b. Srch M. On the significance of carbon monoxide in cigarette smoking In an automobile. Dtsch. Z. f3esamte. Gerichtl. &f:80-89, 1967. Svendsen, K.H., L.II. Kuller, and J.D. Neaton. Elfects of pa.sive sme~king . in the Multiple Risk Fictor Intervention Trial (MRFIT). AHA Circ. Monogr. 114(Suppl.):I1I-63, 1986. (Abstract 210) U.S. Public Health Service. The Health Conmquences of Smoking: Car diovascular Disease. A Report of the Surl;eon General. DHIIS(1'IIS) -- - 1'ubl. No. 84-50204. Washington, D.C.: U.fl. Department of Health and Human Services, Public Ilealth Service, Qllice on Smoking and Health, 1983. 384 pp. Wald, N.J., M. 7die, J. Boreham, and A. Bsiley. Carbon monoxide in breath in relation to smoking and carboxyhemoglobin kvel.. Thorax 36:366-369, 1981. 15 Other Health Considerations in 'Cliildren I Several other health outcomes have beeu studied that relate to the growth and health of children. This chapter discusses studies of the influence of ETS exposure on birthweight of the offspring of nonsmoking pregnant women and its influence on childhood - growth and ear infections. For all postnatal outcomes, it is often not possible to differentiate effects of in utero exposure to tobacco siuoko constituent from subsequent childhood exposures to ETS. ENViItONMENTAb TOBACCO SM1)KB EXPOSURE BY NONSMOKING X`R.EGNAiYT WOMLN The fetus of a smoking mother is exposed in a unique way to the chemicals produced in cigarette smoke. Many studies have docuntented the adverse effect this relationsitip has on intrauter- ine fetal growth, especially during the third trimester of pregnancy (U.S. Department of Health and Human Services 1976). Mater- nal cigarette smoking apparently affects fotal oxygenatic,n, due to high levels of carboxyhemoglobin in the blood of both mother and child (Abel, 1980). However, the effectm on the fetus of a nonamok- jng mother chronically exposed to ETS are not well documented. Solne studies have indirectly approached this problem by evaluat- - ing paternal cigarette smokiug- and birth outcomes in nonsmoking pregnant women. Some early studies of paternal smoking and birthweight dein- - onxtrated a dose-response relatiunship that was discounted as 269

. . - - ....~._.. ~ar;,~.c. 270 "not easily acceptable as meaningful in terms of cause and of= fect" (Yerushalmy, 1962). An interview survey of 982 pregni.n- cies indicated a strong dose-response association between paternal ---- - cigarette smoking and the percent of infantH weighing less thau 5 pounds, 8 ounces (Yerushaimy, 1962). In a litter prospective study of nearly 13,000 births, Yerus-halmy (1971) reported that paterual smoking was more strongly associated with low birthweight than - was maternal smoking. The healthiest low-birthweight infauts were found for couples where the wife smoked and her husband did not; the highest mortality rate was found among infants pro- duced by couples where the husband smoked and the wife did not. These latter couples also had increased risks of producing preuia- ture o[fspring. The possibility that these differences in emoking were associated with differences in social class was not explorud. On the bases of these data, Yerushalmy (1971) inferred that pater- nal smoking may be incidental to birthweight. When the mothor's smoking was considered, the importance of paternal smoking 4lis- appeared. In a study of 12,192 births, MacMahon at al. (1966) con- firmed the negative association between inaternal smoking and birthweight of otfapring and also found thah infants of fathers who smoked weighed about 3 ounces less than those of fathers who did not smoke. They attributed this finding to the correlation between husbands' and wives' smoking habits, or to chance. MacMahon ' et al. (1966) referred to Yerushalmy's (1962) observation ef an association of father's smoking habits with infant weight as "biu- logically nonsensical.'- In a study of 176 normal neonates and 202 neonates with congenital malformations, Borlee et al. (1978) found that pater- nai smoking was independently and significantly associated with reduced birthweight and higher perinatal mortality. They spncu-d lated that the efli~ct occurred through its association with another factor. Gibel and Blumberg (1973) reported on a study of 6,0t10 children in which children of nonsmoking mothers whose fathers smoked more than 10 cigarettes per day had higher porinatul 1uor- tality than children whose parents were both nonsmokers. 'rhe incidence of severe malformations in children of fathers who were heavy smkers was double that of children of nonsmoking fathers, independent of parental age and social claye. BCO4B44B 271 Using code sheets prepared at birth of 48,505 women in world- wide naval installations, Underwood et al. (1967) found that fa- thers' smoking habits influenced pregnancy outcome, However, thie was attributed to the increased numbers of wives who smoked - when husbands smoked. For paternal smoking in the absence of maternal smoking, no association was found. Holmberg and Nur- nunen (1980) and Hughes et al. (1982) also reported no association of paternal emoking with low birthweight in cross-sectioual reviews of several thousand births. Rubin et al. (1986) provide a recent contribution to this sub- ject based on a survey of 500 consecutive births. About two-fifths of the women reported smoking durin re nanc 70 g P g_ _ Yi percent reported drinking. Paternal smoking was evaluated in terms of frequency and quantity of substance smoked, as reported in stan- dnrdized-intorviews. They found that birthweight was reduced an average of 120 g per pack of cigarettes smoked per day by the father. ' This relationship remained statistically significant after controlling for relevant variables, including mother's age, parity, maternal snioking, and alcohol and tobacco consumption during pregnancy. '1'he effect was greatest in the lower social classes. - In_ a prospective study, Martin and Bracken (1986) studied 3,891 antenatal patiente, 2,613 of whom did not smoke during pregnancy. One-third of the nonsmoking mothers (i.e., 906) were exposed to h;TS for at least 2 hours per day. ETS exposure was related to lower birthweight in full-term babies (23.5 g, not signif= icant). A logistic regression to control for geatational age, parity, ethnicity, and maternal age produced a significantly increased risk of delivering a low-birthweight baby, i.e., less than 2,500 g at birth for ETS-expused mothers (relative risk = 2.17, p s 0.06). The re- tardation in fetal growth rate is stnall but appears to be clinically meaningful at the low end of the birthwuight distribution. That is, exlroeure to LTS increases the risk that the infant will weigh less than 2,500 g and, therefore, will have a higher perinatal mortality. GROWTII IN CH1LDItkfN A few studies have exanuined possible relationships between chronic exponure to ETS by children and parameters of growth and development. Many studies have demonstrated that smoking during pregnitncy results in newborns who are lighter and shorter than other infante, even when gestational age has beeu taken into

272 account (Meredith, 1975; U.S. Departmeut of lioalth and Human Services, 1976). This deficit in height and weight appears to persist into infancy and childhood (Goldstein, 1971; Butler and Goldstoin, 1973; Dunn et al., 1976; Miller et at., 19711; Rantakallio, 1983). Current smoking status of the mother also has been associated with decreased attained height (Rona ot al., 1981; Berkey ct al., 1984), although growth rate was not slower among these children (Berkey et al., 1984). These studies, however, did not differenti . ate between smoking during pregnancy and subsequent exposures - during infancy and preschool years. Rona and colleagues (1985) reanalyzed data from the Na- tional Study of Health and Growth (England) for a sample of 5,903 children aged 5 to i1 years, separating the effects of smoking during pregnancy from those of later smoking. After adjusting the data for social class and other social factore, they found that re-d duced height was associated with increasing numbers of cigarettes smoked in the home, regardless of whether the mother smoked during pregnancy and regardless of which parent smoked. There remained a small but significant effect on height-a reduction of approximately 0.05 standard deviations of height (approximately 0.3 cm) for each 20 cigarettes consumed daily in the home. . To verify this small change in height, other studies of com• parable magnitude 'sre needed. Growth is an especially difficult phenomenon to study. Many factors, such as genetics, nutrition, social class, and ethnicity play importmit roles, and it is difficult to assign proportionate causality to each factor. Recall bias in the mothers of school-age ge children regarding their smoking habits during the pregnancy may produce unreliable results, especially in light of the increasing publicity regarding ill effects on the fe• tus of maternal smoking during pregnancy. Moreover, height and weight ratios and other growth measuren are not reliably obtained in standard pediatric surveys. CHRONIC EAR INhECTIDN:3 A number of studies have linked household exposure W ETS with increased rates of chronic ear infections and effusions in chil- dren. Chronic ear Infections or effusions in young children can lead to hearing loss and consequent speech pathology. Kraemer and col- leagues (1983) conducted a hospitaL-bnsotd case-control study of 76 children with persistent middle-ear ol[usions contrasted with 76 273 ` children aslmitted for other types of surgery who were matched for- age, sex, season, and surgical ward. They found that the daily ex- hosure to ETS was greater among cases. They also reported that middle-ear effusions clear loss readily in dhildren heavily exposed to ETS. They concluded that a combination of several factors increased the risk of persisient middle-ear effusions, including re- current otitis media, nasal catarrh, cigarette smoke exposure, and iiasal allergies that chronically inflame the nasal and middle-ear cavities, causing persistent eustachian tube dysfunction. For chil- dren with regular exposure to ETS, atopy, and congestion, the relative risk for PPME was 6.3 (95% confidence interval, 1.9-21.1). In another case-control study of 150 children hospitalized for chronic middle-ear effusions and 150 children hospitalized for other -- - reasons (Black, 1985), the odds ratio for parental smoking was - found to be significantly elevated (1.6). This effect was consis- tent across age grou - _ ps, and became more evident in older children where effusions are less common. Pukander et al. (1985) reported that ETS was.a significant risk_ factor for acute otitis niedia in 2- and 3-year-old children. They evaluated a number of important indoor environmental conditions, including relative humidity, car- bon dioxide, and temperature. In this study, children of smoking parents also had 60% mor_ e_ middle-ear effusions than children of nonsmoking parents. SUMMARY AND RECOMMENDATIONS S ,For all postnatal outcomes among children, it is difficult to dill'erentiate effects of in utero exposure to tobacco emoke con- stituents from subsequent childhood exposures to ETS. However, - for the abovo outcomes, there are indications that exposures to ETS may have effects on the fetus or child. What Is Known 1. Evidence has accumulated indicating that nonsmoking - pregnant women exposed to ETS on a daily basis for several hours are at increased risk for producing low-birthwcight ba- bies, through mechanisms which are, as yet, unknown. Recent studies show a dose-response relationship between the number of cigarettes smoked by the father and birthweight of the children of nonsmoking pregnant women. 6C049448

t Z74 2. A few studies have reported that children of smokers have reduced growth anci development. Those require further corrobo- ration to differentiate in utero exposure from subnequent childlto4 rd exposures. 3. Household exposure to ETS is linked with increased rates of chronic ear infections and middle-ear effusions in young cLiL- dren. For children with nasal' allergies and recurrent otitis media, ETS exposure may synergistically increase their risk of persistent middle-ear effusions. What Scientific Information Is Missing 1. Experimental studies should be developed to articulate possible mechanisms through which paternal smoking adversely effects fetal growth in nonsmoking pregnaut women. Special em- phasis should be placed on identifying relevant effects of preg- nancy on excretion and absorption of E'1'S, including transplacen- tal metabolism. 2. Additional study is needed to corroborate one finding of a dose-response relationship between reduced height of children and increasing numbers of cigarettes smoked in the home, regardless of whether the mother smoked during pregnancy and regardless of which parent smoked. 3. Research should be conducted to explore the mechanisms by which exposure to ETS might adversely affect the functioning of the ear and to study possible long-tecm consequences of ETS exposure for the auditory apparatus. REFERENCES Abel, L.L. Smoking during pregnancy: A review of effects on growth and development of offspring. Hum. Biol. 62:693-026, 1980. Berkey O.S., J.H. Wars, F.E. Speiser, and B.Q. Ferris, Jr. Passive smoking and height growth of preadolescent children. Int. J. Epidemiol, 13s46f• 468, 1984. Black, N. The aetiology of glue ear-A cue-control study. Int. J. Pediatr. - Otorhinolaryngol. 9:121-133, 1986. Borlee, I., A. Bouckaert, M.F. Lschat, and C.B. Mission. Smoking patteras during and before pregnancy: Weight, length and head gireumferean of progeny. rur. J. Obstet. Qynecol Reprod. Biol. 8:171-177, 197N. Butler, N.R., and H. Goldstein. Smoking In prRgnancy and subsequent cklW development. Br. Med. J. 4:673-676, 1973. 275 Dunn, H.a., A.K. MeBurney, S. Ingram, and O.M. Hunter. Maternal cigarette smoking during pregnancy and child's.ubsequent development. l. Physical growth to the age of 6 1/2 years. Can J. Public Health 67:499,06, 1976. dibei, W., and H.-H. Blumberg. Die Aaswirkunpn der Rauchgewohnheiten von El.tern anf das angeborene und neugeborene Kind. Z. Aestl. Fort- bild. 73:341-342, 1973. ' floldstein, H. Factors in6ueacing the height of seven year old children- Results from the National Child 1)evelopment Study. Hum. Biol. 43:9Y= 111,1971. Bughes, J.R., L.H. Epstein, F.. Andrasik, D.F. NeIF, and D.S. Thompson. Smoking~ and carbon monoxide levels during pregnancy. Addict. Behav. 7s271-276,1983. Hol_m_ berg, P.C., and M. Nurminen. Congenial defects of the central nervous system and occupationai factors during pregnancy. A case-referrent study. Ain. J. Iad. Med. 1:167-170, 1980. Kraemer, M.J., M.A. Richardson, N.S. Weiss, O.T. Furukawa, f1.(3. Shapiro, - W.E. Pierson, and W. Bi.rman. Risk factors for persistent middle-ear effusions. Otitls media, catarrh, cigarette smoke exposure and atopy. JAMA 2d0:104k10Z6, 1983. MacMshon, B., M. Alpert, and E.J. Salber. Infant weight and parental smoking habits. Am. J. Epldemiol. 82,947-401, 1966. Martin, T.R., and M,D. Bracken. Association of low birth weight with passive smoke exposure In pregnancy. Am. J. Epidemiol. 124:833-04Y,1980. Meredith, H.V. Relation between tobacco smoking of pregnant women and body sise of their progeny: A compilation of published studies. Hum. - Biol. 47:461-473, 1975. Millet, ILC., lf. Hassaneia, and P.A. Henaleigh. Fetal growth retardation In relation to maternal smoking and weight gain In pregnancy. Am J. Obstet. Clynecol. 146:66-e0, 1976. Pukasder, J., J. Lu.tonen, M. Timore, and P. Karma. Risk factors affecting the occurrence of acute otitis media am_ong two and three year old urban children. Acta Otolaryngol. 100:260-486, 1986. Rantakaillo, antakallio, P. A follow-up study up to the age of 14 of children whose mothers saoke_d_ during pregnancy. Acta Paediatr. Scand. 79e747-763, 1983. Ronb R•J., 0. Du Ve Florey, t3.0, Clarke, and S. China. Parental smoking at home aud height of children. Br. Med. J. 383:1363, 1981. Rona, R.J., 8. Chinn, and O. Du Ve Flor.y. Exposure to cigarette smoking and children's growth. lnt. J. Cpidemloi. 14:402-409, 1985. Rubin, D.H., 1'.A. Krasilnikoff, J.M. Leventhol, B. Welle, and A. Berget. Effect of passive smokin8 on birth-weight. Lancet 2:416-l17, 1986. Underwood, P-.U., K.F. Kesler, J.M. O'Lane, and D.A. Cellagan. Parental smoking nupirically related to pregnancy outc-oiae. Ob.tet. Gynecol. 29:1-8, 1967. U.S. Department of Health and Human Services. The Health Consequences of Smoking. Selected Chapters from the 1971-1976 Report. Report of the Surgeon General Pubi. No. CDC 78-8367. Washington, D.C.: U.S. i)epartment of Health, Education, and Welfare, 1'ub16c Health Service, OIBce on Smoking and Health, 1970. 667 pp. OVU449449

276 Yeruehalmy, J. Stati.tic.l coneideratione and evtln.llon of epldemiological evidence. In a. J.mee, 110d. 'lbbacco and liealth. SpringHeid, Uiinoia Charles 0 Thosase, 1962. Yeruehalmy, J. The reletionehip of parents' Eigarette Nmoking to outcome of preanancy-Implicntlone as to-the problem of Inferring caueation fronn _ -- observed se.ociatione. Am. J. Epidemiol. 93:443 456, 1971. APPENDIXES TVO4g449

Appendix A; Guidelines for Public a,nd Occupational Ciieinicial El xposurgs to Materials That Are Also Founcl in Eilvir--onmenta,l `Pobaccu Smoke Table A-I gives a series of guidelines for public and industrial populations -regarding exposure to chenucals that are also con- - etituents in environmental tobacco smoke (ETS). Not all of the constituents of ETS thought to be toxic or carcinogenic have had guideline levels established. The values in the table are taken from the fourth edition of the Documentation of the 'TAfeshotd Limit Yalucs, published by the American Conference of Governmental and Industrial Hygienists (1986). The NIOSH recommendations and OSI[A standards can be found in the NIOSH Pocket Guide to Chemicat flasarda, published by the U.S. Department of Health and Hunian Services (National Institute for Occupational Safety - - and Health, 1981). - l1','ach of these guidelines and standards has been established with different considerations in miuid. The EPA standards, which apply to outdoor environments, have been established by law to _ - protect the m(sst' susceptible individuals. The OSHA standards - - - and AGGIH, NIOSH, and European guidelines have been estab- - lished for the normal, healthy adult working populations. These guidelines accept some level of risk to some people. They do not consider children, the elderly, or populations with preexisting health conditions who may be at greater risk fc)r health effects of exposure. The appropriate guidelines for susceptible populations -- - probably would be lower. These industrial guidelines also differ from the environmental standards in that they assume that the -- - exposure is limited to a workday period or a 6ime-limited emer- gency. zVoLe444e I 279

zAa1.E A-1 Some Occupational and Pubfic 'Standards forMaterials That Are Also in' Environmental Tobacco Smoke iCp IRL" Public Industrial 0 ` - d d ~ 'S' j" EPA AC6'IR' OS1HA N1OS1#1' at tan s Euam,pean GO ~ Vapor Fhase m s SO' W ppm West Germany--50' ~ Carbon mono:iHe 1 mgAm3-max. g-h 40 mg/m3-max. 1=h TI:V°-5b ppm STBLf-400 pPm pp A 35' ppm--8 h T 200 ppm vee , Sweden-3Sippn' Neither to beexoeeded mioee ithan owwe (no min'time) d xide b 1 pu Sau Noae TL'V-5,000 ppm 10.000 ppm-ll0-h TWA 5,000, ppm on io Car genzene None sSQ,-30.000 ppm TLA'-10, ppm A2 30= ; ppm-10-min ceil. '1,pPm-60-min ceYl. 10 ppm 50'PPm- S.Kden-l0 ppm Wat Gamanr-O ppm Toluene None TLV-100 ppm STEL-+150 ppm 10-min ceil. 100 ppm-10-h' TWA 200 ppm 200 ppm-10•min oeil. 300Ippm aeil. '9vest caermanq- 200 PPm 500 Fpm- Sweden-100 ppm Formaidehyde \osc TLV-1 ppm A2 . 110-min peak '4o.rest feasi'bk Cmsit 3PPm 5 ppm rzi:. Sweden-2' ppm West Germam-1,ppm crolein yrkfi- H*'SrPmen c.anitle oae one None PL'V-0.1 ppm S'1'EL-0:3' ppm 4'I,V-750ippm STII.-'1.0D0, ppm '1T.V-3'FPm STE.-l0,ppm Ca,ing luair- I0 ppm 10 ppm- 30.min aail. 0:1;ppm 250 ppm-10-h TWA 100 ppm None 5 ppm +4:7 ,ppm-10-mut eeil. 4.7 ppm weden-SDO ppm C,ermanY-It'.000 ppm West Germany. Sweden-5 ppm llYat' L;,etmaq7.'6reat B:itain-101 ppm Hydtazine iNone TLV-u ppm 0.04 mg/m3- I ppm A2 120-min ceii Ammonia 'None T1,V-25 ppm . 50',ppm-,S-min ccii. 50 ppm Weat'Germany-50 ppm Meth i mi STEi.-35' ppm S.rsden-25 ppm y a ne None TLV-10ppm None 10 ppm Dimethylamine None '1LV-10 ppm None 10 ppm Nitrogea o:i8e Mione TLV-25 ppm 25 'Rpm 25 ppm- ~ 10=h TWA N uaogen'tliozitle i0:O153 ppm-annual 4'LV-:3 ppm 1 PPm-15 min, 5,ppm ceil. West Garmarny-5 m N;Mttto.o c arithmetic mean None srl:i.-s', ppm A2 None IsateS as a pp 5weden-2,ppm &Mdhyl.mine can"°er'wspeat Formic acid Wooe #geat 917.:V-5 ppm None Sppm Acetic acid None 4I.V-'10', ppm lione . 10 pptn S3'EL-15 ppm , Particulate;phan Particulate matter , 95 pg/m3-amsnal 1f1.V-~10i eqg/m° None 15 mg/m3 gaamauic maa 260 µgiaa'I2+4-h max Nat toi br mxceadetl mote than onae ~ year Nieatiae ~None TGV-+0.5, mg/ms None 0.5 m /m~ Pheaot None TLV-19 mg/m3 20 mg/m3-10-h TWA g 19 mg/m) West'.Germany- 60 ntglrn'-~ 1'S-min cell. 119 /ms Catechol None Ti.'V-5 BPm None None mg Hrtlroquinone None TLV-2 mgxm3 2 mg/m3-15-min ceil 2 mg/ros Aniline None TL'V-2,ppm . None S ppm 2-Tdluidime None TL'V-2,ppm None S ppm West'GermanY-S ppm A2

282 283 The guidelines are given in terms of cumulative exposure over a period of time or in terine of maximal concentrations. The - - Threshold Limit Value (TLV) is the time-weighted average con- contration of a normal 8-hour workday or 40-hour work week. The Short-'Lbrm Exposure Limit (STI~.L) is defined as a 15-minute time-weighted average exposuro that should not be exceeded at any time during a workday, even if the 8-hour time-weighted av- erage is withln the TLV Exposures at the STEL should not be repeated more than four tintes per day, with at least 60 minutes --- - between sui cessive exposures at the S'fEL.'.L'he ceiling limit is the concentration that should never be exceeded. -- -- -- - - - Finally, it should be noted that the guidelines are established fi,r individual chemicals, without consideration of complex mix- - - --- --- - - -- -- tures that may contain these chemicals. The behavior of the chemicals in a complex mixture over time is likely to be com-. plicated.. In summary, the direct comparisons of these guidelines with ambient levels measured in natural or experimental condi- tions shoukl be made with cautton. In some cases, the comparison ------ -------- may be inappropriate. REFEIt.ENCES A t I A E a A:neriean Conkrence of Governmental Industrial Hygienists (AGE31H). Doc- u:nentation et the Threshold Limit Values and Biological Exposure Indices, litth ed Cincinnati, Ohio: ACfi1H, 1986. 743 pp. National Institute for Occupational Safety and Health (1+IIOS11). NIOSH/ OSHA Pocket (3nide to Chemical Hasards. DHEW Publ. No. 85- 14. Cincinnati, Ohio: National Institute for Occupational Safety and Health, 1986. 241 pp. Swedish Board of Occupational Safety and Heaith (Arbetarksyddastyrelsens). HYBieniska ©ransvarden. Stockholm, Sweden: Liber Distribution, 1984. 60 pp. VV04g4LS

Appendix B: Method of Combining Data froni Studies of Environniental Tobacco Smoke Exposurc; and Lung Cancer Consider the following kinda of data tha6 might be reported in an epidemiological etudy of chronic exposure to environmental tobacco smoke (ETS) and lung cancer: Lung Cancer ii Yes Exposure Yes u b to ETS No c d Total; Total Therefore, T is the total number of people in the study, a is the number of people chronically exposed to ETS who also have lung cancer, 6 is the number of people chronically exposed to ETS who do not have lung cancer, e ia the number of peohle not chronically exposed to ETS who have lung cancer, and d is the number of people not chronically exposed to ETS who do not have lung cancer. The marginal totals are ml = a-h 6, m2 = e-I• d, m3 = a + E, and m4 = 6-{• d`1'he data that correspond to these variables from_ all of the studies examined in Chapter 12 are ehown in 'rable CASE-CONTROL 9TUDIE3 In a case-control design, the subjects are chosen on the bisaia of the health outcome, and their exposure history is assessed. 9 "4 285 0 ~ Y ~ ~ Ii8a ~ I ao.~.~...+~.~.. ...~:,~.,..i ~ ~ , ~ ~ JS ~ ~ Np O. N V 3 A ~ v N A h O 1~ v! .~w ^ M/'1 +~ ~ pp ~ 1~ N-.~r iV N~.i V~p~ ryfV - f Vl ~A a6 N :7f• N p S _ H www~w~w~www~ww~ww~w~ ~u~ u u uu~' u~u aa a $ 284 aV04aRUR

286 The expected number of people who are exposed to environ- mental tobacco smoke and develop lung cancer is given by: m' X M' T _ Expected numbers for each of the studies are shown iu Tuble 12-4. The difference between observed and expected numbers of people with lung cancer who are exposed to E`t'$ c-an be calculated, and the variance of this difference is given by: rtsiXr»sXmaXm4 _T X T x (7_. 1) . Therefore, the natural logarithm of the odds ratio (0) can be estimated by: _ Observed - Expected ~ Varinnce(Ob.erved - l;xpec"-) and_ the variance of this estimate is given by: Variance of # = ((Variance(Observed - Expec-ted)]'i (Yusuf et at., 1985). The odds ratio is estimated by exp(+y] attd is ehown in Tables 12-4 (and B-1) with its 95% confidence intervals for each of tbe studies. P13.OSPECTIVE (OR eOHORT) S'1'UDIES In prospective studies, also known as coltort studies, the sub-s jects are classified (or chosen) on the basis of exposure and tlie health endpoint is then asseeaed. In all of the articles the authors have estiutated the relative risk, adjusting for nuch variables as age. Therefore, the published relative risk values were used in the following calculations rather than the estimates of the crude relative risk that could be calr.u- lated from the data given in the text. For those studies where a relativo risk estimate was given for different levels of smokutg by the spouse (Garhukel et al, 1981; Hirayama, 1984), a combined estimate of the relative risk was calculated using the method given below for combining the prospective studies. The number ~wf people who are exposed to ETS who are ox: pected, under tlua null hypothesis of no effect, to develop lung cancer is: m3 - (m3/E) X r, 287 where E is the expected number for ma, bam:d on the published relative risk (RR), that is: E - c =k (uIRR). The approximate variance of the observed minus expected nuntbers of people with lung cancer who are exposed to environ- meutal tobacco smoke is: t .1 xTX (r-1) ml xm2 xrn3 xm4 The variance of the natural logarithm of the relative risk was - -- - calculated using the published confidence limits for the estimate of tlie relative risk, except for one study (Gillis et al., 1984), where the method given above for the case-control studies use used since - no confidence limits were available. SUI4SMING OVER STUDIES The overall values for the case-control studies were calculated by adding the values of Observed - Expected (i.e., U- E) and their variance for the individual studies as follows: ln oR = no -1i")j rVac(0 - E)t and for the variance: Variance(ln OR) = E 1 Vsr(U --~ J:'~ (Yuduf et al., 1985). For the prospective studies, tlte overall value for the In RR. was calculated as: Lr RR ' E 1% (ln RR)r / ~ Var(Lt RR)_ I r Vu(In RR)i and for the variancec Var(ln RR) _ - 1 ~; Var(lu RR)i (Kicinbaum et al., 1982). --- - --- 'Phe overall value, for all of the studies combined, was obtained 9V01.8448

288 using the same method as was used to pool results from the prospective studies using the overall values for the case-control and prospective studies in the above equations. R.EFERENCES Akiba, S., W.J. Blot, aud H. Kato. Passive smoking and lung cancer among Japanese women. Fourth World Conference on Lung Cancer, Toronto, Canada, Aug. 26-30, 1986. Buffler, P.A., L.W. Pickle, T.J. tv/ason, and C. Contant. The eauses of lung cancer in Texas, pp. 83-99. In M. Miseli and P. Correa, 14ds. - Lung Cancer: Causes and Prevention. New York: Yerlag Chemio , International, Inc., 1984. • Chan, W.C., and 8.0. Fung. Lung cancer In nonsmokers in ilong Kong, - pp. 199-202. In E. Grundmann, Ed. Cancer Campaign, Vol. 6. Cancer Epidemiology. Stuttgart; Gustav Fischer Verlag, 1982. Corea, P., L.W. Pickle, E. Fontham, Y. Lin, and W. Hae_nssel. Pa.sive smoking and lung cancer. La_ neet g:698-697, 1983. Garfinkel, L: Time trends in lung cancer mortality emong nonsmokers and a_ note on passive smoking. J. Nati. Cancer ln.t. 66:1081-1089, 1981. - Garfinkel, L., 0. Auerback, and L. Joubert. Involuntary smoking and lung csncert A case-coutrol study. J. Natl. Cancer Inst. 76:409-489, 1986. G_ illis, C.R., D.J. Hole, V.M. Hawthorne, and P. Boyle. The effect of environmental tobacco aece smoke In two urban communities in the west of Scotland. Bur. J. Respir. Dis. 133(Suppl.):121-126, 1984. Hirayams, T. Cancer mortality In nonsmoking women with smoking husbands on a large-scale cohort study in Japan. Prav. Med. 1&880-890, 1984. - Kabat, o.C., and E.L. Wynder. Lung cancer in nonsmokers. Cancer 61:1i14- -- __ 12Z1, 1984. Kleinbaum, D.G., Kupper, L.L., and H. Morgen.tern. Bpidekuiologie Re- search: Methods and Application. New Yorks Nordstrnm Reinhold, 1982. 341 pp. Koo, L.C., J.H-C. Ho, and N. Laa. An analysis of some risk factors for Inng cancer In Hong Kong. Int. J. Cancer 36:149-166, 1985. Lee, P.N., J. Chamberlin, and -M.R. Alderson. Relationship of passive smoking to risk of lung cancer and other smuking-associated disease. - - - - Br. J. Cancer. 64:97-106, 1986. -- Miettenen, U.B. Estintability and estimation In cR.e-referrent studies. Am. J. Kpidemiol. 103:226-235, 1976. Pershagen, G., Z. Hrubec, and 0. 3vensson. Paaive smoking and lung cancer in Swedish women. Am. J. Epidemiol., in prea. Trickopoulous, P., A. Kaiaudidi, and L. Sparros. Lung cancer and passive ' smoking: t3onclusion of Greek study. Lancet l:877-e78, 1983. - Wu, A.11., H.E. Hender.on; M.O. Pike, aud M.E7. Yu. Smoking and other -_ risk faetors for lung caacer in women. J. Natl. Cancer Inst 74:747-76t, . 1985. Yusuf, S. R. Peto, J. I.ewis, R. Collins and P. Sleight. llet& blockade daring and after myocardial infarction: An overview of the randominised trisls. J. Prog. (3ardiovuac. Uis. 37t836-371, 1986. , ~+~'~~8~~ 1I8 Appendix C: Atl justments to E piciemiologic Estimates of Ekcess Lung Cancer in P,ersoiis Exposed to Eiivironmental'1'obia.eeo Smoke Chapter 12 describes 13 epldamlologic studies that estimate ------ -- - the relative riHks of lung cancer in nonslnoking spouses of smokers compared with nonsmoking spouses of nonsmokers. A weighted - average of the relative risks of "exposed" to eunexposod" persons - - - - is 1.34, i.e., a g4% increase in lung cancer risk as a consequence of - -- - environmental tobacco smoke (LTS) exposure. On the other hand, - one can extrapolate in a linear fashion from the relative levels of -- cotinine that had been measured in active sntokers and exposed --- nonsmokers. The expected relative risk for exposed nonsmokers would range from about 1.03 to 1.10. Neither of these estimates has been corrected for possible misclassificatiott of subjects in the epidemiologic studies. The latter risk assumes that the one-time measure is a satisfactory surrogate for lifetime exposure. Misclas- sification problems and problems in estimating actual carcinogen exposure make it very difficult to provide an estimate of the num- bers of cancer cases both in smokers and nonsmokers that might be attributable to ETS. In this section we combine information from several sources to generate crude estimates of the relative risk to nonsmokers as a consequence of chronic exposure to ETS. The computations repurted hero are highly simplificd antl should be looked on as providing only a first approach to risk evaluatiun.'A more detailed approach, including a more explicit statement of the assumptions involved, is given in Appendix 1). A major concern is that persons who have beeu identified as "unexposed" to EY'S may have really been exposed. If this were true, then the risks relative to truly unexposed pi+rsons would be underestimated. To estimate this 289

290 possible effect, the results from studies of urinary Eotinino are used here to adjust for the proportion of sel[ reported "unexposotl" nonsmokers who, in fact, may have been exposed to E'1`S. USING CO`1'INiNE MEASU)t,EMEN3'S TO CORRECT MI3ItEI'O LtTllVG The only source ofcotiniue or nicotine in body fluids is tobacco smoke exposure. '1'herefore, urinary cotinino provides an objective measure of (recent) exposure. It has been reported (Jarvis et ul., 1984; Wald et al., 1984) that urinary nicotine and cotinine are :1 times as high in "axposed» nonsmoking spouses of current smokern than in "unexposed" nonsmoking spouses of current nonsmokers. For example, Wald and Ritchie (1984) report urinary cotinina in the ratio 1:3:215 for "unexposed" nonstnokers, E'i'S-expoued nonsmokers, and regular smokers, respectively. - - Several assumptions need to be made to permit the use of these data before any quantitative risk comiwtatiun can be made: • Current stitoking patterns reflect payt patterns. • Cotinine or nicotine concentrations ln the urine are linearly related to recent exposures to ETS and to the carcinogens iu 14TS among nonsmokers. e All subjects in the various studies liegan to be exposed to ETS at the same age and have continued to be exposed at the same rate throughout the follow-up period, e The excess relative risk for lung cancer in nonsmokers is proportional to the dose (in cigarette equivalents) of ETS alr sorbed. An assumption of a linear dose-Eespouse relationship imliGm that if the risk (i.e., mortality rate) at a given age t) for a specific -- - calendar period (s), given some absorbed dose (d), then ry(t,s~d) is: .1(c, a1d) ='yo(t, a)(1 +/1d). (1) This equation expresses the risk as equal to tho base mortidily risk, rye(t,s), for a truly unexposed person for the same age and calendar period, multiplied by an excess relative risk that incroases linearly with doeut, i.e. (i -1- Pd), where d in tho amount of increai+e 291 per unit dose * Further, the risk for a truly unexposed nonsmoker, i.e., rye(t,s), is assumed to be the sarne for men and women. This assutnption is supported, in part, by the results given in Chapter 12 and, irt hart,'by earlier studies of Garflnkel (1981) and Friedman et al. (1984). Doll (1984), however, gives di(ferent risks for men and women of lung cancer mortality in nonsmokers. -- lf dB is actual dose in the "exposed" persons and dN is the actual doso in persons who believe themselves to be °u nexposed," then we have, from Equation 1: q(f. J1dE) ='Yu(t, a)(1 + Qds). (2A) and 'y(t. sIdN) - /o(t,s)(1+QdN)r (2B) The relative risk for a person identified as "exposed" compared to a person identified as "unexposed" [1tR(d&)+ is given by Equation - 2 A divided by 2B: RR(da) = 1 + QdB 1+QdN' (3) which, front Chapter 12, is 1.34, the relative risk estimated frotn the epidemiologic studies. From the atudies that measured cotinine in "exposed" and °unexposed" persone, we assurne that the operative dose level, da, among "exposed" individuals is 3 times an high as the dose level in the self reported "unexposed" persons, dN, and that the ratio of 3:1 is proportional to a lifetime dose difference. Therefore, Equation 3 may be rewritton as: i -I• 39dN ~ RR(d&) = 1+PdN- (4) Equation 4 can be solved for PdN, which is the increase in risk for persons called ".unexposed " but who, in fact, have been exposed * Work by Doll and Peto (1978) shows that the relative risk for direct aniokers inerea.es as a linear-~iuadratic functioa of dose, rather than the Lnple linear form shown here. A snore sophi.tir.at.d model would take iato account the several stages at which cigarette smoke operates in the multistage development of cancer. At low dosee the linear-quadratic is wai approximated by the linear, i.e., i+/ltd +#2 d2 is close to 1+16d because the d2 term approaches sero. gVO4Q449

292 to some recent ETS, as indicated by their non-sero levels o€urinary eotinine or nicotine. Solving Equation 4 gives: I PdN = 0.20. Thus, the relative risk for a self identified "unexposed" person compared with a truly unexposed person is: 1-h 0.20 = 1.20, and the relative risk for an "exposed" person compared with a truly unexposed person is: 1 + 3(0.20) = 1.60. To see what possible effect these relative risk estiniates would have on the population-attributable risk, i.e., the fraction of lung cancer in nonsmoking individuals attributable to EZ'S, the pro- portion of the population that is exposed to E`l'u needs lia be estimated. Wald and colleagues (1984) have reported that 17% of nonsmoking women and 12% of 'aonsmoking men fall into the cat-y egory of "exposed," i.e., nonsmoking spouses of smokers. By sub- traction, this means that 83% of nonsmoking women and 88% of nonsmoking men would consider themselves "unexposed' Given - this, we can estimate the population-attributable risk, which is given in general form as: pt(RR1- 1) =t (1-pt)(nR2 -1) --- p, (RRl) + (1- p~(RR2) PAR, (8) where pl is the proportion of people who call themselves "e*- posed," .tfRi is the relative risk of self-reported "exposed" persone, and RR2 is the relative risk of self-reported "unexposed" perstins. Thus, for men: PAltd1ea = 0.12(0.80)-1- 0.88L0__lo) _ 0.20. 0.1211.80) + 0.88(1.101 and for women: 0.19(0.60) + 0.83(t1.20) PAR.,oman ° a 0 0.17(1.60) + 0.83(-1.-f0-) 0.21. 'rhat is, about 21% of the lung cancers in nonntnukin,q womon and 20% in nonsmoking nien niay be attributable to expnsure tn k"P`s. 293 REh'EltEN(:ES • Doll, R.D. Fhidemiological discovery of occupational cancers. Ann. Acad. Med. Singapore 13($uppl.):331-339, 1984. Doll, ILD., and R. Peto. Cigarette smoking and bronchial chial carcinogenic: Dose utd time relationships .moRg regular amokers aad lifelong n_ g nonsmokers. J. Lpidemiol. Comm. Heaith-3S:30a-S13, 1978. - - Friedman, G.D., R.D. Bawol, and D.11. Pettiti. Prevalence and correlates of passive smoking. Am. J. Public Health 73:401-406, 1983. • Garfinkel, L.'rime treAds In lung cancer mortality among nonsmokers and a note on passive smoking. J. Natl. Cancer lnst. E8;1081-106(3, 1981. J_ a_ rvis, M., 11. 1Lnstsll-Ped©e, 0. Feyersbend, O, Vesey, and Y. Salloyee. Biochemical markers of smoke absorption and self reported exposure to passive smoking. J. Epidemiol. Comm. Health 38:336-339, 1984. Wald, N.J., A. Bereham, A. Hailey, 0. Ritche, J.E. Haddow, and 0. Knight. Urinary cotinlne as a marker of breathing other people's mioke. Lancet - 1:230-231,1984. Wald, N.J., and 0. Ritchie. Validation of studies on lung cancer in non- smokers n- smoker. married to smokers. Lancet 1:1087, 1984. GV048448

295 Appendix D: Risk Assessment-Exposure to Environmental Tobacco Smoke - and Lung Cancer James Robins This authored appendix was prepared by Dr. James Robins of the Ilarvard University School of Public Health. The material was not considered by the couunittee largely because of lack of time, nor was it reviewed by the Natioal Research Council. It gives an approach to risk assessment that considers both the epidemiologic data and some measures of exposure to the constituents of ETS. It is included as an addendum of this report and is presented here as one possible way to integrate the data contained in the remainder of the report. INTRODUCTION In Chapter 12, the results of 13 epidwniologic studies are Au-ecr mrised. Each study provided an estimate of the ratio of the lung cancer mortality rate among nonsmokers who answered "yen-' to a question like "Is your spouse a smoker7" (hereafter called "ex- poaed' individuals) to the mortality rate among nonemokern who answered "no" to that question (hereafter called "unexpoeed" indi- viduals). A weighted average of the 13 study-specific rate ration is roughly 1.3. In this appendix, we assume that a weighted average of 1.3 is causally related to differences in environmental tolow'.co smoke (ETS) exposure between "exposed" and "unexpoeed" in- dividuals and not to bias (e.g., ., misclasNification of smokers as nonsmokers-see Chapter 12). Wald and Ritchie (1084) have shown that "unexposed" indi- -- viduals have, on average, 8.6 ug/ml of cotinine in their.urine. Since - - virtually the only source of cotinine or nicotine in body fluids is tobacco products, primarily through tobacco smoke exposures, it follows that "unexposed" •individuals are exposed to ETS. For this reason, whenever we refer to such "unexposed" subjects, we place the word "unexposed" in quotation marks. If the "unexposed" subjects have, in -fact, been exposed to E`-l!S, the observed relative risk of 1.3 would be an underestimate of the true adverse effect of --- ETS on "exposed" individuals.. The correct measure of the adverse effect of ETS on "exposed" individuals would be the ratio of the - - - lung cancer mortality rate in "exposed" individuals to the rate in truly unoxposed individuals (which we shall call the true relative risk in the "exposed"). In Section D-1, we use the data collected by Wald and Ritchie (1984) on levels of urinary cotinine in "exposed" and "unexposed" individuals to estimate this true relative risk by two different methods. In Section D-2, we combine the existing epidenuologic data on act'tve smokers with data on nonamokers exposed to ETS to estimate the ETS exposure of an average nonsmoker in cigarette- equivalents per day. Additionally, we compare this estimate to independont estimates of ETS exposure based on (1) levels of resptrable suspended particulates (ItSP), benzo(a)pyrene (BaP), and N._nitFosodimethylamiue (N@MA) in ETS and in mainstream smoke and (2) levels of urinary cotinine and nicotine in active smokers and nonsmokers. In Section D-3, we compute how many of the lung cancer deaths estimated to occur among (lifelong) nonsrnoking persons -- ~ in 1088 might be attributable to ETS. The estimate is made sep- arately for women and for men. Many environmental exposures are regulated to a level where the anticipated lifetime risk of death attributable to exposure is - less than 1 in 100,000 or I in 1,000,000. In Section D-4, we consider whether the lifetime risk of death (from lung cancer) attributable to ETS among nonsmokers with moderate ETS exposure is in excess of I in 100,000. (Although we do not estimate the lifetime risk of death attributable to 1:TS from causes other than_ lung cancer, this does not imply that we believe that lung cancer is the only cause of mortality influenced by ETS exposure. The decision - - to restrict the analysis to lung cancer mortality reflects the fact 09049449 294

296 that the data necessary to perform an adequnte quantitative risk assessment for eausen of death other than lung r,ancor do not exist.) - In discussions of the health effects of LTS exposure, onu should consider the effect on exsmokere of lireathing other peo= - ---- -- ple's cigarette smoke, since exemokers have given up smoking, presumably to protect their health. Therefore, in Suction D-4 wo - -- - - estimate, for exsmokers, the lifetime risk of death frotn lung cancer attributable to breathing other people's cigarette smoko. - The sections D-1 to D-4 give nontechnical expositions pf the id- ---- - - sues. A separate Technical Discussion Section provides additional. - technical support and matheinatical background. In order to make quantitative estimates of the lung caitcnr risk attributable to ETS, numerical values mmst bo chosen for a large number of parameters. When there are either no data or inconsistent data as to the magnitude of an important parameter, results are reported for a range of plausible val uee (i.e., a sensitivity analysis is performed). Summary of Main Results Under the Assumption That the Sumrnary Rate Ratio of 1.3 Ia Causal We summarize our main results. We caution the reader that _ the proper interpretation of these results requires that one read Section D-.1 to D-4 and the discussion section that follows. The estimated true relative risk for "exposed" individuals liie between 1.41 and 1.87. For "unexposed" individuals, the estimated true relative risk•lies between 1.09 and 1.45. Tho number of (an- tively smoked) cigarettes effectively inhaled by a nQUSnioker living with a smoking spouse lies in the range of 0.36-2.79 cigarettes/day. If the spouse is a nonsmoker, however, the estimated tiumber liee between 0.12 to 0.03 cigarettes/day. Of the roughly 7,000 lung cancer deaths estimated to have ocr - curred among lifelong nonsmoking women in 1985, between 1,770 and 3,220 may be attributable to ETS. Of the roughly 5,200 lung cancer deaths estiniated to have occurred among lifelong nonemok- __ - - ing males in 1985, between 720 and 1,940 may be attributable to _ ETS. The estimated lifetime risk of lung cancer attributable to ETS in a nonsmoker with moderate ETS exposure lies between 31)0 and 990 in 100,00t. The estimated lifetime risk of lung cancer attributable to other people's cigarette smoke for an exsmoker who 297 smoked one pack per day from age 18 to 46 and waH* moderately - exposed to uther people's cigarette suioke lir.s between 520 and 2,030 per 100,000. D-1 ESTIMATION 0p'CI-IE TRUE It.ELATIVE RISK Method 1 The first method'for estimating the true relative risk relies on -- - two assumptions: • The excess relative" risk in a nonsmoker is proportional to the lifetime dose of ETS. That is, if an individual's dose of ETS (at all ages) were doubled, his excess relative risk would be doubled. • At every age, "exposed" subjects have been exposed to ETS at a rate 3 times that of "unexposed" subjects. A factor of 3 was selected to reflect the 'empirical observation that the concentration of cotinine in the urine rine of nonsnrokere with smoking spouses is about 3 times that of nonsmokers without smoking spouses (Wald and Ritchie, 1984). These two assumptions iniply that the excess (true) relative risk in "expoHed" individuals ie ;t Limes that of •unexposed" indi- - vidnals. Hence, in the absence of bias, the summary rate ratio of 1.3 equals the ratio of the true relative risk in "exposed" individ- uald to that in "unexposed" individuals. Therefore , -_ , 1.8= 1-H3s 1-1-s' where x and 3x are the excess true relative risks in "unexposed" and "exposed" individuals, respectively. Solving for ¢ gives z= 0.18 and, thus, the true relative risk in "exposed" and "unexposed" individuals of 1.54 and 1.18, respectively. If we used the summary rate ratio of 1.14 from only the U.S. studies (see Chapter 12), we estimate the true relative risk in "exposed" an_d_ "unexposed" individuals to be 1.23 and 1.08, respectively. It is likely that the second assumption above may be inappro- priate (eee Remark 4 in the Technical Discussion). For instance it is unlikely that the ETS exposure in childhood is 3 times greater in subjects who later married smokers, i.e., "exponed" subjects, than in subjects who later married nonsmokers, i.e., "unexposed" sub- jects. If it is not appropriate, then another approach ia nece_ssary. This approach is outlined in Method 2, .vhic-h follows. TSQ4B44e

298 Method 2 Method 2 relies on the following two assumptions: . Assume that (a) cigarette smoke influenccs the rates of tlio first- and fourth-stage cellular events in a Rve-stage multistago cancer process (Day and Brown, 1980; Brown and Chu, in prens); (b) ETS affects the same two stages; and (c) the ratio of the relativo ' magnitude of the effect (on a multiplicative scale) on stage 4 to that of stage 1 is the same for ETS and mainstream sinoke. If we let fi1 and a& represent the magnitude of the effect on the first and fourth stages, respectively, then (c) implies that &/pl is the eaine for ETS and mainstream smoke. . Assume the observed overall sumrnary rato ratio of 1.3 is - the ratio of the true relative risk in "expos«d" subjects to that in - "unexposed" subjects at age 70 (see Remark 3 In the Technical Discussion). It is possible to estimate the true relative risk in `exposed" and "unexposed" study subjects, given two additional pieces of information (see Nsmark 8 in the Technical Discutieion). First, we require an estimate of the ratio p.//li. An estimnto of @./pl can be obtained by fitting the above multistage caur,er model to data on the lung cancer experience of active smoluirs. In particular, an estimate of 0.0124 is obtahwd by fitting the - multistage model to the continuing smoker data among British physicians given by Doll and Peto (1978). Brown and Chu (in press) obtained an estimate of 1.8, derived by fitting the multistago - - model to data from a large European caso•control study of lung -- - cancer. These two estimates of P./P1i however, differ front fine another by 15Qfuld. A third estimate of #s/pi was computod, based on the following considerations. Tho estimate of /!./pi front Doll and Peto (1976) fails to adequately account for the rapid Iall - - olf in relative risk in British physicians upon cessation of smokiug. Since e a larger ratio of p./p1 will be associated with a more raloul fall off of risk when smoking is stopped (especially among smokyrH - of relatively few cigarettes a day), we computed the maximum - estimate of P./Pl that was statistically consistent (at the 5% level) with the continuing smoker data in Doll and I'eto (L978). '1'his - - estimate was 0.225. Rather than choose among these estimates, --- - we performed a densitivity analysis using the three estimates of 299 P. /Pl of 0.0124, 1.8, and U_225 (see Remark 5 in the Technical Discussion). Second, we require, at each age, an estimate of the age-specific. ETS exposure of "expased" and "unexposed" study subjects rela- tive to the current ETS expqsuro of an average adult nonsmoker whose spouse is a nonsmoker. Information does not exist to an- swer questions such as "How many times greater (or less) was the past ETS exposure in average "expoeed" subjects front age 0 to 20 than the current ETS exposure of an average adult nonsmoker with a nonsmoking spouse?" Therefore, a sensitivity analysis was performed using 30 different choices for the lifotime exposure histo- ries of "exposed" and "unexposed" subjects (relative to the current E'L'eu exposure of an adult nonsmoker without a smoking spouse). The choice of exposure histories was influenced by the following general considerations. Smaller differences postulated between the lifetime ETS exposures of "exposed" and "unexposed" individuals will be associated with larger estimates of the true relative risk. (llaving an observed rate ratio as largo as 1.3 when there is truly only a small difference in dose between the "exposed" and "un- ex,iosed" subjects would imply that ETS is a potent carcinogen.) Therefore, we tried to select sonie exhosure histories that would modestly underestimate the true difference in-exposures between tho "exposed" and "unexposed" study subjects snd others that would modestly overestimate this difference. 'fhe rationale for our particular choices of the 30 exposure histories is given in Remark 7. - - Thirty possible exposure histories are given in Table D-1. Remark 6 iu the technical discussion describes how to read the exl:osure histories from this table. Table D-2 gives the maximum and miuimum estimates of tho true relative risk among the "exposed" and "unexposed" for each choice of Q,//11i over the 30 exposure histories. The column olumn denoted "all" gives the overall maximum and minimum as the choice of both Q./pl and exposure history varies. TLe most striking finding is that the estimate of the excess (true) relative risk for "exposed" individuals varies only twofold, froin 0.41 to 0.87, and includes the estimate, 0.54, obtained with Method 1. All-estimates exceed the uncorrected value of 0.30. Estimates of the excess true relative risk in tho "unexposed" range froin 0.09 to 0.45. Because of the possibility that the 30 exposure histories are not representative of those in Japan and Greece, two zqQJ,9449

300 301 Txnl.e E!-2 Eistimated Ranges for the True Relative Risks (R.ft ) in - - - "l:zposed" sud "Unexposed" Subjects 1 ~ R l P./Ot w a ~ a . e Ralio' Grou All 0 0124 0 225 1 8 ~ ~4 p . . . N we ~ "Fspused" 1.41-1.876 1.41-1.87c 1.43-1.32 1.43-1.64 ~ ~. (321)-(113}r (32t)-(113) (•121)-(113) (321)-(113) k s + o oo~Qo ~'s .- en - :~ - 1.3 "Unexlwsed" 1.09-1.45 1.09-1.45 1.10-1.34 1.11-1.27 ~ , 4II (321)-(113) (321)-(113) (•t21)-(113) (321)-(113) ••Fxp,ised" 1.19-1.35 - - - 0 (321)-(113) ! 4 uRo-I o-,41 o A~ 1.14 ••Uneslased" 1.04-1.18 M:, l (321)'(113) N M v z $ 1 'Assume causal summary rate ratio. bItange of RR ewer 30 exposure historks and three values of /f4/(l,. I I I I -u o IF ~ 11tange of RR over 30 exposure histories. o uLxpesure hislorks (a. b, e) at which miniuwm and maximum, respectively, occur Isee hi i d fi i i f t b T bl D 1 f )I ur or e n t on o exposure s es la. . e e - a . A I-N oo,v3~ ~.;....C; 4 oo,o,s, .a...,~ II .,A k d s a ~ .~ ~ ~y y 44 ? d ooao~oo a4o ......,..,_ood•±o ~q;N401;:°4= codoooo_ooo' P of the countrles in which epidentiologic studies were conducted, we repeated the analysis using the ovarall summary rate ratio of 1.14 from the U.S. studies. ln this case the overall range in the - estimates of the true relative risk was 1.19 to 1.8b in the "exposed" and 1.04 to 1.18 in the "unexposed:" D-2 THL CARCINO(IEN-EQl1IYAl,ENT NUMBER Oh' ACTIVELY SMOKED CIGARETTES INHALED DAILY DY PASSIVE SMOKERS: COMI'ARISONS OF EPIDEMIOLOGIC WITH DOSIMETRIC ESTIMATES In this stietion we attempt to eetimute the lulmber of cigarettes, de, that would have to be acLively smoked to deliver to the lung of the smoker a dose of activo carcinogen equal to the daily pul-y munaey dose of carcinogen (attributable to E_TS) of an average adult uonsntoker with a nonsmoking spouse. Roughly speaking, do is the (lung) carcinogen-e41uivalent number of (actively ly smoked) cigarettes inhaled daily by an average adult nonsmoker with a non- amoking eponse. Under the assumptions of Method 2, we saw that knowledge of #4/p1 aud of the relative exposure histories of "exposed" and "unexposed" study subjects was suHlrient to estimate the true wIw wn+1 WIW WiW W$W - M M V w CaO4.B4.4R

302 relative risks. ]f we also have an independent estimate of /11i wo can estimate do as well (see Remark 8). I:ach of our three InuthodH of deriving an estimate for d4/dl from data on active smoketa alsll produces an estimate of /lr. In particular, eatintates of QI of 2.93, 0.803, and 0.14 are associated with ~B.//11 of 0.0124,0.225, and. 1.8, respectively. Some conflicting results need to be resolved, however. For any given level of smoking, the relative risk estimated frotn the British physicians data (1)oll and Peto, 1978) is greater thau that estimated front the American Cancer Society's follow-up data olt a million Americana (Hammond, 1966) eor from the multicenter European case-control lung cancer data (Lubin et al., 1984;11rown and Chu, in press)..'rhe relative risks in these latter two studies are consistent witL one another and will hel e be treated as identical. Doll and Peto (1981) suggest that these ililfcreuces in relative risk may be real differences, attributable in part to the different way cigarettes are slnoked in Britain and other countries. To bring the British data in line with the other data, we adjusted our estilnates of Rl from the Doll and 1'eto data as follows. Separately, h or the d./,O1 of 0.0124 and 0.225 (both based ton Lire British physicians data), we computed the value of pr that would be necet+sary for an individual smoking 26 cigarettes per day since age twettty to have the same lung cancer incidence at age 65 as would follow _ -- if ,0./,Bl = 1.8, fi1 = 0.14 (based on the European cast}conlrol data). This gives adjusted estimates of 1.41 and 0.46 for P1, corresponding to values for P4/iBr of 0.0124 and 0.225, respectively. _ These values are approximately half those previously estimated from the British physicians data. In our sensitivity analynis we use both the ad justed and unadjusted estimatea of fil (see Remark 9). Estimates of do are given in Table D-3. Under the assunlption that the summary rate ratio of 1.3 is causal, estlmates of d„ vary about eightfold froln 0.12 to 0.93 cigarettes per day. For a given pair of values of Pr and P4/p1i the variation in do over the 30 exposure histories is only about twofold. When we use the sutns - mary estimate of 1.14 from the U.S. studies in lieu of the suntmary estimate of 1.3, our estimates of do are diminished accordingly. We next compare the above estimates of d,,, which are based on the epidemigologic data, with estimates baned on the doeiwetriE measurements reported in Chapters 2 and 7. Ih:etimates of do lywed on dosimetric calculations are given in'liable D-4. In Table U-4 we 303 rAewl D-3 Estimated Range fur cl,,, the Carcinugen-Equivalent Nuntbcr tif (Actively Smoked) Cigarettes Inhaled 1)aity by Subjects Without a Smoking Spouse (14/0,: All 0.0124 1.8 0.225 - - - -- - -- . -- /t,: All 2.93 1.41 0.14 0.803 0.46 Rare rutio 1.3• 0.12-0 9J6 0.12-0.27d 0.24-0.57 0.A8-0.t)9 0.26-0.53 0.46-0.93 (311)-(123Y (311)-(123) (311)-(123) (311)-(423) (311)-(123). (311)-(123) 1.14 0.05-0.47 (3u)-(/23) 'Auured causal rate ratio. bRaoge of d,u ln cigarettes/day over 30 expusure histuries uui ali (@t/ti,. (i,). `Exposure history where maximum and ndnimum uccurral. dRange of do over 30 exposure histnrics. TABLB D-4 l,stimates uf da Based on Various Conslituenis of - - ETS in Cigarettes/Day Constituent Range NDM_A_ 0.17-3.75 B.P 0.0084-1.89 RSP 0.0N11-0.005 give an estirttated range for do under the assuutptions that the ratio of the pulmonary (tissue) dcae of active carcinogen in nonsmokers without smoking spouses to the pulmonary duse in active smokers is equal to the ratio of the puhuonary dose of Ba1`, NDMA, or RSQ in the same populations. '.Che estimates in Table D-4 are based on (1) the dosimetric measuremente given in Table 2-10 and Chapter 7 and (2) the daily number of hours of self-reported ETS exposure antong nonsmokers without smoking spouses (Wald and Ititchio, 1984; Friedman et al., 1083). Details of the calculations us111d to produce Table D-4 are given in Remark H of the Technical D'uicussion. The dosimetry of the biomarkers nicotine and 'co_ tinine is more complicated and is discussed in Remark 12. There ia a serious probleln in reconciling the estimate of do (Tiible D-4) based on BnP with that based on RSP, since RSP is often assumed to be a good surrogate for poly4:yclic hydrocarbons such as BaP. The estimate derived from the BnP measurements is V..c~iQfaQ44B

304 several orders of magnitude higher. A possible, although unlikely, explanation is that the measurements of BaP levels in E`1'S (suwn- marized in Table 2-1(/) inappropriately reflect total environmental BaP, which includes contributions from cooking, coal burning, and other sources, and that the contribution of BaP from LTS to total BaP is of the order of 2% or less. The large uncertainty in de seen in Table D-4 restricts the utility of these dosimotric calculations, especially given the lack of knowledge concerning the identity of the active carcinogens in E'.CS and mainstream smoke. ln fact, the limitations of our dosimetric, data may be even more serious than Table D-4 would lead one to believe. Specifically: . the range of values entered in Table D-4 for NDMA could actually be orders of magnitude too high (see step 4 of Remark - - - i1), - 0 the range of values for RSP and Dal' do not reflect dif- ferences between the particulate phase of l:Th and that of main- streain smoke with regard to deposition sites, clearance rates, antl - particle size, a the range of values given for BaP in Table D-4 could bi+ orders of magnitude too high if, as discussed above, the BaP entries -- in Table 2-10 represent the total environtnentnl DaP inhaled by n nonsmoker, and 0 the ratio of urinary nicotine (or cotiuine) in nonsmokers to that in active smokers may not reflect, even qualitatively, the - ratio of the biologically effective dose of active lung carcinogen --- - absorbed by nonsmokers to the dose absorbed by active stnokern (see Remark 12). D-3 EB"1'IMATING THE NUMBER Olr LUNG CANCEIL DEATHS IN NONSMOKERS IN 1985 ATTRIBUTABLE TO ETS An estimate of the total number of lung cancer deaths amonK - lifelong nonsmoking women in 1985 is Eg le(t)N(t), whore N(t) is the number of nontimoking women at risk ut age t in 1985 moil Ia(t) is the age-specific lomg cancer death rate anwng nansmkuig women in 1985. Dnta on lo(t) are given in Garfinkel (1981) fooF 1972; thus, this may be somewhat inaccurate for 1985. National Flealth Interview Survey data on N(t) were made available froon 306 R. Wilson of the National Center for Aealth Statistics. Using these data, the number of lung cancer deaths was estimated to be 7,000, similar to the estimate obtained by Seidman (personal c-ommmucation) using a related_ approac:h.. The total number of lung cancer deaths among nonamokin wonien attributable to ETS in'1985 is g AN ® EAF(t)h(t)N(t), 711 where AF(t) ls the age-specific fraction of lung cancer deaths due to ETS exposure in nonsmoking women. That is, AF'-(t) is - the age-specific average excess true relative rink (i.e., the average relative risk minus 1) divided by the age-speciQc relative risk. In order to estiniate.the age-specific relative risk among nonsmoking - wonien, we require age-specific estimates of the probability of being married to a smoker (i.e., the probability of being 'bxposea") and of the true relative risk in "exposed" and "unexposed" subjects. We obtained age-specific estimates of the probability of being "exposed' from the-' Garfinkel et al. (1985) control population (Garfinkel, personal communication). We estimated the true relative risk in three different ways. First, we use the estimates derived using Method I in Section D-1. Second, we use the estimates based on Method 2 of Section D- -- 1'J'hird, we completely ignore the epidemiologic data on passive smoking and estimate the truo relative risk by combining estimates of fii and P4/P1 extrapolated frour data on active smokers, and estimates of de based on dosinietry (Method a). In a sensitivity analysis, we allow de to equal 0.01, 0.2, and 2 to crudely represent (approximate) exposures to ItSP, BaP, and NDMA, respectively (see Table D-4). The estimates of the attributable number based on Methods 1 and 2 are valid whenever the assumptions justifying those methods hold. For a given choice of de, the estuuatcd of the attributable number based on the third method are valid when the first assumption under Method 2 holds antl the choice of do is correct (see Remark 13). Using the relative risk estimates based on Method 1, we ob- tained - taim:d an attributable number of 2,010. In `1'ablo D-5, estimated ranges for the-attributable number are reported. AN(EP) represents the estimates based on Method 2. AN(0.01), A,N(0.2), and AN(2) represent estimates based on the ilosimetry estimates of 0.01, 0.20 and 2. (Since the estimate ...~iS04B4,LB

306 of the true relative risk based on Method 2 depends only on p4/Qi (and not on P1), the estimate of AN(EP) also depends only on Estimates of the attributable nwnbor of lung -- - cancer deaths based on Method 2 lie between 1,768 and 3,220. These estimates are approximately halved when the summary f ate ratio of 1,14, from the U.S. studies is used in place of the overall summary rate ratio of 1.3.) If the true value of do were 0.01. cigarettes per day, then 259 lung cancer deaths in nonsmoking women would be attributable to ETS. On the other hand, the, - maximum estimate of the attributable number based on Method 3 with do = 0.2 (3,170 deaths) is in agreement with that based on Method l(3,220 deaths). The minimum estiniates, however, differ by approximately threefold. The calculation of the number of lung cancers attributable to ETS in 1985 in nonsmoking males is similar. Garlinkel (1981) and Wilson (personal communication), respectively, give data on le(t) and N(t) for nonsmoking males. Since estimates of lo(t) in males and females are nearly equal and the estimates for females are more stable (Garfinkel, 1981), we use the same estimates of !o(t) for males as for females. Using these data, the estimated number - - of lung cancers which occurred in lifelong nonsmoking males in 1985 is 5200. For males, the fraction "exposed" is taken to be 14% (based on the control series from the Correa et at. (1983) study of males). Using relative risk estimates based on Method 1, it is estimated that 820 of the 5,200 lung cancer deaths are attributable to ETS. Estimates of the attributable number in males based on Methods 2 and 3 are given in Table D-5. Overall, the results fur men are similar to those for women. D-4 LiFETIME 1LISK OF DEATH FROM LUNG CAIYCI+,)ii. AT-T-Ii.IDU'rAULE TO ETS Aiuonb Lifelong Nonsrnokers Permissible exposure limits to environmental agents are oF ten set at levels low enough to reduce the lifetime risk of death attributable to the agent to 1 in 10° or 10°. For purposes of com- parison with other environmental and occupational standards, we have attempted t4i estimate the fractions of all deaths antoug nonsmoking men and womeu who survive past age 45 that are 307 attributable to ETS-induced lung cancer. (This fraction is pre- cisely the lifetime risk of lung cancer attributable to E'1'S exposure - - among persons surviving to ago 45.) Since the risk of lung cancer is nearly 0 before age 45, we have chosen to condition this estimate on survival until that age. (Although years of life lost due to ETS exposure would be more preferable as a public health measure than the attributable fraction of deaths, we restrict our analysis to this latter measure in order to hell) determine whether, for regulatory purposes, ETS is being treated dillerently than other environmental exposures.) Because environmental regulations are generally set with the intention of protecting all (or at least almost all) individuals, we chose to estimate the attributable fraction for a representative subject with ETS exposure history of 2de for ages 0-18 and 4do for ages greater than 18. Based on data from Wald and Ritchie (1984) and Jarvis et al. (1984), this exposure history represents all exposure to ETS that is slightly greater than the average exposure of a nonsmoker exposed as a child to a smoking mother and as an adult to a smoking spouse. We label this expo- sure history as M, since it represents a moderately high lifetime exposure to E'1'S. The fraction of all deaths subsequent to ago to (in our case age 45) attributable to exposure-induced lung cancer is, by definition, AF(M) g E 71&xosss(E)S(tltp)- e>e, whore qgxoaps(t) is the excess of lung cancer 'deaths at age t due to exposure history M and S(t1to) is the overall probability of surviving to age t, given one has survived to to, Given that the assumptions of Method 2 hold, we can obtain an estimate of AF(M) for each value of /94/pl and each of the 30 exposure histo- ries for the "exposed" and "unexposed" study subjects, provided we have data on the age-specific lung cancer rates in nonsmoking wonien, IQ(t), and data on the all-cause age-specific mortality rates among nonsmoking women (which we estimated from data given in Ilammond (1966) (see Remark 14). The maxiinum-and minunum of the AF(Af) across all expo- sure histories for each Q4/p1 are given in Table D-6 in the "never- smoked" rowH for males and females. AF(111) is estimated to be between 390 iund 990 in 100,000. A similar calculation, using the 9SO4B44R

iA=D-S Estimates of ETS-,ACtributabk'Lung Cancer Deaths Among U:S. Nonsmokers:in 1985 (by Sex) 9WA: All ~0.0124 0.2?5 0.0124 0.225 1.80 Bi: All Sex 2.93 :0.803 1.41 0.461 !0.140 Rare IRario : = 1.3 AN tlor F 1768-322& 1768-3220 1820-2800`' 1768-3220 n820-2800 1939-2492 cr=-(113J (323),(113) (3='.1W.'113) (323)-(I113) (32'14-{113) (323)-(113) M 721-1942 721-1942 751-1611 721=1942 750-1611 8550r1390 (321)4113) (321)-(113) (321)-(113) (321)-(113) (321')-(113) (321)+(113) A.'V,(0.01) F 31-_" =-C-250 -14-'102 61-u" 31-49 34.:55 (423')-(211) (473)•(211'J (4?31.L211) i(423),{2',11) (423)a(211) (423H211) M ]4-137 '53-137 24-i0 26-67 14-29 16-25 (473)-(111) (4?3)-{1'11) fC3)-!.21) LC3L'+'"N !~ ==~ (4=-u1311 AI'V (0.2) F b85-3174 1921-3174 978-1695 1059-1939 585-1052 634-988 (42.i)-(211) (+423)rt2'1'1) (423),(211) (423)-(2111) (4?3)-(211) (473)-(211) 1~1 265-1890 908-1890 450-891 425-'1094 265-540 305-465 (413)*(11'1) i(423)-(114) (423k11'1) (433)4111) - (423')-(111') (473)+(111) M7M (2) F 3793-6778 S992-6778 3039-61" 4^!02--Q73 3'~-5163 3E.4-s9-CS (423)-(211) (423)-(211) (423);(2'11) (423)-~(211) (423)+(21'1) i(423)-(21I1) 2016-A80,3 36'12-4803 2904-4060 2758-4057 2016-.3170 2151-2'908 (423)4111) RatrRatio = 1.14 v!'JV IlEPI . (4S3);(111!) (423)4111) (423H111:) (423)-(1'13) 443$)-Q1111) F 935-1730 (3:3 -,(113) M 360-980 (321)-(113) •.1N /EP/ izi bsred.oo epidemiolo;ic dua' in nommolcas exposed, to ETS. bRaale of attn"butsbk nnmber of luaq eaaaerso.a 30 ez,rosure Aisdories and firei choicertlf (01• 94/0i). `Ranp oUANdf lung, eancaes ~over 30 esposwe~ histoiiac in! nonsmol3ng females for 0478, '_0.225.9,i' _0:803. dE;pasuts histary where minimum and ma>®um ocate:.

310 TABLE D-6 Range of I:stimated Lung Cancer Dealhs Attributable to Breathing Other People's Cigarette Smoke per 10,000 Deaths (All Causes) p,,[P,: AII 0.0124 1.8 U.225 Rate Sex Smoking - Status" /t,: All 2.93 - 1.41 0.14 0.801 0.46 1.3b M N J4-99 4a-99 45-95d 4N-99 45-95 39-77 Ex 52-197 62-126 74-149 52-106 100-197 62,115 C 5a-307' 78-157 107-209 5a-t r7 159 707 g6-15a F N 4n-99 49-99 45-96 44-99 45-9L 40-78 Ex 54-20J Irt-130 77-154 54-110 103-203 64-120 C 1t2-3J1 84-169 115-22S 62-125 171-331 92-170 1.14 M N 19-49 Ex 26-99 C 29-159 F N 21-52 Ex 29-109 C 33-182 _- "Smokiug Status: N = never; Ex = smoked I pack Ixc day, age g-45; C = Qonllnuiug smoker. I pack per day fnnn age Ig. 6Assumcd causal rate ratio. - `Itnnge over 30 exposurc histories, S values of (S,. d410,). dItange over 30 exposwc histories. NOTE: All mxima were associated;wilh exposure history 1•123); all minhna with his/ury (311). summary risk of 1. 1.4 from the U.S. studies (Instead of 1.3), halves our estimates for AF114t). Anlong Current and Exnnwkers We now estimate AF(M) for to = 45 for current and exemuk- era of 20 cigarettes per day. 1b clarify the approach, consider a female examoker (or con6inuing smoker) who was exposed to exposure history M of ETS from other people's cigarette smoke. (T'he subject's total ETS exposure re is oven greater, since it consiHts of contributions from her own cigarette smoke, as well.) Then rysxosss(t) necessary for the calculation of AF(1if) ia the dilfer- euce between the lung cancer mortality rate at age t, given her total amoke exposure, and her lung cancer luortality rate at age t, had she had the nitme active smoking history without exposuru to other people's cig,trette amoke. We require the same tmumptiunn and information to estimate AF(M) for oxe+mokers and continuing 311 smokers as we did for nonsmokers, plus an estimate of #1. Esti- mates of S(tlte) are obtained as before, except the exsntoker and continuing smoker all-cause mortality rates given in llammond - (1066) (see lEemark 15). In Table D=6 the maximum and minimunt of AF(M) for each -- of live combinatione of (Al1,,041#1) and all 30 exposure histories for the "exposed" and "unexposed" are given for continuing smokers and exsmokcrs of 20 cigarettes per day starting at age 18 and, in the case of exsmokers, stopping at age 46. For exsmokers, the estimate lies between 520 and 2,030 per 100,000. For continuing - - - smo ers, it lies between 580 and 3,310 per 100,000. A similar calculation, using the summary rate ratio of 1.14 from the U.S. - - studies, halves our estimates. DISCUSSION Exercises in quantitative risk assessment serve several useful purposes. First, public health decisions must often be made with- _ out certainty as to the magnitude of the likely health benefits that would result from implementulg the various policy options. Quan- titittive risk assessment can aid in the Jecision-making process by -- quitntifying this uncertainty. Second, difficulties encountered in providing precise estimates in quantilntive riak assessment high- light areas where scientific knowledge is inadequate. Thus, exer- cises in risk assessment can serve to help focus future research. All quantitative assessments of risk rely un assumptions. In- terval - - terval estimates of quantitative risk are reliable only insofar as (1) the assumptions under which they were deriv4:d are valid and (2) the range of parameter values used In the estimation process in- cludes -- clu<les the truo value. It follows that no quantitative risk estimates can be guaranteed to be reliable. Nonetheless, some risk estimates are more (or less) reliable than others. With regard to point (2) above, it should be noted that, in - - performing the risk assessment presented here, a sensitivity analy- sis was performed only over those parameters for which there were either inadequate empirical estimates (e.g., the lifetime ETS ex- posure history of "expoeed" and "unexposed" subjects) or grossly inconsistent estimates (e.g., the estimates of P4`#1). Thus, the analysee did not account for other sources of uncertainty, such as - statiatical mtcertainty, in estitna/.es of other parametera. If they had, the width of the interval risk estimates may have increased SCD4S44R

312 severalfold. Generally, the more'parameters that. are varied in a sensitivity analysis, the more information that analysis provides; nonetheless, for simplicity, we chose to vary only those paranr _ --- eters with inadequate or inconsistent estimates. It is inevitable that some readere, often with good justification, will foel that we should have used different values for the parameters we treated as fixed or different ranges for the parameters we varied. (Computer programs are available from Dr. Robins.) In our risk aesessment, the most important wiswnption was - that the observed summary rate ratio of 1.3 was causal. If this assumption is correct (below we discuss the puesibility that it is not), we believe that the estimate of the lifetime risk of lung cancer among lifelong nonsmokers attributable to moderate ETS exp« sure JAF(M)] will be accurate to within a factor of 2 to 6. This belief depends on the fact that if the rate ratio of 1.3 is causal, we are not extrapolating outside the range of the datn (for oxample, from high to low dose) in estimating AF(M). (Even though our re- ported uncertainty in estimating AF(M) in iiever-nmokers (Table D-6) is only twofold, nonetheless, as discussed above, our esti- mate of overall uncertainty would likely be larger; we have guessed twofold to sixfold). For any reasonably flexible model, such as the multistage model, the data (when ample) will drive the risk estimates provided one does not extrapolate outside the range of the data. For instance, even though our estimates of p./pi used in the sensitivity analysis differed by 150-fold, the overall variation in the lifetime risk of lung cancer due to ETS in nonsmokers varied only twofold (Table D-6). In contrast, in estimating the lifetime risk of lung cancer due to ETS in ezsrnokers we were forced to extrapolate outside the range of the data. 'Po do this we used sta- tistical models. We found an uncertainty factor of about fourfold (Table D-6) because of the sensitivity of. this extrapolation to the particular coefficients assumed for the multistage model. But even this range of four underestimates tha true uncertainty, because we have little assurance that it is appropriate to use the multistage model to extrapolate. Given that we can know the lifetime risk of k:`L'S-caused lung cancer in nonsmokers within a factor of 2 to 6, is this degree of accuracy sufficient for our purposes? Obviously, it depends on the purpose. If there were a regulatory process through which we could ensure that the lifetime risk of lung cancer attributable to ETS among nonsmokers would be no greater than 1 in 10(1,OOU (or even 313 1 in 1,000), by limiting, if necessary, exposure to environmental tobacco smoke, our risk analysis would appear to be sufficient to drive that process. This is true because, even if the lower estimate of risk of 390 per 100,000 were reduced by factor of 2 or 3 (to take int) account additional sources of uncertainty), it would still greatly exceed 1 per 100,000. In this appendix, we confirmed our risk entimates with those arising under the assumptions that the causal summary rate ratio koni the various epidemiologic studies was either 1.3 or 1.14 (the summary rate ratio from the U.S. studies). In Chapter 12 it was concluded that, considering the evidenco as a whole, exposure to ETS increases the rate of lung cancer among nonsmokers. Further- --- more, it was concluded that our best overall estimate of the causal -- - summary rate ratio from the 13 studies was about 1.3.. In light of this conclutdon about causation, for purposes of making public health decisions for the United States, it would seem prudent to operate under the assumption that the true summary rate ratio was most likely 1.3 and at least 1.14 (even n though values less than - 1.14 cannot be excluded). We therefore did not prepare estimates for values less than 1.14. We also ciid not make risk estimates under the assumption that the causal summary rate ratio was greater than 1.3, largely - because the estimated lifetime risk of lung cancer at this rate ratio of 1.3 was sufficiently large that it did not neem important to quantify how large the lifetime risk might be if the true causal rate ratio were 1.48 (the 95% upper confidence bound for the summary rate ratio of 1.3). Finally, it would have been helpful to be able to compare estimates of risk derived from the 13 epidemiologic studies of nonsmokers exposed to ETS with independent estiniates based on dosimetric- measurements made in active and'passive smokers. Unfortunately, as discussed in Section D-2, uncertainties in the identity and dose of the active carcinogens in E'1'$ and mainstream smoke effectively preclude this possibility at this time. TECHNICAL DISCUSSIONS Estimation of the 1rue Relative Risk Method I The assuumptions presented in Section D-1 above are replaced by niore formul assumptions: ~'iSQ4+9448

I . 314 Assumption la We assume that, in the "low-dose" range repro- sented by ETS exposure, the increment in the mortality rate at age t due to an increment of ETS exposure experienced at age a (u < t) is uninfluenced by any other increment of ETS exposure (whether received at time u or at any other time u'). The mathematical formulation of Assumption la is 7(tI{d(u); u_ t}) = ryu(t)I1 + d(t) f 1(t, u)d(u)dul, (D-1) e 0 where ryo(t) is the mortality rate at t in the absence of exposure to E'rS, d(u) is the dose at age u of the active carcinogen in ETS, ry(tl{d(u); u E t}) ia the mortality rate at t given a history of expo- sure to ETS represented by the curve (d(u); u< t}, f~ /(t,u)d(u),tu tnay be interpreted as a weighted average of an individual's past exposure, and /9(t) is an age-specific measure of the ntagnitudo (on a ratio scale) of the ETS effect. (For example, if there were a 5-year biologic latency period, f(t, u) = 0 fur t-. u < 6). Remark 1 In the above description of Mqw~tion 1)-1, we have implicitly assumed that fk J(t, u)du = 1 e,s that fo f(t, u)d(u)#lu is a weighted average and p(t) is an effect measure. In fact, the -- restriction fe I J(t, u)du = 1 (D-2) Ju is not in general necessary for Equation D-1 to be meaningful, al- though some restriction is necessary to idaintify /!(t). Nonetheless, Equation D-1 cau always be •reparameteriaed" so that Equati(in D-2 holds. If, in Equation D-1, P(t) = P independent of t, we say that we have a linear excess relative risk wodul. If in Equation 1)- 1, p(t)qo(t) =/!', independent of t, we have a linear excess absolute risk model. If there exists a function J(t, u) for which Equation D-1 is a linear excess relative (or absolute) risk model, then Eryuetuin D-1 generally cannot be "rcparatneterized" so that simultaneuusly Equation D-2 holds and O(t) =P or P(t)7o(t) - p'. Remark 2 By extending the argument giveu by Crump ok uI. (1976), one can rthow that xufficient (but not necessary) conditions for Assumption la to hold are (1) the dose of E'1'S.to paiieive smokers at any Lime u has a very stnall influence on risk at I atid Q9QLS449 315 (2) other risk factors for lung cancer. operate through the same mechanism as ETS. Since the true relative risk associated with passive smoking exceeds 1.3, Crump et al.'s argument may not be relevant. In Remark 18, we empirically assess the validity of Assumption la under the further assumption that cigarette snioke a(fecta two stages of a five-stage multistage cancer process. Consider now the subset of the source population of an epi- denslologic study that includes "exposed" individuals at risk at age t. Clearly, the exposure at any time u,u < t, to ETS, say d(u), will vary among persons in this subset. Let d&(ult) be the average pulmonary dose at age u among "exponed" individuals at risk at age t. In a follow-up study in which the data collected includes age, cause of death, and "exposure" status, we can empirically esti- mate the age-specific (average) mortality ratn among "exposed" individuals, q(tIE), and unexposed individuals, ry(tIE). Further- more, it follows from the linearity of Equation D-1 that RR(tIE) s ry(tll:)/-yQ(t) ~ 1 +P(t) ~ f(t, u)d6(uIt)du, u who_ re 1tR(tIE) is the true relative risk (i.e., the ratio of the mortal- ity rate amoHg "expoeed" individuals to that of truly unexposed individuals). Unfortunately, we cannot estimate %(t) land thus RR(t1R)) without further assumptions. Similarly, we are unable to t,stimte RR(tIE), the true relative risk duo to ETS in "unex- posod" individuals.. (Remember, "unexposed" individuals are truly exposed.) But, in the absence of hiaa, from either prospective or caewcontrol data we can empirically estimate ry(t E=. 1+Q(t) fT f(t, u)du (ult)dtu ~ It R t E) (D-9) 'r(tlR) 1+P(t) o f(t,u)d&(ult)du RR(tlB)~ (In a case-coutrol study the left side of Equation D-4 is the agc- epeciHc odds ratio comparing "exposed" to "unexposed" individ- uals.) Assumption lb dB (ul t) _ c(ult) = e(t), D-5 ~ult) ' ~ )

818 where c(t) is a known constant independent of u. Note a(ult) is a ratio of the average exposure at age a of "oxpostid" subjects at risk --- - at age t to that of "unexposed" subjects at rhik at t. Our main result is: Lemina 1: If Assumptions la and ib hold then ftR(tIE) = t+ c(t)s, and RR(tlS) = i+ s where :I'-g -1 y~ l .- -~ •j1 fIOI 6(t) - 7 Proof: Let P(t) f,~ f(t,u)ds(ult)du - RR(tJR) - 1 s z. Then - RR(tIE) = 1-t• c(t)z and RR(tJR) = 1_ + a. Thus, substituting in Equation D-4, !(tIE) ~ 1 + c(t)s implies: '1(tl2) 6(t) - 1=1=s ® ® (D-6) e -, M Exatnple: Suppose c(70) = 3 and ry(701E)/q(701E`) = 1.3, then s= 0.18, RR(701S) = 1.18, RR(701E) = 1.64. We now show that, under Assumption ia, if c(ult) < s for all _- -- u, the previous estimates of RR(tIE) and RR(tJR) uwat, in fact, be underestimates (although the magnitude of the underestimation cannot itself be assessed without further assumptions such as those given under Method 2). First note that even when Equation D- 6 is false, it is still true that, under Assumption la, with s=- RR(tIR) -.1, Equation D-0 holds provided F(t) is replaced by c*(t), where f ~ f (t, u)ds(u)du C' (t) _ -~ to J(t,u)de(u)au' [Note that E'(t) > -y(tIE)/7(tIE').1 Furthermore, RR(tI. PI,) ix still 1+a (by definition) and IrR(tll~J) = t+ s'(e)x. Now it is straightforward to check that 1tR(t1R) antl RR(tJE) are decreasing funcl.ions of c'(t) reaching rospective minima of 1 and ry(tIE)/,y(tIE) when e'(t) -• oo and tuaxiina of ©o wheu 317 c*(t) a~r(tIE)/q(tl~) is greater than 1. a'(t) =7(tIE)/ry(tJE) is the condition of maximum misclassilication between "exposed" and -- "unexposed" groups in terms of exposure to ETS. On the other' haud, when a'(t) = oo no "unexposed" individual is exposed to ETS. Furthermore, it is easy to check that if dB (uI70)/dfi (ul7u) - e(ul7a) ` 3 (D-7) for all u :5 'i0, then c'(t) < 3. It follows that in our previous example, R1t(tl2) = 1.18 and RR(tIE) = 1.64 would, in general, be underestimatos of the true 1tR(tIF.) and RR(tJR) if Equation D-7 holds. Remark d Note that the investigators of the 13 epidemiologic studies analyze their results as if their observed rate ratios were not dependent on age, as evidenced by the fact that none of the - authors reported age-specifc rate ratios. But if the rate ratio varies with age, then the observed rate ratio reported in each -- - study will be a weighted average of varying age-specifirc rate ratios. Since Gar6nkel et al. (1985) found the mediau age of lung cancer in nonsmoking women in his population waM approximately 70 -- (Garflnkel, personal communication), we would expect that this weighted average approximates the rate ratio at 70. This implies that the second assumption under Method 2 in section D-1 is - probably clotie to correct. To be precise, if, in a case-control study, one-to-one matching on age is employed and a matched - pair analysis is performed, the matched pair odds ratio estimator will estimate the following weighted average of the ago-specific rate ratios, ry(tIE)/q(tlR). The large sample expected value of the odds ratio estimator (O1) is E'la3f] = j(q(tIR)/q(tIE)) f(t)dt where I(=) ° ~ (t)fv(t)dW h(t) = p(!t t p(si) t - t1-AR)/70IR)Ip(E'It) + p( ffit) ~ fo(t) is the fraction of all lung cancers in nonsmoking women that occur at age t, and p(Rlt) is the fraction of nonsmokers in the study source population of age t who are "unexposed." Remark 4 W„ now examine the conditions under which Assump- tion lb holdEi with c(70) = 3. We conclude from the following 1904e44B

sia examination that it is unlikely that Assumhtion lb is true, even as an approximation. Wald and Ritchie (1984) estimate that, in 1982 in England, the urinary cotinino concentration of an average nonsmoking male with a smoking spouse is 3 times that of the average nonsmoking male without a smoking spouse. Urinary nicotine data from Jarvis et al. (1984) and interview data from Medman et al. (1983) suggest that similar results would be obtained in women. Given these observations, the following six conditions must, in general, be met in order for Assumption lb to hold with c(70) = 3. Condition 1 The ratio of 3 also applies to exposure to the biologi- cally relevant carcinogen or carcinogens in 1:TS. Condition 1 is likely to hold, at least in our approximate sense, in the United States and England. (Olav Axelaon has pointed out a situation in which it would not hold. Suppose that the - --- - carcinogenic effect of ETS is largely due to the adsorption of -- - environmental radon onto ETS particles. Then, home exposure ~ ' to ETS would be of greater importance if, in general, only home ventilation rates are low enough to allow significant accumulation - -- --- --- of environmental rudon onto ETS particles. Friedman ot al. ('lIahle - - -- --- - 6, 1983) showed that the number of hourn currently-"exposr.d" - - - - women are exposed to ETS at home is 12.7 times the number of hours that currently-"unexposed" individuals are exposed at home. On the other hand, the total number of hours of E'L'S exposure in curreutly-"exposed" women is only 3 times that of currently-"unexposed" individuals. Thus, if radon uptake tather than urinary cotinine had been measured, Wald and ltitchie. nsay have found a ratio nearer 12 than 3.1 On the other hand, in Japan and Grcere the ratio of urinitry cotinine in nonsmoking women with a smoking spouse to that in nonsmoking women without a smoking spouse probably exceadn the value of 3 measured by Wald and Ritchie (1984) in England, since women in those countries are likely to apend less tiane iu - - contact with cigaiette smokers outside the home. It follows that one might expect the observed rate ratio iu the llirayama (19t14) and Trichopoulos et al. (1983) studies in Japau and Greece, re- spectively, to exceed that found in studies in tho United States. Table 12-4 bears out this expectation. '1"hus, we might want• tao exclude liirayamn's and Trichopoulos et al.'s atudies in calculat- _- ing the overall summary rate ratio. We have seen that the 11.5. 319 studies have an overall summary relative risk of 1.14. Assuming Assumption ib with c(70) = 3 holds for the United States, we would estimate the true relative risk in "exposed" and "unex- posed" study subjects in the United States to be 1.225 and 1.075. Nonetheless, since 1.225 is lcsa than the observed rate ratios of 1.45 and 2.01 in the Hirayama and Trichopoulos et al. studies, we must also assume that the ETS exposure of nonsmoking women with smoking spouses in Japan and Greece exceeds that in the United States (if we ignore sampling variability and other sources of bias and interaction). Matsukura et al.'s (1984) data on urinary cotinine suggests this may be the ame for spouse-exposed Japanese nonsmokers. Beca_ use 10 of the 13 epidemiulogic studies were case-control studies, we concentrate on case-control studies in the following. (Most of our remarks would have to be only slightly modified in order to apply to prospective studies such as GarRnkel (1981) and Hirayarns, (1981), in which the follow-up is only 10 to 15 years.) To characterize further conditions sufficient to imply that e(70) = 3, we shall need to be more precise in our def- inition of "exposed" and "unexposed" subjects. We define an ever-."oxposed" (never-"exporied") subject to be a nonsmoker who, when queried in a case-control study in approximately 1982, an- - -- swered 'yes' ("no") to the question "Did you ever live with a smoking spouse?" We defiue a currently-"exposed" (curretitly- "unexposed" ) subject to be a nonsmoker who, in an epidemiologic study in 1982, answered "ycs" ("no-" ) to the question "Do you currently live with a smoking spouse?" Some of the case-control - studies compared ever-"expoeed" and nover-"exposed" individu- als (for example, Garfinkel ct al., 1985). Approximately half of Garfinkel's ever-"exposod" subjects were currently-"unexposed,-`- with the median time since their spouse stopped smoking of 15 yeari ((iarfinkel, personal communication). Other studiea com- pared currently-"exposed" subjects to never-"exposed" subjects. Condition 2 Wald and ltitchie's (1984) ratio of 3 is independent of age and thus applicable to g©-year-olds. Sufficient urinary cotinine measurements have not been made on T0-year-okls to provide enipirical evidence as to whether this coiidition holds. eonditiun 3 Nearly al170-year-old cur-rently-"exposed" individuals married smokers at about at;e 20 in approxims,tely 1932. This is

320 probably a reasonable approximation, assuming little divorce and remarriage in this population. Condition 4 If the study compares evur-"oxposed" to never- "exposed" subjects, the magnitude of the ETS exposure in the years preceeding the study date of ever-"exposed" individuids who are currently-"unexposed" (because theu spouses either died or quit smoking on average 15 years ago) Is the sanke as that of currontly-"exposed' individuals. For this latter condition to hold (even as an approximation), it is necessary either that only a smial proportion of the total E'1'S exposure In currontly-"exposed" in- dividuals is directly from their spouses or that, when the smoking spouse of a nonsmoker either dies or quils smoking, the amount of time the nonNmoker epends with other smokers increases. Our guess, based on '.lbble 6 of Friedman et al. (1983), is that the L'.l'g exposure of an average evor-'exposed" fentale diminishes by a h+tlf or more when her husband either dies or quits. Thus, it is unlikely that Condition 4 holds. Condition 5 The ratio of 3 applies to the EY'-$ exposure of Nix- posed" and "unoxposed" subjects even during childhood. lt sneins unlikely that children who grew up to inarry smoking spounes would have 3 times the ETS exposure in childhood as children who grew up to marry nonsmoking spouxes, although in Remark 7 we consider empirical evidence which suggests it is conceivahle that Condition 6 might approximately hold. Condition 6 Wald and Ritchie (1984) would have found the eaine ratio of 3 if their study had been performed in any year from 1932 to 1982. 1:ven in those case-control studir.s that coinpared currentiy-"exposed" subjects to never-"oxpoeod" subjects, Condi- Lion 6 may well be false. For example, in the 1930s and 1040s, nonsmoking women study subjects (who were then 20 to 30 years old) were presumably less often in contac-t with smokers outside - the home. This would suggest that in the 1930s and 1940s the ra- tio of ETS exposure in nonsmoking women with smoking spout+es compared to nonsmoking women without smoking spouses was closer to 12 than to ;R (provided the resultn of Friedmau ot al. (1983) mentioned in the discussion under Condition 1 can be ex- trapolated to the 19a0s). Remark 5: Estimatee of p4/dl We used three di(ferent estimates for d4/pi in our sensitiviLy analysis. All were obtained from data C.904Bu'g 321 on active smokers. To obtain the first, we f t by the method of maximum likelihood a five-stage mnitistage model, with the first antl fourth stages affected, to the data on continuing smokers given in Doll and Peto (1978) (excluding, as did Doll and Peto, the subgroup who smoked inore than 40 cigarettes per day). This gave - P./Pr = 0.0124 (and P1 = 2.93). To be prec.ise, we f t, as did Doll and Peto, the data enclosed in rectangles in their Tables 2 and 3. We - used the mean number of cigarettes for each "cigarette-per-day" - -- - group given in their Table 2 antl assuaned, for each "cigarette- por-day" group, a variance •that was half the maximum possible variance. We then fit tho data in three different ways. First, we used the reported actual mean age of onset of cigarette smoking (19.2 years) as date of onset anil the means of the age groups defining the rows in Tables 2 and 3 its the age of the event. Secondly, we used age 22.5 years an date uf onset. Thirdly,•we used age 19.2 years as date of onset, but subtracted 3.3 years from the - - means of the age groups delining the age of the event. The first - and third methods both gave essentially the eatiniates reported above, while the second method gave R./p1 = 0.014, P1 =.3.42. The estimates based on the second method are not used in this -- append i x. For our second estimate we used an estimate of p./fis = 1.e, given by Brown and Ghu (in press), based on fitting this same rnul- tistago inodel to data from a large European case-control study. Brown and Chu found that p*/,0l =1.g (and Pl = 0.14) for individ- uals who smoked 21-30 cigarettes per day (see Table 3 of Brown and'Chu). (Brown and Chu find a ratio of 4 for P4/pl for smokers of 1-10 cigarettes per day. We did not use this edtiniate due to its presumed lack of stability.) Note that t the ratio of 1.8 found by - - Brown and Chu was 160 tinues that of Doll and 1'eto. The low - - - - - - -- p4//91 raLio in the Doll and 1'eto continuing smoker data does not appear to adequately account for the rapid decline in risk associ- ated - - -- ated with cessation of cigarette smoking as given in Doll and Peto (1976). This implied that the estimate of 0.0124 was probably too low. Furthermore, the estimate of p./.8r from the Doll and Peto continuing-smoker data was quite imprecise, since the correlation between the estimates of P1 and P1 was -0.93.. Based on these con- siderations, we computed a revised estimate of &/,61 from the Doll and Peto continuing-smoker data by finding the maximum value of ~./~. associated with a point on tlke 2 log likelihoo,l surface that lay 3.87 (chi-squared units) below the value of the 2 log likelihood

322 surface at its maximum above. At this point, the ratio of /i4//fl - had increased 20-fold to 0.2'l5 (and P1 a 0.8i1.3). In our sensitivity analysis, therefore, we used ratios of P./P1 equal to 0.0124, 0.2"5, and 1.8. (One might believe that if the estimate of #4/Ql, which (int+ would hope to be a biological constant, can differ by 150-fold acrt reH - data sets, Method 2 is useless. We actually are not so skeptical. If the sensitivity analysis shows that such large differences in estimates of 8./p1 have little influence olt our estimate of the - true relative risk in "exposed" and "unexposed" study subjeoats, - - this will indicate IL high degree of robustnepH (insensitivity) to tho actual model for lung cancer risk. Therefore, our confidence in tho -- estimates of the true relative risk may therefore be enhanced. An we shall see, we da) indeed find such robustness.) Remark 6: Reading the Exposure Histories front Table D-1 Each of our exposure histories can be representnd by a vector (a, b, e), where the value of a characterizes five possible population-expotiure histories from age 0-20 (a = 1, ..., 6), b characterizes two possible ex- posure histories from age 20-55 (b =+ 1, 2) tutd s characterizes three possible exposure histories from ages 55-70 (s = 1,2,3). Since we can select any of five exposure histories between ages 0 and 211, - - - any of two between ages 20 and 55 and any of three between 55 and 70, we have 6 x 3 x 2 = 30 exposure hbitorics. Each vahte of c gives an exposure history for "exposed" and "unexposed" subjoets between the ages of 55 and 70. The populatiun-exposure hisl.ory between ages 55 and 70 represented by a particular value of e is described by the (up to) six values entered in Table D-1. As an ex- - ample of how to read Table D-1, consider the case a= 3. Readiug Table D-1, we see that paa = 0.5, JloB S3dU, fga6 ° 2do,p.$ - 1.0, Jl"n = ide, and &g is undefined. By definition, paa gives the Irac- Lion of "exposed" individutJs exposed at rate J,an between ages 65 and 70. 1-pos in the fraction of "exposed" individuals exposed nt rate J3aa. Therefore, 50% of "exposed" indivuluald receive a Jone of ETS of 3de from 55 to 70 and 50% receive 1de. Sunilarly, 100% of "unexposed" individuals receive a dose of tde between ages 65 and 70. (Therefore, J,oe need not be defined.) Remark 7: Choice of 30 Exposure Histories ln choosing the ex-. posure histories, we rely heavily on data from the control series in Garfinkel et al. (1985), because similar detailed information is not available for any other study. We made the following assumptions. 323 1. All study subjects were age 70 at the time of the study. (In the Garfinkel et al. study (1985), the average age of cases and coiitrols was approximately 70.) 'L'he choice of 55 as the upper aee cutoff reflects the fact that among controls in Garfinkel - et al. who were ever-"e:cposcd" but not currentfy-"exposed," the median time since their smoking spouses either died or r quit was roughly 15 years. Thus, we chose 70 - 16 = 55 t18 the age at which ever-"exposed" lndividuals who are curreittly-"unexposed" ceased to be exposed to their spoui+es' cigarette sinolce. The choice - s= 2 rnpresents our best guess as to the actual Ii:TS exposure in the Garfinkel et al. study population between ages 55 and 70. The choice c= 3 assurnes the ever-"exlrosed" subjects who are currently "unexposed" receive the same ETS d«se, do, from age - -- 55 to 70 as never-"exposed" subjects Note that c = 1 represents a study in which all exposed individuals are currently- "exposed." 2. For exposure histories between ages 20 and 55, we assitme that all subjects were married at age 20. b= 1 represents a population in which Wald anil R-itchie (1SJ84) would have obtained the sami+ urinary cotinine meusurements had they performed their study in any year from 1932 to 1982. In contrast, G= 2 represents a situation in which from 1932 33 to 1967 never-"exposed" individuals were,exposed to ETS at a rate only 1t~% of that of currently- eunexposed" individuals in 1982 and ever-"exposed" individuals were exposed at a rate half that of a currently-"exl,osed" individual . in 1982.'Phus, b= 2 represents an extreme example of the situation discussed under Condition 6 of Retnark 4 with c(30170) = 10. The true exposure rates between ages 20 and Sli pre$umably lie between those represented by b= 2 and b- 1. 3. 1'Ve next consider ETS exposures between the ages of 0 and 20. 39% (25%) of "exposed" ("unuxposed") individuals in - -- the control series in Garfinkel et ai. (1985) reported that they were regularly exposed to- E'.l'3 in their homes cluring childhood (presumably because, in the large majority of cases, at least one of their parents was a smoker). These controls, who are on average ago 70, had to remember their parents' smoking habits over more - than 50 years. Therefore, sotne misclassification is unavoidable. - As a guess, we suppose that the false negative rate for parental smoking was 0.3 and the false positive rate was 0.15, independent - -- of "exposure" status. D_ eline pae to be the frar.tion of exposed controls with at least one smoking paretit. Then, correcting for - -- - misclassification, our best estimates of p„B and 1,",g are 0.44 and V9D4g44B

324 0.18, respectively (since 0.7 p"a + 0.15(1 -• p"g) = 0.39 implies paB = 0.44). For the exposure histories represented loy a= 1, we used'the uncorrected estimates 0.39 and 0.25, and otherwise used the cor- rected estimates. The uncorrected estimates were used in our sensitivity analysis because we do not know the true misclassilira- tion rates and, if false-positive and false-negativo rates depend tin "exposure" status, the true ratio of psa/p"g may be loss than or equal to 0.39/0.25. To develop estimates of ji, aud /2" we proceed - as follows.. Jarvis et•al. (1985) give mean salivary cotinine levels in children of 0.44, 1.32, 1.99, 3.39 ng/ml, depending on whether nou- ther parent smoked (2(39), only father smoked (96), only mother smoked (76), or both parents smoked (128). (The numbers in parentheses give the number of children in each parental smuk- ing category.) Now, Jarvis at al. conjecture that an average iw-- tive smoker would have a salivary cotinine level of aplnoximatnly 300 ng/ml. It follows that, as a rough approximation, the exlso- sure to ETS of a child with nonsmoking pai ents is approximately o.3de since Wald and Ritchie (1984) found the urinary cotinino levels of currently-"unexposed" individualti were approximately 1/200 that of an average active smoker antl 0.44 x 200/300 €f 11.3. (This result obviously depends on unverified nseumptions about the comparability of nicotine metabolism in adults and children.) Furthermore, the ratio' of ETS exposure in an average child with a smoking parent to an average child without a smoking parent in [(1.32)(96) + 1.99(7a) + 3.59(128)/(98+7s+ 12H))/U.44 = 6.1 (which is a dose of 1.53da = 6.1(0.3)dQ. Similar data from a study of Cuultan et al. (1986) give a ratio of 3.07 rather than 6.1, under the assump- tion that the fraction of children with two smoking parents amung - - children with at least one smoking parent is 128/(7d+56+ 128), as - in the Jarvis study. These results motivated the choice of jI, and JzQ for a equal to I and 2. One would• expect that in the 1920s children living in honies with no parents smoking might well have less 1,`t'S exposure than such children currently have (since in the 19209 fewer caretak-, ers, who were almost exclusively female, smoked). On the other hand, among children who lived in a home with a smoking parent, presumably a higher percentage had only a father who sntoked. (Data on this question was not available horn the Gnrfinkel at al. - (1985) control population.) Thus, the E'1'ti exposure in the 1U'l0s ~;9O4,B4419 325 of an average child with a smoking parent would also be less than that of a similar child in the 1980s. . Theser observations led to our choice of fl" and f2. for a a 4. On the other hauid, it is likely that, conditional on having had a smoking parent in childhood, "expased" individuals are more likely than "unexposed" individu- als to have had a mother who smoked. Furtherniore, it may well be that "exposed" individuals without a smoking parent had, on average, higher exposures in childhood than "unexpuaed" individ- uals without a smoking parent. (Recall that a hig-her percentage - of "exposed" individuals are known to report having at least one sm mg parent.) These observations led to our choice of the ex- posure histories characterized by u- 3. The final choice, a = 5, reflects the ratio of 3.07 found by Coultas et al. and the possi- bility that in the 1920s, when few women smoked, this ratio was even less. The maximum value for the ratio of the childhood ETS exposuro of "exposed" compared to "unexposed" subjects is 3.17 occurring when a = 3. All 30 exposure histories presume that the ratio of ETS ex- posure iui the 1970s and 1y80e in currently-"exposed" subjects to that in currently-"unexposed" subjects ie 3, as_ found by Wald and Ititchie. Thus, the sensitivity nnalysin may not be applicable to studies carried out in Greeee and Japan, for reasons discussed above (although it is possible that in recent 'years the exposure of Japanese nonsmoking wonien outside the honie has increased to United States' and British levels). Thus, one naight wish to use both 1.3 and 1.14 as the suminary observed_ rate ratio in a sensitivity analysis. Remark 8:. Estimating the 7Vue_ Retatiue Risk Under Asauwnp- tiona for Model B Abeve Consider a groupof individuals (i.e., the "exposed" individuals or the "unexposed" individuals in Garfinkel - - at al.'s (1985) study) such that each individual i hus a constant exposure to. ETS, dli, from age 0 to te, exposure dzi from age te to t„ and exposure d3i from age t, to t. The d1i, dz;, dgi may vary - between individuals in the group. Then, the true relative risk at age t for the group compared to a completely unexposed group, when exposure affects the first and fourth stages of a five-stage multistage model, is: RR(t) = 1 + AidulHi1 + p4du1!!v) -i PiP4do'(t1121, D-8

326 (rellccting the magnitude, where P1 and Pj are tinknown constants on a ratio scale, of the exposure effect on the first and fourth stage, respectively), do is av defined in Section D-2 and Hi(t) =(t4do)-11E(dl)t4 + IE(d2) - E(di)I(t - to)4 + IE(d3) - E(ds)I(t - t•)`1 H2(t) =(t4do)-lIE(ds)t4 + (E(Js) - E(da))t.4 + IE(di) ° E(da)Ito'1 H12(t) =(t1do'-)-1{IE(di)1'to4(l + m'(dl)) + E(dl)E(d,)I1 + p(dt, d2)m(d,)m(d2)I(j.' - 10' - (e, - te)4) + (E(dz)la(t. - io)4I1 + m2(d4)) + E(d1)E(d,)I1 + p(dt.d3)m(d,)m(d3)I(t` - t.' - (t - to)' +(t.-to)'I + (E(ds)1'(t - t,)'ll+ m'(d,)) + (1-F- p(da, da)rn(ds)ln(da)II(t - to - (t - t.)'IE(d9)E(d3)} (t, - t Q)4 where E(di) is the average of dl,m(dl) atid p(dl, dz) is the correlation between dl and di. For simplicity, we shall assume that all correlations are 0. This will have little effe~a on our analysis. Now define Fl(t) ffi Ht(t)'k (A4/pi)112(t) and F1a(t) - (P4//li)»is(t)- Then we have: RI~(t) = 1 + ptduFl(t) + pida)%(t). (0-9) Now with to = 20, t, = 561 and t- 70, for any given choice for /1 /Pl and for the exposure vector (a,b,c), we can cOmllute Fl(70),F»('10) for both "exposed" and "unexposed" groulxt. Since 1.3 is assumod to be the ratio of the true relative risk in "exlkosed" subjects to that in "unexposed" subjects at age 70, we have 1 + (pid~JFis(70 + Ltdo'Fitis(7~ )) (1)-10) (#idu)Fis(70) +Vtdo)2i`'t•,s(7U) Equation D-10 is a quadratic equation in plilo. Thus, we ran solve for pldo eveu though we do not know Jilor du seloarately. We 99OLS449 327 then eubstitute this value of Pldo along with the values of Fig(70) and F19s(7u) into Equation D-9 to give an exposure-history-P4/fi1- specific estimate of the true relative risk at age soventy in "ex- posed" individuals. If we substitute Flg(70) and F12,9(70) instead, we get an estimate of the trup relative risk at age 70 in "unexposed" individuals. Note that if we had an independent estimate of fi1, we could also estimate do. Given Q4//11 and (a, 6, c) (and thus fildo by Equation D-10)i our estimate of do is inversely proportional to our estimate of #1. Estimation of do Remark p: InterpretaEion ojfll and do fi1, when estimated from data on active.amokers, is the fractional increase in the rate of the lirst cellular event per actively smoked cigarette. Since cigarettes - - -- dill'ea in carcinogenic potency, neither pl nor do are biological con- stants. Therefore, we must specify the typo of cigarette to which - - we want our ea ' timate of /11 to refer. in this Appendix, we shall lot /tl be the increase in the rate of the first cellular event associated --- - - with one current nonfilter U.S. cigarette containing 20 mg tar as -- -- --- smoked by an average U.S. citiaen. In Set-tion D-2 we adjusted - - - - cour esttmates of pl from the Doll and Peto data (1978) with this tlefinttton of 81 in mind. Even after adjustment, fil will still be - - - - defined in terms of the cigarettes smoked by the study subjects - - -- --- _-- iu the Amarican Cancer Society (Hammond, 1966) and European -- - -- - ---- case-contrul studies (Lubin, 1984), which, on average, contained -- --- - - titore than 30 mg of tar (since most of the cigarette exposure in these studies occurred before the adoption of low-tar cigarettes). Thus, if we wanted to define Pt In terms of actively smoked un- - - Illtered cigarettes with a tar content of 20 ing, one might further divide all estimates of 01 (and multiply all estimates of do) by a factor of 1.5 to 2, although we have not chosen to do so. One tnust still consider the possibility that the lower relative risk found in the European and ACS data compared to the British data is tt consequence of the fact that there was loss misclassification of stnokers as nonsmokers among the British doctors thau among the ACS or European case-coutrol study populations. Since, presunt- ubly, doctors are accurate reporters, such an assumption may not Ile uurealie{tic. If so, the baseline rate among nonsmokers from the ACS study would be falsely inflated upwards and the values of P1 cif 2.03 antl 0.803, as originttlly estimated for the Dritish doctors,

328 would be the more appropriate values to use. For these reasons, we report results for all five of the combinations of ds%al autl Rl given in Table D-3. Remark 10: Adjusting for the ETS Exposure of Active Smtikers In estimating Ql and P4 from active-smoker data neither wo nor Brown and Chu (in press) took account of the fact that in those studies active smokers (and the comparison groups of nonsmokers) were themselves breathing other peoples' cigarette smoke. If 3ite is of the order oP 3 or more cigarettes per day (as in Table I)-a), a proper analysis (and thus proper estimates of /91,/9., antl de) would require relitting the active-smoking data taking account of ETS exposure. We have not done so here. We expect that the effect on our estimates of the true relative risk in "exposed" and "unexposed" subjects using.Method 2 would not be great (bernuse of the insensitivity of these estimates to uncertainty in pr/P.). On the other hand, the effect on our estimates of do may be more pronounced. Further study is required. Remark 11: Estimateon oJdo from Dosimefry The estimates of do given in Table D-4 are obtained in step 5t of the following sequence of calculations. 1. For the ETS constituents BeP and NDMA we estimated` the weight of each constituent inhaled directly by an active sntoker from the mainstream smoke of a single cigarette by using the mid- point of the range given in the mainstream weight eohmin in Table 2-10 (i.e., 25 and 30 ng for NDMA and BaP, respectively). (The weights entered in the mainstream weight column of `I'able 2-10 are averageit based on cigarettes whose maiustream-smoke tar content, as meauured by a smoking machine, varied between 16. and 30 milligraius.) 2. We estimated the weight of each of the above constitnents inhaled daily by a nonsmoker with a nonsmoking spouse by mul- tiplying by 1.07 the range of values given under the ETS weight column in Table 2-10. (1.07 is our estimate of the average num- ber of hours of (aily ETS exposure occurring in nonsmokers with nonsmoking spouses. Nonsmokers without smoking spouses retport that they ere exposed, on average, to C1'S between 5('lhldo 6, Friedman et al., 1983) and 10 hours a week (Wald and Ritclsie, 1984). Our value of 7.5 hours/week (= 1.07 hours/day) in the average of the above estimates. We could have chosen to wultiply 829 the value of 1.07 by a factor of up to 2, since most components of - ETS decay with a half-life of approximately i hour when smoking ceases, assuming approximately one air change per hour and little plating out onto surfaces.) 3.111or each constituent we divided the endpoints of the weight ranges calculated in Step 2 by the weight estimated in Step 1. The resulting range of values is, for each constituent, an estimate of the number of cigarettes that would have to be actively smoked in order that the weight of the constituent in the directly inhaled mainstream emoke would equal the weight of the constituent (at- tributalrle to ETS) inhaled daily by an average nonsmoker with a nonsmoking spouse. We shall call this number I„m. 4. We next estimated for each constituent the number of cigarettes whose mainstream smoke would have to be directly inhaled by an active smoker to deliver to the lungs a dose of the constituent equal to the daily (biologicidly effective) pulmonary dose (attributable to ETS) of a nonsmoker with a nonsmoking spouse. We refer to this number as do,,,. For BaP we multiplied the endpoints of the range for leir, by one-seventh. This reflects - the fact that BaP is in tGe particulate phase antl, as discussed in Chapter 7, a rough estimate uf the deposition rates for particulates in ETS and in mainstream smoke is 10% and 70%, respectively. -- [This calculation ignores important differences between the ETS - and mainstream particulate phases in terms of deposition site, clearanco rates, and particle size. Thus, even if 13a1' were the ac- tive carcinogen in ETS and mainstream nmoke, do,.,, as calculated above, could conceivably be quite different from the true value of --- - do,. deffiied in terms of the biologically eflisctive dose for producing lung cancer.] For NDMA we assumed da,,, = Io,". The rationale for this decision is that NDMA is iu the vapor phase in both_ ETS and mainstream smoke. We therefore aseumed that the pulmonary absorption of NDMA per nanogram inhaled was the same for mainstream smoke and LTS. (This assumlition may be inadequate, since Nl1MA is water soluble and thus will dissolve in mucous _ _ ---- membrauees before reaching the lungs. Therefore, the fraction of inhaled NDMA that reaches the lungs niay well be up to several orders of magnitude greater in active smokers (whose intake is via deep inhalations taken through the mouth) than in nonsmokers (whose ititake is largely via shallow inhalations taken through the noae).. If so, our estimate of dom would need to be reduced by f~9d4,R44e

330 the appropriate factor. We have not made any such adjutitment here. deiri for RSi' was calculated as follows. In Chaptor 7 it was calculated that the amount of tar deposited in the lungs after 8 hours of E'1'S exposure would be about 0.00M0.'l6°6 of that deposited in the lungs of an active smoker of 20 cignretten containing 14 ing tar each. Thus, the upper limit of the range for der,, (in terms of 20 mg tar cigarettes) equals (14/20) x 0.26 x 10'2 x 20 x 1.07/8 = 8.2 x 10-a = 0.005. . The total range is o.0001= 0.005. 5. In what follows we estimate for each of the constituents - NDMA, BaP, and lt6P the number of cigarettes that wouhl havo to be actively smoked to deliver to the smoker a (biologically effective) pulmonary dose of the constituent equal to the daily pulmonary dose (attributable to ETS) of a nonsmoker married to a nonsmoking spouse. This number we will call da. 'rhe + as a symbol serves to distinguish this definition of do from that in Section D-2. du for a given constituent is equivalent to de as defined in Section D-2 if, as assumed in Table D-4, the constituent is the active lung carcinogen in ETS and mainstream smoke or, moro generally, if do for the constituent is equal to do for the unknowii active carcinogen. For the constituents RSP, BaP, and NDMA we first estimated the difference between the total pulmonary dose attributable to a single actively smoked nonfilter cigarette and the fraction of that pulmonary dose attributable to the directly inhaled mainstream smoke. This dilferenee includes contributions from the plume of -- sideatream smoke, the plume of exhaled mainstream smoke, and - - the ETS subsequently derived from the plumes of sidestream and exhaled mainstream smoke. We shall call this difference the non- - - mainstream (pulmonary) dose of the constituent. How does the magnitude of the nonmuinstream (pulmonary) dose to a sinoker - - - compare to the pulmonary dose of the constituent absorbed by a nonsmoker without a smoking spouse in the Wald and ltitchio - study (1984) during that nonsmoker's 1.07 hours of daily expo•• sure? We have no empirical data that directly bear on this ques. - tiou. Nonetheless, we shall assume that the ratio, /, of the dose to the smoker from the nonmainstream smoke of a single cigarette to - the daily dose (attributable to ETS) to n nonsmoker with ,t, non• smoking spouso is between 0.1 and 2. We believe the ratio could be as high as 2 because the active smoker is much more lik„ly tn directly inhale the highly concentrated Plumns of sid_ estream and exhaled mainstream smoke. (In fact, the ratio could possibly he a 331 good deal higher than 2.) This ratio could be as low as 0.1 if active smokers rarely directly inhale the plumes of smoke and during the - - hour in which a nonsmoker with a nonsmoking spouse is exposed to ETS, the average smoker density is 4, with each smoker smok- ing 2.5 cigarettes per hour. (This is a rather high smoker density and 0.1 may therefore be somewhat tot) low an estimate.) It is -- - - a straightforward algebraic exercise to show that the relationship botween do and der,, is , _ 1 do (lyd,m+ f ~ The minimum of the range of do (equivalently, du) values given in Table D-4 (for each constituent) was computed by plugging into the above formula the mininium of the range of de,,, estimated in step 4, and f= 2. The maximum of the de range in 'rable D-4 was computed by plugging in the maximum of do,,, and f= 0.1. The - ranges calculated for d;, essentially equal those for do,,,, with the exception that both endpoints of the dD,„ range for NDMA were reduced by approximately 40% and the upper cndpoint for BaP was reduced 25%. Remark 12: Dosimetry Based on Urinary Nicotine or Cotinine In this remark we consider whether it is reasonable to take the ratio - - of urinary nicotine (or cotinine) in nonsmokers to that in active smokers as a proxy for the ratio of the biologically effective dose (attributable to ETS) of the active lung carcinogen in nonsmokers to the biologically effective dose hi active smokers. In aged ETS, nicotine is largely in the vapor phase. Nico- tuie is water soluble. Thus, presumably most of the nicotine in aged E`1'S dissolves in the mucous membranes of the upper air- ways and diffuses directly into the bloodstream. Thus, little of the inhaled nicotine from aged ETS reaches the lower respiratory tract. Therefore, urinary and blood nicotine in nonsmokers should roughly reflect the total amount of inhaled nicotine. In contrast, nicotine in mainstream and sidestream smoke and in fresh ETS is largely in the particulate phase. Therefore, inost of the nico- tuie directly inhaled in mainstream smoke by a s,noker reaches the lower respiratory tract (and from thei e the bloodstream) since the deposition fraction for particulates in mainstream smoke is 70% with most deposition occurring in thm lower respiratory tract.

332 Therefore, if (1) the true carcinogen is in the vapor phase in both - ETS and mainstream smoke, (2) the true carcinogen is in the particulate phase in both ETS and mainstream smoke, or (3) the true carcinogen is in the particulate phase in mainstream smoke, the vapor phase in ETS, and is, in addition, water soluble (so that the total dose of the carcinogen from ETS greatly exceeds - the pulmonary dose), then serious questions must be raised about the appropriateness of the ratio of urinary nicotine (or cotinine) in nonsmokers to that in active smokers to approximate the ratio of the biologically effective lung dose of the active carcinogens in - nonsmokers to the lung dose in active smokers. Remark 13: Estimating Lung Cancer Deaths Attri6utable to ETS Among Lifelong Nonsmokers in 1985 As in the Garlinkel et al. (1985) study, we use "expoeed" to mean ever-"expoeed", since one cannot calculate a population attributable number from case-l control studies in which individuals who are ever-"exposed" but - not currently-"exposed" are excluded. If we assume t<hat Aesuml;- tion la and Equation D-5 hold with c(70) = y, then RR(701E) and RR(7011:) are 1.54 and_ 1.18, respectively, based on a sumniai y rate ratio of 1.3. We would then need to assume, for examploo, that RR(tIE) and RR(tjS) do not depend on 1. Using this ap- proach we obtain an attributable number of 2,010 in nomsmokiug women. In contrasl., the naive approach, which ignores the ]:TS exposure of "unexposed" individuals by assuming 1tR(701E) = 1.3 - - and RR(701E) a 1.0, givcs an attributable nuniber of 1,150. The second approach supposes that assumptions of Method 2 in Section D-2 hold. We then choose a value for exposure history (a, 6, c) and P./,81 which, given that q(70jE)%,y(70jR) = 1.3, allows us to calculate pido from Equation D-10. Kuuwledge of /lida, theii, - - allows us to calculate, from equation D-9, 11R(t1E) aud RR(tjF.) for all t (not just.t ==70). The third approach is to assume that the first assumptions under Method 2 concerning the multistage cancer niodul hold but not to assume that RR(70jE)/RR(70jE) a 1.3. We thon unust select a value of /41 and d„ in order to estimate Pli/e and, given (a,6,o) and d4/#i, RR(tIE), Remark 14: Estimating the Lifetime Risk of Gung (.ancer Due lo - ETS .4(tlte) ~ ,o(1-A(u)) where a(u) is the all-cnuso mortality rate in 1085 among tionsinokiug women of age u(and we are follow- ing the standard convention (if using current, i.e. 1985, mmortality 333 rate). We estimated A(u) for feniale nonsmokers by multiplying the all-c-ause age-specific mortality rates (for feniale nonsmokers) given in Hammond (1966) by the ratio of the overall U.S. age-specific fe- male death rates in 1985 (all smoking categories) to those rates in 1962. Furthermore, rysxoess(t) = ryo(t)(RRKxaass(t)I, where rya(t) is the incidence of lung cancer death at t in the absence of all ex= posure, and RRsxe$ss(t) is the excess relative risk for lung cancer due to exposure history M: ryo(t) _ 11 -. AF(t)IIa(t) where AF(t) and Io(t) are as defined above. From lilquation D-9 wo can obtain an estimate of RRaxoEss (t) _ 12R(t) -1 for given values of #1,1o and 04/p1 and choice of exposure history 1N. It follows that, under the assumption that the observed rate ratio of 1.3 is causal, we can then obtain an estimate of AF(M) for te = 45 for each choice of exposure history (a, 6, c) and value of JB./101, since, using Equation D-10, we obtain an estimate of flldo from which, in turn, we obtain an estimate of AF(t) and 1tRaxosss (t). Remark 15 In estimating AP(M) in ex-- and current smokers, --- RR$xo$ss(t) can be estimated froui Equation D-9 for a given value of exposure history (a, 6, c), /14//91) and Rl under the as- --- sumption that the rate ratio of 1.3 is causal. (Ii;nowledge of dl is necessary so that we can estimate from_ Ekluation D-10 the value of de rather than simply pldo.) ye(t) is estimated as for nonsmok- -- t,rs. Zb eitimate. the all-cause mortality rate among examokers -- - and continuing smokers we used the data in Hammond (1966) •as described for nonsmokers, except for smokers of 20 cigarettes per ' ,lay we used an average of the age-specific all-cause mortality rates in Hammond for smokers of 1-19 and >19 cigarettes per day; and for exsmokers we used both their smoking rates while smoking - and the number of years since quitting (as a time-dependent co- variate) to enter Hammond's table at the proper place. Missing values in llammond's table were filled in by linear interpolation or extrapolation. Remark 10 Under Assumption la, RRgXQSSS(t) would be the xame for exsmokers and nonsmokers who had the same history of exposure to other people's cigarette smoke. But if we assume that cigarette smoke affects two stages of a multistage cancer model, Lhen, for an exsmoker, the quadratic terms in Equation D-9 cannot Iie ignored. As such, a sniall increment in dose due to breathing other people's cigarette smoke will have a larger absolute effect G904RL48

334 on the age-specific-mortality rate of the exsinoker than of the nonsmoker. Effects of Bias We now consider the following three qiiestions. In deriving our - - - summary estimates of 1.3 we amalgamated studies that compared -- ever-"exposed" to never-"exposed" subJectn with studies that com- pared currently-"exposed" to never-"expoi~ed" subjects. Does this - introduce an important bins? In Remark 17 below, we show that it does not. Second, under the assumption that our multistage model is correct, Assumption la is false, since Equation D-8 him a quadratic dose term. Nonetheless, for calculating the true relative risk in "exposed" and "unexposed" subjects, is Assumptioii la an adequate approximation? Third, should cas+e-control studio+s of the relationship between childhood ETS exposure and lung cau- cer have greater power to detect an ETS effect than case-c:ontrol --- studies of adult LTS exposure? In particular, does the failueo of Garfinkel et al. (1985) to find an effect of childhood exposure cast - - doubt on the validity of our 13 epidemiologic studies of adult 1,I'S -- - exposure? We will show in Remark 19 that when one takes into account the inevitable misclassification of childhood 1:TS exponure occurring some 60 years previously, the oliserved relative risk k ex- pected pected from a cane-control study ofchildhaod ETS exposure could be as low as 1.01 and would be no greater than 1.3. Thus, it is not surprising Garfinkel et al. found no effect 4 childhood exposure. Remark 17 It is clear that the same causal Itarameter is not being estimated in studies in which the "expesed" group is ever- "exposed" as in studies in which the exposed group is curreistllv- "exposed" individuals. Yet, our summary value of 1.3 was based on amalgamating estimates of RR(tJE)/RR(tJE`) from these two dif- ferent types of studies. To estimate the mal;nitude of the bias ioKso- ciated with this amalgamation, we proceeded as follows. Goneoider - -- studies with exposure history of the form (a,b,l). For each chuice of (a,b) and P4/pl we obtain, from Equation 1)-iU, an estimate of d d, o, say, Pldo(c,,b,p4/Q1), if we can assuine 1tR(tlk:)/RR(tll%) is 1.3 for such studies. For each.8ldo(a,b,P.//fi) we estimated, utibig Equation D-10, 1tR(tJE)/1ER(tJE) for a stitdy with faxposure his- tory (a,b,3). The maximuni value of RR(t1t(tIE) estimated in 016140494,4413 335 this way for studies with exposure history (a, b,3) was 1.39 (asso- --- - - ciated with ~B4/,81 = 1.8, of course). Given the confidence interval of (1.12,1.49) reported in Chapter l2 for the amalgannated param- eter RR(tJE)/RR(tIE), it follows that any bias due to improperly anialgainating these two types of studies will be small compared to sampling error. -- - Remark 18 Conditional on the assumption that our multistage model holds for lung cancer, we can test the adequacy of As- sumption la. Let 8ldo and RR'(7o1E) be the estimates of #lde apd RR(701E) obtained by removing the quadratic term (in fildo) frum the numerator and denominator of Equation D-10. Now, since &. luation D-9, modified so that the quadratirc term in #ldo is eliminated, is a linear excess relative risk model, it follows that Assumption la is an adequate approximation if the estimates fildo and RR'(701E) do not differ greatly froni the estimates, 81do and RR(7011,,'), based on the unmodified Equation D-10. We therefore estimated max(RR'(701E) - RR(7111E)i av (a, b, c) aud P,/fi1 var- ied. The maximum was 0.05. Thus, the linear approximation of Assumption la is probably adequate. Remark 19 We now estimate the maximum and ininimwn rela- tive risk (at age 70) we would expect to observe in a case-control study of ETS exposure in childhood (controlling for ETS exposure in adult life) under the assuiuption that our multistage model for lung c-aitcer is correct. `1b do eo, we perform a sensitivity analysis over the possible exposure histories of the "exposed" and "unex- posed" Htudy subjects in such a case-control study. In particular, we assume that (1) for all study subjectti the exposure rate from ages 20 to 70 years was 2d,,; (2) the false-positive and false-negative rates for the exposure "at least one parent smoked" were 0.16 and 0.3, respectively; and (3) exposure rate from 0 to 20 in the truly "exposed" (i.e., among those who did have a anwking parent) tQ --- ,_ the truly "unexposed" was, in units of do, one of the following: 1.53 to 11.3, 0.75 to 0.15, 1.0 to 0.6, or 1.0 to 0.05. It only remains necesaary to choose values for #4/p1 and P1de. For each of our three choices of P4/,81, we let /)1rla range over the values found previoutily (using Equation 1)-10) as (a,b,c) varied. -- - - The maximum relative risk was 1.26, which occurred with - - - ---- - - exposure rates of 1.53 aud 0.:1do in the exposed aud unexposed, re- spective•ly,,84/p1 = 0.0124, and the value of #ldo (cumpute,d using &luatiun D-10) based on (a,b,e) == (1,2,3). The minimurn relative

336 risk was 1.01. Even when we unrealistically assumed that hoth the false-positive and false-negative rates for exposure misclneei- fication were 0, the ntaxintum relative risk was only 1.51. Titus, it is not surprising that Garfinkel et al. (1985) failed to detect an effect of childhood exposure in his case-control study. References Brown, C.C., and 1f.Q. Chu. Use of multistage models to Infer stage - - - - - - - affected by carcinogenic exposure: Example of lung cancer and cigarette smoking. J. Chron. Dis., In press. Correa, P., L.W. Pickle, E. Fontham, Y. Lin, and W. Haensael. Passive - - - smoking and lung cancer. Lancet-2:696-697, 198:I. - - Coultas, D.B., J.M. Sanrot, C.A. Howard. a.T. Peaks, and B.J. Skipper. Salivary cotinine levels and passive tobacco nnoke exposure In the home. Am. Rev. Respir. Dis. 133:A157, 1988.. Day, N.E., and C.Cl. Brown. Multistage models and primary prevention of _ _ cancer. J. NatL Cancer Inst. 84:977-989, 1980. - Doll, R.D., and R. Peto. Mortality in relation to nnoking: 20 years' - - observations on male British doctors. Br. Med. J. 4:1535-1536, 1976. Doll, R.D., and R. Peto. Cigarette smoking and bronchial carcinomat Dose and time relationships among regular smokers and lifelong non-smuker.. J. Epidemlol. Comm. Health 32t303-313, 1978. Doll, R.D., and R. Pets. The E3auses of Cancer: Quantitative Estimates of - - - Avoidable Rieks of Cancer in the United States Today. J. Natl. Cancer lnst. 88:1191-1308, 1981.. F_ riedman, Q.D., D. Pettiti, and R.D. Bawol. Prevalence and correlates of passive smoking. Am. J. Public Health, 73:A01-405, 1983. ElarBnkel, L. Time trends in lung cancer mortality among nonsmokers and • note on passive smoking.. J. Natl. Cancer Inst. 88:1081-10Q8, 1981. Garfinkel, L., 0. Auerbach, and L. Joubert. Involuntary smoking and lung cancer: A case-control study. J. Natl. Cancer I:ut. 75:483-489, 1985. Hammond, B.C. Smoking In relation' to the death rates of one million man and women. Nat. Cancer Inst. Monogr. 19:127-204, 1966. Hirayama, T. Cancer mortality in nonsmoking women with smoking hushands on a large-scala cohort study in Japan. Prev. Med. 13:880-890, 1984. - - - - - Hirayama, T. Non-smoking wives of heavy smokers have a higher risk of lung cancer: a study from Japan. Hr. Med. J. 282:183-185, 1981. Jarvis, M.J., M.A.H. Russell, O. Feyerabend, J.R. Eiser, M. Morgan, P. Gammage, and E.M. (3rRy. Passive exposure to tobacco smoke: Saliva cotinine concentrations in as representative population sample of non- smoking schoolchildren. Br. Med. J. 291:927-929, 1986. smoking Jarvis, M.J., H. TLnstall-Pedos, 0. Feyerabend, 0. Vesey, and Y. Saloojee. Biochemical markers of smoke absorption and self reported exposure to passive smoking. J. Epidemlol. Comm. Ilenlth 98:366-339, 1984. Lubin, J.H., W.J. Blot, and F. Berrino, et al. Patterns of lung cancer risk among filter and nonfilter cigarette smokers. Int. J. Cancer 33:88U-678, 1984. 337 Matsukara, S., T. Taminato, N. Kitano, Y. Seino, H. liamada, M. Uchihashl, 11. Nska3ima, and Y. Hirata. Effects of environmental tobacco smoke on urinary cotinine excretion in nonsmokers. N. Engl. J. Med. 311:828-832, 11184. T<ichopouloi, D., A. Kslandidi, and L. Sparros. Lung cancer and passive nnokinE: Conclusion of Greek study. Lancet 3:877-878, 1983. Wald, N.J., and Q. Ritchie. Validation of studies on lung cancer in non- smokers married to smokers. Lancet 1:1067, 1984. W.QLg4449