Tobacco Documents Online

ORIGINAL UNITED STATES ENVIRONMENTAL PROTECTION AGENCY SCIENCE ADVISORY BOARD INDOOR AIR QUALITY AND TOTAL HUMAN EXPOSURE COMMITTEE - - - - - - - - - - - - - - - - - - -X In the matter of: . ENVIRONMENTAL TOBACCO . VOLUME I SMOKE REVIEW PANEL : ---X Tuesday, July 21, 1992 Main Ballroom Holiday Inn 4610 North Fairfax Drive Arlington, Virginia 22203 The ENVIRONMENTAL TOBACCO SMOKE REVIEW PANEL MEETING of the Indoor Air Quality and Total Human Exposure Committee of the Science Advisory Board was convened, pursuant to notice, at 9:15 a.m. APPEARANCES: MEMBERS: DR. MORTON LIPPMANN, Chairman DR. JAN A. J. STOLWIJK, Vice Chairman DR. JOAN DAISEY DR. VICTOR G. LATIES DR. PAUL LIOY DR. JONATHAN M. SAMET DR. JE'ROME J. WESOLOWSKI CONSULTANTS: DR. DELBERT EATOUGH DR. S. KATHERINE HAMMOND DR. GEOFFR9Y KABAT DR. MICHAEL D. LEBOWITZ DR. HOWARD ROCKETTE DR. SCOTT'T. WEISS ) o4nxEtican cR~o~u 2~e I202i 296-0261 "i/-4 17031644-7636 4

APPEARANCES: (Cont.) SCIENCE ADVISORY BOARD STAFF: MR. A. ROBERT FLAAK MS. CAROLYN OSBORNE EPA STAFF: DR. DON BARNES DR. JENNIFER JINOT DR. FARLAND • DR. STEVEN BAYARD OTHER APPEARANCES: DR. FERNANDO MARTINEZ DR. KENNETH BROWN DR. JUDSON WELLS SPEAKERS: DR. PAUL SWITZER DR. MAXWELL LAYARD DR. PETER LEE DR. STEPHEN SEARS DR. GIO GORI DR. JOHN TODHUNTER DR. PHILIP WITORSCH DR. GARY GIOVINO DR. JOHN BANZHAF DR. ELIZABETH FONTHAM DR. RON DAVIS DR. CLARK HEATH DR. JEFE WAGENER DR. MICHAEL OGDEN o~{me~.Ecan eRe~ioa.tezs ZC 1ZoZi Z96-OZ6, e1/"q 17031644-7636 1-2

1-3 C O N T E N T S PAGE Welcome and Opening a 5 Overview, Introductions and Disclosure 11 Overview of Draft Document Presentation by Dr. Farland 30 Presentation by Dr. Steven Bayard 38 Committee Review of Appendices A - E Appendix A - Reviews of Epidemiological Studies on ETS and Lung Cancer 46 Appendix B - Method for Correcting Relative Risk for Smoker Misclassification 53 Appendix C - Review format for case-control studies 71 Appendix D - Lung cancer mortality rates attributable to spousal ETS in individual epidemiologic studies 75 Appendix E - Statistical Formulae 77 Committee Review of Chapters 4, 5 and 7 80 Chapter 4 - Hazard Identification I: Lung cancer In active smokers, long-term animal bioassays, and genotoxicity studies 81 Chapter 5 - Hazard Identification II: Interpretation of epidemiologic studies on ETS and lung cancer 92 Chapter 7 - Passive smoking and respiratory disorders other than cancer 109 Public Comment Period 145 Presentation by: 146 Dr. Paul Switzer, Tobacco Institute Ln Dr. Stephen Sears, R.J. Reynolds Ln 185 Dr. Maxwell Layard, Tobacco Institute ,`,°, 158 m I-A OD P~112EZCCR/2//112EZCCR/2 <=REliOZtE4s. N 2~e /ZOZ1 Z96-OZ6, Nc4 /7031 644-7636

1-4 CONTENTS: (Cont.) PAGE Dr. John Todhunter, Tobacco Institute 214 Dr. Gio Gori, Tobacco Institute 203 Dr. Philip Witorsch, Tobacco Institute 226 Dr. Peter Lee, Tobacco Institute 168 Dr. John Banzhaf, Action on Smoking OR Health 242 Dr. Ron Davis, Coalition on Smoking OR Health 269 Dr. Clark Heath, Coalition on Smoking OR Health 278 Dr. Jeff Wagener, Coalition on Smoking OR Health 283 Dr. Gary Giovino, CDC 238 Dr. Elizabeth Fontham, Speaking on own Behalf 255 W OTriLE'ZGCQ/2 GRE#o'LtEti 2$c tZOZ1 Z96-0Zbl eUrIq t7031 644-7636

1-5 1 2 3 4 5 6 8 9 10 11 12 13 14 19 20 21 22 23 P R O C E E D I N G S (9:15 a.m.) MR. FLAAK: I'd like to welcome everybody here today and we're going to start our meeting now. Dr. Lippmann will be starting off the meeting in just a moment. WELCOME AND OPENING THE MEETING DR. LIPPMANN: On behalf of the Indoor Air Quality and Total Human Exposure Committee of the Science Advisory Board, I welcome you to this review of the EPA risk assessment document entitled Respiratory Health Effects of Passive Smoking: Lung Cancer and Other Disorders. Bob is with us this year. Last year he might have used his broken shoulder to avoid another assignment. He Pa r4; 4.t-f2. left us to d3ss#pate in the Desert Storm exercise, and Sam Rondberg of the SAB staff filled in ably for him. This year, however, it's up to Bob to help us pull together our deliberations and get out this Committee's report on the review of the document in good order. I hope he doesn't use his bad right arm as an excuse for any delay in getting this done, but I'm sure he won't. This, as I indicated in my opening remark, is a second meeting on a document of this title, or a similar title. We had a meeting previously in which we looked at a N OD A* cn2E2CCR/2 GJ`E#o%EEZS. ZC /zozf z96-o26Y q~vq 0031644-7636

1-6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 first draft, gave our comments and suggestions for improvement. The Agency decided that it was sufficient grounds to do a major rewrite, which they've done. They've made a very serious attempt to address the concerns we identified and have come back with a document that clarifies their position on the state of the science in lung cancer amongst spouses in relation to ETS exposure, and perhaps have done considerably more writing concerning the relationships between ETS and respiratory effects in children, something that we had specifically asked them to do. So this Committee, largely the same committee that participated in the previousf4rround, will be offering comments and suggestions on the adequacy of this current draft for the purpose of the risk assessment. We're fortunate that Dr. Scott Weiss, who was a corresponding member last time, unable to be with us then, is here today to be a full participant. We have added Dr. Paul Lioy to the panel. He's just coming in now, so this is his first time as a reviewer on this issue. Two people who were present the last time had schedule conflicts and could not be with us this time. Dr. Blot and Dr. Benowitz regret that they will not be physically o4inEZF.can :~Re#ozfeas ZC 12021 296-o26, ~V04 /7031 644-7636

1-7 ) 1 2 3 4 5 6 8 9 10 11 12 13 14 19 20 21 22 23 here, but they have provided us with extensive written commentary and in the appropriate places I will read some of their commentary so that we can benefit from that. Two of our members are not here today: Dr. Larson, a recently appointed member,to the Committee; and Dr. Woods, who was a participant last time, are today participating in a review'of the engineering lab's work on indoor air. There was an unfortunate overlapping of interest, and some of our engineering members of this panel are also members ad hoc of the engineering committee's panel on research on indoor air issues. I, myself, spent yesterday with the engineering panel, along with Drs. Larson and Woods, but I needed to be here today and left our participation in that panel's work in the capable hands of Woods and Larson for today. That work will be finished today and they will participate with us tomorrow in the review. And the fact that we're discussing the chapters out of order has something to do that. Their primary review assignments will be covered tomorrow. I will turn the position of speaker for the moment over, first, to Don Barnes, the Staff Director for the EPA Science Advisory Board, and then to Bob Flaak, the Designated Federal Officer for this particular Committee. Don? <vqinEZIcan -.-=-Re#o%EEas 25C /zoz/ zqb-oz6i q'6q /703/ 644-7636

t ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-8 DR. BARNES: Again, on behalf of the EPA and the Administrator, I'd like to welcome everyone to this review. I'd like to say just a few words putting this review in context. , First of all, I know that some of you in the back are having difficulty hearing. Just as it is in church, there are plenty of seats up here in the front pews, if you'd like to come forward before the collection, we'd be more than glad to have you. Please don't wait for the first hymn, just please come forward. The Science Advisory Board was established by Congress in 1978. The purpose of that particular legislation was to make sure that EPA had some outside review of the technical documents that supported EPA regulations. So we always like to say that our rruare on the Science Advisory Board is to provide critical review of the technical underpinnings of EPA's positions. Our role is.not to comment on those positions themselves -- that is the risk management or policy positions which are built upon their technical underpinnings -- but to look at the technical basis for that material. It is then up to the risk managers, it is up to the administrators, it is up to the EPA officials to go ahead and then structure the ,V7I72EZCCQI2 G!`E/2o'LtEY, 25C I202l 296-02b, q'04 [7031644-7636 Ln ~ cn W N OD 1i

1-9 1 2 3 4 5 6 8 9 10 11 12 13 14 19 20 21 22 23 policy on top of that. So our purpose here today is not to examine what EPA is going to do about environmental tobacco smoke. our purpose here today is to look at the analysis that they have conducted upon which they will then build that policy. The Science Advisory Board is used to doing these sorts of reviews. As I say, we've been doing that for about a decade and a half now. The total Science Advisory Board consists of about 80 members. All of these are people who are outside of EPA. They're all outside of the-Federal Government. They are scientists and engineers from across the country drawn from academic institutions, groups that are affiliated with industry, people who have affiliations with environmental groups. Our only concern is we have people who are technically qualified to review these issues. Also, we conduct our operations under the Federal Advisory Committee Act, and that Act explicitly says that the panels we put together should be balanced, balanced in terms of range of technically viable ranges of technical opinion on the issues before us. What you see before us today is one of 10 committees into which those 80 members are divided. We want to refer to the Environmental Engineering Committee which is cqmEZfcan IrRE#ozteu. 1~. C /zozf z96-oz6r 4~ 17031 644-7636

I-10 2 4 10 11 12 13 14 15 16 17 18 19 20 21 22 23 meeting today in Research Triangle Park. This is the Indoor Air Total Human Exposure Committee which is meeting here today. There are eight other committees which meet on a periodic basis throughout the year. Various committees of the Board meet roughly between 50 and 60 times a year. So, roughly, once a week someplace there is Science Advisory Board meeting, if not in your neighborhood, at least somewhere. During the year; these 50 to 60 meetings result in somewhere in the order of 30 and 40 reports. All of these reports are available to the public. I should mention that all of the meetings are open to the public, as is this one. This particular Committee, as I've mentioned, is the Indoor Air Quality and Total Human Exposure Committee. It has been supplemented, in addition to the members of the Committee, by people that we have drawn in for this particular review who have special expertise in this issue before us; that is, environmental tobacco smoke. These people are serving as consultants to the Board, so not everyone is here as a member, but they are all technically qualified to serve with us on this issue. In closing, let me just say a word about the anticipated schedule of.what happens from this point on. For dgm.EZf.can eRe#o%fets1 l~e t202J 296-OZ6, 'U-4 /7031644-7636

1 3 4 5 6 7 8 9 10 11 12 ~ 13 14 15 16 17 18 19 20 21 22 23 ) the next two days, we will be doing this review, hearing public comment and the Committee will be undergoing its deliberations. As has already been alluded to, the Committee will then draft a report which will be reviewed by the Committee in preparation for submission to the Executive Committee at their October meeting on the 27th and 28th of October, I believe it is. Yes, the last week in October. If everything goes well at the Committee level and the Executive Committee level, we would anticipate forwarding a report on this issue to the Administrator sometime in the month of November. OVERVIEW, INTRODUCTIONS AND DISCLOSURE MR. FLAAK: I'd like to make a couple of announcements relative to the conduct of the meeting today, some from an administrative standpoint and others from a procedural one. First of all, this a public meeting of the Board. We've requested, as is our common policy, the members of the press who have any requests for any information from us, please either see myself or Dave Ryan in the back of the room or one of my assistants who are sitting at the back table, Carolyn Osborne or Rasheed Tahir sitting at the back table. ogmezZca.n 4-Rrootteu Ze /2021 296-0261 rlGq /7031644-7636

1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 r 1-12 Typically, Science Advisory Board meetings are not transcribed; however, today it is being transcribed at the request of an outside individual. This is consistent with the policy that we set at our last Executive Committee meeting to allow outside organizations who desire to have a transcript of the meeting made to conduct the transcript, with the understanding that it might be made available to others who so desire it. At the last page of the agenda for today's meeting is a statement regarding the transcript and who to contact regarding availability and the possibility of whatever fee might be involved. This is not an official Science Advisory Board transcript, and consequently, we have no control over the transcript itself. We probably will get a copy for the public record sometime after the meeting. Typically, these take a couple of weeks. I'd like to spend a couple of minutes going over the agenda to make sure that everyone is clear on what we're going to be doing. The primary function of this meeting, the sole function of this meeting, is to review an EPA draft document entitled "Respiratory Health Effects of Passive Smoking: Lung Cancer and Other Disorders." This is a second look at a document this Committee looked at in December of. ~ ~ ~ ~ N a, W ) Anetican ~E~iottev. 17C /zoz/ zqb-ozbr qUo4 /7031 644-7636

1-13 ) 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1990, upon which this Committee filed a report with the Administrator in April of 1991. This report was made public, I believe, around the 22nd or so of June and made available to the Committee members at the same time. It has been made available for public information; however, the Agency has not been taking specific public comments as they did the last time. However, under the Federal Advisory Committee Act, the Science Advisory Board accepts both written and oral comment on an issue before it. In this particular case, I suspect we have about six or seven inches worth of comments that have come in from various interested individuals, some as late as yesterday, and I anticipate some more coming in sometime today. All of these comments have been provided to the members cf the Committee in advance of the meeting. Copies of these comments have also been provided to the Agency staff as well as interested.members of the public who have desired them. Anybody who wishes copies of the comments that we've received, please see one of my assistants at the back table. They have mailing labels back there. If you put your name and address on a mailing label, we will be delighted to 4Am.EZi.can 4-Rootteu. Ze /zo2l zq6-oz6i q/c4 /7031644-7636

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-14 send you a set of that material. However, it will probably be a week or so after the meeting before you get it. The stuff was too big for us to cart over here today. If you need to send a messenger over to our office, please contact one of us, and we can make arrangements to do that later in the week. As far as the agenda is concerned, we are going to have a brief overview by members of the EPA staff before we get started on the review and the rest of the meeting is going to be focused on the Committee's actual review and discussion of the various parts of the document. This afternoon we have scheduled a public comment period which is going to start at 3:00 p.m. I should point out that originally in the Federal Register notice for this meeting, the meeting was scheduled to end at 5:00 p.m. approximately. I suspect it will run later today because of the number of public commenters that requested time and the fact that most of them wanted more time that we originally allocated. We've made every effort to accommodate them in the additional time they've requested, and as you can see in looking at your agenda on page 2, we have some 14 or so public commenters who will be speaking. I've asked each of cAIZEZI.CRK GREl2o'LfE21 2~e Izozl z96-oz& q'df /703/ 644-7636

1-15 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 them to try to limit their comments to about 10 minutes, and in any case, no more than 15, to allow everybody an opportunity to speak. A couple of the speakers have said they would be here both days, and if necessary, we may ask them, rather than to push this into the later evening this evening, to come back in the morning and before we get started, wrap up the last two or so public comments. We'll see how it goes this afternoon, and we'll make a decision a little later on. Bear in mind if we do that tomorrow morning, that would mean the meeting would start just a little bit earlier than we had originally announced. Nine o'clock was the original announced time starting. We might move it up to 8:30 a.m. Again, we'll announce that later on today. What I'd like to do is turn my attention to the Committee members for a moment. One of the issues that we discuss at our meetings, and have done for the last couple of years, is what we call public disclosure. This is an opportunity for you to state publicly for the record your association on various issues that related to the issue at hand. This does not mean a disclosure of the confidential financial statements that are filed by members. However, it Pqq2E'LCCQ/2 G1 \E j7.04.& 7s 2~C /zo2J 2q6-o26, q,o4 /7031 644-7636

1-16 1 2 3 4 6 7. 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 does give you an opportunity to relate to a couple of issues. And there is a document which I sent to you prior to the meeting, of which there's also a copy in your folder, which you might want to glance at if you haven't thought about it already. Basically we're asking people to identify whether they've had research conducted on this particular issue, previous pronouncements made public on this issue, the interest of your employer in very specific areas of this environment tobacco smoke., any financial interest you might have in this issue and any links, such as research grants, perhaps from EPA even, related to research on this issue. Now, we're looking at things that are related to the issue at hand today, not thingsvare related in a very, very broad way. I'd like to, at the same time, ask the Committee members to introduce themselves since we haven't done that yet, and when they do that, to please identify any issues they think are relevant for the public record. I'm going to start -- Geoff, can I.start down at your end, please? Dr. Kabat? DR. KABAT: My name is Geoffrey Kabat. DR. LIPPMANN: That's not a public address microphone. That's just a microphone for recording. So if you leave it back there and speak so that they can hear you, og»zEZCcan lz~?E fso¢tEqs 25e /ZO21 296-oZ6l Nd{ /703/ 644-7636 ~ ~ N Ln

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-17 there will be enough volume in the recorder. DR. KABAT: My name is Geoffrey Kabat. I'm in the Division of Epidemiology at the American Health Foundation in New York. The American Health Foundation is a nonprofit institute devoted to research in the area of cancer, cause and prevention. I should mention that the American Health Wrnder , Foundation and its president, Dr. ind, have played an important, if not a leading role, in linking tobacco use to specific diseases and specific cancers. My own work has focused on a range of issues, including tobacco, alcohol, passive smoking, diet, and other factors in relation to individual cancers. I've gone on record as criticizing some of the methodological limitations of the epidemiologic studies of lung cancer and passive smoking. and this was including our own work on the topic. This was in the spirit of trying to alert people to pitfalls of this kind of research and to try to strengthen the design and the quality of research on this topic. I have no financial interest in the issue at hand, environmental tobacco smoke. I should mention that my funding for my current work comes almost exclusively from the Government, from the National Cancer Institute, and none of N ) 07n2E3.[.CQYL ::RE#O4.fEZd. 2~C /zozf 296-o26r q~v4 /703/ 64a-7636

I-18 1 it comes from MR. DR. University of Center. I'm in the areas toxicology. any tobacco company. Thank you. FLAAK: Dr. Laties. LATIES: My name is Victor Laties. I'm at the Rochester's Environmental Health Sciences an experimental psychologist by training working of behavior pharmacology and behavioral I have not worked ever on passive smoking, and 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 to the best of my knowledge and belief, I have nothing to disclose right now about any contact with any aspect of the problem. MR. FLAAK: Dr. Daisey. DR. DAISEY: I'm Joan Daisey. I'm the head of the Indoor Environment Program at the Lawrence Berkeley Laboratory. The Indoor Environment Program conducts research on indoor air quality, ventilation and energy conservation in buildings. Since Lawrence Berkeley Laboratory is a national laboratory, much of our work is funded by the Department of Energy. At present, I have two research grants and am the principal investigator on two research grants having to do with environmental tobacco smoke. One is funded by the National Heart, Lung and Blood Institute on the physical chemical properties of environmental tobacco smoke, and thee cAnE2i.can eRElzo2tets. ZC /2oz/ p96-oz6r q~Vq 1703/ 644-7636

1-19 10 11 12 13 14 15 16 17 18 19 20 21 22 23 second one is funded by the State of California Tobacco Related Disease Research Program, and that has to do with the behavior of particles of ETS in buildings. I have no research funded by the Environmental Protection Agency, although there is some in our program, and I have no research supported by any of the tobacco companies or tobacco research institutes. MR. FLAAK: Thank you. Dr. Lioy? DR. LIOY: I'm Dr. Paul Lioy. I'm Director of the Human Exposure Division of the Environmental Occupational Health Sciences Institute of New Jersey. I have no vested interest in environmental tobacco smoke. In terms of my research with public pronouncements, I do conduct research on the human exposure pathways and human contact to environmentally-derived carcinogens and irritants and other pollutants, and I also do work on biological markers of exposure and participate and also conduct epidemiologic,studies primarily in prospective studies of human contact with pollution. That's it. MR. FLAAK: Dr. Eatough? DR. EATOUGH: Delbert Eatough. I'm Professor of Chemistry at Brigham Young University. Brigham Young University is a private university sponsored by the Church of d7me zi.can d-Re/i.o 3.te as. Ze /ZO2/ Z96-o261 q,6q 17031644-7636

s 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-20 Jesus Christ of Latter Day Saints. My research efforts have been in the area of chemical characterization of atmospheric pollutants and what's referred to as source apportionment studies determining where materials in the environment came from. The bulk of our work, particularly more recently, has been in ambient studies related to visibility, and some of that work is supported by the EPA. We have had support in the past for the chemical characterization and apportionment of environmental tobacco smoke, and that work in the past was funded primarily by the 14s Center for Indoor Air Research, and prior to CtE'"s formation we had some support from Reynolds Tobacco in connection with that work. MR. FLAAK: Dr. Hammond? DR. HAMMOND: I'm Kathy Hammond. I'm on the faculty at the University of Massachusetts Medical Center in the Division of Environmental Health Sciences. My research has focused on exposure assessment, in particular related to epidemiologic studies and most commonly in occupational exposures. I've had grants in the past from EPA involving environmental tobacco smoke, and from the March of Dimes. PTIf2E'LLCQn GRE#04fE'LS ZC /zoz/ zqb-ozb, q6q /703/ 644-7636

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-21 Currently, I have grants from the National Cancer Institute, it's called Working Well, related to combining both health protection and health promotion; and the National Institutes of Health looking at some health effects related to children's exposure to environmental tobacco smoke. I have two grant applications currently in to the American Cancer Society and the National Institutes of Health related to environmental tobacco smoke. As far as I know, the only relationship I've had with the tobacco industries has been when I attended a conference at the University of Massachusetts, Amherst, on environmental tobacco smoking. They paid for the rooms and the meals there. I've been a consultant to various companies, individuals, governments and consulting firms about measuring exposures to environmental tobacco smoke. MR. FLAAK: Dr. Samet? DR. SAMET: I'm Jon Samet from the University of New Mexico. I'm Chief of Pulmonary and Critical Care Medicine and Epidemiologist at the New Mexico Tumor Registry. My research is presently funded by a variety of agencies, including the National Cancer Institute, the Heart, Lung and Blood Institute, the Department of Energy, and the Ln F" Ln ~ N d¢m.eziean eREhozteu ZC 12021 296-0261 q~vq /703/ 644-7636

1-22 1 3 4 5 6 7 8' 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Health Effects Institute which receives a portion of its funding from the Environmental Protection Agency. I've had funding directly from the EPA in the past. My research has addressed a variety of environmental pollutants and their effects on the lungs, including environmental tobacco smoke. With regard to past positions on the issue, I served as consulting Scientific Editor for the 1986 Surgeon General's Report on involuntary smoking and served as Senior Scientific Editor for the 1990 report on smoking cessation and also as Consulting Editor for the '85 report on occupational issues. I presently serve on the National Air Conservation Commission of the American Lung Association and on their Indoor Air Quality Advisory Group. I'm also on the Editorial Board for Tobacco Control and publication of the British Medical Journal. I also serve in an advisory role to the Health Effects Institute Asbestos Research which also receives funding from.the Environmental Protection Agency. I had participated in a planning workshop when the 'iS Center For Indoor Air Research developed q~~lagenda several years ago and provided my views on research'Cneeds in the area of indoor air pollution. My past work and various literature reviews on the subject of environmental tobacco smoke are c4mEZi.can GRE/zottzU 2~e 12021 Z96-o26/ q'6q t703l 644-763G

1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-23 well documented in the peer review literature. MR. FLAAK: Dr. Stolwijk. DR. STOLWIJK: I'm Jan Stolwijk. I'm in the Department of Epidemiology and Public Health at the Yale University School of Medicine. I have been working in the area of environmental health sciences in a broader sense. Current support is from the National Institutes of Health. Support in the past has been enjoyed from the Department of Energy and from the Environmental Protection Agency. I have, other than the prior review that we did here for this particular document, no statements or positions that affect environmental tobacco smoke. In the course of research, measurements and assessments of environmental tobacco smoke as a contributor to various diseases, is always a factor, but I have not conducted studies that made the environmental tobacco smoke the primary issue. MR. FLAAK: Dr. Lippmann. DR. LIPPMANN: I'm a Professor in the Nelson Institute of Environmental Medicine at New York University Medical Center. I have not engaged in any research directly focused on environmental tobacco smoke. The only interest I have is as a confounder in our studies of the effects of Ln ) PjYI'LE2CCQ/2 GRE#02tEY1 rL1C /zoz/ zq6-o26i q04 0031644-7636

) 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-24 occupational and community air pollution exposures on respiratory health endpoints. I am a member of the Scientific Review Panel for the Center for Indoor Air Research which acts as a secondary scientific peer review for investigator-initiated grant applications in the general area of indoor pollution, including, of course, environmental tobacco smoke. My research support currently is from the Environmental Protection Agency for research in community air health effects, and from the National Institute of Environmental Health Sciences for research on particle behavior in the respiratory system and the use of aerosols as probes of the lung. I serve on an advisory panel for the Health Effects Institute Asbestos Research with Jon Samet in which we are helping spend some of EPA's money indirectly. I have no other ties to this issue that I can identify. MR. FLAAK: I'm Robert Flaak. I'm the Designated Federal Official for this Committee. I'm also the Assistant Staff Director of the Science Advisory Board. I'm an EPA employee and I have no interest in this matter whatsoever from either side of the issue. DR. BARNES: Donald Barnes, Staff Director of the. o4ineti.can eRefiotteas. 17e _tzoz1 z96-oz6r Q/04 /703/ 6qq-7636 %10 N

1-25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Science Advisory Board. Also, I have no interest in this issue either. DR. LEBOWITZ: I'm Mike Lebowitz. I'm a Professor of Medicine and Associate Director of the Respiratory Sciences Center at the University of Arizona. I have had research funds from NIH, EPA in the past, and currently, to study acute and chronic respiratory diseases and the factors that affect it, including active and passive smoking and other environmental pollutants, occupational exposures, et cetera. I've been a consultant and committee member on indoor and ambient pollutant effects for the World Health wNOKV Organization, the Regional Office of WHL, the PanAmerican Health Organization, EPA, CDC, FDA, NIH, et cetera, and CIAR. I do review grants for these various agencies and institutes and review documents for them. I have spoken on the issue of active and passive smoking to various groups and I've published widely on the subject, which is a matter of public record, and I have no vested economic interest vis-a-vis active and passive smoking at any time. DR. WEISS: My name is Scott Weiss. I'm Associate Professor of Medicine at Harvard Medical School and an Amezfca.n cRr_#oa.tEU ZC (2ozf zq6-oz6t q4o4 /7031644-7636 h+ tJt W N a, N m ~

0 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-26 Associate Physician at the Channing Laboratory. I'm a respiratory epidemiologist. My primary research interest has been risk factors in the development of asthma and chronic obstructive lung disease and ETS asthma history. I have conducted research on more risk factors related to those respiratory outcomes and specifically into the effect of ETS in utero exposure to cigarette smoke and their possible influences on lung function, growth and development of children and its influence in pulmonary function in adolescence. I was a co-author of the 1984 Surgeon General's Report, specifically the chapter relating to passive cigarette smoke exposure and respiratory disorders in children and adults. Also in 1986, the Surgeon General's Report, I'm one of the co-authors for that chapter as well. I was also involved in the chapter of the 1990 Surgeon General's Report that Dr. Samet was the editor for. My salary is derived 100 percent from research funds, 90 percent from grants from the National Heart, Lung and Blood Institute. One of those grants is specifically looking at, for which I'm not the principal investigator, but one of the grants that does support my research is investigating the effect in utero and postnatal cigarette Ln J N rn A rilE2LCQ.I2 GRf~2o'L~E'LS N m !bC /zo2/ zq6-oz6, ~ No4 /703/ 644-7636

1-27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 smoke exposure on pulmonary function and respiratory illness in children. I also have a small amount of salary support from an EPA grant that supports collaborative use in the Netherlands which is related to outdoor air pollution. It doesn't have anything to do with indoor air pollution or passive smoking. I have written review articles related to the respiratory and health effects of passive smoking or ETS in children. I've also written editorials about ETS and its relationship or possible relationship to the occurrence of lung cancer. I have no other financial or proprietary interest in this matter. MR. FLAAK: Dr. Wesolowski. DR. WESOLOWSKI: I'm Jerry Wesolowski. I'm Director of the Air and Industrial Hygiene Laboratory, California Department of Health. I'm an Adjunct Professor at the University of California and I'm also Director and Treasurer of the United States Committee on Poland's Environment, a nonprofit organization that assists Poland in helping cleaning up its environment. I believe I have no conflict of interest, either actual or perceived, or governing issue. I have no direct cn Q7/7zE'LCCRl2 GRE/204.fE 21. Ze /ZOZl Z96-OZ6, q~vq/7031 644-7636

1-28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 nor indirect relation with any tobacco company or interest. I have never met or corresponded with Margaret Thatcher. (Laughter.) DR. WESOLOWSKI: My laboratory has not received, is not receiving, money for research, either directly or indirectly from the tobacco companies. The California Indoor Air Quality Program is carrying out research to study the effectiveness of various controlled strategies in office buildings on protecting nonsmokers from environmental tobacco smoke. That research is funded by the University of California which gets the money from taxes which are placed on the sale of cigarettes in the state. I don't believe I have any conflict of interest in this matter. MR. FLAAK: Dr. Rockette. DR. ROCKETTE: Howard Rockette. I'm with the Department of Biostatistics at the University of Pittsburgh. My research has been }n the cancer area and occupational epidemiology doing cancer research. I'm the statistician on several projects that are funded by NIH. On the occupational epidemiology area, many of those studies are funded by industry. None of them are related to the issue here, ognzeucan 1::R1e#otfzu. 1`~C /zoz/ zq6-oz6i QVa4 /9o3) 644-7636

1-29 ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 environmental tobacco smoke or the effects of smoking. I have no funding that's at all related to the tobacco institute and no funding that's at all related to EPA. DR. LIPPMANN: Thank you. It's a shame to take so much time to do this, but in the climate in which this issue was being reviewed, it becomes a necessity. You may have noticed that there are representatives of the press attending, including people with TV cameras in the hall. I have been asked if I would discuss the hearing with members of the press at break today and I declined. They asked if other members of the panel were willing to talk to the press, and that's up to each individual member to make their own decision. At the time of the last meeting, we stuck to the position that the time to talk to the press was after we had concluded our reviews, but that was not binding on any member of the panel. So if you should choose to discuss the procedures and whatevqr else you want to with members of the press, that's entirely up to the individuals. This is clearly a topic of interest to the general public and therefore to the press. On the other hand, it is awkward to state positions before we've heard each other's views on the quality of the science which helps make up AIZEZCCQl2 GRE#o3EE'L1 1`~C /zozf zq6-oz6i q,~vq 17031644-7636 OD

1 2 3 4 5 6 7 8 9 10 11 12 ~ 13 14 15 16 17 18 19 20 21 22 23 1-30 consensus judgment on the adequacy of this report. I won't take any further time from the proceedings. The program calls for introductory remarks from Drs. Bayard and Farland of the Office of Health and Environmental Assessment who are responsible for the preparation of this document, Dr. Bayard, directly, and Dr. Farland, a little bit more indirectly. Dr. Bayard will tell us in relatively few words about what has been done in regard to modifying the document, to address the concerns we expressed at the last review, and Dr. Farland will address, at least in part, one of the issues as how the experience of doing this particular review has given the Agency experience and outlooks about the use of quantitative risk assessment techniques for non-cancer issues. Dr. Bayard. Or, Dr. Farland, do you want to go f irst? OVER,VIEW OF DRAFT DOCUMENT PRESENTATION BY DR. FARLAND DR. FARLAND: I have just a few overheads. I've not made copies, but I believe the Committee will be able to follow along without them. These are really primarily for the people out in the audience and not really as a backdrop . oqinezi.ca.n Z~?efiottEu 2~C tzozJ z96-o26, q,64 17031644-7636

1-31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 to my discussion. It's probably important for me to say that I'm a biochemist and a scientist rather than an engineer at this point. It's always a pleasure to be here to discuss these issues with the Science Advisory Board Committees. As you all know, we value your input on this issue, as well as on all the issues that we come to speak with you about, and we look forward to your comments over the next two days. I think you can see from the document that you have in front of you that this represents significant work and that our organization has been committed to this particular project on respiratory health effects of passive smoking, lung cancer and other disorders for the past several years. We have been responsive to comments that we have received and feel that this Committee has provided a lot of direction and suggestions for improvements to the document which we hope that at the end of the discussion here you'll agree that we've done.quite well. This is a high visibility assessment, and I don't think anyone would argue with that statement. It's high visibility because of its implications for the future of the way that we do business. In terms of our risk assessments, the evolution of ogmezi.can 1::R1e#o%teTs 'L1C /zoz/ zq6-oz6i q~04 j7o3/ 644-7636

) 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 . 17 18 19 20 21 22 23 1-32 the approaches to risk assessment can be characterized through a look at a number of documents that EPA has worked on, but this particular one gives us an opportunity to illustrate a few features, and I'd like to talk just very briefly about that as an introduction to this topic. The focus here really is on this concept of risk characterization. I'm trying to understand the issues associated with potential toxicity of this environmental pollutant and get an understanding of exactly what the implications will be for human health effects. Now, in terms of the features that we might just very briefly focus on, we've had an opportunity to characterize the hazard through a weight of evidence approach for cancer and non-cancer effects, specifically, for lung cancer and selected non-cancer effects. This weight of evidence approach has allowed to take great care in assessing studies directly related to ETS. But at the same time, it considers all of the information, information from other epidemiology data, from laboratory studies, from chemistry and exposure studies that help us to put in perspective the potential for hazard. We attempt to lay out the broad base for our findings which lead to the conclusions that you find in the 44~'Zf.CQK GJ`ElIA7.tE4s. 2~c t2021 Z96-OZ6, N6q 17031644-7636 Ln L" W N IM N F-~

1-33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 _ 17 18 1 19 20 21 22 23 document. It is important to understand that we've not dealt with other cancers other than lung cancer. The information is very sparse on those particular cancers, but that might at some point be an issue. And we have not dealt with cardiovascular effects in terms of another non-cancer endpoint that might be looked at. We have received comments to the effect that perhaps the Agency should consider that at some point in time, but there was a decision within our office not to address that particular issue at this time. In terms of dose response assessment that we are dealing with, I think we have a fairly unique set of information in this particular finding set of data for environmental tobacco smoke. First of all, we have a focus on human data which is fairly unique in terms of dealing with environmental pollutants. This is something that obviously is our best source of information. We need to make the most of it. Despite the problems associated with evaluating human data, it clearly is the important focus of this effort. It does, however, allow us to limit our extrapolation in terms of animal to human types of responses, although we use that as a supporting factor within this assessment. Ln r OT!')2EtLCQ/2 GREf1.o4tE71 L~e 12021 296-026, <vo4 17031644-7636

. 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 . 17 18 19 20 21 22 23 1-34 We have the opportunity of dealing with issues at environmental levels. We're not talking about trying to take high dose occupational studies or high dose animal studies and bring them down to lower levels of risk that might be experienced by the rest of the population. So this is a, situation where our assessment is at environmental levels. Another fairly unique situation with regard to environmental risk assessments is we've taken the opportunity to use some new techniques, meta-analysis for lung cancer that we think will be important to us in terms of combining information from various studies as we do risk assessments in the future. In this case, we're able to use a specific endpoint and look specifically at lung cancer incidents as a subject of this meta-analysis. But in fact, we are particularly interested in trying to extend meta-analysis into a discussion of a number of different approaches. And we hope that you'll find that.this has been an innovative approach and that it bears some merit in terms of future approaches for risk assessment. Finally, we think that a number of the issues related to non-cancer endpoints afford us an opportunity to look at quantitative estimates of impact for non-cancer W ~ Ln dqmlezi.can cRe fioa#ets 2~e 12021 296-026, q~o4 0031644-7636

1-35 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 .17 18 19 20 21 22 23 endpoints that go way beyond our normal approach of being able to use only a reference dose or a discussion of the dose below which one would not expect to see an effect occur based on use of no effect levels or various other indicators of response from animal studies to human,populations. So, we think that there are some interesting and important features and an opportunity to do some innovative risk assessment work in this particular assessment. In terms of exposure assessment, we are particularly fortunate in this case to be able to use the human data that is available to us. Basically, human activity patterns, the type of information that one needs to bring into the assessment of the impact of environment pollutants on humans to focus on the idea of biomarkers of exposure to be'able to support discussions of potential exposure based on activities or job classifications or other types of things that others might be involved in, and to characterize this not,as a bounding estimate of potential maximum exposures, but really to view this as an opportunity to look at a group of individuals with what might be called a high end of exposure for the general population, certainly not out on the very tail of the distribution, but at the high end of exposure. cn ) ogmazi.can :rR-e#ottzu 1~C /zoz/ z96-o261 4~o4 /7031 644-7636

I-36 ) 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 All three of these features are particular issues that we have dealt with in terms of our new exposure assessment guidelines that we see here as an opportunity to really apply to the issue of environmental tobacco smoke. I mentioned earlier our interest in getting into risk characterization and to really focus on that issue. The Academy, when it gave us a framework to do risk assessment in the early 1980s, did not spend a lot of time on risk characterization. We need to look at this as an opportunity for us to discuss both qualitatively and quantitatively the issues surrounding potential risks of environmental tobacco smoke, to provide a range of estimates for exposure and effects that we think will give those interested individuals a full sense of the information that's available and e 14'c C ~.T potential for a€€ee-ts, and to fully discuss assumptions and uncertainties in this document. We feel that we've made good progress in this regard, and hope to hear your comments related to that and express some confidence in terms of our perspective on the database that we have available to us. So these are all features of this particular assessment that we think are going to have a great impact on the way that we do future assessments in the Agency, certainly within my office, and our attempt here is to use . dgm.atf.can cz~?.e#o-ttzts 17C /zo2/ zq6-o26i W.4/703/ 644-7636

I-37 ) 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 . 17 18 19 20 21 22 23 some of these approaches to move the state of the science and move the state of the art of risk assessment in a positive direction. I might make just one comment in clarification, and that is that the Agency will take all of the comments that have come into the Science Advisory Board for this particular session, and will address them, along with other comments that have come to us, as we put the final document together. There will be a copy of those comments in the Agency's Docket and there will be a disposition of each of the comments associated with that Docket as well. Thank you. DR. LIPPMANN: Thank you, Bill. We appreciate all of the voluminous written comments coming from the public and which have been distributed to the entire Committee. These, as Dr. Farland said, will be considered by the Agency and by us as we prepare our report of our review of this document. Clearly, we.don't have the time, as Mr. Flaak said, to have more than a brief discussion from those members of the public who asked for time on the public comment period, and it's impossible for any of us to review the EPA document and all of those public comments in great detail in advance, although, we've all done the best that we can and have made a ogmezican c-R e~.o~ets ZC /zoz/ zq6-oz6i W04 17031644-7636

1-38 ) 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 .17 18 19 20 21 22 23 conscientious effort. We will be looking at both the document, each other's comments and those public comments as we prepare our report on the adequacy of this document. Steve? PRESENTATION BY DR. BAYARD DR. BAYARD: Chairman Lippmann, Vice Chairman °vf y4t Stolwijk, members of the Committee, and members SAB staff, on behalf of the Office of Research and Development and Office of Air and Radiation I sincerely want to thank you for taking the time and the effort to review this. PARTICIPANT: Steve, we can hardly hear you back here. DR. BAYARD: I just want to thank you very much. We have already submitted our reports to the SAB on how we have revised the previous draft to make this revised draft. What I want to do just now is summarize that and discuss the scope of the revisions, how we have revised it. This document is put out by the Office of Research and Development and the Office of Air and Radiation. The technical staff is Steven Bayard, Jennifer Jinot and Aparna Koppikar. The funding and scope was provided by the Indoor Air Division of the Office of Air and Radiation, with Robert Axelrod, Division Director. PTp2EtGCQ.n GREO0dEt5. 2~e t2021 296-o26, q4Vq /7031 644-7636

1-39 ) 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 The officers of the chapters are Steven Bayard; Kenneth Brown; Jennifer Jinot; Brian Leaderer, from Yale did the exposure chapter; Fernando Martinez, Associate Professor of Pediatrics at the University of Arizona in Tucson, wrote these chapters a°n respiratory disorders; Neil Simonsen, who is finishing up his Ph.D. at the University of North Carolina, helped the (inaudible); and Judson Wells, who is responsible for the (inaudible) on misclassification correction. Kenneth Brown is from North Carolina. Jennifer Jinot works at the EPA Office of Research and Development. The history of this, we began work on it in late 1988. Our first draft was released in June of 1990, with a public review period. Our conclusion was that by the weight of the evidence, we believed that ETS was a Group A carcinogen, and we felt that there was some association with respiratory disorders in children. In October 1990, the public comment period ended and this was reviewed.by your Committee in December 1990. We presented a report to the Administrator Riley in April of 1991 which concurred with Group A carcinogen classification, but felt that our conclusions with respect to respiratory disorders ought to be made stronger. We made many suggestions for revision of the document and many suggestions d¢nuti.can '::'RE#oactez1 M 12021 Z,p6-026, l71,14 0031644-7636

1-40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 which we felt were very appropriate, and what you see here is the product of those suggestions and the public comments. The scope of the revisions. I'm just going to run through the chapters to give you an idea of the outline. y4s,2~ ~) We'll go over the chapters and I'll do -r~r-ch quickly. Chapter 3 is on exposure. Our old document had no chapter on exposure, but we had some material in various sections. We pulled that all together in a chapter and expanded it to discuss the physical chemical characteristics and the similarities and differences with mainstream smoke and sections on how to measure and assess ETS exposure by monitoring biomarkers and questionnaires, correlating between questionnaires, monitoring and biomarkers in the models. From Chapter 3, we move on to the lung cancer hazards identification. As suggested by the SAB, we made a totally new chapter on lung cancer and active smoke and included the evidence of long-term animal bioassays and genotoxicity. . In our chapter we emphasized dose-response. We felt that if active smoking caused lung cancer, then we should expect some dose-response to low exposure. We felt that it would be similar to the relationship of passive smoking. C4rneti.can -_Re#ortEqs. `1M /zoz/ 2q6-oz6r No4 /703/ 644-7636 N cn Ln W N o, N ~ ~o

1-41 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 We emphasized dose-response by analyses based on the number of cigarettes, the duration, the age of start, time since cessation and gradient was all for lung cell types. Our conclusion was based on this evidence alone from Chapter 4, we could conclude that environmental tobacco smoke was Group A carcinogen if these were the only data. However, there's a very large dataset on the epidemiology of passive smoking. In fact, for female never-smokers only, the very largest database -- the other databases were far smaller and less inclusive -- we analyzed 31 studies in this report versus 24 in the earlier draft. Nine of the nine were new studies, including two updates. So there were actually nine new studies, but two of them were updates from previous studies. And we provided critical review to all these studies in Appendix A, as suggested by the SAB. In order to include both the positive and negative response studies, we do a meta-analysis, we do extensive trend analyses in all 16 studies which had data by various exposure levels. We pooled data by country or continent into six groupings, and this is different from the previous report which just combined all studies from all countries. We feel 44)2EZLCQ/2 lEf204.tE'ls. 17C /zo2f 296-0261 cga4 /703/ 644-763b

; 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 .17 18 19 20 21 22 23 1-42 we have a better description of power, so that we make it very clear that only 8 of the 30 studies which had suitable data, approximately 31, have the power of at least 50 percent to detect a (inaudible) as low as 1.5. We maintained our one-sided test, as in the previous draft. I put this up for clarification because some of the public comment suggested that we use 90 percent confidence intervals to make the test more significant. In fact, we used 90 percent confidence limits only to correspond to the one-sided tests which we had in the last draft, so that the two would correspond. We adjusted for the smoker classification bias on an individual study basis and we corrected before we did the significance testing for each study. We also provided separate analyses for each study that looked at separate potential confounders, history of lung diseases, family history of lung diseases, heat sources for heating, cooking and occupations and dietary factors. After we did that, we still felt that there was some way that we might not have ranked the studies critically satisfactorily. The SAB the last time suggested that we try and critically rank the studies, and that we did. The idea Co}~^ec7C of being that some studies would d3.raeV for some confounders Ln N Ln ) dgmett.ca.n :::RE#oztets. ZC /zoz/ 2q6-o26, No4 /7031644-7636 9o N m N N ~

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 . 17 18 19 20 21 22 23 1-43 and not for others, some for some biases and some not to others. We tried to judge which studies were a better utility for determining whether or not there was an ETS effect, and we ranked the studies into four tiers. Then we did the analysis by tiers to see whether the higher-range studies provided the better information than the lower end studies. 11~rx Our conclusion after all of this was that-£TA is a Group A carcinogen based on all the weight of the evidence and all our ways of looking at the data, and that includes both the active and the passive smoking studies. On the quantitation from lung cancer identification, we moved to Chapter 6. On Chapter 6, unlike the last timewhen we presented evidence, where we presented sections in both epidemiology and cigarette equivalents, as the SAB suggested, we used only the epidemiology data this time. Furthermore, as the public comment suggested, we used only the 11 U.S. studies to measure U.S. risks as one method of doing it, and as another method, we used only the largest U.S. study which was the Fontham Study in 1991 which included five U.S. centers, with consistent results. The o¢mEZi.can cRefiotfEU Ze (zozf z96-oz6r q/dq /7031644-7636 Ln N N N N

) 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 .17 18 19 20 21 22 23 1-44 study was designed with, in fact, little, if any, smoking irS r,~ S be.,.9 7~'t t~~s classification, biasing only source that we could identify. The Fontham Study was also the only cancer study with actual cotinine data from their study participants. We also used never-smoking females exposed to spousal smoke. Then we separated nonspousal which we declared as spousal risks. We separated males and former smokers of both sexes and we provided sensitivity analyses for the model. These are basic differences from the last draft, and our conclusion, however, was that there's still approximately 3,000 lung cancer deaths among nonsmokers of both sexes per year of children with ETS in the U.S. From Chapter 6 we move on to Chapter 7, and Chapters 7 and 8 represent the ce identification and quantification of non-cancer respiratory disorders. This was completely rewritten following the SAB comments and following the very detailed outline the SAB gave us. The outline is followed almost exactly. We found that we had over 30 additional studies in the earlier draft. We stressed the biological mechanisms and plausibility arguments, including the effects of both prenatal and postnatal exposures. We have a new section on 44mE2i.ca.n eREhoqtaqs. 1~C (zo2/ zq6-o26i <U,_4 (7osl 644-7636

1-45 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 Sudden Infant Death Syndrome. We have a new section on respiratory effects in adults. The conclusions are much stronger, and we feel that there is enough evidence to say that ETS is causally associated with lower respiratory tract infections in infants and young children up to age 3, symptoms of respiratory tract irritation, middle ear effusion in children, small reduction in lung function and aggravation of asthma. Finally, we attempted a quantification of certain non-cancer respiratory disorders. This is also a completely new chapter, as suggested by the SAB at the last meeting. We present estimates for lower respiratory tract infections and asthma in a model which is similar to that in the lung cancer quantitation, one of the models. We present the sensitivity analysis. We present our estimates in terms of ranges and try to discuss the confidence and ambient certainty, as we did in the lung cancer quantitative objective. Our conclusions are that lower respiratory tract infections in children up to three years, which range from 150,000 and 300,000 a year, result in between 7,500 to 15,000 hospitalizations. That it may cause 8,000 to 26,000 cases of new asthma. We don't feel that there's conclusive evidence that it causes new cases of asthma at this time, but we do ~e~.o~ats c4mEtican !bC /2021 296-o26i q~vql7o31 644-7636

) 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-46 feel that there is evidence. We feel there is conclusive evidence that it aggravates symptoms in 200,000 to a million asthmatics per year. Thank you. That's the scope of our revisions. We appreciate your deliberations. DR. LIPPMANN: Thank you. I appreciate your laying out the basic issues confronting us in a short presentation. In reviewing the document, we've concluded that it would be best to quickly go through the appendices. The fact that the Agency felt that the appendices were needed to understand the basic chapters to provide us technical support, I think makes sense for us to look at them first to see if we agree that they provide the technical support for the statements in the chapters. So, let's start with, as on the program, with the review of Appendix A, "Reviews of the Epidemiological Studies on ETS and Lung Cancer," and I'll call first on Dr. Samet. COMMITTEB REVIEW OF APPENDICES A-E APPENDIX A - REVIEW OF EPIDEMIOLOGICAL STUDIES ON ETS AND LUNG CANCER DR. SAMET: This set of Appendices, a set of reviews of the studies characterizes them describing their design components, provides reviews of the individual OT/f2E'L(.CQ/2 GREf2O4.EE'L1 J.`~. C tzoz/ z96-cz61 q04 t7031 644-7636

1-47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 studies, and for the most part, in concluding paragraphs, offer some qualitative comments concerning the contribution of the individual studies to the literature on environmental tobacco smoke and lung cancer. I found that the descriptions of the studies were useful and the studies had in fact been reviewed according to a standardized format, and the form for that review was provided. On the other hand, I found that the reviews of the individual studies lacked a systematic methodologic framework against which potential limitations were addressed. For the most part, these are case-controlled investigations with, in a general sense, well-known sources of bias related to selection, confounding information, and the possibility of selection of an inappropriate comparison group. From my own reading, I would have found the review of these studies to have been more useful if each in fact had been reviewed against,the potential limitations that should be considered in interpreting the results. I think that in fact most of the proper material is contained within the critiques of the individual studies, but it's a matter of systematically reorganizing the reviews of the individual studies so that this Appendix becomes more useful to the 44nEZi,can I:_Re#otfets `L1C /2oz) zq6-o26r q'6q /7031644-7636

I-48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 reader. I think this type of systematic review would be important to buttress the assignment of these studies to two tiers because I think we'll come back to this when we discussed chapter 5. I had some difficulty in being certain, in fact, as to how the studies were assigned to the tiers of finding only a footnote to a table that mentioned criteria, and I would prefer to see more systematic review of the individual investigations and assignment of the studies to tiers. I also found some of the language to be off the mark and distracting. I think for the utilization of this information, the conclusions of the reviewers and, in fact, of the authors, concerning the contributions of these individual studies to the body of evidence are relatively unimportant. These studies emerged over approximately a 10- year period during which there was a shifting summary estimate -- in fact, initially very little precision -- in understanding the risks. So that I think that the historical perspective, as the authors express their opinion and subsequent reviewers disagree with them, is probably unnecessary. I think what is important for this document is a 4vqI)2E'LCCQnGRE#OtEEt, 2~e t2021 Z96-02f1, qGq 0031 644-7636 ti co o1 ti ti V

1-49 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 systematic review of the evidence against biases which may affect, for the most part, case-control, but also cohort data. I think my other comments concerning structured review of these studies and their potential limitations, we'll come back to when we talk about the studies further in Chapter 5. So in terms of this Appendix alone, my summary feeling is that the studies are well described, and that much of the information concerning the limitations of the individual investigations is included in the reviewer's comments, but could benefit from a more systematic overview. And I think some gratuitous conclusions are provided, both by the reviewers and the original authors, that would best be removed, they're distracting. DR. LIPPMANN: Thank you. Dr. Kabat. DR. KABAT: Somewhat along the lines of what Dr. Samet has said. I had trouble with the classification by tier, and I felt that.one could have set up an objective set of criteria, as has been done in certain meta-analyses, which would include points like did they match on never-smoking status; was there the percentage of histological confirmation of the cases; the level of detail of the ETS questions, and so on -- control of confounding. And one could have assigned H dgrnazEcan eRelzottets `LlC /zo2f 296-0261 q~oq 17031644-7636

1-50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 points for each of those criteria and come up with some rating for each of the studies. As it is, the assignment to tier appears somewhat subjective. For example, one might have been surprised that after all the flaws that are pointed out in the Trichopoulos paper, that it was assigned to tier two. But at any rate, the authors acknowledge that their assignment to tier is subjective, and I think it could be improved on. I also found it a bit confusing that their summaries of some of the studies, the ones with higher power in Chapter 5, and assignment to tier is justified there, and you have a complete synopsis of each study and description of each study in Appendix A. What should be the punch line, I think, at the end of the description of each study would be, "you want to know the tier number," and I had to keep flipping back. And it's additionally confusing because in Table 5.1, I think, where the superscript gives.the tier number, some of them are mistaken. So you're wondering whether they changed their appraisal of some of the studies or whether it was just typos from Table 5.1. Then, I've written some comments on specific points that come up in the descriptions of studies but I'll leave O¢rnEZican .-:REfiotteu `L1C t2o2f 2q6-cz6, q~04 /7031 644-7636 %o Ln N IM N N i t0

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-51 that for later. DR. LIPPMANN: Okay. Are there any other members of the panel who elected to prepare comments on Appendix A? Yes, Paul? DR. LIOY: I made notations and I concur about the fact that I got a little bit let down not knowing what the tiers were in these individual studies, and I think Jon's point about logical progression is well-taken. I had a difficult time going back and forth between studies at times. I think in some places it's clerical, and I think it can be easily handled, but I think for consistency and ease of reading and use with Chapter 5, I think that it's very critical that it's put into a consistent basis. DR. LIPPMANN: Yes? DR. STOLWIJK: I was also bothered by the transition from the material in Appendix A and the tier assignment. It almost acts a little bit as if the tier assignments were made.after Appendix A was already completed, and as a result, it didn't reflect the rationale that later on was needed for the tier assignment. The other problem that I think Dr. Kabat mentioned, is that if you start classifying tiers on the basis of strengths or weaknesses in, let's say, bias that might have d¢mezlcan z~?e/iottzu `LlC /zozJ z96-o26, q'6q /703/ 644-7636

I-52 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 crept in, then in cases where people have attempted meta- analyses, they find that each of the tiers -- you get a very large number of tiers and very few members in each tier, and you don't get very far that way. So you have to collapse the tiers -- and so to get more member studies in the tiers -- and it seems that there was not a discussion of what might have gone on in the design of the report as to how many tiers to have and what to base it on. And there were, for the large number of studies, a relatively small number of tiers here. And most experience in meta-analysis attempts I think indicates that you get many tiers because you can't -- it is more and more difficult the more seriously you try to classify for strengths and weaknesses, you get fewer and fewer in a member group. There was no expression of that process having gone on. DR. LIPPMANN: I think that's the commentary the Committee is offering on Appendix A. Do you want to say anything in response?. DR. BAYARD: No, I guess not. DR. LIPPMANN: Okay. I think we will at this point take a 10 to 15 minute maximum break and then continue with the discussion of the other Appendices. There will be an installation of a loud speaker system for the hall during the Ln ogmezi.can I=Rrooztets 2~e t2021 296-0261 w64 /70316a-?036

) 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 r 1-53 break which will facilitate our further discussion. (Recess.) DR. LIPPMANN: We'll continue with our review of the Appendices. We are scheduled next for Appendix B. Dr. Stolwijk, I'll call on you first. APPENDIX B - METHOD FOR CORRECTING RELATIVE RISK FOR SMOKER MISCLASSIFICATION DR. STOLWIJK: The description of the method for correcting relative risk for smoker misclassification addresses a very difficult issue that has to do with misreporting of the smoking status of cases and controls, and goes into considerable detail with respect to the various methods that can be applied, and the baseline data -- the baseline parameters that are to be used in order to make estimates of the over-reporting or under-reporting of smoking in cases. PARTICIPANT: Which is? DR. STOLWIJK: There is some controversy in the area that is largely related to what are the background data that are to be used for the estimation of misclassification, of self-reported misclassification. I think that the method that was used here which attempts to use data from the studies themselves in order to C4rnezi.can cRFqiozte3s. ZC /ZO2l 296-o26, q~04 /703i 644-7636 CA Lq

1-54 l 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 make this correction is an appropriate one because there is a study-specific characteristic of this kind of misclassification. It depends to a considerable extent on the type of questions that they're asked, the setting in which the questions were asked and the effectiveness of an interviewer in such a situation. These tend to be highly study-specific so that using correction factors from the study itself, is clearly a very important element. And I think that was advocated and carried out in this case so that there is a complete documentation of the methodology that was used. There is some controversy, as is clear from the submissions by the public and by other organizations, all of which appear to depend to a considerable extent on what the data are upon which the misclassification is estimated. I think that the method that is outlined in Appendix B is one that I found is supportable. The assumptions are clearly stated, and I think it represents a way of dealing with an important but very difficult issue that is defensible. DR. LIPPMANN: We have written comments from two other members. Dr. Benowitz was also assigned to review this Appendix. He indicates that it's somewhat difficult to read. PTmE 2l.CQl2 GREh0'LtE U 2~e 1ZoZi Z96-OZb, q,6q 0031644-7636

1-55 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 It took several attempts before I was able to understand it. In part this is because the material is highly technical and necessarily abstract, but also some of the terms and mathematical operations are not clearly explained. Overall, the principles of the Wells-Stewart method appear to be reasonable; however, I have some concerns about specific points. I won't take the time to go into some of the various specific technical points that he has that you can consider. I'll just extract a few. On page B.2, the top paragraph: Classifying current smokers according to their levels of cigarette consumption is a good idea for risk assessment; however, the specific classification scheme proposed in this approach needs a better rationale. Why is 30 percent of the average cotinine level in smokers considered to be a definition of occasional smoking? Et cetera, et cetera. He goes into some of the other points that you might consider in looking at that once again. These are all highly specific and technical comments. Overall, I think the Wells-Stewart approach is correct. The misclassification is sensible. The mathematical operations seem reasonable. The basis for selecting particular dosimetrics cutoffs for classifying P/!'I'LEZLCQl2 ~EflARfER1 2~C /zoz/ 296-oz61 Q4d{ 1703/ 644-7636

1-56 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 different levels of cigarette smoking need to be more carefully explained. If possible, this Appendix might be rewritten in terms that are easier to understand by someone who is not immersed in this type of analysis or problem. Dr. Woods includes also comments in his written material. This Appendix provides formulas for two types of calculations. One, adjusting observed relative risks for smoker misclassification and background ETS exposure, and two, estimating the population at risk, which is defined as the ratio of the excess risk due to ETS exposure to the total risk from all sources. It also contains a third part which estimates the number of lung cancer deaths in former smokers. The concepts for these equations are valid extensions of those presented in the NRC report, but the presentation in this Appendix is difficult to follow. Moreover, errors in some of the definitions of terms caused difficulty in verifying equation B.1. And then he goes into considerable detail about how he either thinks there were errors or at least it wasn't stated in a way that clarified how it was done to his satisfaction. Then he goes on to say, "Once these errors were corrected, I was able to verify the parameter sensitivity in Section 4.4.3, as equation B.1 predicted, the correct RRO of UT to 0Tn2EZLCQ/2 GRE f2oR.tE21 17C /zoZf z96-o261 (V ,z,4 /7031 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-57 1.14 when a true relative risk of RRM=1 was issued. He was also able to verify that equation B.1 correctly predicted an RRO=1.41 for an RRN of 1.28, as indicated in the paragraph. So I think in general, he found this useful, but he had some concern about the,accuracy of some of the equations as they might be used to extrapolate to other conditions. And you have a copy of his comments. I think Bob just gave it to you, Dr. Woods' overall statement. Does anybody else on the panel have any comments on Appendix B? Mike? DR. LEBOWITZ: Yes. Thanks. I thought the job that was done was clearly quite impressive, some good positive contributions, especially in terms of the different parameters and aspects of misclassification that had to be evaluated, and I liked the approach. I think also that the evidence was presented very well, indicating the true downward correction for smoker misclassification bias which exists in all of these studies of ETS does exist, but may be smaller than previously used or used in another section of the report. There were a couple of places where I wasn't sure that specific issues had been handled or handled properly. I didn't really see the way, though I think it could be cn ) ogmezlcan cRc#o%tz%s J.~e 12021 296-0261 q'~zOq 17031644-7636 Ln

1-58 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 understood mathematically. I didn't see the way that the correction would be ma ~fpr exposure response or dose- A response and in either tec231years or number of cigarettes per day in the supposed nonsmoker once they were found out to be smokers. And that leads back to Neal Benowitz's comment on the 30 percent cutoff or reference value in cotinine to reclassify someone. I couldn't follow that. I think if it were done by cigarettes per day then, in fact, you would have a set of reference values for cotinine that would be applicable for the reclassification and corrections. I didn't see -- possibly I just missed it -- the correction for past exposure and that might have been due to ex-smokers and the higher risk for ex-smokers both, and how that was really handled. Although there was discussion of a larger likely false/negative reporting with surrogate reporting for medical records, I didn't see how that was handled more different than the population base data, and I think it needs to be. It's a further breakdown of the type of study and the location and time of the study and so forth and the community and culture in which it occurs. In terms of the modeling, I think that use of different models might get slightly different mathematical o'grnEZEcan -Rz#otteu !bC /zozf zq6-oz6Y q':V4 /703] 644-7636

) 1 2 3 4 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-59 expressions, might lead to slightly different adjustments, downward adjustments. I think that especially if models included covariants and confounders which are correlated with the bias and correlated with smoking status per se where smokers and ex-smokers are both likely to also have characteristics that would increase the risk as well, I don't think they were necessarily here. And I'm not sure that within the scope of these kinds of analyses you can do that properly. I'm not sure that any of the studies themselves treated that adequately, so you don't have the data, I don't think, to do that. DR. WELLS: It gets more complicated. DR. LEBOWITZ: Yes. It does, but knowing how well you like to play with these formulas, I would think that that would be the next extension of the issue. Most of my other comments are within the scope of that. There were some issues I had with assumptions when data are available to,know from a variety of settings how long ex-smokers have been ex-smokers, what their smoking habits were, whether they were just occasional or light smokers compared to heavy smokers. I think that's time- and place-dependent also. I'm having some data sent from my study. I hope it OTIf2E2GCQ12 l.E#o2EEZS. 2~C /zoz/ z96-oz6, Wc4 /7031644-7636

1-60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 gets faxed here during the meeting to show you. But I remember that there are a fair number of -- in our 20-year study -- there are a fair number of ex-smokers and former 2 vP smokers who later denied v r having smoked. The distribution is weighted toward the low end of smoking habits and the longer they were off cigarettes, the more likely they were to decide 20 years later that they never had smoked and so forth. But in fact, there were fair number who had smoked heavily. And two-thirds of our ex-smokers, at least males, had quit because of health reasons. So that's a real differential effect in terms of the kind of bias. But there were a number of places, in other words, where assumptions were made based on limited data where in fact data are available to correct those or to account for likely distributions of bias that might be due to actual previous smoking. In any case, I wrote all over Appendix B and will give my version to you to play with later on. But mostly I do want to compliment you on the effort you've made in the time you've had. I think your conclusions are really quite important to the overall perception of what effects ETS passive smoking may have. DR. LIPPMANN: Thanks. Jon? Ln ~ ko ) ogm.ezi.can eREfi.oatezs. ZL' /zozf 2qb-o26, q~o4 17031644-7b36

1-61 ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. SAMET: I think I'm aware of three major efforts having to do with this issue: the work of Judd Wells in earlier versions and the present one; Peter Lee and Nick Wald. And to my recollection -- and I'd be willing to stand corrected -- Nick Wald's work has been published in the British Medical Journal and also was a portion of the NRC '86 report which was peer reviewed. An earlier version of your approach I think was published in Environment International, and Peter Lee's work has been documented in his books. I think that it's important to point this out because I think that the choice of a new method that has not yet gone through the usual peer review mechanism just needs to be made clearer, and I think the links between the portion of Chapter 5 where the choice of misclassification adjustment is described, needs to be strengthened and perhaps the introduction, in fact, to this Appendix should be strengthened on the basis for selection. I read it aAd I would side with reviewers who found that everything was there but they had to work hard to understand what had been done. To the extent that that can be helped editorially, I think it would be useful. The assumptions are there. In fact it might be nice to lay everything out on the table as to what you assume. It's a Lq ~ ~ to ti og=7i.can cRa#o-dzts. ZC /zoZJ z96-oz6, wd{ /7031 644-7636

1-62 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 model where you make the assumptions and then you follow through with it and I think we all understand that. I think the reader will just benefit by a clearer, explicit set of assumptions listed as to how you go through the whole process. I think the example in fact was quite helpful. DR. LIPPMANN: Joan, do you want to comment? DR. DAISEY: Just to raise a question. On the cotinine and urine studies, there seems to be a disconnect between chapter 3 and this Appendix. There are two studies in Chapter 3 -- there's no overlap. And then in Table B.1 there are a number of studies. I'm not sure what the basis -- it may be because two people wrote them, but it's a little disconcerting why some of these were selected for Appendix B.1 but not included in 3. And related to that is that it might not be a bad idea to have a chemist take a look at the sampling and analysis aspects of this. I did not have time to do that, but in my experience,.not all analytical chemistry is of equal quality just like not all epidemiology is of equal quality, and if you're going to use this as a basis for some of the corrections, it's important to take a look at that. DR. LIPPMANN: Paul. DR. LIOY: I have two issues with Appendix B. c0qrnEti.ca.n :::RE1zozfE%s& l~C /zo2/ 296-026, No4 t7a31 644-7636

1 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-63 DR. LIPPMANN: Well, there's still no microphone. Speak a little louder, please. DR. LIOY: Okay. Yes, sir. I have two issues with Appendix B. Maybe it's my background, but I get very disquieted when I start seeing terms show,up before I see the equations. It does not allow me to logically process information and determine what the logic basis is for your arguments. I'll give you an example. On you start "Jh saying, let" designate. What is that related to? I have to go over to the next page to start picking up some tidbits as to where it is and what it means. I think that any formulation that you have for an equation has to set up the argument logically and then proceed to deal with the terminology and make the definitions which I think is clearly defined by other individuals as being out of place and in some cases, difficult to find, and then proceed to deal with the assumptions you have to provide to use that equation. I had to dig around this chapter quite heavily to find out what you were doing. Although it's right, it took me two or three hours, when it should have taken me 10 or 15 minutes. I think that if you present it clearly, you will have the audience be able to understand what you're doing oQmE zi.ca.n eRe,h o t tz qs. 17C /2oz/ z96-oz6r (V04 /703/ 644-7636

1-64 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 much, much better. That's number one. Number two -- and Dr. Daisey jumped me on this, which I think is quite appropriate in some ways since she's a chemist -- is that I didn't find a connect between the cotinine studies between Chapter 3 and Appendix B, which is sort of surprising. I think that it's important because in Chapter 3 you base a lot of your exposure assessment on the cotinine studies, and here you're dealing with biases in your analysis and I think it's clear that you have to at least determine whether or not they're equivalent, and I think expanding on what Joan said, to understand the uncertainties associated with the different studies, if you have a degree of understanding as to whether or not the analysis and also the framework of sampling were compatible between these different studies. I don't see, at this point, making many great changes unless you come up with some kind of surprise, but I think clearly it helps your argument by defining terms, defining uncertainty and showing crossover. DR. LIPPMANN: Thank you. Anyone else? Yes, Dr.- Kabat? DR. KABAT: I didn't read this Appendix in depth, but I want to make two points and maybe you can explain, have d¢mezi.can ::R-.e/2o%tEas. 1~e tZoZi Z96-oZ6t Ndq 003/ 644-7636

1-65 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 enough time to explain. I have a concern that since cotinine only reflects smoking in the immediate past few days, I don't understand how one can come up with an estimate of misclassification of ex-smokers going back people who have c~uit 10, 20, 30 and more years ago. And this might be the bulk of -- it may be a greater tendency for people who quit a long time ago to suppress or deny their past active smoking. So that's a basic concern. A second point has to do with Table B.4 and B.5 where you present data from the Kabat & Wynder publication of 1984. I just wanted to remark that on Table B.5, line 2, the seven*misclassified people who we found out by looking at hospital records had told us in their detailed questionnaire, responding to our questionnaire, that they had never smoked, had never even dabble -- more than dabbled with smoking. We found out when we looked at the charts that they had told the physician, whoever took the history, that they had a history of smoking, and in some cases that they were smokers of two packs a day, but it was 20 or 30 years ago for some of these individuals. I just want to remark that we did that because that was one of the few things we could do to try to improve on our classification, but that appears to me to represent the y OTriLEZI.CQl2 GRE/].O'I.tEY3. 17C /2oz/ zq6-oz6r q'~Vq /703/ 644-76.36 . ,ra

1-66 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 lower bound. I mean, there's certainly more misclassification there. So this one percent from our study can be taken to mean that the actual misclassification has to be higher. DR. WELLS: Can I comment on that? DR. KABAT: Sure. DR. WELLS: There is also a perceived overlap. The cotinine studies are just doing current smokers. Some of those are going to get caught when they would have been in the ex-smoking category. The only tool we have to attack the ex-smokers are these two points in time interviews. So it's the only data source that we have. DR. KABAT: But it's not getting to people who quit 20 years ago because they weren't doing -- DR. WELLS: What we thought was that the two kind of offset each other. I don't know whether I'm making myself clear. Some of those ex-smokers who had also been caught in the cotinine studies -- I'm not so sure I've got it clear in my own mind -- there is an overlap between a cotinine studies and the discord in the answer study, a small overlap there. We thought that would take care of is underlying group that you're talking about. It wouldn't show up in either. DR. KABAT: I think that's whatof the nubs of this l U1 Ln J ~ c~7 »xEZi.ca.n G~e~io~7s N 17C /zoz/ zq6-oz6i q,6q t703) 644-7636

1-67 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 issue. If our experience in our hospitals getting these lifetime smoking histories has -- and Mike Lebowitz referred to something similar -- I think at the very least you need to point out what you can do with the urinary means to detect misclassification and what maybe is beyond the reach of urinary cotinine. DR. WELLS: Well, the cotinine wasn't used at all to judge the ex-smokers. That was for all the -- DR. KABAT: That's just current smokers. DR. WELLS: Even in your case, where you could be took the seven out before you made your analysis, we still have an 11 percent misclassification factor for your study for ex-smokers, so I think we've been pretty conservative, really, in terms of the way we dealt with the data. DR. LIPPMANN: Paul? DR. LIOY: Just one nore point. I had assumed that the cotinine information was just for current smoke and not for historical purposps. DR. WELLS: Right. DR. LIOY: I think that that's the point. It wasn't used for historical. All right? But your point is well taken. It cannot be. At least, I haven't figured out how. Ln F-A Ln ognzezlcan ::Re#ottzss 25C (202i 296-o26, qdf /703j 644-7636

1-68 1 2 3 4 5 6 7 8 9 10 11 12 1 ., 13 14 15 16 17 18 19 20 21 22 23 ) DR. KABAT: But that's got to be gotten by other means. DR. LIOY: By totally different needs. DR. KABAT: And it's got to be made clear to the reader, too, in some of the reviews -- DR. LIOY: Yes. DR. KABAT: -- because Fontham did use that and obviously that's a strength of the study, but it doesn't address the whole issue of smoking. DR. LIPPMANN: I think it's clear in both these Appendices that there's a need for the non-specialist to read this and be satisfied that this is being handled, and they both, so far, have been written more for your colleagues who are intimately aware of this and the logic of presenting it is important, even in an appendix, when you're talking about this particular subject. DR. WELLS: I might say, out of Appendix B, that I wrote a version and I,presented it to Stewart and he said, knrcadq ~ A that's e. Then he wrote a version and then he showed it to Brown, and he said, that's unreadable. And so he fixed it. I think the problem here is that this is such complex problem that it is extremely difficult to write it up o'¢m.ezf.ca.n tRz#oztzU Ze t202l 296-o26, q~vq /7031644-7636 ko N

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-69 in any way that's understandable by people just on a first quick read. DR. LIPPMANN: Still, but it's important to go as far as possible down that road. Steve, do you have any questions or comments about -- we'll give you all -- obviously, not going to take all morning.on this. PARTICIPANT: Jerry has some comments. DR. LIPPMANN: Oh, I'm sorry. Jerry? DR. WESOLOWSKI: I just have one comment as a result of your comment about 11 percent and how you're being conservative -- conservative in the sense that you tend to underestimate the effect rather than overestimate. This happens in a number of places in the document. As a reader, I think it's very important for me that you bring this out'because sometimes we're getting very detailed instructions here about some error your might have made, and that if you actually corrected the error, you would get more cases, rather than less. So, throughout the document there are a number of cases where you're being very, very cautious, so you couldn't be attacked for bringing out estimates that are overestimates, but you didn't really make that clear, that you are being cautious -- as someone said, you've got to be. ogrne7i.can eRrooa#ets. ZC /zoz/ zq6-oz6r q~vq /703/ 644-7636 Ln H

1-70 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 careful of that word because you can use it both ways. So, I think in the document, whenever you do overestimate, in your opinion, that should be stated. DR. WELLS: I think in a general way, Stewart and I have been working on this for about three or four years. It's pro bono work on Stewart's part. He does it when he gets the chance. We had the paper almost ready to submit to the peer review journal in response to Dr. Samet's comment. But what we've tried to do is take the best of what Peter Lee had originally and improve upon it and still maintain the three levels that you have to -- that I think you have to do, that being, the regular smokers and the turned smokers are your key. They're where most of this classification comes from. Then you have the occasional smokers who really contribute relatively little. And then the ex-smokers. And as we said, they're mostly long-term ex- smokers, and that's pretty well documented, but we could strengthen that. So I think what you have here is a synthesis, really, of Peter Lee's work, Stewart's work and my own work. And I think it's about the best we can'do. DR. LIPPMANN: Just a general comment about using material which has not previously been peer reviewed. In the ~ ~ Qn PTIrlEZl.CR/2 ~E~]04~EZ1 L~C t202l 296-0261 f 703] 644-7636

1-71 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 case act context on the ambient air pollutants, there's a general prohibition, but not an absolute prohibition, of that. In other words, when you use a non-peer reviewed work in that kind of a document, you're under an extra obligation to justify its use. Not that you can't ever, but when it's sufficiently and unique and important to use it, you're obligated to provide more justification than you would otherwise do in order to use it, and I think that applies in this case. So, a little more justification of this is warranted. Any further comment? Let's move on then to Appendix C. Dr. Weiss. APPENDIX C - REVIEW FORMAT FOR CASE-CONTROL STUDIES DR. WEISS: Appendix C is the review format for the case-control studies of lung cancers, basically a data abstraction form that's divided into five parts: the general characteristics of the study, data collection which includes study design features; clinical data; statistical analysis; and part 5 which is potential confounders and effect modifiers that were considered. Part 5 was mislabeled; it's called "Dependent Variables." These aren't dependent variables, these are independent variables, covariants or confounders for parentheses sake. So that piece is AniE4.LCR/2 GRE/2o7.tE'i1 Ze [202l Y96-o261 q'6q 17031644-7636 ~ #--A Ln %D N m

1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 - 17 18 19 20 21 22 23 1-72 mislabeled. I think the form is reasonably complete in and of itself. I think the problem that I had is that there's no preamble to this or description of it and how it was used and in the context -- actually going back to Dr. Samet's comments about the review of the studies on ETS and lung cancer, you almost wonder whether something ought to be -- this needs an introduction is what it needs. It needs something that says look, this is how this was used and the people who abstracted the data did this, and so on and so forth. But as it stands now, somebody just reading the Appendix, you're just looking at this form and you don't have any context with whiph to judge it. I think the other issue is that many of the form entries are free form. For example, the study uses the term "nonsmoker" or "never-smoker" to mean, and then there's a space for -- you know -- you could fill in the exact sentence from that particular study. But it isn't clear how this was then translated or reduced in a quantitative sense into some -- to be actually used In a meta-analysie. I mean, there's a finite number of definitions of nonsmoker or never-smoker in these studies. You didn't just take the text here, you know, you had to then categorize ,qrnEZlca.n eRefiotEets 2~e 1zo2j 296-oz6, wdq /7o3f 644-7636

1-73 ) 2 3 4 5 6 7 9 10 11 12 13 14 15 16 . 17 18 19 20 21 22 23 these definitions, so that there's another step beyond this form in terms of how you utilized the data. I think the form is complete. I don't think that that's a problem. I didn't come up with things that were missing. But in terms of actual things that you were abstracting from the study, I think that it needs some kind of introduction as to how it was used and maybe something at the other end in terms of how when you took this, what were the actual number of categories or definitions or so on and so forth. I mean, it needs some sort of -- I suppose the other aspect of this which could be applied to any of the other studies, I think that we used in the meta-analysis or risk assessment, you know, the same -- this form applies to the case-control studies but you could have a similar form that you use for the cohort studies. If you did, how did it differ from this? I mean, I dqn't think you want an appendix that's 20 pages here, but I think you need a little bit more -- it has to be a stand-alone thing. Somebody ought to be able to turn to this and try to figure out what you did. And I don't think you can do that just from this form. I think that's the problem. C4rnezi.can cRr_#oztEts '1.~C /zo2/ zq6-o26i CV04 /703l 644-7b36

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 . 17 18 19 20 21 22 23 1-74 Now, you may want to give some thought to as to whether -- we haven't gotten through all five of the appendices -- as to the order of these and whether they're in the right order. If you think that this -- again, going back to the comment that Jon made, in reading over the summaries of those cases -- that they were prepared from this form -- DR. BAYARD: They were. DR. WEISS: -- then what you want to do is put the form up front and you want to say how you used it and exactly how you did it. And then you can go to the summaries of the studies and it can be organized in a little bit more of a logical fashion. So I think that those are my major comments. There are a few misspellings and some other things. DR. LIPPMANN: Anybody else? Jon? DR. SAMET: Two minor and very specific comments. One on the classification histologic type of tumor. It's probably useful to note whether that was done by a review panel or not because data brought in from either routine clinical records or registries that have not been reviewed are certainly subject to the misclassification that comes with multiple observers. Just another comment. In fact, there's a lot on ~{mc~.i.can dQa~zo~ts 'L1C /zozf 296-0261 qU.4 /7031 644-7636

1-75 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 here for airway proximity, and I think that one or two studies attempted to provide cases as central or peripheral. I'm not aware of criteria for doing that other than radiography, which really has nothing to do with airway generation. So this is probably somewhat of a moot point in terms of utilizing this item once you identify it in the studies. DR. LIPPMANN: Okay. Let's move on then to Appendix D. Dr. Stolwijk. APPENDIX D - LUNG CANCER MORTALITY RATES ATTRIBUTABLE TO SPOUSAL ETS IN INDIVIDUAL EPIDEMIOLOGIC STUDIES DR. STOLWIJK: Appendix D describes the attempts to determine the lung cancer mortality rates attributable to spousal ETS in individual epidemiologic studies. This becomes a necessity because the background exposures and the background rates are very different in different studies and in different countries. And in order to calculate the attributable risk, there are adjustments needed, and the basis of this and the problems with it are described in this Appendix. The tables in the Appendix describe the female lung cancer mortality from all causes and as attributed to spousal smoking and never-smoking. The difficulties in it result . Am.eu:ea.n ::RehoztEas 2$C /zozJ z96-oz6i qUo4 /703l 644-7636

1-76 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 from the fact that if there is not a constant background of lung cancer in females, as there isn't, in most countries, the female mortality from lung cancer is on the way up and in the United States, for instance, it's on the way up very rapidly, so that the time at which the study was made, if there are comparisons that needed to be made, it needs to be taken into account. The way the cases were actually obtained in the population and the study is of great importance. The decisions for exclusion of certain types of cell types or decision inclusion of different cancers affect.the rate of lung cancer mortality in these populations, and all of these factors are dealt with in a proper manner with an explanation of what these difficulties are and what the sources of the data were. I think that in general the attribution of lung cancer risk to two different categories is a thing that is fraught with considerable difficulty, and also the attribution from other causes is not always widely acceptable because of differences in the different populations that were on the study and all of these cause complications that are, I think, appropriately listed in Appendix D, and these results are then later on used in the assessment of population P" l Ln ~ N me~i.can ~~~.o~u 1~.. e IZo2l Y96_oZ6, N6q /7031644-7636

1-77 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 . 17 18 19 20 21 22 23 impact. I have no specific difficulties with the presentation that we have here. The material appears to have been carefully collected to be applicable to the various studies. I think it represents an appropriate way of trying to assess the risk attributable to spousal ETS which is, after all, the major factor of the study in the various studies that are being evaluated in the chapters in the report. DR. LIPPMANN: Would anybody else on the panel like to comment on this Appendix? (No audible response.) DR. LIPPMANN: Then we'll go on to the last Appendix, Appendix E, Statistical Formulae. Dr. Rockette. APPENDIX E - STATISTICAL FORMULAE DR. ROCKETTE: This is short Appendix, two pages, which gives some standard formulas for risk and/or variances in case-control studies and in cohort studies. WaS I felt it as rather straightforward. There is a typographic error, I believe, on the first page, E.1. where a log(OR) is used instead of a log of the odds ratio. But I think other than that, I did not find any errors in it and thought it was acceptable and clear. dgmEZi.ca.n lz~?e/2o%txas `L1C Izoz/ 296-0261 q~vq 1703/ 644-7636

1-78 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 , 17 18 19 20 21 22 23 DR. LIPPMANN: The other reviewer of this is Dr. Woods. Again, I'll read from his comments. The organization and readability of this Appendix is a substantial improvement over the previous draft, which was Appendix B in the 1990 draft. However, a few points of clarification would be helpful. One, the basis for assuming a log normal distribution of data analyzed by the formula in this Appendix should be stated or referenced. Two, a reference or justification of the test statistic given in the last paragraph on page E.1 should be provided. Three, should not the term "RR" in the denominator of the test statistic in the last paragraph on page E.1 be "OR?" Jim thought that that might have been a misprint. Four, a reference or justification for the variants of log(RR) in Section E.3 should be provided. Those were the specific comments he had, and you have the written version of that. Does anybody else have any comment on this particular Appendix? (No audible response.) DR. LIPPMANN: Well, it's clear the Appendices need Lq Oj/72BZLC6tn GREf2odE4s. 2~C /zozf zq6-o26i qUo4 /7031 644-7636

1-79 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 _ 17 18 19 20 21 22 23 some careful re-examination, and in some cases, clarification. I don't think that anybody has identified any fatal errors, although some of the numerical conclusions might need to be adjusted after taking into account these. And if some of the small errors that have been identified are really errors and not just misunderstandings, that might change some numbers in the conclusions. But by and large then, these have been important things to include, just so we can understand the text drafters that come ahead, and we leave the clarifications and corrections, if necessary, to you, and assume then that these Appendix materials enable us then to have an appropriate discussion of the actual chapters and the use of the techniques describing the Appendices. We are then finished with our morning schedule a few minutes ahead of time. We will resume very promptly at 1:00 p.m. since we have a long afternoon. There's no excuse for anybody to be lats for the resumption of this session. I don't need to respond to anybody on the panel to call the morning closed. Whereupon, at 11:33 a.m., the hearing in the above- entitled matter was recessed, to reconvene this same date at 1:00 p.m.) PT~2ELLCQ.I2 GRE/7.o2tz'LS. Ze 1zazJ 296-oz6, q/6q 17031644-7636

1-80 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 AfT B R N O Oh S F. S SiO N_ (1:03 p.m.) MR. FLAAK: Before we get into the meeting, I'd like to ask all of the speakers who are scheduled to speak this afternoon starting at 3:00 p.m. -- and we will take a short break just before the speakers start -- I'd like to you to go to the back table sometime in the next hour or so and let my assistant, Carolyn, know that you're here and to clarify with her what your audio-visual needs might be, whether you need the 35 mm or the overhead. We do have both, but it's a matter of moving the table back and forth for either one of them, and we'd like to know which piece of equipment you will need to help us for our planning. Thank you. COMMITTEE REVIEW OF CHAPTERS, 4, 5 AND 7 DR. LIPPMANN: We're going to review Chapters 4, 5 and 7 this afternoon grior to the public comment. The first Chapter is 4, Hazard Identification I: Lung cancer in active smokers, long-term animal bioassays, and genotoxicity studies. We'll start the review with Dr. Stolwijk. PARTICIPANT: (Inaudible.) DR. LIPPMANN: Oh, okay. We'11 go on then to Dr. ~ ~ ogmEZican cIQEfiozEeqs to ti 17C /zo2J Pq6-oz6r 471 No4 /703/ 6q -7636 N cn 4 ~o

I I-81 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 Eatough and then come back to Dr. Stolwijk. CHAPTER 4 - HAZARD IDENTIFICATION I: LUNG CANCER IN ACTIVE SMOKERS, LONG-TERM ANIMAL BIOASSAYS, AND GENOTOXICITY STUDIES DR. EATOUGH: I will actually have much more to say when we talk about 3 and 6, in spite of the fact that I've been asked to give a formal review on this, because the toxicology that's in the chapter is somewhat out of my area. There were two things that I thought were really missing in the chapter to really bring together the argument. And part of what's missing I think reflects deficiencies, particularly in Chapter 3, that we'll talk about tomorrow in terms of quantifying the differences between the environmental tobacco smoke and mainstream smoke. A lot is made of the similarities between the two but it's discussed very much in a very qualitative nature. I think a number of things could be done to help tie that up. Much of.the data that would tie that up, that would help, I think, to bring that together, is within the chapter, but the comparisons are not very qualitative with yf~ respect to is the xpectan potency for sidestream smoke, environmental tobacco smoke, greater than or less than, what's the sign, and is the difference between the two'orders c4m.ezi.ca.n 'CRE/ioztev.l '1-~e /2021 296-oz6, q,04 /703j 644-7636

1-82 1 3 4 5 6 7 9 10 11 12 13 14 15 16 , 17 18 19 20 21 22 of magnitude difference or less than orders of magnitude difference. And depending, of course, on the conclusions you make, the comparison become -- the similarities between environmental tobacco smoke and mainstream smoke become more or less important. If the prediction is that environmental tobacco smoke is less potent, then mainstream smoke and the differences are orders of magnitude, then you come to different conclusions than if the reverse is true, that you expect the potency to be greater for environmental tobacco smoke. And then the question of the order of magnitude of difference becomes less important. Those were really the concerns that I had as I read through it, is that there really was nothing to let you quantitate the statements that are made about the comparison between environmental tobacco smoke and mainstream smoke that would let you put both the statements in this Chapter and the statements in Chapter.6 into perspective. I think that a good part of that is deficiencies in Chapter 3 that could address some of those questions but those, I think, are things that we'll talk about later when we review that chapter. 23 1 DR. LIPPMANN: Dr. Kabat. N G/7IIZEZLCQI2 GREho'CtE'LS. 25C /zoz/ z96-o261 q/04 /703/ 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 ~ 13 14 15 16 . 17 18 19 20 21 22 23 1-83 DR. KABAT: Most of the epidemiologic material contained in Chapter 4 is drawn from previous reports, the Surgeon General's report and the IARC tobacco report. On reading it through, it appeared to me to hit the major points that I would like to see made on the known effects of tobacco. I'm not sure whether all of the tables that are included need to be included. I could see a more pithy summation of the evidence on active smoking and lung cancer, but that's more an editorial decision. That's basically it. DR. LIPPMANN: My own comments on this chapter are generally that it does the main job it was intended to do. It provides the concise and adequate summary of relevant but indirect evidence supporting biological plausibility for a causal association between long-term exposure to ETS and an elevated risk for lung cancer. The clear dose-related association of lung cancer risk with exposure to.mainstream smoke, the presence of essentially all of the same known carcinogens in both mainstream and sidestream smoke, the evidence that you can get, at least laryngeal cancer is in the Syrian hamster with mainstream smoke, and the greater carcinogenic potency of sidestream tar than mainstream tar in the mouse skin are all ogmeti,can eRE#ottevs. 17C /zoz/ 296-0261 q~o4 t7031 644-7636

1-84 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 consistent with an elevated risk of lung cancer in nonsmokers chronically exposed to environmental tobacco smoke. This is all qualitative, as Dr. Kabat suggested, a kind of chain of the evidence we offhandedly suggested be put in this document, and it's there. In my mind, it's adequate as supporting evidence. I would, however, like to see the text gone through to change some words here and there. For instance, on the first paragraph, a "notable cancer risk," I think you mean an "elevated cancer risk." I'm not sure what "notable" means. In the second paragraph, it talks about these "excellent reports." You might say they were "comprehensive reports." I don't think you want go grade them in the sense of -- with this kind of loose usage of words is a distraction, really, from this report. In the next paragraph, Jan points out "the wealth of" and so forth and so on. DR.•STOLWIJK: Especially a wealth of human cancer studies. DR. LIPPMANN: I think that simply detracts from what basically I think is a sound chain of evidence discussion. There's also in a number of places, a confusion of ETS and sidestream smoke. I think you want to very carefully OT;ME2LCQ.12 GRE f20'LfE21 2~C t2021 296-026i Q4a4 17031644-7636

1-85 go through and say "sidestream" when that was the vehicle that was evaluated. Not that it necessarily wouldn't apply to the other, but to be precise about what it is that you're doing. I guess I have the most significant bone to pick on 10 11 12 13 14 15 16 .17 18 19 20 21 22 23 page 4.10, line 13. There is a statement that "due to the similarity in chemical composition between mainstream and environmental tobacco smoke, and the known human exposure to ETS, ETS would also be classified as a known human carcinogen." It seems to me that statement is not warranted. If it were, there would be no need for Chapter 5. In my view, the material in Chapter 4 adds a strong basis in terms of biological plausibility for the positive associations between ETS and enhanced spousal risk for lung cancer. It is'the weight of the evidence in both Chapters 4 and 5 that justifies the classification of ETS as a known human carcinogen. So I'm not dissenting from your overall judgment, but I find that the evidence in Chapter 4 is not evidence in and of itself. It lends support for the human studies with real environmentally relevant levels. Without that plausibility, the less than conclusive nature of any individual study makes the overall thing look weaker. But I L^ ~ ~ Ln w N ~mezEcan eRE~.o~qs rn ~C /zoz/ 2~-oz6, ~ woq /7o3I 644-7636 'p .

1-86 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 with this plausibility, this chain of evidence, as we suggested back in late 1990, it hangs together. So I think that that statement at the very last paragraph of the text of Chapter 4 on page 4.10 needs qualification. Are you ready, Jan? DR. STOLWIJK: Yes. In the review of this I was pleased to note that the evidence that exists for mainstream smoke being a carcinogen, as has been evidenced in a very large number of studies satisfying almost any criterion that one might wish to apply to it in terms of biological plausibility, in terms of dose-effect relationships, in terms of reversibility of the risk after cessation of smoking, and also in terms of the animal evidence that has been accumulated both for mainstream as well as for sidestream smoke, is all very consistent with the fact that mainstream smoke would be classified as a Class A carcinogen, although to my knowledge, EPA never has actually done so. It has never entered into the risk assessment system within EPA, but it would be an appropriate thing based on its rules of evidence that it be a Group A carcinogen. This, of course, would refer to the much higher concentration and a much higher exposure and higher doses that would be y ti 4472EZCCQK GRE#04.tz'L1 `L1C /zozJ 296-o26r N6q /7031 644-7636

1-87 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 . 17 18 19 20 21 22 23 occurring in people exposed to mainstream smoke. However, the biological plausibility for ETS being a human carcinogen is certainly supported I think very effectively in the report in this chapter. I would, I think, agree with the Chairman that the conclusion that it by itself would be classified as a known human carcinogen in the absence -- if there were no other evidence, is probably not a completely defensible statement at this phase of the document. DR. LIPPMANN: Vic. DR. LATIES: I wish the tables would be more completely explicated somewhat. I couldn't understand Table 4.11, for instance, where the lung cancer death attributed to tobacco smoking in certain countries. It's in the second column, "Number of Deaths." Is that the number of lung cancer deaths? And then what's the expected death in nonsmokers? The first column must be "number of deaths in smokers by lung cancer" and in nonsmokers, is that "ex- smokers" or is that environmental tobacco smoke predicted, or what? I think that Geoff, when I was talking to him -- Geoff Kabat -- suggested these tables come from some other document. And if so, I think more of the other document y cn co ti 44mati,can Z~?E#ottets L`lC /2oz/ 296-0261 Nd4 /7031644-7636

1-88 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 , 17 18 19 20 21 22 23 ought to be brought over into this document so that it would be crystal clear to the uninitiated what they're seeing here. And there's another problem later in "nonsmokers." What does that mean? DR. KABAT: Or else it could be boiled down to something that was more readable by the general reader in talking about the magnitude of the risk and dose response and quitting, and so forth and so on, and anyone would be able to follow it. DR. LIPPMANN: Clarification is always desirable. Joan, you had your hand up. DR. DAISEY: Just some minor things. I certainly concur with all the comments made so far, including a little more precision in language would be useful. One of the places where this applies might be in the use of the word "condensate" in the animal studies. In the old literature, my understanding is that is the mainstream smoke, but.you ought to be sure on each of those studies that that is actually what they tested. It sometimes pays to look at those details. DR. LIPPMANN: There was one comparison where they explicitly used sidestream condensate in comparison to what is apparently mainstream. ,vqm,e-Ucan cRr-#o%tzts 'L1C /zozJ 296-026Y Q404 /7031 644-7636

1-89 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 , 17 18 19 20 21 22 23 DR. DAISEY: Right. Yes. And it has to be clear exactly what it is they're testing. I think you could refine it a little bit. DR. LIPPMANN: Kathy. DR. HAMMOND: In the clarification department, I had a lot of trouble with Table 10 in trying to figure out what the results were because I think they vary. It's a very interesting table, but the results are different results as you move through the different studies. Sometimes they're relative risks and other -- and sometimes I couldn't figure out what they were. DR. LIPPMANN: So, there's a big blank space, and most of the comments, and perhaps comments are needed. DR. HAMMOND: Well, I think that there may be even a -- the columns have different meanings, as you move through the studies. This needs to be clearer, of what's going on. Generally, I concur with most of the comments that have been made so far. I particularly enjoyed seeing Table 4-3 with the dose-response, and also even at the low levels of smoking, the evidence that there were still effects seen at the lowest levels. I'd like to ask a very naive question. For the same level of smoking, the mortality ratios in women are much dgme~~ca.n ~e~i.oz.t~ts , ~ 2~C /zo2/ zq6-oz6, Q404 003/ 644-7636 Ch Co

1-90 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 lower than the men. Is that because the nonsmoking rates are higher or because there seems to be less of an effect? DR. LIPPMANN: No. Cigarette smoke today is not an exact expression of dose. How you smoke certainly can make a difference. DR. KABAT: How long you -- DR. HAMMOND: Well, is the thought that depth of inhalation and how they smoke -- DR. KABAT: Which table are you looking at? DR. HAMMOND: 4-3. The reason that I bring that up is I think that the answer to that would have consequences in some of the later extrapolations. DR. KABAT: This was baseline smoking data from 1959 in the first ACS study. So if you go back to 1959, that was really the'beginning of the taking off of women's smoking. So if you took the average male smoker and the average woman smoker, back then the average male smoker had smoked for 30 years or something, 20 years, and the average woman smoker would have smoked for a much shorter period of time. So I mean, that probably adds -- DR. HAMMOND: Yes, because that's not in here at all. 23 DR. KABAT: Men typically smoke a good deal more OTI)2E2CCQ12 G~?Elt.O'itE4.3. 2~C I202l 2()6-0261 Ndq t703/ 644-7636

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-91 than women. But that wouldn't account for -- this is stratified by that -- and inhalation. DR. HAMMOND: Right. DR. LIPPMANN: I think it's my impression, that maybe they just that they inhaled more deeply, and that would make a difference. DR. HAMMOND: What I was interested in is if there was a gender difference in sensitivity, which, if there were, would be important later in the extrapolation of female nonsmokers to male nonsmokers. That was why I asked that. DR. STOLWIJK: We now have achieved in Connecticut a happy situation where the current incidents of lung cancer in women between 40 and 60 is the same as that as men. DR. KABAT: Full equality. DR. HAMMOND: Yes, right. DR. LIPPMANN: Let's not pursue that any further. Any other comments or questions on Chapter 4? Paul? DR. LIOY: 1Iaybe I should just emphasize Delbert's comment. The beginning of Chapter 4 is going to have to do a lot more to link within Chapter 3, and it's unfortunate that we haven't been able to go through Chapter 3 first because I think there's some logic that has to be put into the beginning of Chapter 4 that is not here at the present time. dgm.E¢i.can eRe/2ottzts ZC /zo2/ 2q6-oz6i qVd{ /703l 644-7636 v m

1-92 1 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. LIPPMANN: Well, that's unfortunate. I did that in part because of the people who were going to be here yesterday and not today. DR. LIOY: We are losing something by not doing that. DR. STOLWIJK: We may need to come back to -- PARTICIPANT: I think that would be a good idea. DR. LIOY: I think it would be excellent to revisit it tomorrow. DR. LIPPMANN: I think in finding that Chapter 4, which is crucial, is certainly minimal, possibly adequate, with -- that you have both these oral comments and written comments, why don't we then move on to Chapter 5. Dr. Samet. CHAPTER 5 - HAZARD IDENTIFICATION II: INTERPRETATION OF EPIDEMIOLOGIC STUDIES ON ETS AND LUNG CANCER DR. SAMET: This chapter overviews the epidemiologic studies of ETS and lung cancer, evaluates these studies and provides a meta-analysis. Summary analyses are presented by geographic region and by tiers of study quality. The chapter ends with an overall evaluation of the evidence and the conclusion that ETS is a Group A human carcinogen. In terms of a few general comments, I support the general approach of reviewing the studies, classifying them oqmezi.can =Re_#o%tzts L1C /zozJ z96-oz6, N6q /7031644-7636

1-93 1 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 by quality, and calculating a pooled measure of association. But I identified a number of major problems and then some minor ones, and I'll take you through those. I think the same concern that I voiced with regard to the.Appendix immediately arises in reading Chapter 5. That is, there's not been a good formal framework of epidemiologic methodologies set out for reviewing the evidence and discussing the various types of bias. In my written comments, I make some suggestions which I really voiced earlier with regard to the Appendix about how you should systematically approach the problems that affect case-control studies and address them. I think, again, while there's some general language in here, for example, on what confounding is and some discussion of aspects of information bias is never, I think, drawn together as well as it could be. I think, again, this problem then returns when the individual studies are discussed in the body of the chapter, just as it did in the discussion of the individual studies in the Appendix. The footnote to Table 5.1 is the description of the development of the tiers of the studies, and because this serves as the basis for selecting groups for into and out of ogrnEZfcan =-Rz#ottzas. `LlC /zoz/ zq6-oz6r Q/04 /7031 644-7636

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 r 1-94 the meta-analysis, I think that this needs better description of how it was done. Was it done by one individual? Was it done by several individuals? Was it reproducible? Or to more fully develop formal criteria and apply them in a blinded fashion to the investigations by several people and see if this classification is in fact reproducible. I don't necessarily dispute the rankings that you've assigned to the studies; I just think that there should be an explicit methodology. So, those were my concerns about sort of the general approach to the epidemiologic literature. I think sort of a corollary of this which I have written as a major concern is that in fact the text just still needs careful editing, and I kept my panel on it while I read it. I think there's some vagueness in the use of epidemiologic terminology that I think introduces ambiguity, and it needs to be corrected. I'm somewhat concerned about the discussion and handling of confounding. There is a discussion of what a confounder is that I feel is subsequently in fact ignored in the discussion of the literature. You turn to -- I guess this is now page 5.17 where there's a formal discussion of what a confounder is and at C4mEZi.can Z~?e#OttEu ZC /2021 296-0261 Q'6q 003/ 644-7636 N

-I-95 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 least without delving into the complexities of confounding, the confounder is a factor that in its own light is the cause of the disease of concern and is also associated with the exposure interest. And there seems to be substantial confusion between variables that are found to be associated with environmental tobacco smoke, which have not been causally linked to lung cancer in never-smokers, but nevertheless, they're discussed as though they were confounders, and they're simply not confounders. They have not met any evidentiary standards that we're now in fact applying to environmental tobacco smoke. So I think it's incorrect to discuss some of these matters such as family history of lung disease in never- smokers where the evidence is not sufficiently amounted to regard them to causes of lung cancer in never-smokers in their own right. So I think a definition of confounding is offered, but it is not then in.fact applied to the literature that follows, and it should be. I think this really goes back to my earlier comment that you need to make sure that your handling of these epidemiologic concepts is precise and consistent, which it is not. I think my fifth major concern really has to do (J) PTN2E'LLCQ.I2 GRE/20R.tE'C,'i 2~C t2021 296-0261 q4~q /7031644-7636

1-96 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 with I think an incomplete handling the studies from China. These studies were placed into the fourth tier. That might be appropriate, but I think a large body of evidence having to do with lung cancer in China is not cited, and that in fact would include work done by the Environmental Protection Agency looking, for example, to work with Mumford and others, looking at the activity of the coal smoke in China. There's also a substantial number of case-control studies published in Chinese that have now been reported in the U.S. literature that should probably be brought in as a basis for better understanding the role of indoor air pollution in China and what the additional contributions of environmental tobacco smoke might be. If you're not familiar with those studies, I can direct you to them. I had a series of other minor comments that will be included in my written material that I will supply, as well as my own sort of editing of the document. I think, again, good attention to the,language would help a great deal. DR. LIPPMANN: Dr. Wesolowski. DR. WESOLOWSKI: Well, I had a few editorial comments which aren't worthy, I don't think, of the Committee's attention, which you have in my written statement. And I certainly don't disagree with anything Jon v Ln .vqrne%i,can z~%_#ottEu 1.`~C /zoz/ zq6-oz6i ql,Vq /703/ 644-7636

) 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-97 says. But I would like to say that I think the EPA staff should be complimented in putting together a lot of studies in a way that was readable. And also, I was impressed with the different approaches they used to analyze, all of which seemed to come to the same conclusion. The basic question asked of the Committee with regard to this chapter was, does any of this new information or this different way of analyzing it alter our previous conclusion regarding the characterization of ETS as an EPA Group A carcinogen. I should say I saw no evidence in that chapter that the Committee should change its conclusion that ETS is a Group A human carcinogen. In fact, I think the way the evidence was presented and the evidence that was presented makes it even more compelling than it was the last we met. So that's my basic conclusion. DR. LIPPMANN: Dr. Kabat. DR. KABAT: There's been a big improvement, I think, in the level of the treatment of the epidemiologic studies on lung cancer compared to the 1990 draft. I have a whole slew of specific comments on the text and on the tables where I felt that the summaries that have pulled out break up the flow of the text and I suggest that maybe the summaries 1:Vqme%i.can Z~?e,fiozteu L`1C /zoz/ z96-ozb, q':Vq /7031644-7636 N

) 1 2 3 4 5 6 8 9 10 11 12 13 14 19 20 21 22 23 1-98 of the studies could just be kept to the Appendix and can be referred to in the text. I felt that the treatment of the issue as confounding by diet was a little bit -- diet is maybe one of the prime confounders that needs to be taken into account, but I'm not comfortable with the dismissal of the role of diet in lung cancer in nonsmokers. I don't think the small number of studies which present information on ETS exposure among nonsmoking women and diet allow one to make a strong assertion that confounding couldn't be going on. So I would rightly point out that Linda Koo, who is very interested in this whole topic of the apparent effect of ETS being due to confounding by diet, that women exposed to smoking husbands have a lot less healthy lifestyle and less healthy diet -- that she hasn't analyzed her own data to evaluate confounding. I think that's a good point. I would like to see that conclusion tempered a little bit because based on the ample studies of diet in relation to lung cancer occurring in smokers as well as nonsmokers, it seems to me that we can't rule out possible confounding. DR. LIPPMANN: Do you want to respond specifically to the way you handled diet? 44mEZi.ca.n cRe#ottEts. `LlC /202/ zq6-o26r "Vlvq /703/ 64q-7636

1-99 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. BAYARD: I just thought the last paragraph back on page 23 really did point to the paucity of data. Based on the studies that we could analyze, we did, and we just said it's difficult to -- DR. KABAT: Right, but then you arrive at the conclusion that there's no fact that it could confound this association. And I'm just saying that that seems to me to be putting a lot of weight on these six or so studies that have any cross-classification of ETS exposure by diet. It's a question of modified emphasis. DR. STOLWIJK: I think what you're pointing out is that the fact there is not strong evidence in these studies ought not to be a reason to throw out -- DR. KABAT: That's it. It is very limited. It's a really important question. It's a potential confounder. That's the point. DR. STOLWIJK: But you cannot assess the confounding in most of these studies. DR. KABAT: That's right, but it doesn't totally eliminate the question. DR. LIPPMANN: Are you finished? DR. KABAT: One of the points that I made in ~1 /'i' G G 1 ~ ' discussing Janeric , there was something in the description ~ C 4 mezi.can o3Eets l~e tZoZI Z96-OZ6, q~v{ 1703i 644-7636

I-100 ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 that I couldn't find anywhere in the paper; namely the statement that "Overall spousal smoking is not positively associated with lung cancer. Whereas, all cohabitant is somewhat associated." I could not find any evidence for that in the text of the paper, and I wondered whether this was based on a comparison of the table showing the effect of smoking spouse versus the effect of total household exposure, but I think it's questionable. DR. LIPPMANN: What page are you on? DR. KABAT: This is in the discussion of the J:,,,ey~c.h S-3Z -aarrer-i-ek paper on page 59S. The statement, "Whereas all cohabitant smoking is somewhat associated." DR. STOLWIJK: Does that have to do with the children's exposure as a child? DR. KABAT: No. It's separate from that. This is in adulthood. Because that's the big conclusion, that people exposed in childhood and adolescents to 25 or more smoking years -- but this is something else. My comment on that_is that this is actually sort of J"t rler,i~ a confusing finding in the Janeri:sk paper. What they're saying is that other household exposures in adulthood are more important -- contribute more smoker years of exposure in adulthood than spousal smoking. That's contrary to what we c7qm.ezi.can ::R--e#oa.Eezs ze t20z1 z96-O2G1 q,6q /7031644-7636

I-101 1 2 3 4 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 see in our case-control study. It seems to be at variance with what you can drive from the Fontham paper, too, where it also appears that spousal smoking is the major source. Anyway, there are other details like that which I've mentioned in my write-up and I don't need to go into them. DR. LIPPMANN: Mike, do you have some comments? DR. LEBOWITZ: Yes. In regard to what Dr. Samet said about the Chinese studies and also in terms of the issue of confounding that Dr. Kabat brought up, I wanted to mention a large number of papers that are published in the proceedings of Indoor Air 1990. I chaired a session which had 18 papers, 15 of which were from China, most of which dealt with lung cancer, in most of those lung cancer in women due to various'indoor combustion sources where a number of them tried to fraction out the complication of ETS versus charcoal wood, et cetera. It does discuss the reason, or it presents the basis for considering these other exposures at least as confounding factors in nonsmoking women in that country. It also shows that the level of scientific investigation there is actually quite good. Again, it's one of the first sets of a series of papers that have actually been translated into Ln Ln dTrnezl.can eRe#oztzq.2 'L1C /2o2f 296-o261 q~Vq t7031 644-7636

1-102 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 English and present quite a bit of data on the situation in that country. So I think that, as Dr. Samet said, it would be worth looking at that a little bit further. DR. LIPPMANN: Okay. Is that it, Mike? DR. LEBOWITZ: Yes. Thanks. DR. LIPPMANN: Before I forget, now that most of the audience that's coming is here, somebody in the audience found this pen. So if anybody is missing it, come and get it. Is anybody else commenting on this chapter? DR. DAISEY: A question, perhaps, on the food S-Z3 chapter back on page 323. The statement is made that the potential for bias, and particularly for confounding, is high. I didn't quite know, is there a reason to think it's high? I wasn't convinced from what was presented here that there was reason to think it was high. I mean, maybe it is; maybe it isn't. Is there a reason to think that there is a high potential for confounding? DR. SAMET: This is what I was trying to say in my comments. I think if you look at page 51-&, there's language for that. I just think it is not correct. A number of other variables have been considered as potential risk factors for lung cancer, and thus confounded to the ETS lung cancer association. Something that is a potential risk factor is c4mezi,can lz~?-e#o2tEas. 2~C (zoz/ z96-oz6, w6q (7o,3l 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-103 not a confounder. I think that's really what Joan is perhaps responding to in relationship to diet. Specifically, we actually have precious little information about the role of diet in people who had never smoked as a risk factor for lung cancer. Much of it is probably contained within these studies, in fact, and I think if we did a detailed review looking at biological plausibility, et cetera, we would be unable to draw a conclusion at this point as to the role of diet in never- smokers. I think it would be hard, then, using our conventional criteria for confounding to conclude that diet is thus a potential confounder. Dietary patterns seem to be correlated perhaps with both active smoking and passive smoking exposure in some cultures. That does not, again, make them confounders. That was my concern in reading the language here. These are correlates.. Perhaps in the future our knowledge on diet should be strengthened, but right now we don't have all the epidemiologic nor other evidence in hand to understand fully the role of diet. It certainly has not emerged as a strong factor yet. So why would it be a strong confounder if it's in fact -~' og=zi.can ~E~o~ts l~C 1202) 296-C26, CI'6q /7031644-7636

1-104 1 2 3 4 5 6 7 9 10 11 12 13 14 19 20 21 22 23 not emerging, at least in the initial studies, as strongly associated with lung cancer risk? DR. BAYARD: It just became a matter of discussing this issue. If we didn't discuss it, then we would have faced the criticism that we didn't talk about it. DR. DAISEY: Yes, but this makes a value judgment that I think is not supported. I mean, you make the judgment that there is a high potential, but it doesn't seem as though there's any evidence to suggest there is a high potential in fact. DR. FiANMOND: Just a potential. DR. DAISEY: There's a potential, but it's not necessarily high. If it were really high, it probably would have come out of these other studies. Do you see what I'm saying? The judgment that it's high is what I'm having problems with. DR. BAYARD: I think we made a suggestion that there was a high poteDtial and then we analyzed the data to see whether or not we could find any. DR. DAISEY: No, but this is sort of in the S-z 3 conclusion at the end of *2'3. Maybe you just need to look at it later. I think it has to do, again, with the issue of some use of language and being very careful with what you Ln cn N oN dgmezlcan ottets N CD 1~C e t2021 296-0261 !a Noq 0031644-7636

1-105 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 say. DR. LIPPMANN: Careful with your adjectives. DR. LIOY: I don't think it's just the language in this case. I don't think we know enough of what variables in diet we're going to have to start thinking about as confounders. DR. DAISEY: But then you can't make the judgment there's a high potential. DR. LIOY: That's right. DR. DAISEY: Or a low potential. DR. LIOY: Or a low potential. That diet has to be investigated to determine whether it is a confounder and then the degree to which it's a confounder will be borne out by the nature of the studies that you conduct. Right now it's just that you know that diet is another actor that could be considered toward, but you're not sure what it is. I mean diet itself is not just a food basket survey, it's how you prepare your food, the types of foods you eat, whether or not it's broiled or fried, or a whole host of other things that one has to consider in diet besides food market. So I think there are a lot of potential problems here in making any judgment at all at this particular time. I think that any judgment at this point is oTmEZi.can eRe#ottzu L~e 12021 296-026, qU6q 17031 644-7636

1-106 1 2 3 4 5 6 very premature. DR. LIPPMANN: Well, I guess what you need is any evidence for a differential diet between smokers and nonsmokers or passively exposed people and non-passively exposed people because if diet has an influence, but it has no differential influence, then it's not a factor. DR. LIOY: That's right. 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 DR. LIPPMANN: Any other comment? DR. SAMET: Just to return to this issue. This may be one way that in handling the evidence from multiple cultures, if you can begin to assess the role of factors that are sort of potential, potential confounders at this point. And if there is, of course, a great deal of variation in diet as you compare across cultures, probably even more than is bounded within the range of diet of many of the participants of these studies, and yet if the association is pooled, you pool in Europe, you pool in the U.S. and you pool elsewhere in the East, then the.picture begins to emerge. So this is yet another way to try and understand the role of theses factors other than whether a variable was put into a regression model in an individual study. DR. LIPPMANN: Kathy. 23 1 DR. HAMMOND: I concur about the comments about the kn ~ ~ ~ N ~ N OD J ~J ~Q ~/YYLExLCCt12 GJ~E#O%tE'L1 25e t202j Z96-O26, wd{ /703J 644-7636

2 3 4 , 5 6 7 8 9 10 11 i 12 13 14 15 16 17 18 19 20 21 22 23 1-107 use or misuse of the word "confounder." I think it is a little too liberally used. And from the other perspective as well, some exposures which may be thought to lead to lung cancer may not be differentially associated with being married to a smoker or not, such as what heat sources are used, cooking with oil and occupations. That's kind of the flip side that Jon mentioned. Again, these may be effect modifiers. They're important but they're not confounders. I think we need to careful with it there. I may be misreading the data, but you make the S " comment on page 5~3 that nobody found beta carotene to be protective. But that seemed to me to conflict with at least the Wu Study in Table 5-13 where you even have a little superscript 3which says statistically significant of the 95 percent level. And there are a couple other studies that look -- they're not statistically significant, but do look as if it's protected. Maybe I'm misreading something here. At least, just check that out because I thought that was a pretty strong statement from other things I heard and the table seems different. DR. WELLS: I think the Wu Study is quite a bit of smokers. c77lME Z LCQrl G,` E#O ztE'LS. M /2ozf zq6-oz6, q~04 /7031 644-7636

1-108 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. HAMMOND: Is what? DR. WELLS: A lot of smokers in that Wu Study. See, she adjusted for so much smoke per day. DR. LIPPMANN: Are there any other comments then on this chapter? DR. BROWN: One thing is the confounding -- as we define confounding, confounder -- and this is a typical definition and it needs to be -- the endpoint that you're interested in -- and on the dietary factors, the diet is of interest for potential protective effects against lung cancer, unlike more typical potential confounders that cause lung cancer. I'm trying to point out though that we're not saying that the diet causes cancer, but it can be confounding because -- potentially confounding, I should say -- because it may have a protective effect so that in the nonsmokers it may reduce the background rate of lung cancer rather than act as a cause that's associated with the exposure to ETS. DR. HAMMOND: I think that a confounder technically can protect or -- it just affects the outcome. (Simultaneous discussion.) DR. LIOY: That's not the [prime] issue. DR. HAMMOND: That's not the problem. Ln ~ cn W N o¢mEti,can ~efiotte~s ZC /202/ Y96-oY6/ q,o4 /703/ 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 } 13 14 15 16 17 18 19 20 21 22 23 1-109 DR. LIPPMANN: I think we've brought out the concerns that will be amplified in detail in written comments. Now it's necessary to move on to Chapter 7, Passive smoking and respiratory disorders other than cancer. For Chapter 7, we'll start out with Dr. Laties. CHAPTER 7 - PASSIVE SMOKING AND RESPIRATORY DISORDERS OTHER THAN CANCER DR. LATIES: I had a problem with the in utero exposure and its interpretation when the mothers also continue smoking and expose their young to environmental tobacco smoke. I don't see in this page -7:73-9 -- I don't see how the case is made to consistently here to keep those two factors apart. Just reading the plain English of these paragraphs, it just didn't come out to me as being a done deal. And once in a while I thought that whoever was writing it was going to jump ahead and assume that the case was proved. This came up over and over again. For instance, on 9•6 f6, paragraph 2, line 1 -- "Activation of the preliminary review of the endocrine system is not limited to ETS exposure but is activated by active smoking as well." But I didn't see where it was -- a function of ETS exposure and how that was separated from the exposure during pregnancy. d¢m.etLca.n :-RE/2o%tezs `1.~e ti02/ Pqf!-OY61 Q'c4 /7031 644-7636

I-110 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Frankly, I can't see how in the heck you can do the experiment unless you talk ladies into having babies and stop smoking during pregnancy and then start smoking right after they give birth and vice versa. I can't see how you can get any human data on this. There seem to be a couple of examples, like Martinez, et al., was cited as having some mothers who did in fact stop during pregnancy and so on, but I don't see the data given in these paragraphs. So the case needs to be made, and I wish you would look at that again because I think it's an important point and I think you want very much to make the case for environmental tobacco smoke alone, although it certainly is important for anyone to be exposed to it in utero. I worry about some of the descriptions of some of 9-Y3 the specific studies like on page 3.43, the Svenson collaborative study. There seems to be clearly no difference between the effects on non-exposed subjects and on subjects who had smoked 20 or more cigarettes per day, but the conclusion seems to be that there's a no-dose response relationship. But there seems to be no relationship. I mean, the data seems to be in poor shape. The other thing is I went and looked at that study and there they had a great problem because they decided which o4mezi.can Z~?E,hottets. W /2oz/ 296-oz6i woq t703/ 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 males had smoking wives and which did not on the basis of the smoking condition at the entry into a cohort study, I guess, where the decision about smoking was made four or five years before the data cited here I think were collected. So it ~ns haue wives who had obviously started smoking again or stopped smoking somewhere along the line. So I think more attention should be paid to some of the descriptions of the studies themselves. The last thing I had to say is on confounders, T1c1ors potential confounders, deo•tars that seem to be affiliated with poor respiratory health in children. I found the article by Hood very interesting. It was sent in from the public. In looking at it, it occurred to me to look at the stuff written by Witorsch & Witorsch in their review, which was also sent in by the public. Witorsch & Witorsch talked about potential confounders and identified 21 different factors which may play a part in dictating the pulmonary health of kids. Essentially, they examined some 40 studies or so and talked about which ones had many confounders, potential confounders taken care of, and which had few. It occurred to me to look at the results in Hood's paper for those studies. If you do this, you find that the studies that took cn dgrnEtZcan cRefiotEeu Ze /zoz/ 296-0261 q~o4 t7031 644-7636 Cn co N

, 1-112 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 into account many potential confounders weren't more likely to show positive results or negative results than the studies which took account of very few potential confounders. And this essentially gives me heart -- I sort of feel suspicious about any case-control studies and even the cohort studies where nothing seems to be fixed in concrete as it would be in a laboratory study with animals. Well, it looked like the number of confounders taken into account didn't matter very much. Now, it may be that the reason for this is that the people who took four or five things into account were picking ones which in fact had strong relationships with the endpoint -- the restriction of range phenomenon, in other words -- and anyone who took any of them into account, took account of most of the variants. So that's why I found myself less impressed by the Witorsch & Witorsch conclusion because most of the studies were lousy. I'd be interested in hearing from Dr. Witorsch if he's going to be up later on that point. DR. LIPPMANN: Okay, on Chapter 7 then we'll go to Dr. Lebowitz. DR. LEBOWITZ: In general, I though it was an excellent chapter. It did to a great extent what I hoped it would -- from our last meeting, what I hoped could occur in AmE7.l.can GJ?E`2o'LtEU 2$C /zoz/ zq6-oz6, CUd{ 17031 644-7636 c4 I

1 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-113 an appropriate evaluation of endpoints that I felt were very important in terms of environmental tobacco smoke and what we do about it. Obviously, there are a lot of minor suggestions that I have that I'm not even going to go into. But let me deal just with the major ones. One of them is that you have to go back and make sure your conclusions match the data at all times. Sometimes they're understated, sometimes they're overstated and sometimes on the same page or one page to the next, you change your mind about whether it's causal or whether it's risk factor or whether it's potentially causal or risk factor or whatever. I think that just requires editing. And then there are cases where you go into unnecessary conjecture, which might be fine if you were writing an article and you wanted to fill your discussion with what you thought might or might not be happening for which you had no data: But sometimes you have the data, whether it's there, and in some cases you just didn't look at it and it's there. And other times you don't have the data. And sometimes the conjectures are what I might describe as a little bit off-the-wall, meaning that they're great thoughts that might be deemed as hypotheses for some future testing ti Ogmezi.can --fRe#o,ttEU J.7C /zo2/ 2q6-oz6r qGq 003/ 644-7636

1-114 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 somehow, but really they might not have a place in this chapter. There is actually a reasonable amount of evidence that is lacking from the paper, although obviously, as shown this morning, a lot of studies were added. It would certainly strengthen the case in some instances to add those references, and I've noted some that I was aware of, and will provide references to that. The one bit of information that I have provided both to the Committee -- and then I will provide to the EPA -- is the study I presented the last time showing lung function changes in children related to ETS exposure. That's been published and was not included. So that's an example of the kind of study that actually strengthens the conclusions about the effect of ETS on kids' lung function and brings in the issue, by the way, of effect modifiers. In this case, it added something to that section in the conclusions about.what happens when kids start life with low lung function. Are they at greater risk? And yes, they are at greater risk, even when you control for other factors that might be associated with both starting life with low lung function and with parental smoking. So, having looked at those in controlling form and still seeing this major Am.aUcan .-:RefiottEqs Ze 12021 Z96-OZ6, q,6q 17031644-7636 ~ #--A cn W N m N w

, 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-115 effect I thought contributed a lot to our understanding of the effects of ETS on kids' function. It also had an impact on what the conclusions might be in terms of the maximum effect on at least a subset of children in terms of the decrement in function they will have experienced. And that data are longitudinal and therefore meet Dr. Samet's criteria of not being based on recall bias or what have you. And so other studies of that sort I felt contributed. Now, at the same time, to balance that out in a way, I added some information from the studies quoted, including Dr. Martinez' study on asthma in kids, looking at the differential odds ratios or relative risk ratios based on what he controlled for and didn't control for in difference cases. And using that basis, plus some additional studies on asthma, I came up with two new conclusions. One is that the strength of the association with asthma is much stronger, but that the number of cases of new asthma is probably much smaller than he provided, based on, in fact, his own numbers from my study. I tested that out in a couple other studies where I had data and came to the same conclusions. On the other hand, the proportion or the frequency <Z4mEtt.can eRe#otEzts. 17e 12021 2qf!_02f1, qad{ /703/ 644-7636 51592 6294

J 1 2 3 4 5 6 8 9 10 11 12 13 14 19 20 21 22 23 1-116 of exacerbations of asthma produced by ETS based on the data provided in the articles quoted would indicate a much larger frequency, a much greater proportion of cases affected, plus more frequent times they were affected. And so I've modified that in this case to make a stronger argument. DR. LIPPMANN: Excuse me. Is that in your published paper, or is that something you've concluded separately? DR. LEBOWITZ: Well, it's the change in the number of exacerbations -- is in some of my other published papers, for which I've provided reference, some of which are already quoted, because I've looked a lot at the exacerbations of asthma and in association with ETS. And a few others have, as well, that I've included. The frequency of new asthma or -- sorry, the incidence rate of asthma actually will be discussed more in chapter 8, and I'll hold off on that. But in general, that's sort of the type of tl}ing -- and I talked about other mechanisms and other factors that were important. I also think that the authors have to pay attention to Appendix B more, and the statement about how biases tend to overestimate the risk, and the misclassification bias when corrected is a true downward correction. That is, that the c,qrne%i.can eRefi.o%tzzs `L1C /2ozf 2q6-oz6; Woq 17031 644-7636 uf

1-117 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 biases aren't operating toward the null hypothesis; they're operating toward the alternate. That comes up a number of times in the chapter and I felt that that needed attention. There are aspects also of the impact of some of the conditions mentioned. For instance, the impact of lower respiratory infections in children and what might occur later on; the impact of bronchial responsiveness in children on what detriment in lung function they were going to have in early adult life. And I know Dr. Weiss has presented data from his study showing they start life with, what, a 26 percent decrement, or something like that, in one of the lung function measures. They had hyper-responsiveness. I know you've presented that; I don't know if it's published or not. Well, that is a major finding, and in fact I've found that if they had asthma in childhood, they start life with a 26 percent decrement both in my study and in Dr. Weiss' population. I thought that that was critical to add because if you start Odult life with a 26 percent decrement, the risk of getting chronic airway obstructive disease and what that represents in terms of disability, economic status, early death, et cetera, is quite great. So there are a number of cases where I say that that should be -- discussion should be extended in terms of. W ) N l 47qmEZtca.n,-Ra,& ot t z u Zc tzoz1 Z96-ozbi Ch Q4o4 /7031644-7636

1-118 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 the impact of ETS. The other aspects. One of the questions I had, too, and it may be inappropriate in this kind of scientific discussion -- but economists who think they are scientists also have made estimates of the annual impact of each of these conditions, like asthma. And there was an excellent argument by Kevin Weiss -- no relation to Scott -- in the New England Journal of Medicine about that the cost of asthma is $6.2 billion a year. I made a note that, if it were legitimate, to add some of these types of well-documented evidence about impact. M.9RTiA/£Z,-' DR. LEBAVITZ: We discussed doing that and decided against it. We decided it was going to be based purely on health and science. DR. WESOLOWSKI: If I could to make a comment. Actually, one of my comments was going to be that on page-73 1 you do have a sentence on the additional costs for emergency care for asthma. And,I was going to recommend that, although I don't agree or disagree with it, it shouldn't be in this document because this is not a document on dynamic problems. So I would say that probably you're correct in your judgment to state it for science. 16,1,0 7j A/d'L " DR. L~EBOWITZ:~ Ordinarily we wouldn't have done ,vTrI2E2ECQrl GREhO'LfER.S. 2~C /zoz/ z96-oz& CV-1q /7031 644-7636

1-119 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 that. It just happened to be in the same paper that we pulled other data from. DR. WESOLOWSKI: You mean it just didn't get blocked off and -- M-'1/? r 1 ~6 z ' DR. Z: No. It was in the same papers. I mean, we used that paper for other data. DR. WESOLOWSKI: I'd say to throw that out. I mean, i just jars -- it jars the reader to suddenly see $68 in the midst of this multi-text document. DR. LEBOWITZ: I didn't talk about the willingness to pay to get rid of certain symptoms that might be produced. But that's fine, I'm quite willing to accept that. Most of the technical comments you'll find in the edited version of Chapter 7 that I will provide you with the additional references. Getting back to some of the conclusions, though, partly because of the way it was written and I think some confusion in the mindr of the authors that led to the way it was written, I don't think that we feel comfortable, or that I feel comfortable, with some of the conclusions about, for instance, ETS and SIDS here. I think sometimes the way we rely on the NRC report and the Surgeon General's report would also apply to the way we may want to rely on the CDC report ~ r Ln tb N AnEti.ecsn -.Re#ottEt& O1 N ~ 1~C t2021 296-0261 OD q~vq 1703i 644-7636

1-120 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 on SIDS and ETS, and which in fact would strengthen the discussion. Now, I know that you reference the CDC report in terms of the number of cases that might be attributed. I think one has to go into that a little bit'more, and again, in the same way that you did the NRC and the Surgeon General's report. I think that if you missed studies in the other areas that I'm aware of, then you probably missed studies in that area. And I don't know how far one should plumb this, but if one is on the border line and you're sitting on the fence -- and as uneasy as you appear to be sitting on the fence -- then it may be worthwhile trying to find some more of those studies to make either a stronger or weaker argument. But at least get off the fence. Whether you're able to do that or not, whether they exist or not -- in the case of SIDS I don't even know -- but those are some of the broad feelings I've had about it. Now, in terms -- I, of course, also as an epidemiologist, always have problems with the way effect modifiers, confounders, interactions, et cetera, are discuss. And I can just tell you that I've added a lot more on that as well. y y ) ogrnEZtcan :::RE# ozfEU. 2~e jzozj Z96-oz6t q'~V4 f7o31644-7636

1-121 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 In conclusion, I would say that from what I saw, with some work it could be not only more readable but a stronger discussion and argument and also provide greater information about the impact of ETS in kids, which is -- in other words, you haven't quite come up to the expectations in terms of saying how important this is. And that's the bias I bring to my review. DR. LIPPMANN: Thank you. Dr. Weiss. DR. WEISS: Let.me first say that I think that this is substantially improved over the chapter in the initial document. I think that -- and let me confirm my comments to four areas. I think that most of these have been mentioned so I'm going to be brief. As was mentioned by Dr. Laties, I think this issue of in utero exposure versus ETS exposure is a critical issue, and it's a very difficult one. I think that you might as well admit up front that it's very difficult to separate these two. I don't know that the ultimate implications for disease are all that different, but in terms of trying to separate out whether the exposure is truly in utero or postnatal is a very big problem. I mean, just in the area of SIDS, I think it's particularly important to acknowledge this because I think Ametf.can --RE#oa.tES.s ZL' /zozf 296-o26, W 1703] 644-7636

1-122 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 that from the biology of things in utero smoke exposure is associated with placental insufficiency, smaller placenta, size, placental abnormalities. And in addition, one of the things that happens in developmental biology -- and this has been documented with ultrasound -- is when women smoke during pregnancy, the respiratory rate of the fetus drops and the fetus has aptic periods. It's well known that there's feedback -- thought to be feedback -- between respiratory efforts in utero and the development and the control of breathing apparatus in the fetus. You know, I'm not sure that changes one's view ultimately from the scientific point of view as to whether the SIDS phenomenon related to this particular exposure -- how much of the weight you would place in the in utero exposure versus the postnatal exposure, but I think that it's a serious problem for almost everything that's discussed in the chapter, particularly about things that happen during the first year of life. qo I think that that's an important issue. I shared Mike's feeling that there were references that were still missing here. Part of that I think relates to the fact that this is an area where there's continued literature and it's still coming out at a fairly impressive 1VqrnEZi.can -RE#otteU L~C t202] 296-0261 CV~q /7031644-7636

1-123 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 . 19 20 21 22 23 rate, and there is a cutoff time for this. But I would encourage you to go back and be sure -- and in my written comments I try to provide articles that maybe are missing here. I guess two other final comments. One is that I think that the whole phenomenon of the ETS effects here have to be placed in two separate contexts: the immediate health effects and consequences to the child, which I think are documented here; but also then the potential for increased morbidity and mortality as an adult. And I think that although lung function abnormalities that one observes here in terms of maximum lung growth may be relatively small, they may still be significant in terms of the ultimate development of disease in adult life. And then the final point I think is that it's clear that not all -- in any of the subject areas discussed here that not all the studies have taken into account all of the confounding variables; And I think in the previous draft of the chapter there was excessive emphasis on the methodologic issues to the detriment of the biology. This has the biology up front so that, I mean, it makes sense in terms of -- but I think that some of that, you know, the critical covariants and whether they were or were not controlled, I think it cn ~ cn W c*n2EZLCRI2 GRE#oZEE4s. 1~C /zoz/ zq6-ozb, q~vq N031 644-7636

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-124 would be helpful to go back through the document and just, particularly with the most important studies, and really look and see whether all of that is in there. I personally don't believe that, you know, there's no such as a perfect study, and no study can control for everything. So I don't think that that's any particular utility, nor is simply a cataloguing of all of the various potential confounders worthwhile. But I think that the emphasis here is in the right place. The emphasis should be on the biology and what's going on. I think that there has to be a little bit more attention, though, to some of these methodologic issues in these studies that are discussed. But I think that it's not -- I mean, this is something that can be worked with. I think the document itself is clearly improved over what the initial draft -- substantially improved. So I don't think that these are major concerns. DR. LIPPMANN: I will read some, but not all, of Dr. Benowitz's written comments on this chapter. "This chapter is an important component to the report and it substantially improves the previous draft. Section 7.2.2 presents evidence that the in utero exposure to cigarette smoke via the mother significantly affects immune A12EZLCQ.12 ~--RE~2o'LfE43 Ze t2o2l Z96-o26, q,04 t7031 644-7ib36

1-125 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 and pulmonary functions both in animals and humans. While the evidence is fragmentary, the effects are consistent and are overall convincing. "The last sentence of the section states that animal studies suggest that postnatal exposure to tobacco products enhances the effects of in utero exposure to the same products. This sentence needs to be referenced. "Section 7.2.3 indicates the reduced airway size and reduced respiratory flow rates in neonates are associated with increased risk of pulmonary infections during infancy and possibly at predisposition to chronic obstructive lung disease in adulthood. This section provides an important theoretical link between in utero or early life exposure to cigarette smoke and abnormal lung function and pulmonary morbidity in later life. "Sections 7.2.4 and 7.2.5 provide evidence that parental smoking -- at least heavy smoking -- enhances the likelihood of bronchial hyper-responsiveness and other atopic illnesses in children. ~~hese findings are even more striking «ieKv in light of the healt~y smoker effect. Overall, Section 7.2 provides an excellent overview of biological mechanisms that may underlie the ETS respiratory disease association." He then goes on to mention a recent study by Young, o4rnexi.ean v?e#ozfEu 25e [zo2l Y96-o26, q,6q 17031644-7636

1-126 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 et al., from the New England Journal of Medicine in '91. He thinks that should be mentioned. "In this study, airway responsiveness to inhaled methacholine was studied in 63 healthy infants two to 10 weeks of age. Airway responsiveness was increased both in infants with a family history of asthma and equally in children without a family history of asthma but who had one or two parents who smoked cigarettes during pregnancy. Whether this is an effect of prenatal or postnatal exposure to cigarette smoke cannot be determined. However, this paper strengthens the arguments presented in Section 7.2." Section 7.4. "The title of this report is Recent Studies of Acute and Chronic Middle Ear Disease. However, some of the studies address upper respiratory tract infections and'snoring. This discrepancy is first distracting" -- he said "destructing," but I think he means distracting -- "to the reader. An introductory sentence might be included to gxplain why upper respiratory tract infections and snoring are relevant in a chapter on middle ear disease. "The study of Strachan is particularly important because it quantitates ETS exposure and shows a dose-response relationship between saliva cotinine concentrations and risk oqmezfcan =~?rOo¢tEts 2~e t2021 Z96-oZ6, W /7031644-7636

1-127 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 of middle ear effusion. The importance of this type of study should be emphasized." He wrote it, I guess we ought to read it. "The recent studies on cough, phlegm, or wheezing, and ETS are particularly impressive in that so many studies performed in different countries and with very different methodologies have arrived at similar conclusions. Potential confounders are thoroughly discussed. The issue of "susceptibility," that is; of greater reporting in children when the parents also have symptoms -- they are more likely to be cigarette smokers -- is a difficult one to resolve and will probably never be fully resolved. Nonetheless, the conclusion of ETS exposure being associated with increased prevalence of respiratory symptoms in infants and the young is highly likely." Section 7.6. "The new studies in this area offer support to the conclusion that ETS exposure aggravates asthma in children. The disgussion that it is more likely chronic rather than acute exposure to ETS that increases airway responsiveness, the importance of high levels of exposure, and the socioeconomic influences on exposure are important for our understanding of the nature of the relationship between ETS and asthma, as well as for developing prevention OTmE2LCQIZ GRE,7A2fE2s `1.~e Izozl zq6-oz& Q4d{ 17031644-7636 W N W m M

1-128 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 strategies. The contributory role of airway hyper- responsiveness due to a neonatal exposure might also be mentioned as predisposing to later ETS-induced asthma. One implication might be that it is more important to reduce ETS exposure in children whose mothers smoke during pregnancy than for those whose mothers did not." Section 7.7. "The studies on ETS and risk of sudden infant death are remarkably consistent in showing an increased risk, even though patients were studied in different areas in the U.S. and in different countries around the world. The association between ETS and SIDS persists after controlling for a variety of potential confounders, including low birth weight. While it is true that the relative risks of prenatal and postnatal cigarette smoke exposure cannot be distinguished and, therefore, the causative link between ETS and SIDS cannot be established, it would seem prudent in view of the substantial risk and the large total number of,lives at risk, and other studies indicating that ETS may impair pulmonary function, to conclude that ETS exposure should be reduced in infants as a way to reduce the incidence of SIDS. . "A causal relationship between ETS exposure and reduced air flow in childhood seems well justified. In ti to ti 4AnEZican cRE/zotfzas !M tZOZj z96-oz6, q~cq /7031 644-7b36

1-129 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 support of the idea that there are susceptible populations, the study at Rubin in 1990 found that in children with cystic fibrosis there was a significant correlation between the number of cigarettes smoked in the home per day and peak expiratory flow rate." Section 7.8. "As noted, the data on ETS and pulmonary health in adults is quite variable from study to study, and is potentially confounded by exposure to other respiratory irritants that are linked to socioeconomic class and/or to a greater tolerance of ETS-exposed individuals to other noxious airborne chemicals. "Not mentioned in the discussion of potential confounders is the issue of parental or childhood exposure which could influence pulmonary function in adulthood. It is likely that children whose parents smoked, even if they did not become smokers themselves, are more likely to marry a smoker compared to children whose parents did not smoke. .h ~2rG Also relevant to the issue is the study of ,Zantric indicating that a substantial fraction of lifelong environmental ETS exposure occurs during childhood and adolescence." Are there any comments on this or any other issue? Kathy? c4n2£LCCQ12 GRE`1ATt£%1 SM fzo2/ 2q6-oz6r w64 /7031644-7636

1-130 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. HAMMOND: Not on that, but on the chapter. DR. LIPPMANN: Is it on this issue, Kathy? DR. HAMMOND: This is a small stylistic point. I found it very difficult dealing with the new term "true never-smokers." There are quite a few new terms introduced in this chapter. And "true never-smokers" is defined ?-yl q F'VS differently on page C4 and 7-4,5. In one case -- and I originally thought I -- I know how difficult it is to deal with multiple authors, but the work does need some editing, going through and -- there are very difficult concepts in terms of how much exposure people have and don't have, passive and active. I don't know. Is "true never-smokers" supposed to include people who have childhood exposure but are married to nonsmokers? DR. WESOLOWSKI: It just means never-lying never- smokers. DR. HAMMOND: Never what? DR. WESOLOW$KI: Never-lying never-smokers. (Laughter.) DR. HAMMOND: That was my first thought as to what it meant, exactly. That's exactly what I thought. And that is the, you are married to a nonsmoker and you don't smoke. And then it's that you've never lived with a smoker. And Ln ti (n ogm.etica.n z~?e#otteaa 2Se t2o2l 296-O26, q'~o4 003] 644-7636

1-131 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 those are all different meanings. And it is difficult because there were all these different possibilities out there. So I don't have a solution as to how you define something for that.. Similarly, "ever-passive smokers" has some ambiguity to it. So they are minor but they are confusing. DR. STOLWIJK: As opposed to a never-passive smoker. (Laughter.) DR. HANIIMOND: I wasn't sure if it was only those nonsmokers married to smokers, but not nonsmokers who are otherwise exposed. DR. STOLWIJK: And there's the ex-passive smokers. DR. LIPPMANN: John. DR. SAMET: I think this has come a long way, and I think there are two things that could help the reader. One may be a more specific introductory discussion of the issue of confounding in loo)cing at these respiratory health outcomes. I think it's possible to develop lengthy factors that might be of concern. But clearly in a biological model of how these factors act, some of them are potential confounders of concern and some are simply correlates of passive smoking exposure and not confounders in the sense rn 1~ Ln W N I 44riLELCC6tl1 :rRE#0ZEE'LS. 2~e t2021 Z96-m6, q,o4 /703) 644-7636

1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 I-132 that we've been trying to deal with. And I think to deal with that up front and orient the reader towards what these factors are and how they may be important and why it may or may not be important to control for SES, which in fact may not be necessarily appropriate in looking at the effects of environmental tobacco smoke. So I think some help there. The other thing that might help is some discussion of the very difficult issue of looking at respiratory tract infection in this age range. Throughout the studies that are cited, there are in fact a variety of outcome measures that are used that have variable relationships amongst themselves. There's physician visits, there's parent reports. And another thing, in some introduction for the reader, the fact that these different studies are looking at the same phenomenon but using different approaches. There are different factors that may be important in determining the degree of misclas$ification associated with those outcome measures -- would be useful. And the third minor comment, but one I'd like to see addressed because it comes back in the introduction, is there is a sentence or two on dosimetry issues that is left unreferenced. I, for one, would like to see some references Ln ) ,=;nzEti,can 1:-R-e#o%tEu ZC /zo2/ 296-oz61 Q4Wq /703/ 644-7636 ~ Ln %D N

1-133 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 for this first sentence at the top of page 7-3 where there is some discussion about likely differences in particle deposition and gas absorption in the circumstances of active smoking and passive inhalation of ETS. And I think that probably needs to be at least looked at or referenced. DR. LIPPMANN: I wonder if you might benefit from a visual -- if you made a matrix diagram with the names you've used -- ever-smoker, passive smoker, et cetera -- and all the possible routes of exposure -- spousal, at work, childhood, et cetera, with checks for us. And somebody would say, well, what do you mean by that term? You could go back to this -- it means precisely that. It either had -- or a question mark. In other words, a check if you know that a certain kind of exposure is included in that term, a question mark if it's impossible to determine, and so forth. It might be a road map through this jungle that we're dealing with. Jerry? DR. WESOLOWSKI: Just a comment along those lines. In reading this chapter I came across a number of terms you normally don't come across, like otitis media with effusion and tympanograms. They aren't defined. There are other places in the report, like where you talk about atopic or (n ogmE7lca.n lZROozteas. 2~e I2U2l 296-o26, q,6q /7031644-7636 N Ln aN w ti

I-: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 things like that, where you've given a nice definition. Since this report will be read by more than just medical people, I think when you introduce a medical term or an illness, you should define it the first time, or, alternately, have a little glossary so that nonmedical people can look it up. DR. LEBOWITZ: I forgot to mention one other review and study, both that I thought you should include. I mention that you should consider more the CDC evaluation of SIDS. NN~S The CDC, through has also evaluated children's exposure to ETS before and after birth by a variety of characteristics, and they've also discussed the impacts that these different exposures may have and so forth. I've included that in the packet to the panel. I'll also include that in the pack to you. But I think it's sort of a general statement when other recognized federal agencies who are NAS and NRC, because of requests by federal agencies, have done these reviews, then they should all be utilized to the extent possible. PARTICIPANT: Was that a -- DR. LEBOWITZ: It's Advanced June 1991. And it does go into some of the points and some of the references ~7mE2i.can GRe~.o4~E4s W t2ozl z96-oY6, q,oq /703/ 6q4-7636

1-135 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 backing up the points that Dr. Weiss made about the importance of in utero exposure. I'd like to second everything he said. DR. LIPPMANN: It's mislabeled. Dr. Woods also had comments on this chapter. I'm trying to extract things that would be most useful to this discussion. "In some of the studies previously reported by NRC and the Surgeon General and in some of the 26 subsequent studies evaluated in this report, the proximal relationship between the child and the source was reported, but this may be an under-evaluated factor in determining the magnitude of the ETS exposure and the relative risk to the children. "It would be helpful if an explicit definition of ETS was provided in this report. One was provided in the policy guide which we reviewed last time, along with the risk document," but Dr. Woods said he could not find such definition in the risk assessment report per se. I'm just reading what Jim said. He could not find it; that's what he said. Any other discussion on this chapter? Paul? DR. LIOY: I've looked at this for the first time and I find a lot of the information quite interesting. However, at the end of it, maybe it's because I haven't been ~ ~ Ln to AJ w Ogm.ezi,can C-Re fiottEts. !be tzozl Z96-o26, 'UZ4 /703/ 644-7636

1-136 1 2 3 4 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 as close to it as everyone else, I feel sort of let down because I'm given all the different aspects of non-cancer respiratory diseases and I'm not,left with a feeling of which one I should be focusing on first. On the compelling arguments I think that Mike Lebowitz made about the exacerbation of asthma, to me I think are very crucial in terms of dealing with Chapter 8. And I don't get that feeling of comfort when I come to the end of this because each section just summarizes information, and it doesn't give me a capsule view at the end or a perspective so that it leads me into Chapter 8 as to what I should be looking for. I think a good dichotomy here is SIDS versus exacerbation of asthma. I should be looking more toward a detailed analysis of asthma and less emphasis on SIDS because maybe -- and more recommendations about SIDS research, rather than looking for equivalency between what the calculations will be between exacerbation of asthma and SIDS incidents. I think that is necessary to couch things so that in Chapter 8 I'm able to deal with things more clearly. I don't know if I'm the only one who has that problem, but I think that it would put things, again, in the light that a person who would want to read this can P/IrlE2LCQI2 GRE0AZfE4s ~ . W 122021 z96-oY6, w ~ qUo4 N03] 644-7636

1-137 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 understand the relationship because I'd hate to see somebody lose their guard on SIDS, because they assumed that they stopped smoking in the house that SIDS may not occur again. I think there are a lot of other things that we don't know about SIDS. I really want to make sure that we're vigilant in how we, you know, weight these particular factors in this discussion because there are some very crucial issues here. Secondly, I guess -- and maybe this is something that's kind of inherent in the problem or the system -- is I really don't like the title of the chapter saying Respiratory Disorders other than Asthma -- I don't think that -- I mean, other than cancer. I don't think that asthma and chronic obstructive lung disease should be put in a secondary role or a nonequivalent role. I think we should ask our pulmonologist for maybe a better definition for use in the title of this chapter because I really don't think, again, it provides a weighting factor -- it puts a weighting factor here where I don't think one should put a weighting factor. Again, I would go back to the argument that Dr. Lebowitz brought up, that there is a significant financial burden on our health care system each year, even though we ogmexi.ea.n Z~?e#oztes.l Lq Lq ti 2$e t2021 Z96-oYb, ,Vd{ 1703/ 644-7636

1-138 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 can't discuss it in this document, in terms of a child's admission, or at least visit, to a doctor for asthmatics attacks. And so I think clearly we have to be very careful in how we weight different arguments. DR. BAYARD: I'm sorry. I'm just not clear on what you really think we should be focusing on here in terms of the more important effects. I can see where -- DR. LIOY: Not necessarily important. Information, too. I mean, with SIDS we don't have a significant database, as much as we might have for exacerbation of asthmatic attacks. There's going to be a certain amount of weight of evidence in terms of those things that you can prove conclusively versus those that are inconclusive and those which may be in between. DR. BAYARD: We do deal with that in the summary of conclusions. Now, maybe you think that a summary section at the end of chapter 7 is warranted. DR. LIOY: Yes. I think that we have to -- that's the point, to consolidate. DR. BAYARD: In terms of which evidence is more compelling for which to do than others. DR. LIOY: Uh-huh. Yes. Because that leads to Chapter 8 where you start doing the calculations, and Ln ~ ~ ~ N W ~ ~ ~J ~Q PTI9ZEtLCQ.12 G/~E#0'LtE'L1 17e Izozl 2q6-o2G, qUoq /9031644-7636

1-139 1 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 obviously there's a strong -- there's a significant section in there on calculation of asthmatic attacks, et cetera, et cetera. DR. BAYARD: Okay. DR. SAMET: One approach that might help would be to simply separate the material on children from that on adults. The patterns of exposure are different, the patterns of finding are different. And if you had a chapter that was on exposure during childhood to ETS and its effects, and in there there was a strong summary that might address Paul's concern, because I think -- I agree, by the time you read through the material and then you read the section on adults and then it lacks a sort of a focusing summary, it might be better as two separate chapters in fact. That's one approach. DR. LIPPMANN: Essentially we have that. I think it may be one of just labeling and organization. When you discuss adults, you're essentially discussing the data on spousal smoking. And you make the exclusion that you decided not to address cardiovascular effects for these reasons. We're not disputing that you have good reason not to. We're just saying that if you -- I think following up on what Paul's saying, you could talk about the effects of ETS ogmE-ti.can cRz/xodexs 2C 12021 296-026/ -'UIIq /703/ 644-7636 F- Ln N w ~ 00

1-140 ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 on adults, which essentially is your discussion of spousal cancer and the things that you're not discussing, such as cardiovascular. Then you get to the next chapter, which would be the effects of ETS on children, and you say that we 2lave no • ~/A ner"C`7 strong evidence for cancer, except possibly in the eaneric study being related to prenatal or for infant exposure. Just put that up forward, and we can't do anything with it, and that's one of the things we're putting aside, as we put cardiovascular aside in adults. And then you're into other effects. So rather, I think, in terms of the way the chapters are titled -- one being cancer, one being other things -- it could just as well be one being adults and the other being children. Then, to the extent that it's a problem -- and I didn't myself think it was a problem, but to the extent that it's a problem -- that cancer gets prominence just because other things don't have the same scare value, you avoid putting it into that context. So it's worth considering, the effects of ETS on adults, which ends up being a cancer chapter, and the effects of ETS on children, which ends up not being a cancer chapter. Right? Ln N ~ ko N ogmezi.can t-Reiiostz2s Ch w 1`~C' /2oz/ Y[~-OY6/ ~ q'~V4 /7031644-7b36

1-141 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. LEBOWITZ: Yes. I think that would help separate -- whatever organization that separates, it and allows one to draw conclusions that would help. I think that also if you were to separate adults a little more, you could discuss some of what I mentioned about the impact in terms of disability. One of the questions that I was reminded of that I had -- reminded by Dr. Lippmann -- in the issue of scare effects, was also the psycho-social impact of some of these factors on the kids themselves and on the parents, and there was no attempt to have a psychological impact discussion within this in which ETS would play a strong part. In other words, having just reviewed the psychological impact of asthma, of hospitalization for LRIs and so forth in children, I would come to the conclusion that in fact for those families that experience it, the scare phenomena is extreme, or can be extreme, for the kids who experience certain of,these phenomena. The scare is really extreme. There is a lot of psychological sequelae to this as well. Again, not that you should have everything in the world in the chapter, but at least a small paragraph or something somewhere within the discussions of asthma or SIDS, dgrn"f can --Re1so ztE U 17C /zoz/ zq6=oz& W 003/ 644-7636

1-142 1 2 3 4 5 . 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 LRIs, et cetera, with appropriate references might be very important for the public to get a feeling for what these diseases mean. Otherwise, I have the impression from talking a lot to the public about things like asthma, they don't know what the hell I'm talking about, and what the disease means and what kind of impact it has and so forth. Until I started using some of the information gathered on psychological impact, they had no sense of what these things meant.to them. The cancer, they may have a better sense of. DR. BAYARD: Do you think we should have an introductory paragraph about what that disorder is or how it's defined? DR. LEBOWITZ: Yes, or somewhere, even toward the end before the summary of each of each chapter, or something a paragraph'on that. Now again, I don't know that that was in your original charge and whether that's appropriate or not. But, in thinking about it,,it would make much more sense to the public if they were aware of that aspect of it as well. NNL Jv.l There is readily available information on that through ~, or NIH, and so forth, that would help to put that in and to give it a broader perspective. DR. STOLWIJK: I had a question on the whole ,vqmEZ(.cQI2 1:fRroo'LtE'Li. 2~C /zozf 296-oY6r Q1,Vq t703] 644-7636

1-143 1 3 4 5 6 7 8' 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 discussion of SIDS that is present in the report. The discussion starts out by saying that neither the Surgeon General's report nor the NRC report addressed this. Then it describes some of the literature and some of the statements that have occurred around it. The final couple of lines in that three or four pages indicate that the cause of SIDS is still unknown and that it is not possible to assess the biological plausibility, and consequently at this time the report is unable to assert whether or not passive smoking is a risk factor for SIDS. And even that discussion gets propagated into summary and one begins to wonder what all of it is really doing here because it's also causing Paul some uncertainty as to what it all'does and what it means even with respect to other things. And I wonder whether SIDS is being dragged in by the hair a little bit, and then dropped. DR. BAYARD:. I don't think so. The record for maternal smoking is pretty strong, so it seems reasonable to see if there's a connection between ETS and SIDS. DR. STOLWIJK: There are some things where there's a dose-effect relationship between the risk for SIDS and the amount of smoking that mothers do. That's essentially the cn 164 Ln AmEZfcan =RehozEets. l~e tzOZj Z96-oZ6i cgcq /703/ 6q4-7b36

. 1 2 3 4 5 6 7 9 10 11 12 13 14 19 20 21 22 23 1-144 evidence. But the biological plausibility for that is really not immediately apparent. DR. LIOY: I guess the problem I have is -- although we can't discuss Chapter 8 yet -- but it's in the conclusion of 8-13 that really has me worried. DR. STOLWIJK: Well, it propagates all the way through the summary as well as into the population risk assessment. DR. LIOY: Right. Because you say that you don't have evidence, and then all of a sudden it appears twice, and it gets stronger as it goes through, and I think the propagation -- it gets me worried at this particular point. DR. BAYARD: I'm sorry. Be more specific. DR. LIPPMANN: Why don't we come back to this in discussing Chapter 8? DR. LIOY: I'll come back. I'll leave it at this point, but in Chapter 8 it's the fourth paragraph. DR. LIPPMAN$: When we come back to it, and in discussing Chapter 8, let's be more explicit. And we'll leave it at that. DR. LIOY: But the thread of Jan's argument I think is logical, and I think you should consider it at this point, because you're making some jumps in logic that I don't think Cn ti Ln OgrnE,Uca.n ::Re/SOZtEZs 2~e t202J 296-0261 q/o4 /7031644-7636

1-145 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 you have enough weight of evidence to support. DR. LIPPMANN: We don't have to start discussing this chapter now just because we end this phase of the -- DR. STOLWIJK: I was just asking a question. DR. LIPPMANN: Okay. Do you want to do this about dinner? DR. FLAAK: Let me do an administrative thing here with the Committee members, check and see how many would like to go out to dinner together this evening. So I can just get a head count and make a reservation. 7:30. Okay, I have a count. Thank you. DR. LIPPMANN: Okay. We'll take a break now until three o'clock, and then we'll hear the public presentations. (Recess.) PUBLIC COMMENT PERIOD DR. LIPPMANN: We have quite a few speakers. We want to give everybody their chance to get their main points across. We have arrapged them in order, I guess, in part to relate to their projection needs. The first speaker will be Dr. Paul Switzer on behalf of the Tobacco Institute. If you want to use this microphone, this is the only workable microphone we have. Just a couple of quick comments for the speakers. Ln r Ln C4m.eziccan --Re#o%tZqs. ZC t202J 296-0261 No4 /703j 644-7636

1-146 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 We have some changes in the speaker schedule; it won't reflect exactly what is in your agenda. We'll be calling the speakers up as we've changed the list at the request of several of the speakers. I'll ask that speakers bring up a hard copy, if they have, and give them to me. Just please give them to me and I'll get them distributed. I'd like to ask the speakers to try and keep their comments to under 15 minutes so we can get everyone in this afternoon if possible. Thank you very much. PRESENTATION BY DR. PAUL SWITZER TOBACCO INSTITUTE DR. SWITZER: My e is Paul Switzer. I'm ~aql'~r Professor of Statistics a University. I was asked by the Tobacco Institute to come here to make a presentation to you. (Inaudible. PA system completely fades out.) There were several specific questions, and I trust that the Board will a¢dress those questions in a serious manner. My comments are geared to several of those questions, in particular with the material in Chapters 5 and 6 of the EPA draft report referring to lung cancer. One question which the Committee was asked to address is: Is the approach used to derive estimates of the o4rnezi.can lz?FoA4.tE'ls. 2~e tzozI z96-oz6t W /703] 644-7636 N

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 U.S. female never-smoking risks scientifically defensible? I want to make several comments about this. The 95 percent confidence interval estimate derived from the pooled analysis of U.S. studies is 1.01 relative risk4 to 1.38. These are for supposedly exposed populations. And there are several problems here, even though one would hesitate to make far-reaching public policy decisions, even based on such small and uncertain estimates. I believe the panel members, the chair, and vice-chair of the panel earlier in their discussions of both Chapters 4 and 5 indicated that without the epidemiology there would be no basis for the action taken, the classification of ETS as a Class A carcinogen. So it's to that point that I'm speaking. First of all, there's a wide variety of selection effects which tend to bias reported studies in an upward direction. These include the natural tendency of authors to report in published studies of positive results, to advance publication of uncompleted studies when results appear deA positive, as in the Fontham study, or, conversely, t publication, as in the second ACS study. There is also a tendency to make selective cnv4r,~a~t_ varia adjustments, a tendency to make selective he 1 23 1 comparisons, similar selection effects that tend to originate Ln I.- Ln W 0T112E4,CCRn IJ?E f2.O'LiE7.S N 01 Ze t2021 296-026I W N 01 q,or{ /703] 644-7636

1-148 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 with authors, and as I said, additional examples. For example, Fontham's study, on which heavy reliance is placed, dNe v)OCQ/'ti'nO+W<S emphasized the apparent significance for s as opposed to all cancers and emphasizes exposure measured in pack years. It's because of the selected emphasis and the multitude of choices that investigators have in reporting their studies that there are inevitably selection effects. Now, all these effects can be directly measured; but because we know they are present, it is for reasons such as this that scientists are skeptical of small relative risk numbers in general and why we should be seeking to not try to identify relative risks in this range so close to one simply because they are always (inaudible) selection effects. Also, post-study corrections. The EPA draft report makes various post-study corrections. Whenever such post- study corrections are made, new sources of uncertainty are introduced, which should be but were not recognized in the uncertainty estimates: Examples of post-study corrections which introduced additional uncertainty to relative risk estimates which are not reflected, for example, in 1.01 to 1.38, are the misclassification adjustments that occur in corrections. In terms of the actual meta-analysis which was aqmEZf.can cJ?-e#o%ta%1 ZC tzo2/ Zqb-oz6, qUd{ /703l 644-7636

1-149 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 done, the variability between study populations was ignored. The U.S. meta-analysis quoted in the draft makes no allowance for variability between study populations. No allowance is made for differences between characteristics of the U.S. population female nonsmokers and those used in the epidemiologic studies. In other words, the analysis is done as though the individuals were randomly sampled from the U.S. population. Instead of treating the individual studies in an observational sense, as is appropriate for meta-analysis, they were treated and analyzed as other random samples. Allowance for the additional source of variability would naturally and appropriately broaden the combined confidence interval estimate. So the interval estimate I gave you before, 1.01 to 1.38, does not include any y on the variability or uncertainty with regard to generalizations of the U.S. population. I'd like to now point to the fact that there are serious errors in the calculation of P-values for dose- response. An important part of the argument hinges on the consistency of dose-response relationship. As it is repeatedly stated in the draft, the epidemiologic data showed consistent results. This is not the case, especially not the case when one looks at the statistical evidence for dose- c7qrnExEca.n tRE_#ottEts Ze t2021 Z96-oZ6, q'o4 t7031 644-7636

1-150 1 2 3 4 5 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 response relationships in individual studies. This is a serious conceptual error in the calculation of significance probabilities, that these include the categories for exposure. According to this faulty logic, a single exposure category would then suffice to establish a dose-response relationship if you follow the methodology used in the report. A striking example of this error are the reported 1-AM r highly significant values for the LIMT study, Table 5.9. I quote the -- that's at the bottom of that table, if you could raise that a little bit. Could you raise the whole thing quite a bit more? That's it. KAzA You will see there, for example, in both the QO results, that if you remove the zero exposure category, there is obviously not even a monotonic relationship, not to mention a significant relationship. And yet the P-values are reported as less than .01. So, this is just an example of what is done and how # skews the results. Next I come to another question posed to the Committee, which I hope they will address here. DR. LATIES: May I ask a question? DR. SWITZER: Yes. DR. LATIES: On the previous slide, why did you P7/f2EZl.CQI2 Z~?EjhA2fE'is. Ze t2021 296-0261 q~o4 /7031 644-7636

I-15. ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 kAlo~- remove the 1.0 value in the ~JW study? DR. SWITZER: If you're trying to establish a dose- response relationship, the normal way to do this would be to look at several dose levels, and the actual level of dose levels are actually just three, which are the ones below it. In other words, if you following logic and include it, you will only need one single exposure comparison in order to say you have a dose-response relationship. DR. LIPPMANN: The question is really why isn't the zero a valid point on the dose-response relationship. DR. SWITZER: It's not a valid -- well, it's valid if you're trying to, say, do tests for trend, which they did, which is not the same as a test for dose-response relationship. As I point out, that's the difference between establishing an effect versus establishing a dose-response relationship. If you're trying to salvage an effect, you only need one -- you need pxposed and unexposed, that's sufficient. Correct? But it's not sufficient for establishing a dose-response relationship. To establish a dose-response relationship, you need additional levels of exposure. DR. LATIES: But aren't those levels of zero, 1 to 1:4mezfcan -_rRefioa.tEZ1 l~C e t2021 Y96-oY6, wo4 /703] 644-7636

1-152 1 2' 3 4 5 6 7 8. 9 10 11 12 , 13 14 15 16 17 18 19 20 21 22 23 10, 11-20 -- DR. SWITZER: Well, yes. Those three are valid categories. The point I'm making is that if you take a look at the relative risks, they're not even going to one to -- DR. LIOY: Well, supposing we had 1 to 5, 6 to 10, 11 to 15, and so on. Would that satisfy the requirement for getting a dose-response relationship, if they were re-sorted --~ DR. SWITZER: As they are. I mean, as they are. There are three categories here. DR. LIOY: Well, make it four. Go 0 to 5, 5 to 10 -- 1 to 5, 5 to 10. DR. LATIES: Why wouldn't the same objection obtain for the other studies? DR. SWITZER: As I say, if you're just looking for an effect, it's sufficient to have exposed and unexposed. But if you're looking for an increasing effect, at increasingly higher dpse levels, you need more levels. DR. LATIES: But one zero dose is not acceptable. DR. SWITZER: Because it's not actually -- I mean, you'll see there that a zero dose is put in at a risk of 1.0. So in some sense it's not actually data. LAr~ i DR. LATIES: How does the IQM study differ from cvqm.EZEcan t~?E#otteas `Ll C (zo2/ z96-oz6i qc4 (7o3f 644-7636

1-153 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 the KOO study? L AM r DR. SWITZER: They don't. I only highlighted the ® study because it's an extreme version of this criticism. Now I'd like to address another question posed to the Committee which is the approach used to extrapolate risks s~o Nsa ~ in speci-al studies, is it scientifically defensible? And so this is an extrapolation question. And there are several serious problems associated with the extrapolation of spousal risk obtained from epidemiologic studies. First is a quantification of the background risk. The first problem involves this quantification issue. It's based on cotinine ratio data for spousal exposure versus background exposure. This ratio is highly inconsistent in cross-studies. But the serious error is to assume that cotinine levels are truly zero for the unexposed population. This is an implicit assumption and it's used in a very strong way to obtain the background risk. And this is despite empirical data which appear to contradict this assumption. If you'll offer even a modest non-zero value consistent with the empirical evidence, will substantially , increase a eeping ratio and result in sharp reduced background risk. And I give a little example at the bottom of the slide. You might move that up, please. Ln ~ cn ko o¢me zi.ca.n GRE~io a EE7s Ze tZ021 Y96-oZ6, q,6q /703/ 644-7636

1-154 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Whereas an apparent Z-ratio, which is the ratio that's used in the calculation of background risk -- Z-ratio of 6 over 2, which is 3, if you subtract out the non-ETS cotinine in this little illustration, we'd come to find, it's very easy to go from 3 to 05 with an adjustment for non-ETS cotinine. The second problem, extrapolation relative risk to background exposure levels. An approach is used to extrapolate not just, of course, exposure, but to extrapolated relative risk as well. The claims that were made earlier that the studies provide information and actual exposure levels, has to be tempered somewhat in that they do not provide direct evidence for lower relative risks and what are called background exposure levels. Those are calculated, and those account for more than 70 percent of all the estimated lung cancer deaths. The existence of a threshold -- again, sustained resistance of a relative risk threshold, dose-response threshold -- could again substantially reduce the cotinine background risk because those assumptions strengthen all the down to zero. Another question for the Committee is the degree of confidence in the estimates of the population which is appropriately characterized. 'So this is an important Ln OTI)ZE,CCRI2 _-RE#0ZfE4S Ze 12021 296-OZfJ, q,04 /7031 644-7636

I-155 ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 question: The degree of confidence, is it appropriately characterized? Now, let's take a look at the population risk estimates. The draft concludes, on this page 6-2-3, that the uncertainty range for the U.S. population ETS lung cancer risk, when we get a lower estimate of 2,500, upper estimate of 3,300 LCDs. So I think this is a very unrealistic statement of uncertainty. ' This range cannot possibly convey all the uncertainties that are present in the information from which these calculations were made. And particularly you can do this -- although phonics experiment -- because if the entire U.S. population consisted only of spousally-exposed women, namely the kinds of people you have in the epidemiologic studies, comparable to the risk group in these studies, and the uncertainty propagated from the pool of U.S. studies would substantially exceed the reported uncertainty which I said is plus or minus 400 lung cancer deaths. Even using the underestimated uncertainty relative risk, and I mention that -- even the epidemiologic uncertainty has been underestimated, for the reasons I mentioned earlier. But in addition to that -- in addition to the c~7 mEtica.n ~efso~as T~e A-021 296-0261 q'6q /7031 644-7636

1-156 1 2 3 4 5 6 8 9 10 11 12 1 13 14 15 16 17 18 19 20 21 22 23 ) propagation of the epidemiologic uncertainty -- there is substantial additional uncertainties which are not reflected, and which rate the appropriate statement of this range of 2,500, 3,300 LCDs very inappropriate. These include the speculative background exposure calculation. The data extrapolation of the possible dose-response function to presume a background of levels, the extrapolation of nonsmoking female spousal exposure risk calculations, groups, for example, of former smokers. If we account for all major sources of uncertainty, we can easily include an estimate of zero for the population. there's no question. And I don't think you'd get any statistician to disagree. The fact that some sources of uncertainty are difficult to assess does not imply that these uncertainties are absent. By the way, I mentioned at the bottom of that transparency, just to put things in perspective, the EPA reports, cites, four percent of all lung cancer deaths due to ETS and 72 percent of those are due to so-called background exposure, for which no direct epidemiological evidence is available. I should mention that in terms, again, of the degree of confidence of standard methods of error propagation c4»zetlca.n t-RE#oztet& lSe (Y021 296-026, W 17031644-7636

, 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 have not been used. In other words, they did not follow standard procedures. And even though some of the components of the error calculation may be difficult to establish, in their calculation they just didn't even try. Finally, my conclusions. Number one, to the question, Is the female never-smoker risk scientifically defensible? I put to you that it is no. Without proper ~n~cr accounting for i-ntra-study variability in possible selection and misclassification errors, confounding, or any of these, the relative risk cannot be statistically distinguished from 1.0. Two, the approach used to extrapolate risks in spousal studies, is it defensible? The answer, again, no. The approach does not recognize the assumptions and uncertainties associated with the extrapolation, either with respect exposure extrapolation or dose-response extrapolation. And third is, the degree of confidence in estimates, population was appropriately characterized? Once again, the answer to the question put to you is no. The range of lung cancer deaths is not even recognized. The propagation of epidemiologic uncertainty, not to mention greater uncertainty associated with extrapolations, and by its finding a value of c4rnEti.can d-Re fioztzas ZC (202J 296-026I q6q (7031 644-7636

1-158 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 zero lung cancer deaths is not incompatible with our current information. Thank you. DR. LIPPMANN: Dr. Layard, I believe you are our next speaker. , PRESENTATION BY DR. MAXWELL LAYARD TOBACCO INSTITUTE r>9,tx.rt1) DR. LAYARD: My name is Mi-cha€1 Layard. I'm an independent consultant in statistics. I'm speaking in behalf of the Tobacco Institute. The opinions I express are my own. I would first like to address the question of whether the ETS lung cancer epidemiologic studies support the causal inference. I believe that they do not because the reported associations are generally weak and inconsistent. So the possibility that they are due to uncontrolled bias and confounding cannot be ruled out. First slide. Inconsistency is particularly evident when the studies are grouped by a country, as the EPA report does on Table 5-8 and as is s1}own in this graph. The data in the graph are those used in Table 5-8, except that the U.S. relative risk, when correctly calculated from the data in Table 5-8, is 1.17 rather than 1.19, and is further reduced vKYlcr.~y to 1.15 after including the result from the Sarcerick study for all respondents, rather than for self-respondents only. W N m OT1'12Etl.CRl2 eRE#04tE2S. W IZ0YJ 296-o26, q~vq /703/ 644-7636

1-159 j 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 This change is necessary for consistency with the data from other studies, including the Fontham study, in which 34 percent of the case respondents were surrogates. Another difference from the EPA report is that the conference intervals in the graph are 95 percent intervals, rather than 90 percent. The unexplained switch from the revised EPA report to 90 percent intervals in one-sided significance tests is not a justified departure from generally accepted methods for assessing whether a reported association is unlikely to be due to chance. The heterogeneity test for the country-specific results is highly significant, indicating that the differences aren't due to chance. Unless an explanation for the inconsistency can be found, the substantial differences between the countries' specific results I view strongly against a causal interpretation of the data. Though the EPA report rightly focuses on U.S. data in its assessment of V.S. risk, the U.S. relative risk of •J 44 t, r.'v-C dl:- cey,~( 1.15 is not statistically significant t level, and so it does not provide a sound basis for an ETS lung cancer risk assessment. Moreover, as Peter Lee and others pointed out, the smoking status misclassification adjustments made to the U.S. studies are too small in view of ~ co ' rn .~J ~J w P/1'12E'LI.C'Q.12 GREhOYtE2J. ~ CD Ze t20ZJ 296-C26, NGq /703/ 644-7636

1-160 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 available data on this classification and concordance rates for U.S. women. Reasonable assumptions about smoking in this classification would result in a considerably smaller and quite nonsignificant U.S. risk estimate. Even if causal association is assumed, the substantial uncertainty involved in the EPA report's misclassification adjustment is not adequately reflected in its estimates of ETS-attributable lung cancer deaths in the U.S. In fact, even the sampling variability in the U.S. relative risk estimate is not adequately considered in the report's estimated range of attributable lung cancer deaths. The U.S. relative risk is heavily weighted by the Fontham study. Without that study, the full relative risk would be 1.09, with 95 percent confidence interval, 0.9 to 1.29. The EPA'report devotes considerable attention to the Fontham study, considering it to be of high quality, although in many respects it does not appear to be better than other studies. _ For example, the EPA report notes with approval that cotinine testing would help identify cheaters among self-reported never-smokers. But that testing was incomplete in the study, and was largely useless for identifying ever- smokers, particularly among lung cancer cases, some of whom Ln H Ln to OTIKEZLCQIZ ZiE#07tE4.s. !be I202/ 296-OY6/ q,6q /7031 644-7636

1-161 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 may be recent or long-term quitters. More importantly, the Fontham paper was an interim report of an incomplete study, and the analysis presented did not take into account potential confounders such as diet and occupation, although data were collected on those variables. In calculating the summary of relative risk for China, the EPA report used a crude relative risk of 0.78 for the Wu Williams study. Although the relative risk adjusted for age, education, and study area was reported to be 0.7, s4a,4io1 the adjusted relative risk n have been used since it was the stated policy of the report to use adjusted results where available. Doing so reduces the summary relative risk for China from 0.82, which, of course, increases the inconsistency among the country's specific relative risks. Next'slide. In its commentary on the Wu Williams study, the EPA report dismissed the negative ETS result as inconclusive primarily on the grounds that the presence of indoor pollutants frozq cooking and heating sources might have obscured the effects of ETS. However, that is not a credible explanation for the negative ETS lung cancer association. There is no indication in the study report that ETS exposure was negatively correlated with the pollution risk factors, and ETS was still negatively associated with lung cancer when ~ 0T/12E'LLCQ.12 ~J~E~lO'L~E2S !be tZOZI Y96-DY6I C11oq t703l 644-7636 51592 6340

1-162 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 it was considered simultaneously with pollution sources in the logistic reversal analysis. This means that cooking and heating pollution was not a statistical confounder for ETS. For pollution to have completely obscured the ETS effect, it would be necessary for it to have biologically protected against such an effect, which is quite implausible. On the assumption of an additive model for the joint effect of two agents, the apparent relative risk of each agent would tend to be reduced, but not eliminated, by the presence of the other agent. However, in this study, such an attenuation would be small because the reported pollution relative risks were generally on the order of only 1.5, and not all subjects were exposed to the pollution risk factors. This question could be investigated by examining the association between ETS and lung cancer in the subgroup that was not exposed to the pollution risk factors. In just the same way that most ETS lung cancer studies have confined their attention to never-smokers, since no ETS pollution interaction was reported, it is clear that there was no ETS lung cancer risk elevation in the subgroup. The EPA report's vague speculation that ETS effects N ogrnEZi.ca.n e-Rehos.tets. L~e t202J Y96-OY67 q/6q 17031 644-7b36

1-163 1 3 4 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 were obscured by cooking and heating pollution is unfounded. The only explanation the EPA report offers for the inconsistency in the country-specific relative risks is that the ETS exposure from spousal smoking may be a higher proportion of total ETS exposure in some countries than others. This remark relates to the question of adjustment or for background where non-spousal ETS exposure. The EPA report estimates the U.S. background exposure ratio of 1.75 using unpublished U.S. cotinine data from three studies. Applying that ratio increases the U.S. summary of relative risk from 1.15 to 1.44. There is little available information concerning background ratios in other countries, but the published SAn c. . d results of the t_.-R). 10-Country Cotinine Study doas~suggest the foreign cotinine-based background ratios are much more than 3. A ratio of 3 increases the summary relative risks to 4.06 for Greece, 1.95 for Hong Kong, and 1.8 for Japan. The difference between thg U.S. and foreign relative risks isn't decreased by these background adjustments, so differences in background ratios do not explain the geographic inconsistency in report of the ETS lung cancer associations. A more likely explanation is that the report of associations are the product of uncontrolled bias and G1 // W OTI)2E2LCQI2 Ef2oLEEtS N 17L' /zo2/ 2q6-oz6t W /7o3J. 644-7636

1-164 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 confounding linked with spousal smoking status and operating in varying degrees in different countries. Work place ETS lung cancer results, which aren't discussed in the EPA report, support this conclusion. The summary relative risk from 11 workplace studies is 1.01. Applying a background adjustment with a workplace/non-workplace ETS exposure ratio of 1.35 derived ~RC from the I-rR. 10-Countries Study, gives nonsignificant relative risk estimates of 1.04 for ETS exposure for at work adults versus no exposure, and 1.03 for exposure other than at work versus no exposure. This implies that risk elevations in spousal smoking studies are the result of bias and confounding arising from living with a smoker. Smoking status misclassification is one example of such a bias. The background adjustment, as applied in the EPA report and elsewhere, depends on the assumption that the relative ETS dose between subjects who are spousally exposed and those who are not can be estimated accurately from cotinine concentrations in body fluids. There are several reasons for doubting that cotinine based background ratios provide an unbiased estimate of relative ETS dose. One reason is that analytical methods -~;I12E4.l.CR/2 GRE,&o4.tERS. 2~C /zo2/ 2q6-o26, IR,r{ t703l 644-7636

I 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 for determining low level cotinine concentrations are not reliable. Another is that nicotine intake can occur from sources other than ETS, including dietary sources and off- gassing from nicotine absorbed on the surfaces. Several studies, including the Cummings study in the U.S. and the 3-R. 10-Country Study, have reported cotinine concentrations in nonsmokers married to nonsmokers that are close to those of nonsmokers reporting no ETS exposure. This suggests that the measured cotinine concentrations in nonsmokers married to nonsmokers are largely due to factors other than ETS exposure. Background t)j ETS exposure ratios derived from measurements made with personal nicotine monitors are much higher than those derived from the same studies from cotinine measurements, again suggesting that cotinine measurements overestimate background exposure. The implication is that cotinine based background ratios are too small and background adjustments based on them are excessive. The background adjustment applies a powerful lever to the EPA report's risk assessment, since 72 percent of the estimated 3,060 annual lung cancer deaths are attributed to nonspousal ETS exposure. Since a causal interpretation is not supported by o¢meti.can eRElSottEU. Ze t2021 296-0261 (Vo4 /703/ 644-7b36

1-166 ) 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 the epidemiology, a background adjustment which presupposes causality is not appropriate. The EPA report, of course, does assume causality, but although the estimated population impact is highly sensitive to the value assumed for the background ratio, as indicated in this graph, the report's risk assessment does not adequately reflect the uncertainty involved in that assumption. Other uncertainties in the EPA report's risk assessment concern the extrapolation of risks estimated from female never-smoker data to males and ex-smokers. These population groups account for 51 percent of the estimated ETS attributable lung cancer deaths. In the case of males, the report rejects its own misclassification method because it eliminates the associations reported in the U.S. male studies, and instead simply assumes that the estimated female never-smoker risks apply also to the male never-smokers. In the case,of ex-smokers, for whom no studies report an association between ETS exposure and lung cancer, the EPA report arbitrarily assumes that risks are the same as those for female never-smokers. Neither of these extrapolations is scientifically defensible. To summarize, the U.S. summary relative risk of 44na-Ucan z~?e/iotEaTA J..W /ZO2J 296-0261 N::;Iq (703/ 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-167 1.15 is not statistically significant and was not adequately adjusted for smoker misclassification. It cannot provide a sound basis for risk assessment. Secondly, the geographic inconsistency in reported ETS lung cancer associations is not explained by differing background ETS ratios. Thirdly, a more likely explanation for inconsistency is that reported spousal smoking associations are due to uncontrolled bias and confounding arising from having a smoking spouse and varying between countries. This conclusion is supported by the negative ETS workplace epidemiology. Fourthly, the ETS lung cancer epidemiologic studies do not support causal interpretation. Therefore, a background adjustment which presupposes a causal relationship is not justified. Lastly, even if a causal association is assumed, the EPA report's treatment of uncertainty is quite inadequate with respect to the s4mpling variability of the epidemiologic results used and with respect to the many assumptions involved in the risk assessment, particularly in misclassification and background adjustments. Thank you. DR. LIPPMANN: Dr. Lee, I believe you're next. (A ~-A L" ) o0911tEtl.C6tl2 GRE/20'LtE'iS 2~e IZ0Z1 Z96-026, N~ 17031644-7636 %o N

1-168 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 PRESENTATION OF DR. PETER LEE, TOBACCO INSTITUTE DR. LEE: My name is Peter Lee. I'm a statistical consultant working in England. I've been involved in ETS and lung cancer over the last 10 years quite heavily. The opinions I'll give today will be my own. I'd like to concentrate on lung cancer in the EPA report. As you know, the EPA's main conclusions were as summarized on the first and second slides. I won't read them out. Next slide, please. Oh, all right, sorry, I'll do it. That's the second main conclusion. These are the EPA's preliminary findings taken from their report. Some three months or so ago I published a book, Environmental Tobacco Smoke and Mortality, which has an extensive discussion of the evidence in lung cancer. And my conclusion from the evidence was that, considered in its entirety, the evidence did not convincingly demonstrate exposure through ETS increases risk of lung cancer among nonsmokers. I'd like to highlight some reasons why I believe I think the EPA have come to the wrong conclusion, and why my conclusion is more appropriate. I'd like to concentrate on the epidemiological evidence, but first I wish to comment N W ~I7ZELLCRI2 GJ~E~20 i~E'LS •j L~e t2021 296-0261 Nd{ 1703] 644-7636

1-169 ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 very briefly on the arguments put forward in EPA's Chapter 4. This was in fact made by Dr. Lippmann, really, today, that the evidence in Chapter 4 is not -- it merely provides some plausibility and evidence generating a hypothesis to be tested. It doesn't of itself prove cause and effect. ETS and mainstream smoke aren't the same as exposure to many chemicals. It's very much lower from ETS than from mainstream smoke. And active smoking doesn't demonstrate -- the fact that you can't see a threshold for active smoking doesn't demonstrate that it doesn't exist for much lower dose levels. In any case, just to add to that point -- I mean you could have made that argument 15 years ago and the scientific world didn't really react until (inaudible) published their epidemiological data. The epidemiology is really the key to the whole issue. In my evaluation of the epidemiology, in Chapters 5 and 6, there were four major types of weakness I found. One was failure to characterize uncertainties. The second was underemphasis or ignoring of data that didn't fit in with the hypothesis that ETS caused lung cancer. The third was overemphasis of data that did seem to fit. And the fourth PTmE4CCQ.I2 UREf20'LtE'LS. 21e tzo21 2q6-oz& q~Vq 0031 6aa-7636

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17. 18 19 20 21 22 23 1-170 was underestimation of the possibilities of bias. So if we consider failures to characterize uncertainties first, in the first draft of the EPA report a major criticism of it was that the uncertainties weren't characterized. You can illustrate this. I mean, this is so sharply illustrated by the second conclusion -- and this has already been pointed out by the previous speakers -- where it's stated there's a range of 2,500 to 3,300 lung cancer deaths and the confidence in this range is medium to high. Now the man in the street, even the statistician, would read this as being a confidence limit. And this in no way resembles a confidence limit. What it is, the 2,500 and 3,300 -- if you actually look at the details of the report -- are merely two point estimates based on one study, Fontham, varying one parameter. This cotinine-based background correction factor zed. It takes into account no element of a sampling variation for which the Fontham stuoy, even the 90 percent limits, and I agree with Layard that I would prefer 95 percent limits following precedent. The lower limit is very close to 1, so the lower limit of deaths must be about zero. And then one has the -- there's enormous sources of variation -- they are ignored. I mean, the whole estimation o¢m.ezl.can 4-Refio%EeTA !be t2021 296-0261 q'6q 1703) 644-7636 cn 1- Ln

1-171 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 procedure is a slew of assumptions and parameter estimates, and no attempt is made in the report to show how the results depend on these assumptions. Now turning to the underemphasis of data that don't fit and overemphasis of data that do fit. Let's just think of how this estimate of 3,000 deaths is made up. There is an element of 470 deaths attributed to spousal exposure among women who never smoked. And this is totally what comes from the epidemiology. This one cell is then extrapolated to seven additional cells where one extrapolates from women to men, from never-smokers to former smokers, from spousal exposure to non-spousal exposure. If one actually looks, is this justified? The answer to this is clearly no. If you look at the male data, it doesn't show any evidence of an increase, according to the EPA's own analysis, yet one ignores the fact the male data show nothing. One extrapolates the female data, and assumes it's the -- the male Is the same as the female. There is admittedly very little evidence on former smokers. Such as there is from the ~/A rela/`/an e r i ~ ~ study shows no evidence of an increase. I mean, they say ones -- really, the numbers for former smokers are just really plucked out of the air. They don't bear any resemblance to science. cn ~ cn co N c4m.ezi.can ~e~.o~eas. Ze /zoz/ 2q6-ozbi q04 /703i 644-7636

1-172 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 And then this extrapolation from spousal exposure to non-spousal exposure is extremely weak. I mean, there's ~10r~c )tLCQ Anp~ extensive data on w~rat~g3aee-in childhood exposure now. And it shows no evidence of an increase at all for workplace exposure. There's 14 estimates I've got which give you a relative risk of 1.02, which is nowhere near significant for childhood exposure. There's 12 estimates which give you a relative risk meta-analyzed to .94, which is nowhere near significant. So, one is in a position where one is saying, well, we've got lots of data that doesn't show anything. Let's ignore that, choose the data which does, and inflate that up to make the numbers as big as we can. I mean, that's really what seems to have happened, to me. I mean, the report doesn't even mention the data on workplace and childhood exposure. How anyone is supposed to actually come up with an interpretation of the evidence without even being told this vital fact seems beyond belief to me. Some other major things that are in the report. One isn't told that particulate matter retention is very low for ETS, sort of three or four orders of magnitude less than for mainstream smoke. I'm not sure why this is kept secret ~172E4LCR12 GJ`Lf2.o'L~E'i~ 2se 1202i 296-DYb, q'6q 17031644-7636

1-173 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 in the report when it could be argued that it's a very relevant index of exposure. There is nothing in the report about the inconsistency of the evidence on histological type. One knows in active smokers there's a strong association with squamous cancer and only a relatively weak association with p a carcinoma, and you would expect the same in ETS. You don't see it. You see that some studies show relationship with squamous, some without, and some with neither. And this, again, is important for evaluating the evidence. Perhaps a less important point is that the zed factors in the background exposure correction -- there's no discussion at all of the problem of misclassification of active smoking in them. One takes the mean cotinine level in people married to smokers, and the mean cotinine level in people married to nonsmokers in many of the papers without any adjustment for mieclassification. You only need one or two smokers in there with enormous cotinine levels, and your zed factor has become enormously unreliable. DR. LIPPMANN: Could you clarify the basis for your first statement? DR. LEE: The first statement? What particular? Ln ~ Ln ~ hi o, ~{rnEZican ~E~iottazs vW i N 17C' /2021 296-o26, q~oq t703J 644-7636 pdon

1-174 / 1 2 3 4 5 6 7 9 10 11 12 13 14 15 What, the basis? It is very low. I mean, it's quite -- DR. LIPPMANN: Not just because you say so, but what evidence have you got? DR. LEE: It's in my book. It's in my -- there's -- DR. LIPPMANN: Can you recall what it is? DR. LEE: Can I recall what factor it is? DR. LIPPMANN: What is the evidence? DR. LEE: There's published papers from -- oh, I'm not very good at remembering papers, but there's messes of material on this. I mean, there's studies done t Harwell. ~rtin of~ There's studies done -- there's stuff cited by 4ran u and Sterling, which is cited in the EPA report. retention of particulate matter as a factor There's DR. LIPPMANN: You're talking about percentage 16 retention. When you say 3 or 4 orders of magnitude -- 17 DR. LEE: I'm not saying retention is 3 or 4 -- the 18 retention factor is about 8 or 10 times less. It's about 80 19 or 90 percent retained from acting smoking and about 10 20 percent from passive smoking. When you then multiply that 21 with the actual amount of exposure, that's what I'm -- 22 perhaps you misunderstood. 23 DR. DAISEY: Is that based on experimental L, r tn N W ~ ~ ~J ~Q W ~7112Et1.CQ/2 ~JXE/].O'CtE4.S. 2~e f2021 296-0261 N6q 17031 644-7636

1-175 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 determinations of retention? DR. LEE: The relative retention, as I understand -- DR. DAISEY: No. For ETS, not for mainstream -- or modeling, experiment modeling. DR. LEE: It's not my area of expertise, but I understand the data to be fairly solid, that retention is lower -- low from ETS. Dr. Lippmann would know more about it than I would. It's not mentioned in the report -- getting on to my fourth point -- that various sources of bias can cause an artificial trend. I mean, there's a sort of assumption that if you can find the trend, it must indicate cause and effect. There's evidence that the heavier the smoker, the more likely they are to be'married to a smoker, which increases misclassification bias. There's evidence that the heavier the smoker the more likely they are to be exposed to other risk factors, which i0creases the bias from confounding. This deserved mention in the report. DR. LIOY: Could you give me a citation of where they actually misclassified active smoking in their Z- factors? 23 1 DR. LEE: All I know is that -- In In W 1VqmEti1-,a.n lz~?e#ottcts ZC /zoz/ 296-o26Y Nd{ /703/ 644-7636

1-176 ) 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. LIOY: I'm asking you a question. You cite that they are biased. Could you give me a citation of where EPA did that? I would like to know where that is so that I can review it and determine what they did. DR. LEE: As I understand, there are a number of sources of zed-factor estimates given in the EPA reports. DR. LIOY: I understand that, but where -- DR. LEE: For nearly all of them, as I understand it, there is no mention of removing -- they nearly all talk about mean cotinine levels in people married to smokers, and mean cotinine levels in people married to nonsmokers, and the ratio of these is the zed-factor. DR. LIOY: I understand what the zed factor is. I'm asking you where do you show the misclassification occurred? DR. LEE: I mean, one knows misclassification does occur. DR. LIOY: We can argue that all day. I mean, there are notations of misclassification that can be ascribed to even in epidemiologic studies done in the workplace. But do you have evidence? DR. LEE: Well, the zed factors. What do you disagree with? I'm not clear. I mean, it's apparent that if cn ~ Ln W N OTIKEZCCQ/2 1V?E#0?.tEZi Ze 1202] 296-0261 qUdf /7031644-7636

, 1 2 3 4 5 6 7 8 9 10 11 i 12 13 14 15 16 17 18 19 20 21 22 23 ) ,v'/Il2EYl.CQ/2 GRE/30'CtEU l.~e tY021 Z96-G26, (V,:'# t703) 644-7636 1-177 you have a scattering of smokers' cotinines making up part of your mean, of your two groups, you're going to -- DR. LIOY: That's a hypothesis. DR. LEE: Well, one knows. There's abundant evidence. I've written a book on the subject, that people misclassify smoking. DR. LIOY: I'm asking you in this particular instance. DR. LEE: Why in this particular instance -- I don't understand what difference is the instance between measuring cotinine studies in order to determine a zed factor and measuring cotinine in studies in order to determine a misclassification rate. DR. LIOY: Well, I'm looking at it in terms of the document itself, and would like to know where there are biases and I'd like to be able to review it. DR. LEE: That's all right. Well, let me say, the section on -- DR. LIOY: Or are you just giving me a general overall argument that we have to consider? DR. LEE: I'm giving you a general statement which I would think that, given one knows there are misclassifications, that -- F.. cn N Ln

I-178 ) 1 DR. LIOY: All right. I understand what you're saying now. DR. LEE: Sorry. DR. WESOLOWSKI: But you would have made that statement regardless of whether you had read the EPA document or not? 2 3 4 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 N DR. LEE: Yes. I mean, I did in fact read the EPA document. But it didn't mention -- DR. WESOLOWSKI: I just wanted clarification. DR. LIOY: There's no real information. DR. LEE: That's right. It's true, I mean, it's self-evident, given that there is misclassification of active smoking, so that usually means it's dangerous. The last one is the question of publication bias, and it's rather -- I found it surprising in the report that they didn't even mention the possibility whatsoever. I mean, I thought any paper, report, doing meta-analysis would consider publication bias, but this was unusual that it didn't. So if one now goes on to the final general point, major potential sources of bias -- well, the question of publication bias -- in my book I actually demonstrate some evidence which suggests it may exist to a small extent. I ,vqmetlca.n eRefsotteu ZC /zo2f 296-0261 N~ /703/ 644-7636 r

1-179 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 think there is some indication in the analysis in my book that there's a tendency not to publish small negative studies. I'm not saying this is a major source of bias. It perhaps shouldn't be under major sources of bias; it's a source of bias. The question of specific study weaknesses. I mean, it's an interesting one. It was much discussed by the SAB committee and there's quite a lot of attempt in the report to mark studies as having or not having specific weaknesses. But there's no attempt to quantify how those weaknesses relate to relative risk. In my book you'll find an analysis, where I took what I thought was the most likely source of bias from study weaknesses where this is failure to compare like with like, akd where cases ~ controls differed on vital status, place of interview, next of kin, respondents -- this sort of thing, where you weren't clearly looking at the same thing. The studies,which I observed, the studies which were weaker in this respect, had a very much higher relative risk than the studies which were stronger in this respect. My analysis came up with 1.74 for what I classified as poor studies, and 1.17 for ones classified as better studies. So I think it will be interesting to extend this work and look CAf2E21.CQ/2,zRE#A'LtELS ZC /202) 296-o261 wa4 003) 6"-7636

1-180 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 . 17 18 19 20 21 22 23 at how specific study weaknesses do affect the actual relative risk estimate. Going on to confounding, there's a question of -- EPA seemed to be seeking to explain the whole association in terms of one confounding variable in their discussions, which I don't think is right. I wouldn't claim that for any confounding variable. I think they are of moderate importance. I would have thought they ought to have organized a proper analysis of the Fontham studies since they lean so heavily on it, and it hasn't been looked at for confounding. And they don't. The report ignores massive external evidence of occupation and all these family history of lung cancer, diet, are lung cancer risk factors. I don't see why they concentrate just on the spousal, on the lung cancer ETS studies when there's a lot of evidence outside. DR. SAMET: Peter, I hate to erode into your minute and 46 seconds, but since the evidence is massive, I'd like to ask you about, again, the issues of confounding and never- smokers, which is the group that we are discussing now. I'm just ~ aware of massive external evidence; that is, evidence external to the studies at hand that describes, for example, family history of lung cancer as risk factor. Here c41f2E'LI.CQ.Ii Z~?rOO'CtE'CS. Ze f2021 296-0261 w41q /7031 64a-76s6

1-181 1 2 3 4 5 6 7 8. 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 I'd like you to provide specific citations for literature on never-smokers on this, details on -- DR. LEE: I think it's a peculiar hypothesis that diet would affect lung cancer in smokers and not in never- smokers. I think it would -- DR. SAMET: No. I asked you for family history of lung cancer in never-smokers, the specific citation. 1 0 DR. LEE: I'm claiming that Q" cancer -- that I don't think this is -- DR. SAMET: I'm not asking for a claim. I'm asking for a reference to the massive external evidence. DR. LEE: All I'm saying is that isn't relevant. What you need -- I mean, why should effects of diet be different in smokers than nonsmokers? DR. SAMET: Peter, I'm sorry. It was your slide that says there's massive external evidence. DR. LEE: I don't say that for nonsmokers. I said there's massive -- I think it's highly relevant evidence that all these factors are risk factors for people, not -- people are people. I mean, why do you want to restrict to never- smokers? I mean, it seems a phenomenally unlikely hypothesis that diet would affect smokers and not nonsmokers. It just doesn't make sense to me. It seems like you're restricting ogine zi.can d QE#o tEE u Ze (2ozl 296-oz6, C1~oq /703] 644-7636 51592 6360

1-182 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 the evidence -- it's just not sense. DR. LIPPMANN: But confounding is not a global phenomena. And I want a specific to the study and the population at hand. Dr. Samet's question is still relevant here. DR. LEE: I mean, I'm not saying -- there's not massive -- there's clearly not massive evidence that there are risk factors for nonsmokers, no. DR. LIPPMANN: I think we're interested in whether there's any evidence, not even whether it's massive or not. Is there any? DR. LEE: Well, indirectly from biology and from smoke -- from overall people. I mean, that's pretty strong evidence as far as I'm concerned. Can I just -- I've been heavily -- I've got four or five quick slides on misclassification. I'll skip this on confounding. There is a report on this and some general conclusions on confounding which you can see the;e. I won't read them out to you in my report. Can I have just two minutes on misclassification? In fact, it's obscurely written but lots of people have said -- let me just go on to the fact that Wells and Stewart find they are biased, which is less than mine. Looking at this, dgmezi.can :::Re#oatgats ZC /zoz/ zq6-oz61 q~vq (703/ 644-7636

1-183 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 there are various possible sources. The date assumes, which is slightly different in some cases -- the methodology is just clearly different, and the misclassification rates. Looking at methodology, I list five things I think are wrong with the Wells-Stewart method. That one assumes I can't see any logic for assuming ETS exposure, not while risk in former smokers has no effect for current smokers. It just doesn't make any biological sense to me. And the other, the next two or three points are technical ones. The last point is the actual method of correction, given the observed rates and the misclassification rates, is mathematically wrong as far as I can see. I was surprised when I checked this through. I've got all the details in my submission. And there's a difference in misclassification data, where the data I cite from my book shows about twice the rate that Wells cites. I've got rather more data. Admittedly, some of it is non-U.S.; some of it is for males. But that isn't the whole explanation. I think there are studies with more data. I won't get into the question of why I have it; it was admitted; it doesn't make much difference to the outset. My final point on misclassification is I think one should use a clearly defined method, which one hasn't done. . cn Ln W N o¢rnEZi.can cRz,fiorLts Ze tzoz1 Y96-o26, (Vc,4 /703) 644-7636

1-184 1 2 3 4 5 6 7 9 10 11 12 13 14 19 20 21 22 23 It's very obscure. You should use a mathematically correct method, which I don't think it is. I think you should use a simpler method. It's really rather difficult to see what's going on with the very large number of assumptions. One should have a section in the report showing how the bias depends on varying your assumptions and varying your parameters. In London we have about 12 assumptions of parameters. You have one value for each and no indication for uncertainty. And the last idea I would like to suggest is that one might involve independent statisticians. I mean, one has a situation here where really -- you have Judson Wells on sort of one side of the fence producing one set of results. You have me on the other side of the fence producing another set of results. But, really, I don't think many people in the world actually understand what's going on at all otherwise. I think it would be quite nice to wheel in some independent statisticj.ans who actually haven't got any axe to grind on ETS and let them get on with it and see what they come up with. Well, finally, that's my main conclusion. It's that I don't think it should be classified as a Group A carcinogen. ogmEttcan GRE/20ttEt1. !be tZOZ/ Z96-oZei q'64 /7031 644-7636

1-185 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. FLAAK: Dr. Sears, I believe you're next. Let me just make an observation. We've spent about an hour going through the first three commenters. At that rate we'll be here all evening in all likelihood. I am allowing 15 minutes. I'm timing the presentations, and I'm adding on the comment times from the committee members. So in some cases there may be a little bit more than 15 minutes; in some cases there may be less. I ask the remaining speakers, please try to get to the point and focus on your issues so we can get everybody in this evening and get a fair opportunity to hear everyone's comments. PRESENTATION BY DR. STEPHEN SEARS R.J. REYNOLDS TOBACCO COMPANY DR. SEARS: Good afternoon. My name is Sears. I have a Ph.D. in Theoretical Chemistry from the University of North Carolina at Chapel Hill. I work as an applied mathematician.at R.J. Reynolds Tobacco Company. The EPA draft document Respiratory Health Effects in Passive Smoking, Lung Cancer, and Other Disorders is in my opinion seriously flawed in a number of areas. My comments today will be directed only toward two of these areas, specifically misclassification bias and o~ucan eRehotEau ~ rLl e (zOZI Y96-OY61 ~ 4~or{ (7031 644-7636

1-186 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 background adjustments, and more specifically how these two topics impact specifically the American epidemiology. Before I begin considering each one of these topics in detail, let me recall for you that the 1986 EPA Risk Assessment Guidelines require that for a causal association to be established between an agent and cancer there can be no identified bias which could explain the association. I contend that not only has the EPA failed to rule out the single type of bias that they chose to look at -- that is, misclassification bias -- but they've also done an entirely inadequate job of dealing with many uncertainties in this risk assessment, particularly those that influence the misclassification bias and the background adjustments. DR. LIPPMANN: Could I ask you a question? DR. SEARS: Sure. DR. LIPPMANN: One can read that several ways. One could say that the EPA has not been able to identify any bias that could explain the association, and I think that's a fair statement of the literature. What you're saying is something different. You're saying that they have not specifically explained IT. DR. SEARS: Perhaps you misunderstand. I'm reading this exactly as I believe it's presented in the Risk L" L" ko N PTmE'LLCQ/2 GRE/20'tEE'LS. l.~e 12021 Z96-026/ w64 IxsJ 644-7636

1-187 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Assessment Guidelines. That is, one of the three criteria for establishing a causal association is that there is no identified bias that could explain this association. DR. LIPPMANN: Exactly. And so it's up to the people who doubt the EPA report to show which bias could explain the association. DR. SEARS: Well, perhaps we disagree to this extent, Dr. Lippmann. I don't think that there has to be one bias. There in fact may be more than one bias that could a J ol : 3~ vt have an g~ effect which would result -- DR. LIPPMANN: Right. And the argument would be more convincing if you came in with evidence that one, two, three, or four biases could explain the association. DR. SEARS: Well, if you will allow me to proceed here, I promise you that at least on the evidence dealing with the misclassification bias, perhaps I'll be able to shed a little light on that. Let's begin,with a discussion of misclassification bias. As an overview, in the next several minutes I hope to demonstrate to you that there are uncertainties and assumptions which are multiplied in the current model dealing with misclassification bias in this report in comparison to both the EPA's previous work and in other published models. L" cn %o <AmEqi.ca.n cReltottets Ze 12021 296_DY6, QUo4 /703j 4a-7636

1-188 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Secondly, this method is mathematically indirect and unnecessarily complicated. Small changes in a single parameter yield nonsignificant results and the use of more *o realistic misclassification rates reduces the risks 93 statistical non-significance. Let's consider each one of these in a little bit more detail. First of all, the current model, the model presented in Appendix D contains a host of adjustable parameters and assumptions. Depending upon the way you count them, there are easily between two and three dozen parameters, variables, and assumptions required to make this model run. Each one of these, I think it's fair to say, has a good deal of uncertainty associated with it. The EPA has made absolutely no effort to propagate that uncertainty throughout the calculation. The result is an overall statistic which I believe is rendered useless by the failure to propagate these uncertainties. Many of the,parameters, unfortunately, have been obtained by private communication. This, of course, subverts the peer review process, and I think even more importantly, for a document of this source, the public review process. Some of the data -- hJ E L~S DR. : It all came from peer reviewed ogrne4i,ca.n GRe~.ottetrs !bC /2021 z9f1-oP61 q'oq /703] 644-7636

1-189 1 papers. St~ DR. giNGER: I understand that. JE t, L-5 DR. : And I wrote to the authors to get 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 the.detail that was required in order to make an accurate assessment. DR. SEARS: Yes, but it's impossible to get the data that's required specifically to make the parameters that you need to insert into this model to make that work directly from that peer review literature. So an additional analysis of the data has to take place. WrtLS DR. : What I was trying to do was to use Peter Lee's original model to get the data that would make it run. DR. SEARS: I understand that. Wk1-Ls DR. : That's what I got. DR. SEARS: I understand that. W 1= L.LS C,ouj4kS 60,,,t rLC DR. WESAL9W6K-I: And from PeeI-tzs, F4erce, and Cummings, those three.. I didn't get it from Lang, a~j~ kn~~~ eanrreet, and Pron and a couple of others. DR. SEARS: Let me continue my discussion. I agree with you to the extent of what you've stated. But nevertheless, there had to be new analysis of this data which is not available for me to review. c4mEZican Z~?E#ottEqs. Ze f2021 296-0261 W /7osl 6a4-7636

1-190 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 NrcL.LS DR. WF69E9WSI4I: You could write to the same authors. DR. SEARS: Well, in fact, I probably will do that. The method itself also has problems. At the beginninc of Appendix B there are four, actually five, operating principles which are set out as.a sort of standard against which all misclassification models should be judged. These are not followed, or they are followed incompletely in carrying out the analysis. The algorithm, the mathematical algorithm, is I think obscured by a poorly written appendix. It's very difficult to tell what's going on here. And finally, this methodology, to the best of my knowledge -- and I has been think confirmed earlier today -- has not been peer reviewed. This model is sensitive to a number of parameters, but it's particularly sensitive to one. This is the misclassification ratq at which regular smokers report to be never-smokers. In a moment I'll show you a demonstration of how sensitive it is to that parameter. One of the most important things today that I have to share with you, though, is some new data. This data comes from a recent study that was commissioned by R.J. Reynolds Ln V-% Ln %o N O7/f28'LLCQ/2 --RE#0'LtER.S Ze 12021 296-C26, W r7o3j 644-7636

1-191 I 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 %O W Q ~ E2l.CQ/2~ 7/f2E2l.CQ/ 2CQ/2 ~JYE#0'LtE7S. c;7/ f 2 Ze t202j 296-0261 /7O3l 644-7636 and carried out by an independent survey firm. It involved n.'nt- approximately 900 randomly selected subjects in bn diverse ~ cities across the United States. This is, I believe, the only study conducted to date that can truly claim to be representative of the American population at large, and simultaneously derives misclassification data as a function of marriage to smokers or nonsmokers. I'd like to give you just a very brief overview of the results from that study. I will point out to you that this misclassification rate that I've computed on a population- weighted average from those nine studies is -A Z computed exactly in accordance with @P analytical chemistry cotinine requirements specified in the EPA draft document and by the definition used for this parameter in Appendix B. The result of these nine city studies for the misclassification rate is approximately five percent. Compare this to the EPA estimate used in Appendix B of 1.02 percent. Now I will tell you at this point, this data, the study has only recently been completed. It's in the process now of being written up for submission for publication, and I'll be happy to discuss any more details of this with you if you require them. Now, the purpose of doing this is because when the V m

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-192 use of more realistic misclassification rates are added to the Appendix B model, it has a very big effect. And in fact a re-analysis of the United States data, including data from the nine city studies easily yields misclassification rates in the range of four to six percent. What I've done here is to plot in a sort of sensitivity analysis the pooled relative risks of the American epidemiology as a function of this critical misclassification rate. The first bar on this curve corresponds roughly to the EPA's result, a summary relative risk of about 1.175 with a lower confidence interval hovering just above statistical significance at 1. If one increases this one parameter only out to a value of about 2 percent, 2.0 percent misclassification -- actually a little bit more, between 2.1 and 2.2 -- then statistical significance is lost. And going out to more realistic misclassification rates, it's clear that the relative risk has dropped sharply and they are all statistically nonsignificant. I'll also point out to you that these are based on the new 90 percent confidence intervals described in the document. DR. DAISEY: Could you just go back so I can see the cities in which they did the study? Ln ) CAnE tl.can 1V1?e#o tte u. W tZ0ZJ Y96-OZ6, q~64 17031 6aa-Tb3G

t ) 1 2 . 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 •22 23 1-193 DR. SEARS: Yes. Freehold, New Jersey; West Des Moines, Iowa; Muncie, Indiana; Denver, Colorado; Fort Myers, Florida; Dallas; Los Angeles; afld-eme-la Buffalo. And in fact we have some more data to show in just a moment. DR. STOLWIJK: Could I ask, what questions were asked, in what setting were they asked, and were they comparable to the questions that are asked in the study? DR. SEARS: This was based on a standardized questionnaire. It took about 20 or 30 minutes for the subjects to fill out. The overall strategy of this study was a mall intercept type of study, where you intercept subjects in a mall and haul them aside, ask them to fill out the survey and then ask them to run some tests. The tests involve salivary cotinine (inaudible) for determining the cotinine levels. In summary, the misclassification model is ob'tKS e logically r-edtced. It requires numerous adjustable parameters and assumptions as input. The EPA has, I think, inadequately accounted for uncertainties in the misclassification calculation, and the use of realistic parameters in the model quickly leads to statistical nonsignificance. I'd like to turn to the background adjustment, dTmcu:can fRc fiottats. !Ne 1ZoZi Z96-oZ6, wd4 hosl 644-7636

1-194 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 which is the adjustment that's done after a pooled relative risk is calculated. As an overview, I will argue that the EPA's background adjustment is based on selective, nonrepresentative, and unpublished data. New data exists which drastically alter the conclusions, and finally the application of & more realistic data makes a background correction irrelevant in the first place. Very quickly, background -- ETS exposure is all ETS exposure other than that due to spousal smoking. The critical parameter in the background calculation is called a Z-factor. The Z is defined simply as the ratio of ETS exposure from spousal plus other sources to exposure from other sources alone. The justification for carrying out a background correction in the EPA document is the so-called ubiquity of ETS in our environment. Now, the contention of this ETS ubiquity is in fact supported on the basis of a single study. This single study is 9f cancer screening clinic patients in Buffalo, New York. This is the Cummings study. Well, let's consider each one of these shortcomings in more detail. First of all, once again, in the case of computing the Z-factor for the background correction, the EPA has obtained data entirely from private communications. None Ogm.ati.can I-REfio%.Ee:s. 2~e 12021 N96-G26, q~~q 1703/ 644-7636 51592 6373

1-195 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 of this data in the form that it's needed exists in the literature. All of it comes from private communications. In the end, Z is derived from only three studies. Those three studies are Cummings, Fontham, and Riboli, and in each case the EPA has treated the data from those studies in a questionable manner. As an example, in the Riboli study, one of the American cities reported in that study was entirely dropped from the analysis and in another case, the case of Los Angeles, half of the data was excluded on the basis of an abnormally high urinary cotinine level. Now, the thing that's wrong with that is that that abnormally high urinary cotinine level is right in line with other studies that the EPA did include in their analysis. Once'again, I have an extension of the study that I've told you about before. This in fact is another study or an addition to the former study which was commissioned by RJR and it involves approximately 800 randomly selected nonsmoking subjects in 10 cities across the United States. Here again, these are the only studies done to date that can truly claim to be representative of the American population at large. DR. SAMET: I'm sorry to interrupt you, but because o9rnEti,can Z~?s#otEr-zs ZL' /zo2f zq6-ozG, W 1703/ 644-763b Ln ~-A Ln N

1-196 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 this could be important and you're claiming representative- ness, can you tell us something about the age distribution, the sex distribution, and whether the misclassification value that you're using is overall for both sexes, whether there's any age adjustment, and,just tell us a little bit more because it's clearly important in our arguments. DR. SEARS: Sure. I can answer that question partially. I'm not the principle investigator on this. However, this is a study of when exclusively -- once again, it's a mall intercept study, which involves a questionnaire and subsequent workup of salivary chemistry. If you require more data than that, I'm going to have to refer you to the principal investigator. DR. SAMET: Specifically, what proportion of the subjects were in the age range that would be comparable to those considered in the case control study? I think more information is clearly requisite. DR. SEARS: Sure. I'll be happy to provide that. I'll also point out that in one of these studies, the one in Columbus, Ohio, we actually used three separate ETS markers, or three separate markers were used. In addition to salivary cotinine, also airborne nicotine and 3-ethenylpyridine. To the best of my knowledge, this is the P7mEtLCRI2 GJ`E#A'LtE4S Ze [ZOZI 296-026, w6q I70sl 6aa-76sb N n ~ ~ ~ to

1-197 ) 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 only study to have had three markers. I've summarized the results of that study on this slide. I'll just point out a couple of things for you. rA'N es Notice the range in Z-bodi-es. They range from a low of 1.7 in Freehold, New Jersey, to a high of 14.6 in Los Angeles, California. The two numbers that are subscripted there are based on airborne nicotine and 3-ethenylpyridine. I'll point out just a couple of these numbers to you. The number for Buffalo, New York, which is one of the lower numbers in our study, is in rough agreement with Cummings' data from Buffalo, New York. The number from Los Angeles, California, which is one of the higher numbers that we determined, is roughly in agreement with Riboli's higher value found in Los Angeles, California. DR. DAISEY: Are there equal numbers of subjects in each of these cities? DR. SEARS: No, there are not equal numbers of subjects, but they're,roughly the same. Compare these numbers to the estimate used by the EPA of 1.75. Notice that the EPA estimate is higher than the lowest value that we found. I'll point out to you now, and you'll see graphically later on, that the lower the Z-value, the greater the inflation of the risk estimate. og»zexi.can CRe#otEets 2~e tZ0Z1 Z96-oY6, W t703/ 644-7636

1-198 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 In fact, the Z-value employed by the EPA -- it's equal to 1.75 -- inflates the risk estimate by over 200 percent. A more conservative estimate based on the EPA cited data, plus new data, but still based solely on cotinine, is in fact -- gives a Z-value of about 4.11. If we extend this to include other markers, other than just cotinine alone, then the Z-values fall roughly in the range of 8 to 22. Based on the overall evidence, we estimate that the best estimate here is something greater than 10. Now, what I've done here is to plot the risk inflation due to background correction. The Y-axis is the incremental relative risk, that is, that portion of the relative risk that's due only to the background correction. It's a function of this critical Z-parameter over here. Notice the steepness of the curve for low values of the parameter Z. I'll point out three points here. The first point is the value assumed by the EPA, 1.75, this occurs in the most steeply varied section of the curve, where the inflation is by far the greatest, hardly a more advantageous point to be picked to inflate the risk. As we go out to more realistic values, that risk is dramatically reduced, and in fact, by the time you get out to our best estimate value, the N o0qIf2EZl.CQ.I2 GRE#O'LfE'L1 17C /zozf z96-o26i W t703l 644-7636

1-199 1 2 3 4 5 6 7 9 10 11 12 13 14 19 20 21 22 23 risk is truly insubstantial. DR. DAISEY: Can I ask another question about that? DR. SEARS: Sure. DR. DAISEY: The experimental data, I think, is very important, and these are people who are in malls. Do you have any information on the ventilation rates in those malls, because that's going to affect the exposure numbers, and in the sense that you're taking a salivary cotinine measurement at a time when they are only in a mall, it may not be representative of what those exposures are. I mean, it may be elevated. My impression with malls is that the ventilation rates are very poor. DR. SEARS: To answer your question directly, I don't have the information available on the ventilation rates in those malls. We allowed the outside survey company to make decisions of that sort. I presume that that data is available, and if you're interested, I can try to get them. DR. DAISEY:. We sort of assume that a mall is representative of kind of all exposure locations, which may not be correct. DR. SEARS: Well, the best answer that I can give to that is that the survey companies that have been doing this sort of work claim that the population that they Ln ~ Ln kD J o09m.8ti.ca.n GRe,ho'rtEu 2~e t202j 296_0261 Ndf t703l 644-7636

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-200 intercept in those malls is truly representative of the broader population of that -- DR. DAISEY: But that's not the question. The population could be representative. I'm not asking that. What I'm saying is, the exposure during the time that they are in the mall may not be representative of the population exposure, because a mall is a rather specific location as opposed to -- DR. OGDEN: I had one comment. These are cotinine measures which reflect exposure of material in the previous three to five days. With the ventilation in the mall, at the time they're intercepted, it's really not relevant. DR. DAISEY: But the nicotine and ethenylpyridine are air measurements. DR. OGDEN: They are air measurements averaged over a period of five to seven days. So it's not simply a measure of the point of contact. It truly is measured over an extended period of time. DR. SEARS: This is Dr. Michael Ogden; he's a collaborator on this study and a principal investigator. You misunderstood or I misspoke. The 3-ethenylpyridine and the airborne nicotine measurements were made only in one city, and that one city was an extended 07l~2EZl.CQ12 GRE/]AtfE4~. 2~e t2021 296-0261 W /'703/ 644-7636 W U? W N ~ W V t0

1-201 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 period of time over a seven-day period. DR. LIPPMANN: Where in that city? In their homes? In the 800 homes? DR. SEARS: Throughout the hour, throughout the day. DR. DAISEY: But in the mall. DR. LIPPMANN: In the mall, or in the homes, or what? DR. SEARS: No...The Columbus, Ohio study -- am I right, Dr. Ogden -- in the Columbus, Ohio study there were given passive nicotine monitors which are worn on the lapel -- DR. LIPPMANN: For a period of five days. DR. SEARS: -- for a period of five days. DR. DAISEY: So they went home -- DR. SEARS: They were sampled four times during that period. DR. DAISEY:. They would go home and come back to the mall. DR. SEARS: Yes. And even when they slept it was kept with them. DR. DAISEY: That wasn't clear. DR. LIPPMANN: That makes it clear. c4matlcan 1=_RE#ottets W tZ0Z/ Y96-oZ6, q':::Oq /703/ 644-7636 ~ ~ ~ N

1-202 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. SEARS: Okay. In summary, the EPA's background adjustment is based on selective, nonrepresentative, and unpublished data. New data exists which drastically alter the conclusions, and the application of more realistic data makes the background correction -- DR. SAMET: I'm sorry. Just one more point of clarification. These Z-values are -- I can remember Appendix D where we were talking about people who would classify themselves as nonsmokers or never-smokers. And when you examined -- when you calculated these data and used the means, did you exclude values of cotinine compatible with active smoking? You showed us already a 5.0 percent misclassification rate, and at the low values of cotinine typically observed in the general population, misclassified active smokers could substantially impact the Z-values. What did you do to deal with that? DR. SEARS: Well, first of all, of course the background correction is separate from the misclassification correction, which is what's going on in Appendix D. Appendix D addresses -- DR. SAMET: These are data that assume nonsmokers -- Ln N cn W dTm.ezi.can GRe#o2teu 2~c tZO21 Y96-o26, W /703144-7636

1-203 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 DR. SEARS: But if it were determined -- DR. SAMET: Did you remove, then, values that exceeded -- DR. SEARS: Yes. The answer is yes. DR. SAMET: Okay. That's fine. DR. SEARS: All right. And finally, concluding with both of these topics, measured uncertainties exist, uncertainties that haven't been taken account of. The EPA has used unrepresentative, unrealistic data. In the absence of realistic data there's been little effort to reduce the uncertainty. And I believe new data invalidate the EPA's K TS i3 G~. Gl-~~no Q') contention that :P.,hes-kts Group A DR. BAYARD: Could you submit the data within part of the published comments? DR. SEARS: I will be happy to provide you with this data as it is put in a form that can be provided, as soon as we get it written up. DR. FLAAK: Dr. Gori? The next speaker is Dr. Gori. There you are. Do you need any overheads or -- PRESENTATION BY GIO GORI, TOBACCO INSTITUTE DR. GORI: No, I don't need anything. 6.0 Mr. Chairman, ladies and gentlemen, my name is JAe 23 1 Gori. I'm a consulting toxicologist, also trained in C4lrtSZCCQ/2 GREl2o2tE4.l. ZC 1zo2/ 296-o26i q'or{ 1703/ 644-7636

1-204 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 epidemiology. Up to 1980 and for 12 years I was with the Nn+i o ri cj Ruaaian Cancer Institute. Among other duties I was Deputy Director of the Division of Cancer Cause and Prevention, and Nr4 ) -;; Founding Director of the Smoke and Z1rhe&e Program, and of the Diet, Nutrition, and Cancer Program. I'm here on request by the Tobacco Institute. I will take questions -- given the high number of sharpshooters in the front row, I will take questions at the end of my presentation. In proposing the ETS and ~ung Pancer }~ssociation, the EPA draft document fails as a scientific assessment and even as a credible dialectic proposition. Its principal thesis is that, 1) ETS is the equivalent of mainstream smoke in terms of chemistry and specific biologic activity; 2) there is no threshold for mainstream biologic activity; and therefore, 3) ETS is a Group A carcinogen. ~ In proposing the ETS VSS equation, the draft status, it is plausibXe. More precisely, I should say it is con c t ;vaA" veagid~b,le because it would no longer seem plausible after examining the evidence. ETS derives from the dilution of sidestream smoke and from the small residues of exhaled mainstream smoke. The smokes have intuitive similarities, but also marked differences in chemical physical composition o4m.azi.can eRe1aoitzss. 2~e Izoz/ pq6_oPey q'cq /7o3j 644-7656 Ln ti W N rn w

1-205 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 and behavior. Mainstream smoke is concentrated and confined to the high moisture environment of the mouth, throat, and lung. Sidestream smoke is significantly different from mainstream v++ ETS. In the end, the ordinary ETS is over 100,000 times more diluted than mainstream in an atmosphere of much lower humidity and extremely low concentration of volatiles. Water and volatiles evaporate more efficiently from ETS particles, which, within fractions of a second from their generation, attain sizes that are 50 to 100 times smaller in mass and volume than their mainstream particulate counterparts. While the lifetime of mainstream smoke is counted in seconds, that of ETS is counted in minutes and hours. While aging, ETS undergoes substantial modifications due to dilution, oxidation, photochemical processes, linkages, and due to water and volatile losses, reactions 4hQ~ with other environmental components, absorption of-+n 'd saoreo ~,;,7 selu64on to and from environmental substances, and so on. Some 4,000 mainstream components have been characterized, but only about 100 from sidestream, and less than 20 in ETS, due to the extreme dilutions. Thus, it is clear that besides the measured and known chemical and physical differences of mainstream of ETS, many yet Ln ~ cn W ti ogmmzi.can ~Qefio~t~~s Ze tZOY1 p96-oZ6, W /7031 644-7636

i 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-206 unmeasured and substantial differences must occur. The evidence for the mainstream ETS equation in specific biological activity is virtually nonexistent. In vitro genotoxic activity has speculative but unknown relationship to the carcinogenicity in general, and has no demonstrable power to sustain the ETS mainstream equation. Other evidence cited by the EPA draft are the studies of Staunton and Grimmer. However, these studies relate to ETS but not to -- I mean they relate to sidestream -A -, dose but not to ETS. In reality, t4ros% that reacIttarget cells in the lung must be vastly different between mainstream in ETS, both in composition and •in local concentration. The only components likely to reach single cells with some qualitative similarity may be some high molecular weight, known volatiles in the particulate phase. Even this, however, must remain an unverified conjecture, given that ETS derives mainly from sidestream, and sidestream and mainstream noH- kAown volatiles are known to be different. We simply do not know what the composition of ETS particulates might be because of the difficulties in collecting and separating what may be ETS particles from other environmental dusts. Again, the mainstream ETS equation is I:'z7q»zeti.can :::Re#ottets '~L1L' /zozf zq6-o2& q4Vq 1703/ 644-7636 <on C t, watl/0 51592 6385

1-207 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 but its plausibility is not sustained by the evidence. Nevertheless, even if we assume that 1the equation is valid, rt /.a 7r #vt Qr o$~ 0'1 GL consideration of v evidence of no-effect ~~ircS~olJs ra#.ies deny the notion that ETS could be a human lung carcinogen. ETS dosimetry has been an issue of contention for a long time. Only a handful of ETS volatiles have been measured. For particulates, there is a contemporary agreement that prevalent exposures to ETS occur at concentrations below 50 micrograms per cubic meter in ambient air. Now, straightforward considerations that include differential, lung retention, and clearance rates leads to the conclusion that average ETS doses are between 10,000 and 100,000 1-~6 times less than mainstream doses. Elementary calculations involved have been published in a recent paper Ma ., h / by Ment-e-1 and myself, which is attached to a separate submission to this group. As for the gas phase, simple dilution considerations arrive at estimates of comparable magnitude. For instance, the 1986 Academy report on ETS calculated that the nicotine exposure for ETS is between 57,000 times to 7 million times below mainstream exposure. c4rna.acan ':rRE#ottzU Ze /ZOZ/ Z96-oZ6, q/04 /7031 644-7636 w CD C,

1-208 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 There is ample evidence for thresholds in all experiments in animals dealing with tobacco smoke and its fractions. Results show very steep gradients passi~g from 30- 11 maximum effect to no effect in a matter of 10 to 3-4 dilutions only. None of these studies are cited in the EPA report. SEP Aha yle, My second written submission gives several references and examples that are a common experience for anyone familiar with this literature. All the epidemiologic studies cited by the Surgeon General a~id the )&Lsociation of S d igarette Pmoking and ~isease give evidence of no significant effect thresholds. Pe'I' oThresholds are especially evident in the Doll and o Eater cohort of British doctors, widely regarded as the best do se rlw nft / data set for t*ese response analyses. Dr. Mentel and I again reported this in 1991. Just a week ago I had the opportunity Vyr,o/ly, of discussing again this issue with Dr. Wynd am and other prominent colleagues who fully agree with the reality of thresholds for cigare%te smoking and lung cancer, as well as for cardiovascular diseases and respiratory diseases. It appears that smoking up to three to four cigarettes daily may not be associated with a significant risk. This concept coincides with the general agreement that cigarette smoke is a weak initiator, acting primarily in c41'I2EZiCQn ,:zRE/2o'LfEYs ZC 1202j 296-0261 qaq /7031 644-7636 v

1-209 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 a promotional role, a concept further supported by the dt C rt k Se- apparent degrees and eventual erasure of risks following smoking cessation. As we have seen, ETS doses are thousands of times below thresholds for active cigarette smoking. Overall then, neither logic nor evidence sustained the proposition that mainstream and ETS are equivalent, that no threshold exists, and that ETS is a lung carcinogen in humans. Q ; ~It ,.,,'o Therefore, it is no surprise that the e~i~oq~c evidence is as weak today as this panel found it in December 1990. If anything, the evidence for confounders is increasingly stronger, although still curiously belittled by the EPA draft document. I was flabbergasted this afternoon to hear that diet may not be a confounder. The'draft continues to state that no single confounder has been identified that could explain the slight meta-analyzed blip in the relative risk. The draft, however, fails to give evidencp that major known confounders have in fact been controlled for, and seems unaware that several a~ p/a, confounders can be simultaneously ex~~:a,i#red. This would require that their aggregate and cumulative contribution be accounted for in addition to corrections for misclassification and possible background exposure. ~ N ~ ~ N Amati.ca.n cRc fiottzqs 1X' /zo•r/ s96-oz6i (l/04 (703/ 644-7636

1-210 1 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 V'hnder Again, talking recently with Dr. Wyrndem and other colleagues, we asked what happened to the axiom of epidemiology whereby relative risks below 2.0 cannot be reliably interpreted one way or another, especially in situations where confounders are known to exist and they are WQ r1uStd : not accounted for. Could it be bemt:sed that epidemiologic sensitivity has improved suddenly, unbeknownst to us? A nKlI cardinal rule of scientific conduct is that the keewn hypothesis must be defended. Obviously the EPA draft does not follow this principle. I shall not dwell on the factual errors in the draft, on its lax criteria for meta-analysis, on its glib ease in switching from 95 percent to 90 percent confidence intervals, on its reliance on published personal communications, the inclusion of data from unfinished study, dose the exclusion of negative results, the contrivance of these response trends that are not apparent, the contradictions whereby ETS sometimes.is said to equal and sometimes is said not to equal mainstream, the Byzantine logic of the models and appendixes and so on. Frankly, I would be concerned if my graduate students kept producing papers of this caliber. Unquestionably cigarette smoking is a legitimate interest of r Ln 16.4 ~ to w ogm.E%l.can cRe#ottZzs 'DC /2ozJ 2q6-oz6Y q/04 0031644- 7636

1-211 ) 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 public health, but the best intentions hardly justify improper means. This obviously is a policy proposal, not a scientific analysis. As in too many instances before, the Agency has elected to become an advocate for a cause being pursued with wrongful methods. Indeed the current interest in the ethics of scientific conduct provides a fitting and timely framework for judging this draft report. To whom should we listen? Last time, some panel members were criticized for accepting research funds from CIAR, a foundation underwritten substantially by the tobacco industry. Today nobody objected that members of the panel are also grantees of contractors of EPA and other agencies. Personally I find the whole matter irrelevant. Killing the messenger may be a political sport, but should have no place in a scientific context where only the issues count. In this vein it may look as if some of us are not here to praise Caesar, but neither us nor you members of the panel would survive as scientists without a healthy respect for science. To any observer the corruption of science in the frenzied rush to incriminate ETS as a health hazard ought to be cause for concern; more so when this occurs at the hands of an agency empowered by public trust and funding. If not at EPA, where should a citizen look for objective and ogrnEti.can 'V1?.ehor&u 2~e t2021 Y96-CZ6, qd{ 17031644-7636

1-212 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 accountable information? Personally I am comforted that some of you, Dr. Samet and Dr. Kabat in particular, have written about the lack of scientific support for the association of ETS with lung cancer and other diseases. Hopefully the entire panel will see necessary to act in defense of science and will resolve that promoting policy without sound science is not good policy at all. Thank you. No questions? DR. DAISEY: I have a question. DR. GORI: You have a question. DR. DAISEY: You mentioned modifications of environmental tobacco smoke due to photochemical and oxidation processes which are certainly plausible. Are you aware of published data on these? DR. GORI: There are several publications that show that smoke undergoes oxidation by its form when it is exposed to the air. DR. DAISEY:. The nitrogen oxides change. DR. GORI: Nicotine and other things. DR. DAISEY: Anything on the carcinogens? DR. GORI: Like what? DR. DAISEY: The polycyclic aromatic hydrocarbons or the nitrosamines. Are there data showing there are %D Ametl.can cRelzotEeu !be t2021 296-0261 q~Vq /7o3f bq4-7636

1-213 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 chemical changes in those that occur in relatively short time frames relative to air exchange rates? DR. GORI: I am not aware that benzopyrene or hydrocarbons, which are relatively inert, will undergo rapid changes, but they probably get oxidized like anything else by photochemistry or whatnot. DR. DAISEY: It is an important issue in terms of chemical transformation and how fast it would occur. Based on the intensity of light-indoors compared to outdoors, you are talking about a factor of 100 to 1,000 lower light levels. The other factor of course is you have an air exchange rate, and how quickly you are introducing fresh smoke and so forth is going to complicate the issue. I am just asking if you have any experimental data on how large that effect might be. DR. GORI: I would be able to provide you with a few inches of literature on that subject. DR. DAISEY:. I have seen the N-0, but I am asking about the carcinogens in particular. DR. GORI: The carcinogens, I'm not familiar now with the individual components, but as I said, I would be able to provide you with the literature for reference. DR. DAISEY: I probably have the NOX. ~ ~ ~ N w c4rneti,can cRE#otfZts W t202l 296-0261 W /703f 644-7636 W ti

) 1 2 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-214 DR. LIPPMANN: Dr. Todhunter. PRESENTATION BY DR. JOHN TODHUNTER, TOBACCO INSTITE DR. TODHUNTER: Good afternoon. Thank you for the opportunity to comment on the Agency's Draft Risk Assessment for Environmental Tobacco Smoke, or ETS. I have prepared these comments in response to a scientific question about ETS, not smoking, which was posed to me by the Tobacco Institute, which I will refer to as TI. I was approached by TI as a recognized expert in risk assessment and also due to my high degree of familiarity with EPA's carcinogen classification and risk assessment processes. As you may be aware, I served previously as EPA's Assistant Administrator for Pesticides and Toxic Substances and also as a member of the interagency working group which developed the Office of Science and Technology policy document on principles for carcinogen risk assessment. The present EPA carcinogei3 assessment guidelines are of course based on those principles. I will note here in regard to Dr. Lippmann's query about where the burden of identifying biases lie, it is a self-imposed burden assumed by EPA within the guidelines as they were developed. (n 4Amezi.ca.n fRE#otEets. !bC /zo2J zqb-oz& q~04 003] 644-7636 W N m w w

1-215 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 The question which I was asked was, are ETS and active smoking sufficiently comparable that one can draw valid conclusions about ETS based on the active smoking epidemiology? It's the same question that was posed by SAB to the Agency in SAB's review of the first draft of the ETS Risk Assessment document. Of course since there are many major and obvious differences between active smoking and exposure to ETS, SAB actually asked the Agency a more limited question which was to first consider at least whether or not ETS and MS, or mainstream smoke, were sufficiently comparable so as to allow one to draw for ETS the same conclusion that one might draw about MS based on active smoking epidemiology and other data specifically relevant to MS. Why was the question posed? Because as several SAB members commented, at that time on its own the ETS epidemiology data were clearly inadequate for purposes of carcinogen classification. The carcinogen classification process requires that one have sufficient evidence in humans in order to assign an agent to Group A or known human carcinogens. That is, one must have a clear signal from epidemiology studies. There is no such clear signal from the ETS OTIrtE'LI.CQ.12 GRE f1.0'CtE't.'i Ze (ZO2~ z96-o26, qUvq (7031 644-9636

1-216 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 epidemiology database. This was recognized in SAB's previous review, and the Agency was asked to go see if it couldn't come up with a link to some other epidemiology base which might provide a clearer signal. Before moving on to my comments, which are specific to the question of ETS and MS comparability -- and I would like to mention that I have submitted two somewhat detailed written comments which are referenced, and some of your questions if you do raise them, I will probably ask you if I may just be allowed to give you the relevant parts of that. I would like, as I said, to address the process of generating a risk assessment like this. As you are probably aware in developing such risk assessments the Office of Research and Development, or ORD, does act as a service unit within the Agency to the office which requests that the assessment be prepared. In such a capacity, ORD's task is basically to take the database and extend it, push it, see what can be done with.it as far as it will go in terms of supporting policy decisions which may be of interest to the requesting unit. In this light I will say ORD has done its job as well as can be expected given the weakness of the database too begin with. The ETS epidemiology data have in particular y L9 t0 ~ tv 01 1:Vqmezi,ca.n GRE/20RfESS, ZC /zoz) Zq6-o26, Noq 17031 644-7636 w to y

1-217 ) 1 been massaged to an extraordinary degree so as to try to get some semblance of a carcinogenic signal out of them. Just as one example of this, consider the unorthodox shift to a 90 percent statistical confidence in this document instead of 95 percent statistical confidence which is_used in every other 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 EPA risk assessment and is used as a standard benchmark by the scientific community. Even so, frankly, the required signal is still not there, which brings us back to my comments on ETS and MS comparability. If I may have the first slide. As highlighted on this slide, once we focus it, the biggest single problem is that one frankly doesn't know if one is going to make a comparison between MS and ETS, what are the relevant comparison factors? I have put on this slide some different factors, whether they are comparable or not between the two, and as you face it, the right hand column, my comments, which are extensions of my submission, as to are they relevant or not? Even the Agency clearly doesn't know what the relevant factors are. Take for example the contradictions between Chapters 3 and 4 and Chapter 6, if I can have the next slide. In Chapter 3 of the draft document, the Agency goes to some y 44mc-Uean GReoottets Ze (ZOYl 296-o26, W (7031 644-7636 g

1-218 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 significant effort, particularly in Table 3.1, to compare ETS and MS on the basis of chemical composition and draws the conclusion that the chemical composition is similar. This conclusion is also repeated at the end of Chapter 4 where the Agency states that if MS is Group A, then ETS is also Group A due to the similarity in chemical composition. I think the weakness in that argument was already commented on by members of the SAB today, but let's go to Chapter 6 where, particularly in Section 6.2.2, the Agency goes into some considerable detail as to why ETS risks cannot be reliably estimated on the basis of extrapolation from active smoking data. At pages 6.6 and 6.7 of the draft document most of the same points are raised which are expanded on in my written submission. In fact there is an agreement at that point between my written submission and the draft assessment that ETS and MS are not comparable in terms of carcinogenic risk. In this section the Agency basically repeats what it stated to Congress in '89 and also cites the Surgeon's General report to the same effect. The contradiction exists because ETS and MS are in fact two quite distinct items with respect to carcinogenic risk assessment. o'g~%i.can -,fRE#o%teu Ze 12021 296-026J q~o4 /7031644-7636 N

1-219 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 The Agency of course is looking for similarity between the two because it was asked to do so and points out in Chapters 3 and 4 that there is a chemical similarity. But as the Agency recognizes itself in Chapter 6, these similarities are not the relevant factors upon which to make a valid comparison with respect to carcinogenic risk. I will point out that in all fairness one may argue that Chapters 3 and 4 -- the Agency at that point is really concerned strictly with hazard identification, whereas in Chapter 6 the emphasis is on quantitative risk assessment. And we must say that in all fairness. But this, I do not believe in any way, negates the plain fact that if ETS and MS are sufficiently distinct biologically to preclude a comparison with regard to quantitative risk assessment even at "equivalent exposures" as the Agency points out, well then they are also sufficiently distinct to preclude comparison with regard to hazard identification, If I could have the next slide. I don't want to dwell on this slide too much because these are some of the same points raised by the Agency. But I have summarized here some of the differences between ETS and MS which are likely to be relevant to carcinogenic risk. It is a summary only. ogrnEZi.ca.n ::rRefi.ottEts 2~e fZo2l 296-o26, CM ho3l 644-7636 Ln i

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-220 These points are discussed in detail in my written submission, and they basically are repeated in Chapter 6. I will point out that some of these different areas potentially admit of a threshold of biological action and those are the ones which are asterisked. In any case, as I have already stated, there is a similarity in constituents between ETS and MS but that's about all that they have in common. Is that a sufficient basis for classification of ETS? Clearly it is not. With this next slide, I am going to go into DR. LIPPMANN: Excuse me. I don't quite see where asterisks fit a threshold distinction. DR. TODHUNTER: In fairness to the remaining speakers, if I can come back to you specifically on that slide and reference it to my submission, would that be okay? I can go into that. The top one obviously has to do with I mixture interactions, which you can affect. The others have to do with rates of dglivery to target tissue and background risk process rates. DR. LIPPMANN: I think those are relevant to the main heading of the slide. I just don't see where they relate to threshold issues. DR. TODHUNTER: That's what I said. I would like to ogm.Ea.tcan c/QE/i.otEzas. ZC 12021 296-0261 W 17031 644-7b36

1-221 ) 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 send you that slide and specifically index that to my written submission so it's easier to determine. With the next slide this is a concern which is related here; that is that there is too much emphasis on this question of constituent similarities. Why do ETS and MS have a comparable constituent content? Quite simply because they are both derived from the combustion of vegetation-derived matter. So of course they are going to have similar constituents. In fact there are a large number of studies in the literature showing that smokes from any combustion of vegetation-derived matter contains respirable particulate matter, has many if not most of the same chemical components Y1 ; T r e,S 6 including n44res4we compounds and PAHs as does MS, and like MS can be active in the Ames mutagen test. Now by any vegetation-derived matter I mean here wood, leaves, coal, and so on. I'm not putting this forth as a trivial comparison -- DR. DAISEY: Your comments on the chemical composition of these as being very similar, I think you can't justify. DR. TODHUNTER: Oh, yes, I can. I can provide you with literature. 01 071'12E2GCQli GRE# OZtz 4.S. m m 17C (zozJ rq6-oz6, q,04 (7031 644-7636

1-222 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 DR. DAISEY: I think I've seen that same literature and spent a chunk of my research life looking at it. DR. TODHUNTER: We can discuss that at some point, but you will find PAHs in these smokes -- DR. DAISEY: There is a difference between finding PAH -- you can certainly find PAH in all of those, but there is a difference between that and looking at the overall chemical composition of it. If you look at the overall chemical composition of these things, there are not identical. DR. TODHUNTER: Well, neither are ETS and MS. DR. DAISEY: And many of them are not all that -- DR. TODHUNTER: There is not an identicality -- DR. DAISEY: Many of them are not all that similar, from a chemical point of view. DR. TODHUNTER: If you would like, I would be more than happy to -- DR. DAISEY:. And in terms of their biological potency, they have been shown in animal tests to be different. DR. TODHUNTER: Well, you also have the same difficulties with MS and animal testing. I will be more than happy to provide the literature and just put it out for the d¢inexica.n 4-Roottzu Ze jzo2f 296-oz6, wo4 hosl 644-7636

1-223 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Committee to look at. In any case it's not a trivial comparison either since human exposure to smokes and smoke particulates can be significant. Just as two examples, Dresser has shown that wintertime nonattainment of EPA's of Colorado is mainly shown that wood smoke the total PAH load to automobiles do. DR. DAISEY: due can air Is to wood PM-10 standard in portions smoke and Mele et al have easily contribute up to 10 times in rural areas there a in winter than reason to think that there is a differential effect on nonsmokers who either live with smokers or do not live with smokers for these exposures? DR. TODHUNTER: Yes, actually, if you'll allow me to go on, I'll cite you one study in particular. For that particular thing I was going to add that you might note the Sti-~he <J study of 6i-blrdu in 1990 from Japan which finds a relative risk of 1.9, which is statistically significant at the 95 percent confidence level for the use of straw or wood as a cooking fuel in women. DR. DAISEY: That was not my question. DR. TODHUNTER: Well, I thought it was your question. But in any case, no increased risk in that particular study for ETS. The whole point of this is not c4ma-ti.can ::RE#otfEu Ze /Zo2l Z96-o16, NC4 003/ 644-T636

1-224 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 identicality but rather that there is not identicality between even MS and sidestream smoke, let alone between ETS. And the best that was come up with in the draft risk assessment is that there are many shared constituents. I think the whole point of this is that shared constituents is not a sound basis to make a classification. I think the SAB recognizes that. It just leads you into some logical absurdities. You just can take chemical analysis so far. If you took me and analyzed me, I'd analyze in terms of chemical components about the same as a sea urchin, but I'm not; biologically quite different. The only relevant issue is what's biological. This does raise some concerns, particularly in light of Dr. Farland's comments that this may be a precedent setting process. I'm a little concerned. Since I'm running short of time, if I could go to the very last slide. I would just like to remind the Science Advisory Board what a Group A agent looks like or historically has lookgd like and what one does not. In this slide I've provided relative risk ranges reported for the marker tumors for arsenic, benzene, and vinyl chloride which have been classified as Group A, and for electromagnetic fields which recently was classified as inadequate, and ETS for comparison. u, ~-A cn W ogme~Ecan ~a~so~tets 17C /zozf zq6-ozb, W /703j 644-7636

1-225 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 I think quite clearly an agent that gives a relative risk of 1.19, and that after considerable torquing around of the database whose statistical significance is highly dependent on which test is used and what assumptions are made about the uncertainties in the data, just doesn't belong in Group A. Since ETS and MS really aren't biologically comparable, as the Agency itself admits, the case for ETS classification really has to rest squarely on the ETS epidemiology. I don't think that is a strong case. It's certainly weaker than that for EMF, and I don't think there is any question that the proper classification for ETS would be Group D and not Group A. Thank you. DR. WESOLOWSKI: I have a question. You have the slide there taken from Chapter 6, page 6.6, which said the basic assumption that lung cancer risks in passive and active smokers are equivalent, is not tenable, et cetera. But there were some dots in ther,e, and the dots were rather important. The quote says "the basic assumptions of cigarette equivalence procedures is that the lung..." The point is that EPA did not use the cigarette equivalence procedures. If you were just agreeing, then EPA did a biased thing in not using that procedure. Ln r PTr12EZICQ.12 G~40ttE4s `1.~C /zoz/ zq6-o26, q'~Vq /7031644-7636

1-226 1 DR. TODHU why I bring it up 2 that there is a difference in C hapters 3 and 4 and Chapter 6 3 in terms of what EPA was lookin g at. The point I'm making 4 from that is that in Chapter 6, the reason they don't use the 5 cigarette equivalence is becaus e they don't think they are 6 biologically equivalent. 7 DR. WESOLOWSKI: Hold on just one second. The 8 audience, I'd like them to see the full quote which is on 9 page 6.6, and I think that'll c larify it. Thank you. 10 DR. TODHUNTER: And I think it's consistent with 11 what I said. That's why in fai rness I said, you are talking 12 about QRA in Chapter 6 and you' re talking about hazard 13 identification in Chapter 3 and 4. But as a -- and I think 14 as you are well aware, if you h aven't got sufficient 15 biological comparability to mak e QRA comparisons, then you 16 really ought not to be making h azard identification 17 comparisons. 18 MR. FLAAK: Dr. Witor sch? 19 PRESENTATION BY DR . PHILIP WITORSCH, 20 TOBACCO I NSTITUTE 21 DR. WITORSCH: While we're setting that up, my name 22 is Philip Witorsch. I'm a pulm onary physician and clinical 23 professor of medicine and adjun ct professor of physiology at ~ ~ ~ Anezi.can lz~?rootEetl N ~ ~L1C (zo2J zq6-oz6r ~ ~Uo~{ (7o3I 644-7636 Q cn

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-227 George Washington University Medical Center in Washington, D.C. I am here at the request of the Tobacco Institute, and •my comments will specifically address the issue of parental smoking and respiratory disorders in children covered in Chapter 7 and part of Chapter 8 of the EPA review draft. I must state at the outset that the opinions that I am expressing are my own and those of my coauthor and they do not necessarily reflect the opinions or positions of either of our respective institutions or for that matter necessarily of the Tobacco Institute. The EPA review draft concludes that exposure of children to ETS from parental smoking is associated with increased prevalence of respiratory symptoms of irritation, increased prevalence of middle ear effusion, and a small but statistically significant reduction in lung function. It also concludes that exposure to ETS from parental smoking is a risk factor for asthma prevalence and severity and is associated with increqsed risk of lower respiratory tract infection in young children, particularly infants. We have conducted an independent analysis of the published studies pertaining to the association between parental and household smoking and respiratory illness and pulmonary function in children. This analysis included with OTI'rZCtCCQ.l2 G!`E#O'LfE4.S Ze (2021 Z96-oZ6, W (7osl 644-7636 1~_ Ln W N

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-228 few exceptions all the papers reviewed in the draft and was in fact more comprehensive and in our opinion more thorough than that conducted by the EPA. The results of our analysis do not support the conclusions stated in the EPA review draft. As can be seen from this slide, if you can see it, we reviewed and analyzed 41 studies addressing respiratory clinical endpoints in preschool children, 45 studies addressing respiratory clinical endpoints in school-age children, 17 studies addressing middle ear disease, and 38 studies addressing pulmonary function endpoints in school-age children. In this analysis we considered both consistency of specific endpoints from study to study and certain important aspects of the design of these studies, namely validation of clinical endpoints by medical records or physician's examination and treatment of potential confounding variables. While a consistent association between maternal smoking and respiratory symptoms and certain diseases was noted in the 41 studies involving preschool children, little consistency was evident with regard to specific clinical endpoints in the 45 studies regarding school-age children. Even when an association was evident in older children, there ,vq17LEtLCCt/2 .-:REhO'LtEYS. W 1202] Z96-OZ6! W1Vq 003/ 644-?636

1-229 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 was considerable variation from one study to the next with regard to the particular clinical endpoint, such as asthma, cough, wheeze, or bronchitis. Thus tabulation of all the data from the relevant studies as shown here reveals that an association between parental smoking and a specific endpoint was generally confirmed no more than 50 percent of the time. Similarly a tabulation of the 17 studies dealing with parental smoking and middle ear disease reveals that almost two-thirds of these studies fail to show a statistically significant association for this particular endpoint. When the epidemiologic studies dealing with the association between parental smoking and pulmonary function in nonasthmatic school-age children were analyzed as shown here, specific spirometric measures were not consistently associated with parental smoking. Furthermore reported decrements in spirometric measurement were usually small in magnitude and within the range of normal pulmonary function. With regard to the one possible exception, FEF,,_,,, where an association of a relative decrement with parental smoking was reported in two-thirds of the studies, it is important to note that this particular parameter has a very high coefficient of variation. In many of the studies the d¢rnetlcan '=Re1tottZt.s. ZC /zozf z96-0261 q/a4 17o3164a-7636 W N

1-230 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 differences between groups, particularly with respect to the FEF25-75, were well within the range of variation generally seen for this parameter, even if they were statistically significant. We next looked at certain aspects of the design of the epidemiologic studies of parental smoking and respiratory endpoints in children, both in the preschool and in the school-age groups. Using a specifically designed protocol, each of the studies was examined carefully and thoroughly to determine whether or not clinical endpoints usually derived by parental responses to a questionnaire were validated by medical records or a physician examination, which potential confounders were considered, what specific criteria were employed for each potential confounder, and whether or not a statistically significant association was found between a potential confounder and a clinical or pulmonary function endpoint. Although verification of clinical endpoints in epidemiologic studies of preschool children was relatively greater than in those of older children, it is noteworthy that even in this preschool group, almost 50 percent the studies lacked such verification, while in the school-age c4ff2E2CCQ/2 GRE#02fEY.i !bC (zozJ z96-oz6i Q~a4 (7031 644-7636 i

1-231 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 group verification was lacking in almost 85 percent the studies. As shown here, and I'm not going to go through all of those just to illustrate, a list of 21 potentially relevant confounders was derived from variables considered in these and other epidemiologic studies. DR. SAMET: Each of these then you regard as causally associated with child respiratory health. DR. WITORSCH: Potentially. DR. SAMET: Potentially? Actually or potentially? DR. WITORSCH: Potentially. DR. SAMET: Potentially. So these have not -- DR. WITORSCH: Some there is evidence; some there is no data on. DR. SAMET: So these are, as we were discussing before, sort of potential, potential confounders, not each having sufficient evidence to be causally linked -- DR. WITORSCI3: Some are real potential confounders in that there is sufficient evidence. Some are potential, potential; that is correct. DR. SAMET: And in the studies of concern we have asthma incidence and exacerbation, childhood lower respiratory tract illness, we have middle ear effusion, we c7n2E2GCQ.12 GRE,h0YtE4.3. ~M tzozl Pq6-o26Y `v~q hosl 644-7636

1-232 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 have respiratory symptoms and reduction of spirometric lung function with an array now of 21 factors. In fact the evidence for each of these factors in relationship to those various outcomes to judge their causality is limited. DR. WITORSCH: It is. If you look at it that way, that is absolutely correct. Now, our analysis revealed that there was considerable variation in the extent to which potential confounders were considered in both clinical and functional endpoint studies, in both younger and older children. Those considered infrequently or not at all included indoor pollution, day care use, animal exposure, stress, dampness and cold, heating and air conditioning, season, quality of housing, nutritional status of the subject, have particular importance in the case of preschool studies; maternal smoking during pregnancy, which was considered in only 4 of the 41 studies. Furthermore,there was no apparent standardization or consistency with regard to the particular confounders considered in individual studies, as well as poor representation of the potential confounders in many of the individual studies. On the average, each paper considered only 7 confounders, with a range of from 1 to 16 confounders Aaezi.ca.n ::R'e#o%tzu Ze /zo2J zq6-oz6, q/6q /703/ 644-7636

1-233 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 per paper. As can be seen here, very few of the papers considered 11 or more, or more of the half of the potential confounders. There was also considerable variation in the specific criteria used to characterize potential confounders, most notably among such categories as socioeconomic status, family health history, and personal health history. In the case of socioeconomic status, for example, no fewer than 10 different individual criteria were used among the various studies. Controlling for a confounding variable was also not standardized. Depending upon the study, the adjustment of a confounder would vary from not at all to sophisticated logistic regression or stratification. Our analysis also revealed that a number of these potential confounding variables were in fact consistently associated with increased prevalence of respiratory symptoms and diseases. Although this analysis has not yet been completed -- and that's just within these studies -- we have so far found six such variables in the studies of preschool children, three of which were also evident in the studies of school age children. They are listed on the slide. Some other variables were found to be consistently ,Vqh2E4CCA.rl cz~?S#A'LfE4.S. 17e tzo2f sq6-o26, No4 /703j 644-7636

1-234 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 well as inconsistent with other published literature. as While this may reflect a true lack of any relationship between socioeconomic status and respiratory health in children, it is more likely a reflection of the lack of standardization of the specific criteria used for each confounder, the specific clinical endpoints considered, their lack of validation, and the variability of treatment of the confounders. While parental smoking may be consistently associated with increased prevalence of respiratory symptoms and certain diseases in young children, especially.infants, no consistency of association is found with middle ear disease at any age or with respiratory symptoms and disease or impaired pulmonary function in school-age children. The EPA review draft attempts to convey the impression that the issue of confounding variables is resolved. On the contrary, our analysis indicates that for studies of parental smoking and children's respiratory health -- both in preschool and school-age children -- attention to confounding variables lacks rigor in a variety of ways. Ln ~ Ln 1 not associated with an endpoint, or to provide equivocal results in these studies. These included socioeconomic status, a finding that we thought was counter-intuitive W N ) OTI72E2GCQ.12 GRE#0'LtE'La 'L1 e tznz/ 2q6-ozG, ~df /703) 644-7636

1-235 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 There is also deficient rigor reflected in the relative lack of verification of clinical endpoints. It is of interest that several confounding variables were found to be associated with increased prevalence of respiratory symptoms and diseases in children, particularly in preschool children, while others that are potentially important, such as maternal smoking during pregnancy, have been greatly underaddressed. The foregoing inadequacies in the treatment of confounding variables raise a compelling question as to whether or not the apparent consistent association between parental household smoking and respiratory illness in young children is legitimately attributable to ETS or is the result of bias due to confounding by one or more or even a combination of such factors. Thank you. DR. LEBOWITZ: Well, I appreciate the fact that you can count things like confounders, but part of the question is whether for each study you went back to all the reference sources to that study to see if in some way at some time they have examined those confounders and other factors you say were ignored, or have examined such things as medical records or clinical examination for validation; number one. Number two; I assume that each time you said there , Lin l ~o N c~rnEUCa.n ~e~t.o~.tats a, rL1C /2o2f 296-o26, ~~. ~ No4 /7031 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-236 was a lack of a clearcut finding of association that if the authors did not calculate power, that you yourself calculated the power to show there was sufficient power to have such an association if one did exist. DR. WITORSCH: As far as the first question, we,did not always go to the sources referenced in the paper, but we went through the papers exhaustively and really bent over backwards in terms of assuming any treatment at all being considered, being a consideration of the confounder. In some cases, but not all cases, we may have gone back to references or other papers from that group, but otherwise, for the most part we went through the specific papers. My coauthor has got a comment on that. COAUTHOR: As far as going back, we attempted to give the authors in terms of confound the benefit of the doubt, which meant that usually when a paper is cited for previous reference or methodology, we went back as far as we possibly could, so thot we attempted as best as possible to leave no stone unturned with regard to consideration of confounding variables. With regard to whether or not an observation was statistically significant, we based the judgment on the conclusion of the author, not necessarily on our assessment. ~ ~ ) ogrneti.ca.n :zRr-#oq.tEas !be I2o21 29f1-o26, ql6q /703j 644-7b36

1-237 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 If the author had no statistics that stated that there was an observation, again, we attempted to leave bias in favor of what the author was attempting to conclude. The only time that we went against that was when an author concluded that there was a statistically significant,effect and there were a 95 percent overlapping of confidence levels. DR. LEBOWITZ: So you went one way and not the other. Well, I would suggest, knowing most of the studies that have been done that you did not do a sufficient job of going back to the references that documented evaluation of what was done in those studies, and that you cannot claim that someone didn't show an association if he didn't calculate the power. However, I would like to see all the documentation you have on these different studies to confirm my conclusion that that was the case. I know reading my own papers, for instance -- and I have written quite a few -- if you did not go back to non-ETS pagers and studies, of which there are about 200, then you will not have absorbed that in fact the other confounders were evaluated and you will not have seen that power calculations have been performed. So you would draw the wrong conclusions. If that were the case for some of the other studies PTmE'LGCGt12 4z~?Poo'LtE'L1 25C /2oz/ zq6-oz6r q,6q /703/ 644-7636 ti

1-238 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 of which I am very familiar, then I would have to say that you have to return to your study of these other studies to determine if your conclusions are correct. DR. WITORSCH: I think with respect to your papers, I suspect we rarely ever mischaracterized them. We have read a lot of them. As far as some of the other papers, certainly we're open to look at that again obviously. Did Dr. Samet have a question? I guess he stepped out. MR. FLAAK: One the one hand I'm encouraged to take a short break and on the other hand I know that's the equivalent to one or two speakers, so I thought we might just keep on rolling along for a little while. Let me identify the next couple speakers. The next speaker is Dr. Gary Giovino, followed by Dr. John Banzhaf, Dr. Elizabeth Fontham, Dr. Ron Davis, Dr. Clark Heath, and Dr. Jeff Wagener. We'll see if we're either out of time at that point or everybody has fallen on the floor. Gary? PRESENTATION BY DR. GARY GIOVINO, CENTERS FOR DISEASE CONTROL DR. GIOVINO: My comments will be brief. I'm Dr. Gary Giovino. Currently I serve as the Chief of the Epidemiology Branch of the Center for Disease Control's Office on Smoking and Health. Contributions to these ogaztica.n z~?e#otEets `1.~C /zo2/ 2q6-ozG, W /703/ 644-7636

1-239 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 comments came from the National Institute for Occupational Safety and Health, the National Center for Environmental Health, and the Office on Smoking and Health. Each year about one in every five deaths or more than 430,000 deaths in the United States is caused by cigarette smoking. The cost of smoking-attributable disease to our nation was estimated at $65 billion in 1985. Since 1964, the year of the first Surgeon General's Report on the Health Consequences of Smoking, more than two and a half million people have died of lung cancer alone. Since then research has shown that the health risk from inhaling tobacco smoke is not limited to the smoker but also include those who inhale environmental tobacco smoke. As you know, the 1986 Report of the Surgeon General on the Health Consequences of Involuntary Smoking focused on the health effects of exposure to ETS. One of the three major conclusions of this report was that involuntary smoking is a cause of disease,, including lung cancer, in healthy nonsmokers. This conclusion was based on four main findings: the absence of a threshold for respiratory carcinogenesis, the presence of the same carcinogens in mainstream and sidestream smoke with some carcinogens being more highly ogme7i.ccsn eRz#ottet& ZC /2oz/ 296-0261 q~d4 /703/ 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 1-240 concentrated in sidestream smoke than mainstream smoke, the demonstrated uptake of environmental tobacco smoke constituents by nonsmokers, and the demonstrated increase in the risk of lung cancer in nonsmokers with long-term household exposure to environmental tobacco smoke. In the same year, the National Research Council concluded that exposure to ETS can cause lung cancer. According to a 1988 Federal survey, 17 million U.S. nonsmokers experienced at least some discomfort from ETS exposure at work with 4.5 million of these experiencing great discomfort. Further in 1991 the CDC's National Institute for Occupational Safety and Health, NIOSH, found that the collective weight of evidence, that is from the Surgeon General's reports, the similarities in composition of mainstream smoke and ETS, and the recent epidemiologic studies is sufficient to conclude that ETS poses an increased risk of lung cancer and possibly heart disease to occupationally-exposed workers. NIOSH recommended that ETS be classified as a potential occupational carcinogen, as defined by OSHA's carcinogen policy, and that occupational exposures to ETS be reduced to the lowest feasible concentrations. The draft EPA Ln r ) c7qm.zti.can cRs~G.oa.tEas L~e t2021 Z96-OZ6Y q,d{ 1703] 644-7636

1-241 ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 risk assessment updates and strengthens the science base regarding ETS. As stated in this draft, the individual risk of lung cancer from exposure to ETS does not have to be very large to translate into a significant health hazard to the U.S. population because of the large number of smokers and widespread presence of ETS. The report estimates that the relative risk of lung cancer in nonsmokers exposed to ETS to be 1.19 and the number of annual U.S. deaths resulting from such exposure at 3,000. By using only studies conducted in the U.S., the EPA estimates are conservative and appropriate for the U.S. population. The 1986 Surgeon General's report also concluded that children of smokers compared to the children of nonsmokers have an increased frequency of respiratory infections and respiratory symptoms and slightly smaller rates of increase in lung function as the lung matures.. More recent literature reviewed in the EPA draft risk assessment confirms that conclusion and also finds that children's exposure to ETS is a risk factor for chronic middle ear effusion and asthma exacerbations. These new findings have enormous and important public health implications for American children. ~ ~ N 0T1rlE'L/CQ/2 GRE,&04tzts L~e t2021 P96-OZ6, W /7o3f 6q4-7636

1-242 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 We find this report to be a comprehensive and objective summary of the available evidence. It uses a sound analytic approach to synthesize the vast body of literature on health effects of passive smoking. The CDC supports the classification of ETS as a Group A or known,human carcinogen, according to the EPA carcinogen classification system. In summary, we support the conclusions of the draft risk assessment and we commend the EPA for the important work it is doing in the area of environmental tobacco smoke. Thank you. PRESENTATION BY DR. JOHN BANZHAF, ACTION ON SMOKING AND HEALTH DR. BANZHAF: Members of the Board, my name is John Banzhaf. I'm Executive Director of Action on Smoking and Health, sometimes known as ASH, a national nonprofit scientific and educational organization concerned with the problems of smoking and with the rights of nonsmokers. I've absolutely no criticisms or even suggestions concerning the substantive conclusions of the EPA's tentative draft. Indeed I would like to convey our appreciation to the EPA, its staff, its consultants, and to the members of the SAB for doing such a fine job under what are very trying conditions and lots of pressure from the tobacco industry. 44m.EZican z~?ehotfz4s. 2~C (2ozf zq6-o261 Q04 0031644-7636

1-243 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 However, while I believe the report, especially the section on lung cancer, is the most comprehensive yet produced, much of t's importance and fact is likely to be lost because it is also one of the most incomprehensible. I say this as somebody with a scientific degree from MIT, several published papers, two U.S. patents, a well-known mathematical construct now known as the Banzhaf Index to my credit. I've also been following the medical and scientific literature on the effects of ETS ever since 1969 when we first summarized it and used it to persuade the then-Civil Aeronautics Board to require smoking/no smoking sections on airplanes. Yet even with this background, which I think is a little bit more than the general public has, I must confess to you that vast parts of your report are incomprehensible. I'm not just talking about the formulas or the appendix. I'm talking even about thq summary and conclusion. Far more importantly, let me tell you that I spent most of the day yesterday on the telephone with reporters who were generally confused -- and in many cases I think deliberately so by the Tobacco Industry -- by various pseudo- scientific arguments which they continued to make. You heard ko N ogm.etiean z~?e#ottets. Ze 12C21 Z96-o26, 464 /7031644-7636

1-244 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 some of them today. The ones they try to put off on the public are far, far worse. This is not the first time this problem has occurred. It happens every time the topic of the EPA's report becomes newsworthy. In fact, it occurs every time the question of the carcinogenicity of environmental tobacco smoke is raised in any public forum; that is, before a legislative body, a regulatory proceeding, a talk show, or any other. Many members of the press who I think in general have more education, greater background than the general public, for example have real problems understanding how a study which is statistically insignificant -- as some of them put it, or not statistically significant -- could possibly have any value'at all. They ask how the EPA could possibly have concluded that ETS causes cancer in nonsmokers based upon only 30 studies, many of whicb don't show this at all, or how we can be sure that the wives in the spouse study didn't contract lung cancer as a result of other dangerous habits of their smoking husbands. Now although many of the answers to these questions are available in the report somewhere, they are buried in CA V /mE2(.CQ.n V \E/(l0'i.LLt1 W t2oPl 296-o26, "Kc4 t703/ 644-7636

1-245 1 2 3 4 5 6 7. 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 discussions of confounders, type 1 and type 2 errors, predictive power, tiers of studies, and other jargon largely incomprehensible to the intelligent layman to whom I think the report is ultimately addressed, and who ultimately will be making the important decisions based upon it. As a law professor who spends some time trying to teach first year law students how scientific questions concerning the public health are approached, I can tell you that many people, even those regarded as very intelligent -- our incoming law students -- don't have any idea how these decisions are made. And I can tell you from personal experience over many years the same thing is true for legislators, regulators, and judges in a growing number of legal proceedings which now involve the issue of environmental tobacco smoke. For example, many members of the public simply don't understand that many decisions made to classify or regulate chemicals as.carcinogens are made primarily, if not exclusively, on the basis of animal studies or highly elevated levels of exposure. Indeed many think that a single study either proves X or disproves X, and there is nothing in between. They have no real inkling of statistics or epidemiology. o'¢in.etica.n eRa/zotEEts Ze 12o2i Y96-oZ6, qor{ t7031 644-7d36 ~ ~ ~

i 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 1-246 As a result, I feel that much of the force and effectiveness of the report, and therefore much of the effort which has gone into creating and perfecting it, will be lost if the information is not presented to the public in a form that it can understand and have confidence in. Publishing a report which is accurate but unintelligible, comprehensive but incomprehensible, complete but confusing, undercuts much of its value, does not provide the information governmental officials and others need to make crucial decisions, and only serves to further encourage the tobacco industry to hold these preemptive press conferences to attempt to confuse the public with pseudo- scientific arguments. Remember that unlike many other reports which seem to be addressed to narrow chemicals in a particular industry, this is one that affects virtually all nonsmokers. Thousands of legislators, regulators, governmental officials, judges and lawyers are going.to see it, read at least the summary and conclusion, and try to make intelligent decisions based upon it. For these reasons I would like to suggest five simple suggestions to substantially improve the readability and persuasiveness of the report. None of them affect the ti ~ ~ a'gineti.ea.n eReliottr-ti ze (zoz/ 2qb_ozb, N4 t703j 644-7636

1-247 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 . 17 18 19 20 21 22 23 substantive conclusions; therefore I don't believe they should hold up the report at all. First, I would rewrite the all-important summary and conclusions chapter -- which is the section most likely to be read by these judges, legislators, regulators, and so on -- with the help of a professional science writer to state exactly the same conclusion in a form which is easier for laymen to understand and appreciate. For example, in summary and conclusions we read, "The studies are individually assessed with strength of association exposure response trend. They quantitatively value potential confounding, bias, and likely ability to provide information about lung carcinogenity and ETS." Well, I think I know what you mean. I can tell you the great majority of the public does not. Unfortunately scientists, health professionals, even lawyers tend to write in a particularly stilted manner that we've become accustomed to, and of course we write in jargon. Fortunately in the science area there are skilled science writers who can make that more intelligible. Unfortunately for lawyers, nobody can do anything with legal writing. So I would strongly suggest that you see if you can't get a science writer to help. Ln ~ cn W o'g»zzzican I:_Re1sotfzts `1.~C /zoz/ z96-oz6i q,~vq /7031 644-7636

1-248 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Secondly, I would include a section listing all the governmental and other scientific/medical bodies, both here and abroad, which have likewise concluded that ETS is a dangerous carcinogen which causes lung cancer and lung cancer deaths in nonsmokers. Why? Let me give you an example of what the tobacco industry said just yesterday, from a UPI story. They said, "The EPA report is distorting facts to support political viewpoints. It is subjecting science to tortuous statistical manipulations in order to say what may be politically correct." I think Dr. Gori said some of the same this afternoon. I think for most members of the public one of the most persuasive things is this is not something the EPA has suddenly discovered, but that well over a dozen different scientific, medical, governmental, nongovernmental bodies, not just here but in other countries in the World Health Organization, have all reached exactly the same conclusion. Your report.mentions a few, and cites a little bit from them. Why not take all of these different bodies which have made these conclusions, set out their major conclusions, and show that the EPA, and by inference the SAB, is not off politically grandstanding and subverting science to policy. Obviously, all of these different bodies can't all dTrneucarn tRe#ottets Ze /zo2i 296-o261 q~aq 1703) 644-7636

1-249 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 be biased against the tobacco industry. Indeed with that evidence people might begin to suspect it's the other way around, that it is the tobacco industry which has their bias and that they are using it to perhaps change their scientific viewpoint. Indeed I would respectfully ask and suggest that this panel, this afternoon, ask each of the experts who have testified here today on behalf of the tobacco industry if they will state on the record whether or not they agree that the evidence today overwhelmingly supports the conclusion that mainstream tobacco smoke causes lung cancer in smokers. Many of these people are still in the room. They can be asked before they leave. Now, this is not in any way an unfair question, particularly in view of the various insinuations that have been heaped upon the EPA, and by inference, I believe, the SAB, which supported the conclusion, because they claim that although they are being paid by the tobacco industry, their payments have not affected the objectivity of their scientific views. Obviously they have to be familiar with the literature because some portion of the EPA report deals with mainstream tobacco smoke and its effect upon smokers. So ask them. Do you agree that mainstream smoke o'grnEtlcan ~Z#ottets `1.~C /ZCZf z96-cz6, W 1703J 644-7636

1-250 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 causes lung cancer in nonsmokers? Now, if their answer to that question is no, this should undercut if not destroy whatever scientific credibility they may have come into before this Committee, as well as the weight the Committee might be inclined to give to their assertions that ETS evidence is simply not strong enough and that more study is needed, because after all, if they are not willing to admit today that there is evidence that smoking causes lung cancer in smokers, they will never admit that tobacco smoke causes lung cancer in nonsmokers at the other end of the room. It's a simple thing the Committee can do to clear the record. By the way, if any of them do admit that mainstream tobacco smoke causes lung cancer in nonsmokers and they are testifying here on behalf of the tobacco industry, I think it will be a front page headline in tomorrow's news. My third suggestion would be to include within the report a question and answer section, in which doubts about the statistical significance, the weight of the evidence, and other similar questions are answered briefly in lay language and using simple examples for those who are not familiar with statistics and epidemiology. I have to do this all the time because people ask me what the EPA report is all about. How 23 1 could something statistically insignificant amount to oqmezican -Re fsottEts. Ze 12021 296_026, qoq /7031 644-7636

r 1-251 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 anything? How could you push studies together? Well, I explained to a reporter, let's make a deal; I'll take out a coin, we'11 flip it, and if it comes down "heads" I'll get a hundred dollars; if it comes down "tails" I'll give you a hundred dollars. You say to me, well, let's •test that coin. So I flip it a hundred times and it comes down fifty-six times "heads." Well, that's not statistically significant to 95 percent probability; only about 90 percent, so you're a little bit suspicious. I'm almost certain you wouldn't desire to inhale a chemical which.has been proved to cause lung cancer only to a 90 percent probability. Now, suppose we do it again and again and again and again. We do it four times, four hundred flips, and each time it comes down on an average of 56 times. No one study is statistically significant, but you put them all together and statistical significance is overwhelming. I think it's about .99. So yes, you can in some senses take individual studies which are not.statistically significant and with care -- obviously this is an oversimplification -- put them together. These are the kind of explanations that people want and need if they are going to trust the report that you give them. 3 23 1 Number four; I would add a table showing the number co 1:Vqmezf.can 1:: IRE# ottcts '1X' /zozf zq6-oz6t qUo4 /703j 644-7636

1-252 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 and different types of studies, the types of evidence upon which other substances have been classified as Class A carcinogens. A lot of people think the EPA is going way out on left field on this. I mean, after all, you've only got 30 studies, plus animal studies, chemical studies, in vitro studies, so on and so forth. Don't you require much more to classify all these other things? I think if they saw the evidence laid out on other common substances that they know, they would have a much better appreciation as to the strength or weakness, and again this does not affect the substantive conclusions at all. It shouldn't be something that should require review. This is a matter of record. Five; I think it would be important to prepare a chart showing the number of lives which are estimated to be saved by the EPA or other agencies' regulation of carcinogenic substances. A lot of people say what's 3,000 a year? We've got a lot of people in the country; what's 3,000 a year? Well, I think if they sit down and look at the other substances, whether they are carcinogens or whether they are components of outdoor air pollution, look at the estimates as best we -- we, your agency and the others -- 1Vq/72E4LCQ.I2 GRE`20'CtEZ1 17C /zo2f 2qb-oz6, 4/04 /7031644-7636

1-253 1 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 have made them as to how many lives are placed at risk and why we're spending billions of dollars to reduce exposure to them and compare it with the estimates you have here, they would appreciate your report all the more. In summary then I think your report -- I've no criticisms of your report and its conclusions -- but I am respectfully suggesting given all the time and energy and reputation that you have put into it will be of little value if it is simply scientifically correct but is neither believed nor understood by the thousands, tens of thousands, of people without your scientific background who will have to make the decisions as to whether or not to ban smoking in this place, to limit smoking in that place, whether to award damages in ETS cases in courts and agencies, and to make all the other decisions. You cannot ignore the impact of the report. Make it not only the best scientifically, but the most readable and the most persuasive. I thank you for the opportunity and would welcome any questions. DR. BAYARD: One question. I know you want to get a professional scientific writer to write the summary and conclusion, but do you think you could get three or four professional scientific writers to agree if they had to write N P/wE4CCR.l2 GRroo4.tE'LS. 'L1C (zozJ 296-0261 q'dq (703] 644-7636

I-254 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 the conclusions together? DR. BANZHAF: I didn't say I wanted to get a professional scientific writer to write the conclusion. What I would like to suggest is that the EPA and particularly the people who are drafting the all-important summary and _conclusion section, which is the only part the overwhelming majority of the public will ever read, employ the assistance of one or more science writers who will assist them in a variety of ways in simplifying the writing style. If you get two or three of them, yes, I think they will all agree on certain things. You have got to shorten the sentence length. You've got to avoid a lot of the jargon which is in there. It's all right to use "confounders" somewhere in an appendix or off in Chapter 10 after you've begun to explain it. Using them up in front, nobody knows what they mean. I think you'd find that many science writers would agree as to the basic things that have to be done. It's been my experience that they can improve it, and one good example would be some of the articles which have appeared for example in "Consumer Reports" and particularly a very early book by "Consumer Reports" which explained the then very thin evidence that active smoking caused lung cancer in smokers. That is the first Surgeon General report 04m.e7l.can v?e#ottzzs 2~e tZoZI Z96-OZ6y w11q ho31 644-Tb36

1-255 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 in 1964. Many of these same questions were then raised by the tobacco industry, and this report took you step by step: well, how do we know it isn't something else, and what does a dose response mean, and why is it important that all these different studies all came to the same conclusion, and so on. And once again, I hope you really will ask the tobacco industry what they think about the evidence relating to mainstream smoke. I think that's part of the record. I think the public deserves to know it. I think it would be very interesting to get the answers. Thank you. MR. FLAAK: Dr. Fontham is next, but in the interests of comfort, why don't we all take a short break for about seven minutes or so. (Brief recess.) PRESENTATION BY DR. ELIZABETH FONTHAM, SPEAKING ON OWN BEHALF DR. FONTHAM; In my prepared remarks -- sorry, excuse my voice, it's going -- I have written down good afternoon. At this point I'll say good night, good evening. I'm Dr. Elizabeth Fontham. I currently hold an appointment as Associate Professor in the Pathology Department at Louisiana State University School of Medicine. W N ogrnEZi.can lz~?e#OttEts. Ze tzo21 fq6-o26, qd{ t703] 644-7b36

1-256 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 I was invited to come here today by the Office of Smoking and Health, but I'm speaking on my own behalf. I serve as the National Coordinator and the principal investigator for the Louisiana component of a multi-center case control study of lung cancer in nonsmoking women, which has been included in this report of the assessment of lung cancer in adults and its relation to passive smoke exposure. Our study was initiated in 1985 to evaluate the association of environmental tobacco smoke exposure with risk of lung cancer in lifetime nonsmoking females. The study was designed to minimize -- although it couldn't possibly eliminate -- some of the methodologic problems which have been concerning previous studies of ETS and lung cancer; specifically misclassification of smoking status, accuracy of case diagnosis'and cell type, recall bias, ETS exposure from sources in addition to spouse, and potential confounders. The findings of the first three years of our five- year study were published in November 1991, and our report has been included in the present EPA draft. I would add parenthetically that our study was originally designed to be a three-year study, but because of the unique histologic type distribution, we applied for additional funds to do two additional years. Ln ~ ~ kD 01 ,=PIMEZI.CQlZ GREf1A4.tE4& Ze t2oZl 296-0261 N-lq t703] 644-763b W U1

, 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-257 Because our study is one of the more recently reported ones, I would like to briefly describe our study and its findings. I'm clicking for the first slide, but perhaps in the wrong direction. The study is a population-based, case-controlled study of lung cancer in women who never used any type of tobacco. Tobacco use was defined as 100 or more cigarettes smoked or any other type of tobacco use for six months or more. Geographically the study included five metropolitan areas: Atlanta, Houston, Los Angeles, New Orleans, and the San Francisco Bay area, representing a population base of approximately 18.5 million people. Rapid case ascertainment procedures, which included weekly or biweekly review of pathology reports from study hospitals, were used to identify potentially eligible lung cancer cases. Upon completion of case ascertainment, at the end of the five years of study, over 17,500 potentially eligible cases were sgreened. Eligible cases included English, Spanish, or Chinese-speaking females ages 20 to 79 who had a histopathologically confirmed diagnosis of primary carcinoma of the lung made prior to death, had no history of previous cancer, and were determined to be a lifetime nonuser of 4Amezi.can zRe#oztzts Ze tZOY1 pq6-OZb, CUd{ /703/ 644-7636

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-258 tobacco according to procedures which I will shortly describe. Two control groups were selected in order to evaluate response bias which is always a consideration in comparisons of ill cases and healthy controls. The first control group, the population control group, was selected by random digit dialing and supplemented by random sampling from- the files of the health care financing administration for women ages 65 and older. Controls for frequency matched the cases on age, in broad age groups; less than 50, 50 to 59, 60 to 69, and 70 plus, in a two-to-one control to case ratio. They met the same eligibility criteria as cases for age, residence, language, and tobacco use. A second control group was selected for females ages 20 to 79 with a diagnosis of primary carcinoma of the colon -- also histologically confirmed -- who met the same language, previous caucer, residence, and tobacco use criteria as the cases and were matched by 10-year age groups and race. A multi-step procedure was used to determine lifetime smoking status. The medical record of each potentially eligible lung cancer case and colon cancer Ln ~-A Ln %D ) tz;I12EZCCQ./Z 'Z~?E f2.04.&Sa 1~C t2021 296-026! -q,6q 1703/ 644-7636 N

1-259 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 ,z4rnE2i.ca12 rE#o'LtE2S M /202) Y96-02e, qUo4 /703) 644-7636 control was reviewed for information on tobacco use. Patients identified in the medical record as current or former smokers were considered ineligible. The physicians of lung cancer cases and colon cancer controls still considered potentially eligible were then contacted for additional information on tobacco use. Potential study subjects with lung cancer and colon cancer identified as current or former smokers by their physicians were considered ineligible. This screening procedure not unexpectedly eliminated a larger proportion of lung cancer cases than colon cancer controls. All remaining potentially eligible study subjects, cases and both types of controls, were contacted by telephone to elicit information on tobacco use and ask specific questions. Women who reported ever smoking 100 or more cigarettes or using any other form of tobacco for more than six months were considered ineligible. At the time of interview, these same,questions were repeated to confirm each study subject's reported non-use of tobacco. About 4 percent of otherwise eligible women with lung cancer were found to be never-smokers after this multi-step screening procedure; that is, 96 percent were considered ever-smokers. At the time of interview and not prior to that N W Ln ~ cn

1-260 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 time, a urine sample was requested from each study subject by the interviewer. All interviewers were nonsmokers to eliminate the unlikely possibility of contamination of the sample. Specimens were stored at minus 20 degrees centigrade until shipment to the American Health Foundation for analysis of cotinine and creatinine. This biochemical determination was used as a verification of current smoking status to exclude study subjects likely after these many questions still to be current active smokers. Overall about 2 percent of the study subjects were excluded from further analysis with cotinine/creatinine ratios of 100 nanograms per milligram or higher. This level was chosen to eliminate persons most likely to be active smokers, while allowing for the possibility of very high ETS exposures. Point-8 percent (.8 percent) of the lung cancer cases, 2.6 percent of the colon cancer controls, and ~ percent of the population controls tested were determined to be ineligible by this criterion. Had a lower cutpoint of 55 nanograms per milligram, which was used earlier in an international multi-center validation study by IARC, been selected, then 2.4, 3.3, and 3.9 percent of the lung cancer cases, colon cancer and Ln N Ln W c7qmezi.can V'?e#orfxu !he tzvzl zq6-o26, q,oq /7031 644-7636

) 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-261 population controls respectively would have been excluded. And this had minimal effect on the study findings. The proportions of study subjects with cotinine levels greater than 55 nanograms per milligram and greater than 100 nanograms per milligram are virtually identical to that in the IART study which also purportedly included only nonsmokers. The lower proportion of cases with cotinine levels above 100 nanograms per milligram in our study as compared to the colon cancer controls and population controls, may be the result of prior medical record reviews and physician queries used to screen out ever-smokers from the lung cancer and colon cancer control series. The information was more commonly available in the chart and from the physician for lung cancer cases. A larger proportion of potential cases were excluded as ineligible prior to obtaining the urine specimen. Cotinine analyses are now complete for the entire study. A higher proportion of cases have no determination because approximately 34 percent were deceased at the time of interview with proxy respondent. However among all study subjects who were self-respondents, a higher proportion of cases were able to provide a urine sample; 88.5 percent c4I)?E'LI.CQn GRE#04.fE'LS. 2~e t2021 Z96-o26, N6q /703/ 644-7636 cn u,

1-262 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 versus 82.5 percent of the controls. Unless vital status is strongly associated with smoking status among cases, which I do not believe to be the case, then only one case without a tested specimen from among these proxy respondents might have been excluded as a result of urine analysis. The extensive procedures taken in this study to screen for tobacco use served to minimize misclassification of smokers as nonsmokers. In addition to obtaining pathology reports for each case, representative slides were requested from the hospital to confirm and to uniformly classify each case. Slides were available for about 85 percent of the cases and were reviewed by one pathologist specializing in pulmonary pathology. Of the cases in our published report, as shown in this slide, 2 percent were found to be ineligible and were excluded from the analysis. Upon completion of the review for the five years of the study, no additional cases were found to be ineligible and we have an overall 1 percent ineligible by path review. The histologic type of the reviewed and unreviewed cases overall at the end is quite similar. A very high proportion of the cases were adenocarcinomas -- around 75 percent of them -- and this was consistent in each of the Ln N Ln ka oqmEa.tcan e-R'e'hoXfe%1 2~e (zo2l zqb-c26y q,04 /703l 644-7636

1-263 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 o09i-2EZLCQ.I2 GRE`20'LfE'LS. ZC t2021 296-O261 4'0q 17031644-7636 five study centers. In our report, cases that had not undergone review were included in analyses of all lung cancers combined but were not included in the analysis of a specific histopathologic type. One of the specific aims of this study was to evaluate the histologic specificity of the ETS/lung cancer association by examining the relationship for each of the main histologic types. At the time of our three-year report, only the number of pulmonary adenocarcinomas was sufficiently large to achieve reasonable statistical power in histological type specific analyses. Therefore our report considered the association of ETS with all primary lung cancers and with primary pulmonary adenocarcinoma. The estimated risk of lung'cancer in nonsmoking women associated with ever having lived with a spouse who smoked was approximately 30 percent, regardless of which control group served as the comparison. An approximately 50 percent increased ris$ was observed for adenocarcinoma of the lung, a gain consistent for each control group. Separate analyses were conducted for subjects who personally responded and for those for whom information was obtained from surrogate respondents. The findings were consistent for self and proxy respondents. Ln %o N Ln

1-264 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 All odds ratios were adjusted for age, race, geographic region, respondent type, income, and education. Other potential confounders including diet, occupational exposures, and family history of lung cancer were not considered in the published three-year report. Additional findings from this data set were recently reported at the American Society of Preventive Oncology. The approximate 30 percent risk of lung cancer associated with spousal ETS exposure persisted after adjustment for vegetable consumption, which was the most significant food or nutrient factor, family history of lung cancer and employment in high risk occupations or industries. Household radon was determined in a sample of case and control homes under separate funding. Radon levels are quite low in all of the areas included in our study and less than 1 percent of all homes tested had levels of 4 picocuries per liter or higher. The observed ETS associations in our public report seem unlikely to result from confounding by these factors. A positive dose response was observed for all lung cancers and pulmonary adenocarcinomas with several measures of spousal ETS including dose, duration, and pack/year. Trends were similar but tended to be higher for pulmonary ,vqh2E4.GCRl2 GRE#ATtEZ1. Ze tzo21 Yq6-vz6, NP4 /703j 644-7636

1-265 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 adenocarcinoma than all lung cancers combined. Exposure to ETS from various sources during adult life was evaluated in addition to spouse-related exposures. Exposure to cigarette smoking from other household members, on the job, and in other activities of adult life, we call that "social exposure" -- exposure of two or more hours a week from sources other than occupational or household -- were associated with an overall 40 to 60 percent significant elevation in risk of adenocarcinoma of the lung and of all lung cancers combined. No association was found between risk of lung cancer and childhood exposures to cigars, pipes, cigarettes, or all types of tobacco combined. These exposures were limited to the first 18 years of life, and after that age exposures were'attributed to adult life. The findings of our study which included methods to evaluate recall bias, minimize classification of smokers as nonsmokers, ensure acquracy of diagnosis and classification of lung cancer, and adjust risk estimates for potential confounders, are consistent with and extend the findings of the numerous published reports that were not able to address all of these issues. The overall 30 percent increased risk associated Ln 1~ Ln N C*12E4.l.CQI2 GREf202tE'LS 2e Izo21 Y96-o.-6t q/o4 0031 644-7636

1-266 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 with ETS exposure from a smoking spouse is remarkably close to the 25 to 34 percent estimates of the evaluation of relevant studies in the 1986 report of the National Research Council. The significant positive dose response to exposure to tobacco smoke within households, in occupational and social settings during adult life, in this study strongly supports an etiologic role of ETS in lung cancer in nonsmokers. Thank you. I would be happy to answer any questions. Yes? DR. SAMET: Can you say how large your series'will be when the study is completed, and also just something about the timing of the interim report that you've presented. How did you decide'when you were ready to present the totals? DR. FONTHAM: The number of cases in the interim report was.420. The number of cases in our final report will be 650 and I forget the last digit; 53, 56. The study was originally intended to be a three-year study and we had a very large case series at the end of three years. Because we had such a large•series, we thought -- and because it is an important public health issue -- it was important for us to publish what we had rather than wait for 4Am.EZEca.n IZRe#ostz4s. 17e tioz/ zq6_ozb, N,_4 t7o3J 644-7636

1-267 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 two additional years hoping we had enough squamous and small- cell carcinomas to look at, which is essentially what the additional two years will achieve. DR. SAMET: What percentage of potentially eligible cases were actually considered cases after you stepped through your procedure for identifying whether they had any personal smoking history in the past? You gave a number; I just couldn't remember it. DR. FONTHAM: I'm sorry; 4 percent. Yes. I thought you meant -- DR. SAMET: Of all the potential cases after you went through your -- 4 percent actually had no personal smoking exposure after you went through all of those potential sources? DR. FONTHAM: Right. Which was quite different from the proportion of never-smokers that we used in our sample size calculations at the time we wrote the grant. Three of the study ceuters of the five had done previous studies of lung cancer fairly recently. We'd done one from '79 to '82; Anna Wu had done one in Los Angeles; Pat Butler in Houston. The range in those studies was 8 percent, 10 percent, 12 percent. We were at 10 percent. We used that AnEtl.can eRzhoa.tEu 25e tzoz/ 2q6-ozbi W t703) 644-7636

1-268 1 2 3 4 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 middle figure and we -- DR. STOLWIJK: I can confirm that. In Connecticut we are finding 5 percent. DR. FONTHAM: It's not just L.A. or just San Francisco; it's very consistent for all five areas. PARTICIPANT: What are your results just for squamous cell cancer over lung cancers other than adenocarcinoma? DR. FONTHAM: Flat. It was a very small group. PARTICIPANT: But for active smoking, those cancers that are implicated, and adenocarcinoma is not in -- DR. FONTHAM: Oh, adenocarcinoma certainly is. All sorts are. Any other questions? Thank you. DR. KABAT: I have a quick question. I was curious as to whether you looked by duration in childhood and adolescence to see whether there was any confirmation at all of the Janerick findings, in that subgroup that was most heavily exposed. DR. FONTHAM: We found when we looked ever/never for mother/father or other childhood exposures we saw nothing. When we looked at duration we likewise did not see an increase at the high end exposure. But in our study, unlike his -- in the Janerick report the only elevations in 0 J dgrnez~ca.n ~e~SO~.tEu 2~e t2o2j Y96-026, q~04 [703/ 644-763b

1-269 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 childhood were after 25 smoker years. Seventy-five percent of the cases in our study were 17 years of age and older. Few if any of them had mothers who smoked. It was quite uncommon. We consider childhood to be from birth to 18 years of age. In order to get 25 smoker years in our population you'd have to have a father smoking for 18 years and a couple of series for other smokers. So, no, we did not. Thank you. MR. FLAAK: The next speaker is Ronald Davis. Let's see if this thing will work. PRESENTATION BY DR. RONALD DAVIS, COALITION ON SMOKING OR HEALTH DR. DAVIS: Good evening, I'm Ronald Davis. I'm the Chief Medical Officer of the Michigan Department of Public Health and I'm a former Director of the Centers for Disease Control's Office on Smoking and Health, and I also serve as editor of a new international journal called "Tobacco Control" whigh is published quarterly by the British Medical Association. At this meeting I am representing two organizations. First, the Coalition on Smoking or Health, which is comprised of the American Cancer Society, the American Heart Association, and the American Lung 4A=4.CCQIZ eRE#OdE'L!. ZC /Z0Z/ Z96-026, Ncff /7031 644-7636

J 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-270 Association. I am also representing the Association of State and Territorial Health Officials, or ASTHO, which represents our nation's state health directors. Before I give my own prepared remarks, I wanted to mention that Dr. Doug Dogri has submitted a written statement on behalf of the Coalition as well. Dr. Dogri is with the Harvard School of Public Health and we hope that his written statement would be entered into whatever official record there is of this meeting, and I would just like to read the conclusion in his statement. I quote, "The EPA has provided an objective and complete review of the evidence of the association of environmental tobacco smoke with lung cancer. To me the evidence of a causal association is overwhelming, in fact far more compelling than for any other environmental carcinogen. Exposure to this carcinogen is probably the most common environmental exposure. It is time to state the obvious, that environmental tobacco smoke is a cause of lung cancer and other disorders." At the outset, I would like to make clear that the Coalition's three organizations as well as ASTHO endorse the conclusions of this draft EPA document. This treatise represents the most thorough and up-to-date review of the C4inEti.ean 1::Re#otEet& 2Ne tzoz1 Y96-o26, No4 /7o3f 644-7636 Ln ~ Ln ko N

1-271 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 evidence on the respiratory health effects of passive smoking. The evidence shows definitively that ETS is a known human carcinogen and a cause of many serious diseases and symptoms in children and in adults. We commend the Science Advisory Board, this Committee, and EPA staff, for its careful consideration of this critically important public health issue. While the Committee so far today has received a heavy dose of technical discussion and presentations, at this late hour what I'd like to do is to provide a brief chronological review of the evidence of the health effects of passive smoking which helps put this report into a bit of perspective. That chronology reveals that over time passive smoking has been indicted as a cause of a growing number of diseases and conditions in a growing number of populations based on a growing number of scientific studies for each population and disease. It's interesting to note that the health effects of passive smoking were reviewed twenty years ago in the 1972 Surgeon General's report on smoking and health. In fact, that report devoted an entire chapter to "Public Exposure to c09rnEZi.ca.n Z~?ehottz4,s. Ze /ZO21 Z96-o261 q'dq t7031 644-7i636

1-272 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 .17 18 19 20 21 22 23 Air Pollution from Tobacco Smoke." That report concluded in part that one, an atmosphere contaminated with tobacco smoke can contribute to the discomfort of many individuals, and number two, exposure to carbon monoxide in tobacco smoke "may on occasion, depending upon the length of exposure, be sufficient to be harmful to the health of an unexposed person. This would be particularly significant for people who are already suffering from chronic bronchopulmonary disease and coronary heart disease." So as far back as two decades ago we had evidence to conclude that passive smoking is a threat to health. Of course the evidence has advanced significantly since then. The health effects of passive smoking were again reviewed in the 1982 Surgeon General's report on smoking and health, which was devoted to the topic of smoking and cancer. At that time, only three epidemiologic studies of passive smoking and lung cancer had been published. The report concluded, "Although the currently available evidence is not sufficient to conclude that passive or involuntary smoking causes lung cancer in nonsmokers, the evidence does raise concern about a possible serious public health problem." As this Committee is aware, and as many people have d¢rnetf,can eRa#os.tets 2~e 12021 296-0261 qU64 003/ b44-T636

1-273 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 already pointed out, in 1986 the Surgeon General released a report devoted entirely to the health effects of passive smoking. By that time, the number of epidemiologic studies on passive smoking and lung cancer had grown to 13; 11 of which showed a positive correlation. Along with other lines of evidence, those epidemiologic findings allowed the Surgeon General and the U.S. Department of Health and Human Services to conclude that passive smoking is a cause of lung cancer in healthy nonsmokers. That report also addressed the health effects of passive smoking for children. That same year, as has been mentioned, the National Academy of Sciences issued a report that came to similar conclusions. A third report issued in 1986 by the prestigious International Agency for Research on Cancer, or IARC in Lyon, a branch of the World Health Organization, concluded that, "Knowledge of the nature of sidestream and mainstream smoke, of the materials absorbed during passive smoking, and of the quantitative relationships between dose and effect that are commonly observed from exposure to carcinogens, leads to the conclusion that passive smoking gives rise to some risk of cancer." 23 1 So three key reports came out in 1986 on passive Ln cn W cAnetEcan cz~?z#ottau ZC /zozf zq6-oz6, q'04 1703166q4-7636

1-274 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 smoking, coming to the same basic conclusions. In June 1989, the EPA issued a fact sheet on environmental tobacco smoke. In the first sentence of that fact sheet, it states that ETS "is one of the most widespread and harmful indoor air pollutants." The fact sheet also states that ETS "is a known cause of lung cancer and respiratory symptoms, and has been linked to heart disease." So some people may not be aware that EPA already has policy stating that ETS is a cause of lung cancer. This report obviously looks at the evidence in more detail and looks at it quantitatively, but the EPA has taken a stance on this issue as far back as 1989. A year later, as everyone here knows, the EPA released for public comment the first draft of the report that you're now considering. By that time, the number of epidemiologic studies on lung cancer and passive smoking had grown to 24; 18 of which showed an increased risk of lung cancer in those who are chronically exposed to ETS in the home. In June 1991, as has been mentioned, again, NIOSH published "Current Intelligence Bulletin 54, Environmental Tobacco Smoke in the Workplace: Lung Cancer and Other Health Effects." NIOSH reviewed the previously published studies PTrilEYI.CQ.I2 GRE#o'LfE24. '1.~e tzvzf 296-oz6i wGq 17031 64a-7636

1-275 ) 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 and reports and concluded using OSHA legal terminology that ETS is a "potential occupational carcinogen." By the time the current draft EPA report was completed, the number of epidemiologic studies on passive smoking and lung cancer had grown to 30; 24 of which showed a positive correlation. Moreover, as you all know, the report reviews the evidence now available linking passive smoking more definitively to conditions such as asthma and chronic middle ear disease for which data were felt to be inconclusive when the 1986 reports on passive smoking were issued. I've not even mentioned the accumulating evidence that shows an association between heart disease and passive smoking. Heart disease is not considered in the EPA report, so I will not address it except to mention two points. One, that nine epidemiologic studies have been published on heart disease and passive smoking, most of which show a positive correlation between the two. Two, risk assessments by Wells, Glantz, and Parmley, and Steenland indicate that passive smoking may cause anywhere from 25,000 to 35,000 deaths from heart disease each year in the U.S. That toll is 10 times that for lung cancer attributed to ETS in the current draft EPA report. So we d4m.Eti.can 1:Re#otEets Ze rzoz1 zq6-nz6i q~o4 1703/ 644-763b

1-276 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 hope that EPA will follow its risk assessment on ETS and cancer and other respiratory conditions with a risk assessment on ETS and heart disease. In summary, the evidence on the health hazards of passive smoking has grown steadily and substantially through the years. The number of epidemiologic studies on passive smoking and lung cancer has grown from 3 reviewed in the 1982 Surgeon General's report, to 13 reviewed in the 1986 Surgeon General's report, to 24 reviewed in the first draft of the EPA risk assessment, to 30 reviewed in the current draft EPA report. There is a striking parallel between the growing evidence on the dangers of passive smoking and the growth of evidence in the 1940s, '50s, and '60s on the hazards of active smoking. Every major health and medical organization that haE reviewed the evidence has agreed with the conclusion that passive smoking is a serious threat to health. With few exceptions, the only challenge to that conclusion comes from the tobacco industry and its consultants. This is the same tobacco industry, as Professor Banzhaf pointed out, that continues to deny that smoking causes lung cancer in smokers. The industry probably even denies that smoking causes so much as a cough. ogmcti.ca.n 1::R1-e#o%tZu Ze [zo2i z96-oz6, N6q 0031644-7636

1-277 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Given the tobacco industry's continued attempts to deny the undeniable evidence linking passive smoking to death and disease, it's instructive to quote from the preface to the 1979 Surgeon General's report on smoking and health. Then-Secretary of Health, Education, and Welfare, Joseph Califano Jr., wrote, "In truth, the attack upon the scientific and medical evidence about smoking is little more than an attack upon the science itself; an attack upon the epidemiological, clinical, and experimental research disciplines upon which these conclusions are based. Like every attack upon science by vested interests, from Aristotle's day to Galileo's to our own, these attacks collapse of their own weight." That comment is as accurate today about the industry's statements regarding passive smoking as it was in 1979 about the industry's statements about active smoking. I conclude by expressing our hope that EPA approve this report with dispatch. The evidence is already in. Formalizing this document will expedite the passage of appropriate public policy throughout the country by states, by local jurisdictions, hopefully by OSHA, and by private employers, which we hope will protect, finally and comprehensively, nonsmokers from this serious health threat. N %o Ln 1-4 Ln Ln o~{m.a~ican ~e~sott~u Q1 l~e j202l 296-026, q/04 /7031 6aa-T636

1-278 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Thank you. MR. FLAAK: The next speaker is Dr. Clark Heath. PRESENTATION BY DR. CLARK HEATH, COALITION ON SMOKING OR HEALTH DR. HEATH: I'm Clark Heath. I'm the Vice President for Epidemiology and Statistics at the American Cancer Society. I'm speaking this afternoon on behalf of the Society as well as on behalf of the Coalition on Smoking or Health. I submitted a brief written statement, which I believe you have. I would just like to make four general epidemiology-oriented comments from my review of the massive draft document, or at least those parts that have to do with epidemiology, particularly of lung cancer. My first thought is an overall impression of the document, which is that it's extraordinarily thorough, that in a scientific way it's well written although I would agree with the comments about its lay interpretability, which would be well to revise in terms of the introductory parts, the summary parts. I think it's entirely reasonable in its conclusions. As Ron Davis has just recounted, this is not a new issue. It's been growing over the past decade or longer in terms of the accrual of epidemiologic data, as well as data (n AnEZi.can ~e~osEet~ Ze 12021 296_C26, qoq t7031 644-7636

1-279 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 from other sources. The conclusion that environmental tobacco smoke is a lung carcinogen has already been clearly stated several years ago now by the National Research Council and Public Health Service in this country. Those conclusions, as Ron indicated, were based on 13 studies, and here we have a review now of 30 epidemiologic studies to date. The most impressive one I must say to date is the one reviewed in detail by Dr. Fontham. The evidence that environmental tobacco smoke causes lung cancer is not an epidemiological one alone, however. It comes from a mosaic of information, not the least of which is the background information with respect to active smoking itself, the nature of tobacco smoke and its constituents, whatever forms it takes, and finally, laboratory studies and these more recent epidemiologic studies on environmental tobacco smoke itself. Which leads me to my second general comment, which is to remark upon the,appropriate inclusion in the EPA review draft of the seven so-called recommendations of Sir Bradford Hill, which were enunciated in the '50s, with respect to the kinds of thought process that people in policy-making situations and scientists, epidemiologists in particular, need to go through in considering what are the causes of a PjIYLE'LLCQ/2 GRE#0'CfE'LS. 2~C 12021 296-02f)1 W 0031644-7636 H Lq to o,

1-280 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 chronic disease, not necessarily cancer, but of course his thinking was derived from thoughts about cancer and active smoking at that time. All seven of the recommendations that are listed in the review draft I think are nicely fulfilled, or adequately met, with the data that has been reviewed in the draft. I would particularly emphasize the three recommendations that have to do with the interdigitation of epidemiologic data with data from other sources and the need for multiple epidemiologic data for an issue like this. The first is the consistency of information of cross studies, and here I think this review of the epidemiologic data shows a strong consistency among the studies. It's not perfect of course. One would hardly expect that. Epidemiologic studies come in all shapes and sizes, and it is very important in judging our consistency to weight them according to their value. Without goipg into how that's done, although I think the draft outlines that quite adequately, it is necessary to create tiers of studies or a hierarchy in terms of the better, more comprehensive, extensive, larger studies, going on down the tiers. I think that the judgment of the studies in the draft according to their tiers is well drawn. 4Anazi.can GRr-#o%tZU 17e /zozf z96-oz6i q~vq 0031644-7636

1-281 ) 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 The second of Bradford Hill's general criteria had to do with biological plausibility and there's certainly no doubt that this is a situation where a potential carcinogen is clearly biologically plausible. The third is its coherence with existing knowledge. It fits in well with what we know from many sources and have developed from many lines of scientific work over the years and it is utterly coherent. This is perhaps the most important element in drawing causal conclusions and making policy thereupon of all of the Bradford Hill recommendations if I may say so. My third general epidemiologic point has to do with how the report handles the issue of comparability. I think for any epidemiologic study the principal concern is comparability. Epidemiologists compare exposed people to unexposed people, or people with disease to people without, and if these groups in these comparisons are not comparable -- ideally in every way other than what you are comparing them for differences in -- then the study will go awry. So it is very important to look at and consider clearly the ideas of confounding and bias, and of course the question that has been perhaps beaten heavily in this forum, misclassification; not just misclassification with respect to O7n2E4f.eQl2 `ZRO0R.tE'LS. ZC /zoz/ 2q6-oz6i qUd{ /703/ 644-7636 i

1-282 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 smoking, but misclassification with respect to diagnosis if you wish to pinpoint the type of cancer you're talking about. For the general conclusion about cancer however, misclassification from one cancer type to another is a relatively optional point. The last of my general points would have to do from an epidemiologic point with how this draft deals with the question of population risk, attributable risk. I think it does it in a sensible manner, clear at least in the way it is presented scientifically, and to the extent that I understand statistics, being more of an epidemiologist than a statistician. But the elements that go into making -- the assumptions that go into making -- a population estimate of how many cases, what is the burden of human cancer that environmental tobacco smoke produces, I think is important to lay out clearly and I think the draft does that. And I do believe that even the most conservative, careful estimates of this attributable burden for human lung cancer for environmental tobacco smoke will prove to be, as someone pointed out earlier this afternoon, a substantial number of cases, particularly when stacked up against other environmental carcinogens for which Class A status has been QTI'rtE%LCQ.Ii lZRFO0'LfEtS 2~e 12C2i 296-o26, q6q /703/ 644-7636

1-283 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 granted and for which we have societal safeguards. Those are the four comments I care to make; the general idea of the overall impression with respect to this report, the reference to Sir Bradford Hill's recommendations, the way the report handles comparability, and finally its dealing with population risk or attributable risk. Again, I think it's a thorough report and I would consider it entirely reasonable in its conclusions. Thank you. MR. FLAAK: Your last speaker for the evening is Dr. Jeff Wagener. We will defer the other two speakers to tomorrow morning, and we'll start a little earlier tomorrow. PRESENTATION BY DR. JEFF WAGENER, COALITION ON SMOKING OR HEALTH DR. WAGENER: Thank you. As the last speaker, I'll try to keep this relatively brief. I appreciate all the members of the Committee staying here long enough at least to hear this. My name is Jeff Wagener and I'm an Associate Professor of Pediatrics at the University of Colorado. As a pediatric pulmonologist I want to speak on two facts. One is, as an academician, I am impressed with the data as presented. I was asked by the Coalition on Smoking or Health to come and look at the data and to discuss this before you today. Ln N CJT W ogm.ezican 1V1?a#ottzts Ze [2021 296-OZ6i W /7og1 6qq-7636

1-284 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 As a pediatrician I would like to make one major criticism. That is that when you read the title of this abstract or this draft copy what you find is that they refer to lung cancer and other disorders. I would point out that the pediatric population of the world is fairly large and to be considered an "other disorder" may not be appropriate. The one suggestion that was made to separate this into adult and pediatric or childhood lung problems I think is very appropriate and may increase the validity of the study. Nearly 10 million children in the United States are exposed to cigarette smoke in the household, so we are talking about an extremely large exposure rate. I would like to discuss just the Section 7 as it pertains to these 10 million children. It's also important to recognize that 25 percent of pregnant women continue to smoke through pregnancy. So when we discuss prenatal effects of cigarette smoke, again we are referring to a very large population. Involuntary,exposure to environmental tobacco smoke affects children in three different ways. There is a prenatal effect of exposure which has a marked impact on total body weight, lung growth, and brain growth. There is secondly a postnatal exposure which may have significant impacts on the incidence of respiratory illnesses, may affect Ln OTIrLE'Ll CCi/2 GRE/20'itE'L1 ZC /zoz/ z96-oY6, q'04 003/ 644-7b36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-285 ears, brain, and finally may have a direct effect on survival as is demonstrated in some correlation with SIDS. Finally, and I think the third aspect that I would like to point out that is not covered very well in this draft -- and possibly most important -- is that childhood exposure to ETS impacts on the child's decision to engage in active cigarette smoking as an adult. That in its own right may be one of the largest causes for lung disease related to ETS in childhood. Now, there are both nonrespiratory and respiratory effects from environmental smoke The nonres irator effect . p y that I think is discussed to some degree in this topic would be SIDS. We don't know the etiology of SIDS. We're not sure whether or not it is going to be related to the respiratory system. But I'would still point out that it should be included in this document. There are a couple of reasons why. One is it's not included in other doc~ments. Everybody avoids the relationship that's been shown between ETS and SIDS, and it needs to come out someplace. The second is that there are plausible reasons such as prenatal exposure affecting control of ventilation of the fetus that may in some way eventually prove to show a OTIrLELCCQ.12 fRE/2OtfE'LS Ze t2021 296-0261 q~vq 1703j 644-7636

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1-286 relationship, a direct cause and effect relationship, between ETS and SIDS. On the respiratory effects the chapter does an excellent job of reviewing various articles. Unlike some studies where they simply take all the articles published and weigh which one comes out on one side and which one comes out on the other side, not looking at the validity of individual studies, I feel this chapter does an excellent job of reviewing individual strengths and weaknesses within individual studies. However at the conclusion, as was mentioned by some members of the Scientific Advisory Board, one is not sure which of the major impacts is of greatest importance. While Chapter 8 begins to address this, I would suggest that the impacts be separated in two different fashions. One is that there are a group of effects from ETS which are strongly supported by scientific data currently available. Those include abnormal lung development following prenatal exposure, increased nonspecific airway hyper- responsiveness in children raised in a smoking environment, an increased frequency of lower respiratory infections in infants living with caregivers who smoke, an increased risk of severe asthma and frequent acute asthma episodes in 'VqmEa.i.can 'ZIRFOottets l.~e t202l 296-o261 qlc4 /7031 644-7i636

1-287 ) 1 2 3 4 5 6 7 8 9 10 11 12 1 13 14 15 16 17 18 19 20 21 22 23 children with asthma who live in a household where someone smokes, and lower lung function throughout childhood in children of smoking families compared to nonsmoking families. There is very little data to refute these claims. However there is only suggestive evidence in a few other areas, specifically the higher risk of atopy in ETS exposed children still needs some increasing information to make a very strong statement. More importantly the acute ear infection and chronic ear infection data I think can be questioned because it, unlike the data related to lower respiratory infections, includes a variety of study design many of which are of questionable value. Finally the higher rate of chronic cough and phlegm has been well shown but the correlation with wheeze and ETS is not as clear. There is a higher rate reported now of diagnosed asthma, increased incidence of asthma, in children exposed to ETS, but I would point out that there is one weakness in that. If you assume that a child has increased hyperreactivity of the airways because of ETS and if you assume that a child has an increase in amount of cough based on record, the pediatrician is faced with making a diagnosis. Since we use cough and increased reactivity as the diagnosis d¢rnE-Uean '=-Refi.ottets Ze I202i Y96-OZbi No4 /7031644-7636

2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 1-288 for asthma, we're now taking an environmental toxin, its side effects on the individual, and equating that with a disease called asthma. It may be simply by removing the toxin you can remove the disease. There were a couple of comments made by the SAB or questions that came up. I would like to address just one of these. That has to do with abnormal lung development. What I would point out to you is that while, yes, there is not a clear etiology for what aspect of tobacco smoke causes abnormal lung development, what is clear is that human studies demonstrate abnormal lung development detected soon after birth in children who are exposed prenatally to cigarette smoke. These findings are identical to the pathology and physiologic findings noted in animals. So as far as does cigarette smoke affect lung development, I think it is absolutely clear based on both human and animal studies showing identical results of this impact. Whether or not subsequent exposure to environmental smoke has further impact on the lung is still in question and obviously still being reviewed intensively. In conclusion what I would like to suggest on this report, beyond renaming the report at the beginning, is adding an aspect to the childhood risk and that is the risk C4 ti ) OTrl2Etl.CQ/2 GRrO0'LtE'C1 Ze tZ0Z1 Y96-OYb, <v64 1703164a-7636

1-289 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 of a child exposed to ETS becoming a smoker as they become older. There is data out there to support this. Dr. Lebowitz has published this in several places. There is a high likelihood of a child exposed to ETS becoming a smoker subsequently. That's the overall probably highest risk of developing lung disease in adulthood and causing problems. Thank you very much. Are there any questions? Everybody wants to go eat. MR. FLAAK: We will reconvene the meeting tomorrow morning in this room at.8:30 to close out the last of the public comments and then continue with the rest of the scheduled meeting at 9:00 a.m. tomorrow. So at 8:30 tomorrow morning we will start again. Thank you. (Whereupon, at 6:40 p.m., the meeting in the above- entitled matter was adjourned, to reconvene at 8:30 a.m., Wednesday, July 22, 1992.) Ln OTI/zE%LCQK GREl204.tE'LS Ze t2021 29&026i q,og /703J 644-7i636

ORIGINAL UNITED STATES ENVIRONMENTAL PROTECTION AGENCY SCIENCE ADVISORY BOARD INDOOR AIR QUALITY AND TOTAL HUMAN EXPOSURE COMMITTEE - - - - - - - - - - - - - - - - - - -X In the matter of: : . • ENVIRONMENTAL TOBACCO . VOLUME II SMOKE REVIEW PANEL : -X Wednesday, July 22, 1992 Main Ballroom Holiday Inn 4610 North Fairfax Drive Arlington, Virginia 22203 The ENVIRONMENTAL TOBACCO SMOKE REVIEW PANEL MEETING of the Indoor Air Quality and Total Human Exposure Committee of the Science Advisory Board was convened, pursuant to notice, at 8:35 a.m. APPEARANCES: MEMBERS: DR. MORTON LIPPMANN, Chairman DR. JAN A. J. STOLWIJK, Vice Chairman DR. JOAN DAISEY DR. TIMOTHY LARSON DR. VICTOR G. LATIES DR. PAUL LIOY DR. JONATHAN M. SAMET DR. JEROME J. WESOLOWSKI DR. JAMES E. WOODS, JR. CONSULTANTS: DR. DELBERT EATOUGH DR. S. KATHERINE HAMMOND DR. GEOFFREY KABAT DR. MICHAEL D. LEBOWITZ DR. HOWARD ROCKETTE DR. SCOTT T. WEISS l PT/nE2LCQ.1L GRE/20dERS 2~e t2oz1 z96-oY6, q~oq t7031 644-7636

11-2 APPEARANCES: (Cont.) SCIENCE ADVISORY BOARD STAFF: MR. A. ROBERT FLAAK MS. CAROLYN OSBORNE EPA STAFF: DR. FARLAND DR. KOPPIKAR DR. STEVEN BAYARD DR. JENNIFER JINOT OTHER APPEARANCES: DR. FERNANDO MARTINEZ DR. KENNETH BROWN DR. JUDSON WELLS DR. LEADERER FROM THE PUBLIC: DR. GARY FLAMM DR. MAURICE LeVOIS PTmE'LI.CQl2 GRE#A'LtE'CS. '11 C 1202) 2q6-o26, w6q 17o316aa-7636

11-3 , C O N T EN T S PAGE Public Comment Period (Continuation) Presentation By: Dr. Maurice LeVois, Tobacco Institute 5 Dr. Gary Flamm, Tobacco Institute 23 Committee Review of Chapters 3, 6 and 8 Chapter 3 - Estimation of Environmental Tobacco Smoke Exposure 40 Chapter 6 - Population Risk of Lung Cancer From Passive Smoking 88 Chapter 8 - Assessment of Increased Risk for Respiratory Illnesses in Children From Environmental Tobacco Smoke 117 Committee Review of Chapters 1 and 2 153 Other Issues/Future Business 167 C091y1EtCCQ/2 GRF~2ot~CZs ZC /ZOZf z96-026, N6q /7031 644-7636 N ko Ln Ln 01 v

11-4 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 P$ Q CZERIHG Sz (8:35 a.m.) MR. FLAAK: Good morning, ladies and gentlemen. Welcome back to the second day of our meeting. For those of you who don't know me, I'm Bob Flaak. I'm the Designated Federal Official for this Committee. For the benefit of everyone here, including the Committee members, the EPA has gotten around to installing some microphones for us this morning, and they seem to work pretty well. Yes, I know, they're kind of loud. So hopefully everyone can hear everything that's said for a change. It's also a little cooler in here this morning, and I think we'll probably try to keep it that way. It kind of keeps things lively. This morning we have some carryover business from yesterday. We have two additional speakers to finish out our speaking part of the iueeting yesterday, and they are Dr. Gary Flamm and Dr. Maurice LeVois. Gary, are you -- there you are. Maurice? I'd like you to go right now. We have about 30 minutes or so. I'll V give you both 15 minutes. Do you need some overheacl? DR. LeVOIS: No, I don't. cn ~ Ln ko ) PT/)2ELl.CQ.lZ GRE#O'CtE41 Ze tZ0ZI Y96-oZ6, w6q /703l 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-5 MR. FLAAK: You're all set. PUBLIC COMMENT PERIOD (Continuation) PRESENTATION BY DR. MAURICE LeVOIS, TOBACCO INSTITUTE DR. LeVOIS: I'm Maurice LeVois, and I'm here at the request of the Tobacco Institute to comment specifically on some of the epidemiologic issues. I would like to thank the Committee for taking time and giving me this opportunity to speak. I certainly hope that some of the things that I say will be of use to you. I'll be brief. I would like to start out by considering where we have been. The first EPA draft presented primarily G, statistical argument based on epidemiologic data, and buttressed by analogy to active smoking. The Committee has properly, I think, asked them to go back and present a much more thorough analysis of the epidemiologic data upon which the statistical argument was based because it's really incorrect to think that you can use statistics alone to interpret epidemiologic data. I believe that the EPA has done a credible job of adding to the base that they laid down with their original report. It was deficient before, and I think that in many ways they have tried credibly to address the deficiencies. The new draft has a lengthy section on (A ) P~'Lf.CLt/2 GREoo4.tEYS. `LlC /2021 296-o26, NGq t703l 644-7636 %o N

11-6 J 1 2 3 4 5 6 7. 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 epidemiologic interpretation. I think that there are still some problems with that interpretation, and it relies, I think this time, too heavily on analogy. I'd like to speak very briefly to the issue of analogy because I think that it goes beyond simply the correction which I think I understood yesterday from the Committee to the staff which was that you can't really push the analogy so far as to say it constitutes sufficient human evidence. The analogy to active smoking is certainly an important consideration because it gets to the question of biological plausibility which is certainly relevant in the interpretation of the epidemiologic data. So I would just point out that there are shortcomings to the analogy. It's not an either/or sort of thing with biological plausibility. It's a matter of degrees. And there are limits to how far you can go with the analogy. Within the report the EPA tries to say up front that the analogy is sufficient, but then later it says that it's improper to use dosimetric modeling based on cigarette equivalents as another method of predicting the level of risk that you might expect. I submit that you can't have it both ways. If the Aneti.can --RE#o3tcts ZC tZOY] 296-0261 4~04 t7031 644-7636 51592 6474

11-7 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 analogy stands, then the dosimetric predictions have to be seriously considered. If the analogy is not that good, as I think that you agree that it is not, then it's important to talk frankly in the draft about how far it goes and what its shortcomings are. I would submit a few of the shortcomings are as follows: Regardless of what constituent you would model on -- and we must admit up front that we don't know what constituent to model on -- one of the problems that I have with the epidemiologic level of risk is that most people were surprised, and I think should continue to be surprised, that it is much higher, as you would suspect, based on the differences between the magnitudes of exposure. Whether you use cotinine, which I'm not saying you should use, or some other material, the predictions that you get from dosimetric low dose extrapolation are orders of magnitude lower. Now, that's not saying that that's where the real risk is; but.that's saying that based on what we know, the risk is surprisingly high based on dosimetry. Another problem with the analogy, another point where it breaks down, is that you don't have an ETS exposure, the heat, nor the astringent characteristics, nor the destruction of the primary defense mechanisms of the lung o09inESiican C=Re#ottEts. Ze tZO21 Y96-026, W11q 17031644-7636

) 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 _17 18 19 20 21 22 23 11-8 that you have in mainstream smoking. Given current models of carcinogenesis that rely much more heavily on self-proliferation, it's difficult not to wonder how this effect that we're seeing, that is larger than you would expect on the basis of analogy, is manifested despite these fairly important differences. One alternative explanation that I think has to be seriously considered, and I think that IARC seriously considered the alternative explanation, is that the epidemiologic studies themselves are flawed. You do have problems with both an imprecise measure of exposure that is basically spousal smoking and the problems of concordance, not only with cigarette smoking activity, but also with diet and the household environment, the occupational exposures, and so forth. It's not sufficient in my mind to rule out the possibilities of diet or other types of confounders based on a narrow view of the relevant literature. Clearly the most important literature is the type of literature that was presented yesterday by Dr. Fontham. When you have studies the specifically address the question, that's an important consideration. I should point out that while other studies have og=ti.can V,?z#ostzts ZL' /zoz f 2q6-oz6i q'04 /703] 644-7636

II-9 ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 not effectively, as Dr. Fontham has, shown that there is not a confounding, they have shown that there is ample opportunity for a confounding. There are unquestionably correlations between active smoking and spousal smoking, and diet, as shown in two or three papers, and a variety of other health-related factors. So to give short-shrift to the possibility that when you're dealing with a weak epidemiologic association and you're dealing with a design that includes a measure of exposure that is prone to bias and confounding, that you have adequately taken care of those, and the small studies in particular that we have seen that are particularly in observational designs prone to the effects of bias and confounding -- because you don't have enough numbers to control for them; you can't stratify; you can't use even very good measures because you just don't have enough data -- I would guess that the effects of misclassification are not adequately represented by one percent error. This is based on my own experience in conducting epidemiologic surveys and in personal discussions with some of the authors that have done ETS studies. I would agree with what Dr. Kabat said yesterday, which is that one percent might be considered a lower limit. dgrne-ti,ca.n -RE#otLeas 2~C /zozJ zq6-oz6, q~~q /7031 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 II-10 I don't know what the number is, but I think that, again, it's important not to limit your consideration of the relevant evidence to this very rarified set of ETS studies in which there are not often the measures that you would like to see. If you were conducting an epidemiologic survey or an epidemiologic study of your own, I think that you should -- and most people would -- consider all of the evidence when they decide what measures they're going to include and what measures they're not going to. The same standard applies here. You don't have to have a study that convincingly shows that something is a confounder. If you do, you should definitely include it. But if you have studies that suggest that there is a correlation between the variables that you're concerned about and the outcome that you're concerned about independently, then there is reason to suspect. And all it takes is, I think, a reasonable database -- and there is such a reasonable database -- to conclude that there is some problem that should continue to attract serious research attention. I think that the presentation by Dr. Fontham addressed many of my concerns about that paper. Initially. Ln LD ) C741'r2E'LLCQ/2 GRE#O'LtE'iS 1~, e I202i 29&026, ~a4 /7o3l 644-?636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 the paper was presented in such a way that it appeared to come out prematurely. Nowhere in the paper was it explained that this was initially a shorter protocol that had been extended. So it appeared that it had come out early. We were told that it collected information on confounders, and we weren't told how the confounders behaved. I think that the problem of misclassification a~d her paper is not fully taken care of by the use of the information that she was able to obtain. And some people spoke to this yesterday. For one thing, cotinine is going to detect deceivers once they're hospitalized in very high numbers. Most people, once they're hospitalized, have enough disincentives for smoking that they will have probably quit relatively recently. And dietary cotinine, if people are deceivers, you can't rely on that to reveal what their true status is. I also have.a feeling that people at least reasonably -- and this may not apply at all to the study in question -- but I think that as a practical matter, studies that take place from this point on have to consider that there are not only social incentives to deceive, but there are also financial considerations. PTN2E'LICQl2 GRE`2otEE4s. Ze t202j 296-0261 W /7031 644-7b36 w ti b cn v b

11-12 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 The status that you declare has an impact now on your insurance rates. And,although some insurance companies take cotinine measures, I think that there's probably a database that would be very interesting to explore, finding out a little bit more about what these measures are showing in general populations concerning deceivers. It's not a convincing argument by any means, but there are reasons to believe that there is still a problem with misclassification. Even in Dr. Fontham's study, if you are a deceiver, the chances are that the various sources of information that she collected on misclassification are all really information that comes from the same source. They're not independent sources of information. And if a person has denied smoking, it's not necessarily likely that they would have told their physician or the researchers. So, again, I have a feeling that this question, while it is minimized in the existing report, should continue to bear some serious consideration and some serious research because I think that the fact that we're dealing with a very low association, an association in this country below 20 percent excess, has to give pause to anyone in terms of the strength of the conclusions that are drawn from those data. I would like to make one last comment, and that cjqrnE z,i.ca.n zi Efio ttE u. Ze t202l 296_o261 a,:.z4 /7031644-7636 N W kn ti Ln

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-13 pertains to my own personal view about the use of epidemiologic data in meta-analysis and whether or not that's really the answer that EPA is looking for in terms of a better, more reliable method. I think that it can be, but I think that it's prone to some serious problems. The very conditions that make epidemiologic studies desirable candidates for meta-analysis -- that is, their small size and their lack of power individually -- undermines your confidence in the basic underlying assumption of meta- analysis, which is that the component studies provide an unbiased estimate of some true effect. You have reason to accept that assumption when you're dealing with randomized control trials. Even very small randomized control trials that adhere to a protocol have met that assumption, whereas small epidemiologic studies, as I said, are very difficult to rule out underlying sources of bias and confounding, and you can't rely on the assumption that they're unbiased. Pooling both studies then, may or may not give you a better point estimate. I think that the confidence interval around those point estimates are meaningless. I don't think that there's any basis, therefore, for using that type of confidence interval as a test of causation or a d9mzti,ccsn I=Re#ottZu ZC tYa21 z96-oZe, 4~vq 17031 644-7636

11-14 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 bracket for risk extrapolation when you're dealing with a collection of small epidemiologic data. And I particularly believe that if EPA continues to pursue, as I think they will and will have to and should, epidemiologic data in this and future risk analyses, that they recognize that it's not a simple matter. They have been told that by the Committee in its response to the first draft, and I think they need to continue to be told that by the Committee because there are not very many epidemiologists at EPA. And I.don't get a sense from reading either draft that they fully appreciate what it is to conduct a balanced criticism of epidemiologic research as a basis for risk analysis. Thank you very much. DR. LIPPMANN: Thank you. I appreciate your careful, reasoned analysis. Could you stay there a minute? DR. LeVOIS:. Sure. DR. LIPPMANN: I think we'd like to engage a little dialogue on some of these points that you've made. You've included the issue of coherence of the results, and let's explore that a little more. We start by coming to a judgment, I think, on this P/If2E'Ll.CQ/2 v?E#o'LtEYs ZC /zoz/ z96-oz6r q'04 /703/ 644-7636

11-15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Committee largely on weight of evidence. Clearly, as you indicate, any single piece of evidence has its real limitations, and our obligation is to come to a balanced judgment on the evidence as a whole. I think the analogy argument starts with the established carcinogenicity of mainstream smoke. I don't know if you agree on that. Do you? DR. LeVOIS: You know, I'm actually amazed that the Chair would succumb to the pressure of the lawyer who got up here yesterday and demagogued. DR. LIPPMANN: If you don't want to answer the question, let's proceed to the next speaker. DR. LeVOIS: I'll tell you, my answer to the question is very much as Gio Gori's yesterday. My interpretation of all of the evidence is that it's a weak initiator and acts primarily through promotion. DR. LIPPMANN: Thank you. Okay. Let's move on then to the next point I wanted to address because I'd like to walk through some of this evidence with you and make sure that we're talking about the same scientific basis for our decision-making process. Dr. Gori did spend a lot of time talking about the physical chemical differences and their impact on the o*nE %i.ca.n V'?0o tfe ts ZC /zo2/ 296-oz6i q'cq /703/ 644-7636 V? W ti

) 1 2 3 4 5 6 7. 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-16 dosimetry of mainstream versus ETS. It strikes me that there's a coherence in the issues that trouble you from the differences in the physical/chemical properties of mainstream and ETS. Mainstream smoke, as he indicated, is a highly concentrated smoke in which you have mass effects and in fact it's well established that you get very high deposition efficiency in the lung, as he said, moreover, very highly efficient deposition in the large airways, very little penetration to the small airways. He also indicated that the ETS is in much more dilute, small particle aerosol. And he didn't mention, but it's also well established, that nicotine is essentially aerosol in mainstream smoke in fairly high deposition in those large airways, whereas nicotine is a vapor when inhaled as a component of ETS. So that helps explain, in my understanding of the physics, the chemistry and the toxicology and dosimetry of why adenocarcinoma is associated with ETS and why squamous cell carcinoma is much more associated with mainstream smoke. It also helps explain why cigarette equivalents are less likely to be appropriate, as you suggest, in that you'll get much more nicotine deposition with mainstream smoke than o7rneti.ccsn GRz#oa.tzzs !6L' tzc2f z96-oxbi W /7031 644-7b36 (n N as 4 OD AN

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-17 with sidestream smoke, in which it's inhaled as a less than perfectly soluble vapor. And so I don't see any problem with your conclusion that you can't use the cigarette smoke equivalent basis, or with EPA's conclusion, because of these differences. Where in your view, since you've considered these things very carefully -- and I admired your balanced presentation of the facts -- where we should be persuaded that these differences which are all coherent with the observed facts need to be disregarded? DR. LeVOIS: Well, I think that part of what you said is speculative. I don't really see any evidence that would suggest that the materials of concern are reaching more distal or more deeply into the lung, and there's a reason to suspect that they might not. I presume that you're talking about some active gas phase constituent, and I would suggest that people who breathe nasally and bzeathe lightly don't breathe deeply enough to have the penetration of those gas phase constituents that a person that drags on a cigarette very heavily would. So I don't think it necessarily follows that what we're seeing is all that consistent, as you've just c4ine,aza.n z~?Eliottass Ze tZOZl Pq6-0P6, W /70316444-7636

11-18 ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 suggested. I think that we don't know, but I think that much of what you've just gone through is speculation. DR. LIPPMANN: Well, that's based on most of my research career on the dosimetry and deposition of inhaled materials, so I think it's a little more than speculation, although we don't have -- DR. LeVOIS: Well, what I'm saying is that -- DR. LIPPMANN: -- I acknowledge we don't have data specifically on the uptake of nicotine and vapor in aerosol form that I'm aware of. DR. LeVOIS: Nor do we actually have in any of the ETS studies an adequate discussion of where the cancers are. Most of the information that I'm aware of is very, very sparse in the epidemiologic literature. It is, as you say, not -- well, you've implied that we're seeing almost an exclusive association with adenocarcinoma, first of all, and that's not the case. About half of the studies show an increased risk for squamous and small cell. So there is not this complete consistency that you suggest. But a further shortcoming of the data right now is that we don't know where these cancers are in the vast majority of these studies. So you are, I guess -- again, I suggest that you're speculating beyond what we know right now. (n N ,~7qm.E-ti.ca.n z~?e/sotteva. 1~C e t2021 296-0261 W t7031 644-7d36

11-19 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. LIPPMANN: Any other questions? DR. LATIES: It occurred to me while you were speaking that we may have a different kind of misclassification or deceiver coming up on board now. Because it's getting possible to sue tobacco companies and perhaps gain large amounts of money, people dying of lung cancer may start swearing that they smoked when they haven't smoked. What would that do to -- DR. LeVOIS: I have no idea. (Laughter.) DR. LATIES: I mean, would we have to start collecting for that? DR. LIPPMANN: You don't have to answer that, if you don't want to. (Laughter.) DR. KABAT: In reference to the site within the lung, the location within the lung, my understanding of it is that most adenocarcinqmas and the bulk -- whether it's in the Orient or in the Fontham study or in our study, the bulk of the cases of lung cancer occurring among nonsmokers are adenocarcinomas. My understanding of it -- and maybe someone can confirm that or disconfirm it -- is that adenocarcinoma 1=4m.EZi.ca.n z.-REliottat, 2~e tZOZ1 296-o26i q'oq /7031 644-7636 Ln ~-4 Ln ti 1%j

) 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-20 arises in the periphery of the lung. So Jon Samet spoke to the problem of reading radiologic radiographs and actually determining in case controlled studies that the glandular cells in the peripheral -- DR. WEISS: Another comment here. I mean, cell type is not immutable. There's a famous study that all epidemiologists know and pathologic slides have been submitted blind to pathologists of lung cancers, that the cell-type reading -- there's a substantial amount of observer misclassification as to cell type. They can say it's cancer, but they can't say which cell type it is. I view this as somewhat peripheral to the question that we're really addressing here today, which is the overall question of does ETS pose a cancer risk? DR. SAMET: A comment on what Geoff said. At least whenever I've been involved in looking at this evidence, and this was most recently in the Radon Dosimetry Report for the National Research Council, the issue of cell of origin, I think, of lung cancers, remains an area of research and some controversy, and the question of whether all dividing cells have the potential to become malignant and the actual histologic expression may then depend on multiple factors is, o4mttEcan cJQa#ottet& Ze tzo2l z96-oz6i wo4 hosl 644-7636

11-21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 I think, very likely. So it's probably not so simple as peripheral carcinomas or adenocarcinomas originating somehow from cells with mucin secreting potential. As Scott said, there are many examples of multiple -- expression of multiple histologic types of careful looking or transformation as cancers evolve. So I think the question is complicated, and one that has not really been approached in the epidemiologic studies, as we already mentioned. DR. LeVOIS: And I agree entirely that it's hard to predict what you would expect to see because we don't know what the different concentrations are and what the differences might be in the active constituents. But I think that it's -- I may have overinterpreted what was said earlier, but the cellular specificity is still an important question because it does help with the understanding of ideology, which is what we're trying to get at. DR. SAMET: Can you give me some examples of -- DR. WEISS: I would argue exactly the opposite to that question. I think that the answer to that is exactly the opposite from what you said. DR. SAMET: Can you give examples of links between 44m.EZi.can ::Rehoztats. Ze IzozJ zq6-o26, w6q /7031644-7636

11-22 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 specific agents and histologic type that are well described? DR. LeVOIS: I'm not a pathologist. I would defer. Probably the next speaker could do a better job of that. DR. SAMET: I mean, I think we have surprisingly few examples, which is really my point. DR. LIPPMANN: The reason I raised it was that some of the other testimony in public yesterday implied that we shouldn't believe the ETS could cause cancer because it caused a different distribution of cancer types and mainstream smoke. And so I was reacting somewhat to that testimony and trying to explore whether in fact the differences in the cell types of the lung cancers was really different and, therefore, breaks the relationship between -- DR. LeVOIS: It certainly doesn't. I don't think you can push it that far or whether it's useful information. DR. LIPPMANN: Or whether in fact, if there is a difference, it would be consistent with the expectations on dosimetry because of the different dosimetry of mainstream and ETS. Mike. DR. LEBOWITZ: Along that too, I mean, you're dealing with a complex mixture, so you should expect the different cell types to arise in different locations in the lung related to that. Certainly anything in the bronchials, ,:;4inEtica.n =~?a#ottzts ZC (zo2f z96-ozbr q~Vq (7031 644-7636

11-23 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 the alveolar region, certainly could occur because of deposition patterns regardless of the cell type. The fact that we don't know exactly what chemical produces what cell type and the fact that even with active smoking we're getting a shift towards the adenoids and so forth would indicate that we're not dealing with something very simple here that could just be determined on the basis of a cell type. DR. LIPPMANN: I think we need to move on. Is Dr. Flamm here? PRESENTATION BY DR. GARY FLAMM TOBACCO INSTITUTE DR. FLANIIM: Good morning. That was an interesting discussion. My name is Gary Flamm. I have a Ph.D. in medical biochemistry, and I've been asked by the Tobacco Institute to address general scientific and policy issues raised by this document. I appreciate very much the opportunity to make these comments this morning. The first draft of the document submitted to this Committee approximately two years ago was sent back for revision as it did not make a convincing enough case for thee claimed causal relationship between ETS and lung cancer deaths. Ln ~-A Ln W PT1rLEZGCQ,lZ GRE#A7.tCt3. ZC t2021 296-026J QU64 0031 644-7636

11-24 J 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 While the second draft differs significantly from the first, providing, as you've heard, much more detail and attention to individual studies, it too in my judgment falls short of establishing in a scientifically defensible way a causal relationship between ETS and lung cancer deaths. As acknowledged by EPA's guidelines, establishment of causality requires elimination of plausible biases which can explain an observed increase in relative risk. The principal bias examined in depth in the present draft was misclassification of smoking status. Other potentially important sources of biases were disregarded, such as recall bias, proxy reporting bias, publication bias, diagnosis and autopsy bias. Though misclassification bias was addressed, the data relied upon produced a very low estimate of misclassification. Using what we now believe is more representative data, the percent of misclassified subjects increases approximately four-fold, causing a loss of statistical significance in the relative risk of ETS exposed spouses. Similarly, use of more representative background data, instead of the apparently extreme values used by EPA, leads to lower relative risks among ETS exposed spouses, ~ ~ tA~'LCC6L/2 GRE#o4.tE'LS 1.7C /zozl 2q6-o26, q'64 17031644-760

11-25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 further reducing any statistical significance in relative risks. Indications that biases are affecting the results of the epidemiologic studies is apparent from the geographic variation in relative risk which cannot readily be explained in terms of differences in background exposure. A reasonably lengthy and substantial list of uncertainties are identified throughout the report on many subjects, but in general, and unfortunately, these were not factored into the conclusions drawn by the document. The estimated range of 2,500 to 3,300 lung cancer deaths per year is extraordinarily narrow by any measure, but is unrealistic as it ignores uncertainty and ranges in background exposure and ranges in misclassification of smoking status. Consideration of these factors, all of these factors, leads to, it seems to me, the inclusion of zero lung cancer deaths in the estimated lower bound of risk of ETS. Turning to EPA's assertion that chemical similarity between ETS and MS smoke suffices to conclude that ETS is carcinogenic to the human lung, I agree with the Committee members who expressed the belief that such information is insufficient. It is insufficient for several reasons. The AnEZi.can 1--:RE#ottEts. 2~e JZOZJ Yq6-o26y 'Va4 /703/ 6qq-7b36 Ln r Ln to N

11-26 4 10 11 12 13 14 15 16 17 18 19 20 21 22 23 components responsible for mainstream smoke's biological activity are not known. Only a small fraction of the substances in ETS has been characterized, as opposed to thousands that have been characterized in mainstream smoke, and also, the large but imprecisely determined dilution of ETS compared to MS smoke. In summary, the lack of realistic animal studies demonstrating carcinogenicity of ETS and the lack of significant increases in relative risks among studies that have been appropriately corrected for misclassification of smoking status, leads to the conclusion that there is no basis for classifying ETS as a Group A carcinogen. In regard to classification, I find a sharp contrast between EPA's classification of ETS and previously listed Group A carcinogens which all have relative risks appreciably higher than those of ETS. Further Group A classification of ETS appears to be in contrast with EPA's position on electromagnetic fields, which in general is supported by stronger evidence than ETS; at least epidemiologic evidence. Such inconsistencies where caution begins by the expert panel on the role of science in EPA in their recent report, the expert panel stressed that biases should be d*mEZi.can cRe12o%tEU. rLle /2o2f 296-oz61 q6q 17031644-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-27 avoided, as should the temptation to adjust science to fit a policy, as appears to be done in this case. Thank you. DR. LIPPMANN: Dr. Flamm, are you serious when you say that there's a more firmly-established basis for electromagnetic radiation than for ETS? Are you really saying that in public? DR. FLAMM: Well, I'm saying there are more positive epidemiologic studies, and they had -- DR. LIPPMANN: Well, there are far fewer studies. I think the point more correctly was that the calculated relative risk is higher, but that doesn't mean anything if the relative risk is meaningless, as it probably is in the electromagnetic field with such weak epidemiology as compared to the ETS epidemiology. The relative risk is not the issue. The issue is how much you can believe it and how important it is relative to the size of the population exposed. Relative risks for chemicals that are seen by hundreds to thousands of people may have less public health impact than smaller relative risks that affect the whole population. That's the problem we're really addressing here. DR. FLANIIri: I was talking about the strength of the Ln J QTI'I2E2CCQ.n GRE#o'CtE'ts. `Lle (zozf zq6-ozbi q~vq 1703/ 644-7636

11-28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 evidence as opposed to what the public health implications were. DR. LIPPMANN: Well, both you and the previous public speaker made much out of the small magnitude of the estimate of relative risks; therefore, it can't be important if it's a small number, and contrasting it with other established Class A carcinogens which have higher relative risks. So it clearly is an issue in terms of whether this deserves a Class A designation. Its public health impact is exactly what we're here to evaluate. DR. STOLWIJK: Could I ask Dr. Flamm how he would expect to see a high relative risk in an exposure with these ambiguities and when it affects as many people as it affects? How would you expect to see a high relative risk there? DR. FLANM: Well, I would not expect to see a high relative risk for several reasons. But as far as ambiguity is concerned, I think that takes us back to estimates of the Z-factor and the background factor. And I think it's clear from the studies that have been done that the spousally- exposed subjects have an exposure which is several times greater than spouses of nonsmokers. DR. HANIIMOND: I would disagree with that, and I think there are a couple of important points within that. dTnmti,ccsn cRFoottztl 2~G' /zozf zq6-oz6, qU6q t703i 644-7636

11-29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 First of all, I've seen other data that does not indicate that it's a factor of several differences between them. Secondly, I think that there have been significant changes over the last decade in ETS exposure in general. What is important is the exposures in the past, not the current exposures, so that today to look at what the distribution of, for instance, work place exposure to ETS is, is quite misleading in terms of interpreting work place exposure ten years ago. And so I think we have to be very, very cautious in using Z-factors established today to interpret prior ETS exposure which would be relevant for chronic effects such as lung cancer. DR. LIPPMANN: Can I explore something else you brought up? And since your expertise is clearly in animal toxicology, I think it's relevant. I had been struck over a long period of time with how ineffective mainstream smoke is as a respiratory tract carcinogen in rodents. The only study the EPA document could come up with produced laryngeal cancers in golden hamsters. And so certainly the absence of established carcinogenesis in animal models with ETS is by no means surprising if mainstream smoke hardly can produce it at all with the Ln ~ ~ ko c77m.E-ti.ccsrs --:RE#osfzU `L1C' /zo21 aq6-o26i wd{ /703/ 644-7636

11-30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 conventional laboratory tools. One wouldn't expect ETS to do it, and clearly, very few studies have been done, which further convinces us that it's unlikely that you would get a positive animal response. So, in a sense there's a concordance, a coherence, again between the lack of response in those few ETS studies which have been done and the very large body of hundreds of thousands of poor little rodents who have been sacrificed to mainstream smoke studies. Would you comment on that? DR. FLANIIK: Well, what you say is a fact. There are no good animal models that have yet been developed for mainstream smoke that I'm aware of. I haven't seen them if they have been developed. But I think that is not inconsistent with the statement that there isn't supporting animal data here. But I think also we would agree that we have very little understanding of what the dilution factor is of the substances that are really of interest to us. DR. LIOY: One point. We keep talking about dilution as if this is a constant mixture, and I think it was established yesterday that this may not be a constant mixture. That there's going to be changes in composition. So.I would refrain from using the word "dilution" as a model Ln H U) N o1;nEUCan .-:R,.e#ottzts. Ze 12021 296-026, (Vlv4 /703] 644-7636

11-31 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 between mainstream and ETS. In some cases you'll have enhancement. In other cases you'll have decreases in the chemical composition, individual components of the chemical composition. DR. LIPPMANN: Well, one last comment, and we'll move on. DR. LeVOIS: Dr. Flamm, you brought up the comparison between the electromagnetic fields and environmental tobacco smoke. I would like for you to comment, if you wish, on the biological plausibility of these two. DR. FLP,NIlM: Well, I think the biological plausibility of electromagnetic fields has been looked at fairly carefully by some physicists, and I think it's very difficult to construe a relationship to those factors such as DNA damage and other cellular damage that we would associate with either the initiation of carcinogenesis or the carcinogenic process.. With tobacco smoke, currently there is information concerning animal carcinogens. There is, on the face of it, a greater, clearer plausibility except we really know so little about mechanisms. We know so little about what components are really of interest that while on the face of OTIYLEYI.cQ.n GRE#0'LtE4S 1~. C [2021 296-OZ6, q'6q [7v3] 644-7d36

II-32 ,.\ 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 it there is the greater plaueibill.ty, what that really amounts to I think still eludes us. DR. LIPPMANN: Thank you very much. I dppreciate your comments and your response to bur questions. We'll resume our review of the documents, but prior to that, we're glad,to see our two additional members join us today. They both were working for the SAB today on other indoor air issues, and are now available to partivipate in our conterenoe. I'll ask each of them to.introduoe theiajelves and then go through the soul-bearing that we have come to do on this Committee in terms of responding to the quidslines on disolosure of potential conflicts of interests. Dr. Larson. DR. LARSON: Thank you. It'd nice*to bohere today. My name is Tim Larson. I'm from the University of Washington. I currently am not doing any research in the area of environmental tobacco smoke, nor am I revaiving any grants from EPA on the subject. So I don't construe that I have a vested interest-in this subject •- a oontllot of interestt excuse me. I just have a general long-standing interest in the subject of air pollution. DR. LIPPMANN: Thank you. Dr. Woods. DR. WOODS: Thank you. I also am glad to be here ~met.i.can ~e~sotkets (so:/ s96-a8b, "V44 l9n!160-rb96

r 1 1 2 3 4 5 6 7 9 14 15 16 17 '18 19 20 21 22 83 Ir-33 this morning and to be able to participate in thiO conference. I think we'll make it a rather long week as.far as review is concerned. I have a statement that X believe Mr. Flaak will have available. It's similar to the one that I presented in December of 1990 at the review of the first document. one difference is that since I was yartipipating in an 6AH conference review for the previous two daye, it's probably a little more comprehensive than it needs to be for this, but I thought I'd just write one and just d6al with it. I'm a professor of building construotion at Virqinia Polytechnic Institute, the state university. I've been there for about three years. Previously I was at Honeywell Corporation, where I was responsible for the investigation of buildings around the country. 8efore that, i was a professor of mechanical engineering and arChiteoture at Iowa State university. I am a registered engineer. In addition to my teaohing, since 1973 I:have been involved in some consulting work, including design engineers, architeots, insurance companies, attorneys, utility companies, state agencies. There's a long list, including some work with EPA. In 1990 i served on an engineering review panel on which indoor air quality engineering researoh O~-aot.ft eRE#0'GtC;d 21C' IsosJ spb-osb1 y44 hos! 0a4• r6ya . ti

11-34 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 programs were reviewed. I received a fee for about two days of service and travel expenses. Otherwise, I have not received funding from EPA except for services on the Science Advisory Board. Currently I am a member of three scientific or professional societies for which I serve without fee; the American Society of Heating, Refrigerating, and Air Conditioning Engineers, the International Society of Indoor Air Quality and Climate, and the American Society of Testing and Materials. I've just retired as a member of the Building Research Board of the National Research Council this month after completing the maximum six-year tenure. I'm also a member of the Indoor Air Quality Technical Advisory Committee for the American Lung Association. I have been so since 1987, for which I serve without fee but am reimbursed for travel expenses. I have served as a peer reviewer of research proposals for the Center for Indoor Air Research from 1989 to 1991. During that time I reviewed nine research proposals in three cycles, and received an honorarium of $100 for each proposal reviewed. I became a member of CIAR's Science Advisory Board in 1992, and in this capacity have met once with the QTr12E4.(.CQ.I2 G~%#O'CiE'LS. Ze /zo21 2q6-oz6i w64 /7031644-?636

11-35 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Committee to review research proposals. An honorarium of $800 a day and travel expenses were received for that meeting. At Virginia Tech I'm involved with the development of a research program on indoor environmental control. The program was announced in January of 1992, of which I'm the director of the program. We're currently focusing on three technical aspects; ventilation effectiveness, dilution and removal control strategies, and building economics. So far, Virginia Tech has committed more than $300,000 of its resources to construct a research and demonstration facility to support this research. In addition, we've received funding from industry, including Union Carbide Chemical and Plastics Corporation. We've received approximately $600,000 over the last two years to investigate air cleaning technologies for the removal of gases and vapors from indoor environments. We anticipate receiving a continuation contract this fall from them. We have a research contract with Philip Morris. We received funding last year from $1.2 million, which was broadly mentioned in the press, to study alternative methods of air distribution for improving indoor air quality control, including environmental tobacco smoke in office spaces. ogmEtlccsn 1zRz#o%tzas. !be Izoz1 Y96-oz6, q,6q 17031644-7636

) 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-36 In addition to the newly-constructed research and demonstration facility, the cost of which is included in this contract -- it's a Phase II of the building -- the addition has been designed and is being constructed as the primary site for this investigation. We anticipate receiving a continuation contract of approximately $500,000 this fall. We also have funding from Interface Chemical Corporation and Porter Paint researching mitigation techniques of antimicrobial agents. We have a small amount of funding from the United Technologies Corporation, in which we're studying the validity and reliability of several field survey forms to investigate human responses to indoor exposures. We anticipate funding from both the Electric Power Research Institute and Johnson Controls. With regard to communications or correspondence, the only thing I've received in addition to the package of material that I believe the whole Committee received from Bob Flaak, is I received a letter from Carol Thompson on 13 July 1992. I'm currently involved with one consulting project that might be perceived to bear on reviews with this Committee. In June 1992 I was deposed as an expert witness Q7IrLE'LI.CQn GJ`E#0%tz41 25e (YoZl 296-o26, q'cq (7031 644-7636 y ti y 44

11-37 in a case involving the indoor air quality exposures of EPA I 2 ~ S 3 4 5 6 7 8 9 10 11 12 : 13 14 15 16 17 18 19 20 21 , 22 23 employees at the Waterside Mall, which i:t the U.B..EPA Headquarters. To the best of my knowledge, this is a odmplete disolosure of my activities that might be perceived as a potential conflict of interest in the review of thi subject documents. DR. LIPPMANN: Thank you. There seems to be a lot of interest in vr. Woods' activities and how it influences participation on this panel. I regret it took so much time to go through it, but if anybody wanted to know what the potential was, it,s been laid out. In that regard, I've been.advised by me4ers of the audience that there's an interest in the potentidl:oontlict of interest of authors of the parts of the EPA dooument, and this morning I ask Dr. Bayard to briefly tell us something about how he selected people to help him put the document together and how he dealt with potential bias or oynfliot of interest among the authors. DR. SAYARD: Why is it when you always ibdress me, I always think my last name is "briefly"? (Laughter.) 0 DR. BAYARD: The two authors from EPA: X am one ~ ~ ~ N o¢»utl,oa.n CRtIxottcts !:6C' /rwc/ rqb-os6, v-4 17051 baa-rbsb ~ ~ ~ L"

11-38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 and Ms. Jinot is another. We have received no money from either the anti-tobacco -- the anti-smoking folks or the tobacco folks, and, indeed, receive only minimum wages from EPA. (Laughter.) DR. BAYARD: Most of the work here was done under a contract from ICF Incorporated. I don't know ICF's connections with the industry or with the anti-smoking people. However, all the-authors were subcontractors to ICF, and I do know their connections. The major author, Dr. Brown, has no financial or other interests that might be construed as a conflict of interest. Dr. Fernando Martinez also has none. Neil Simonsen at North Carolina, I am assured by Dr. Brown, also has none. Dr.'Wells, who like Neil Simonsen, is a subcontractor to a subcontractor, Kenneth Brown, tells me that he is a volunteer for the American Lung Association and receives no compensation from them. Brian Leaderer from Yale, do you get any money under the table? (Laughter.) DR. LEADERER: I could say something about current research and funding that might be useful to get into the record. I'm currently a co-investigator in an NIH grant Ln r Ln ~ N oginezi.ca.n d?,e#otEet& l~e [Z0Z1 Y96-oY6i wo4 0031 baa-7636

1 2 3 4 5 6 7 8 9 10 11 12 ) 13 14 15 16 17 18 19 20 21 22 23 11-39 looking at environmental tobacco smoke exposure and pregnancy outcome. Currently I have an additional grant from the Environmental Protection Agency in looking at the characterization of organic emissions off of materials and human sensory responses to those emissions. I participated in the -- was the principal investigator in the EPA Library of Congress Symptom Study in the offices there. In the past I've done work for EPA on assessing emissions from environmental tobacco smoke in controlled environments, and some work in assessing concentrations of respirable particulates in nicotine in home environments. I've served on a National Research Council committee on passive smoke exposure and health effects. I've served the Health Ministry of Canada on an advisory committee for health effects of exposure to environmental tobacco smoke. I think that exhausts the list. I didn't come prepared with a detailed list, but I certainly will be happy to DR. LIPPMANN: I think that will suffice. Thank you, Brian. DR. BAYARD: I have one other addition. On the o¢meu;ca.n -Rchottzas. 2NC' /zozf zq6-o26r W 1703] 644-7636 Ln L" N

11-40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 V-1 previous draft, one of my authors, Dr. Thorsl$nd, disclosed to me -- and I disclosed this to the Committee the last time -- that he had done some work for the law firm of defending in the Rose Cippilone suit. So he had done some work indirectly for the tobacco industry, but we used his material anyway. DR. LIPPMANN: Let's now turn to reviewing of Chapter 3. I'll throw the first one back to Daisey. COMMITTEE REVIEW OF CHAPTERS 3, 6, AND 8 CHAPTER 3 - ESTIMATION OF ENVIRONMENTAL TOBACCO SMOKE EXPOSURE DR. DAISEY: There was a good deal of work that was done on epidemiology, and the epidemiology chapters I thought were quite comprehensive and a very substantial improvement over what we had seen in the first draft. So I was very favorably impressed by that. Relative to that work, I found Chapter 3 to be something of a disappointment. I'll give you some specific examples. One is the entire discussion of physical and chemical properties is really less than a page and a half and it doesn't really bring forward some of the more recent data. Not to say that length of presentation necessitates quality, l Ln r (A N 0% ~IfZE tLCQ ~2 GRE~lO'LEC'LS G Ze t2021 296-0261 OD Na4 /703) 644-7636

11-41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 but it really didn't seem to address things in a relatively comprehensive way. In the discussion, for example, of the particle size distributions, figures are mentioned, but their significance to the exposure; that connection wasn't made explicitly. It should have been made. There should have been another sentence or two. More critical in this chapter is that there are data presented, and a major theme of this chapter is the chemical similarities of ETS -- actually sidestream smoke -- and mainstream smoke -- and I do not think they adequately support the conclusion that the two are chemically similar. I mean, they may be chemically similar; I think there are other reasons for thinking they're chemically similar, but the data that are in there, speaking as a chemist, they simply don't make the case. If you look, for example, at the table which is really taken from the NRC report on the specific compounds that are in there, le$s than ten percent of the mass of the particles is really accounted for. And I find that a serious defect. I did spend some time trying to see what kind of data were available. Can I just show a transparency? DR. LIPPMANN: Sure. P4mE4Gc[i-L z~?d#04.tG'LS Ze tzozl 296-o26, W 003] 644-7636

11-42 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. DAISEY: This is the kind of information I think you need to make the case in a sound scientific way that the overall composition of mainstream and sidestream smoke are similar. And I put this on as an example. This is taken from some work EPA has done on cigarette smoke condensate, coke, oven and roofing tar. Basically they've fractionated these mixtures into some broader classes. And you can see that there are some areas -- they're not all that similar. Nonetheless, each of these is a combustion mixture, and each of these mixtures in animal skin-painting tests has been shown to induce tumors. And I think that's important, that combustion mixtures that we have tested have been shown to be tumorigenic in animals, but it's not necessarily a fact of inhalation. I spent some time trying to see if this kind of data were available for passive smoke. As far as I can tell, they're not. Now, I contacted Joellen Lewtas, and she said they had tried to get some support to do that kind of work. They didn't get it. I did talk to Dietrich Hoffmann who, of course, is a well-known expert. He did some very pioneering work on cigarette smoke condensate using sort of the classical ko 071'12E 2l.CQ.12 GRE/20'LtE'L1 L~l C /zozJ 296-o26, q'a4 hosl 644-7636

11-43 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 methods of fractionating things. He said to his knowledge it hasn't been done. So that we issued in -- now, I'm not saying I've contacted everybody whom I haven't. There may be unpublished data that we simply don't know about. But I think that this is what you need to have to make the statement that the two are chemically similar. I don't think ten percent is enough. That's not to say you can't make a case of combustion mixtures having carcinogenic properties and containing carcinogens. All of the ones that we know about have these kinds of properties. And certainly passive smoke has carcinogens in it. There have been some animal tests that show -- even toxic effects. I think this has to be addressed in some way, and I think that chapter has to be modified because it simply is not correct scientifically to say that it has been shown that they're chemically similar. That also brings you to an issue of what you mean by "chemically similar," which is not so simple to discuss. And since I think you can't address it anyway, perhaps we don't have to consider it. But in a broader sense, the chapter often talks about sort of vague quantitative terms, that there are OTI'r2EYl.CQ.12 GREf207zR,9. rLIC' /zozf zqb-o26, q,dq /903/ 644-9636

1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 1 19 20 21 22 23 11-44 differences between mainstream and passive smoke. But you don't get a sense of how big they are in statements like that. I mean, there really is a difference. If you're talking about differences of a factor of 20 percent or 2 or 3 or 4, that's different than a factor of 100 or 1,000. And that's one of the places in -- I think those kinds of things have to be refined in this chapter so that you have more of a sense of the magnitude of those differences and the magnitudes of similarities where they are there. Dosimetry was brought up earlier, and I think most of this Committee is agreed that it is very difficult to do dosimetry, a comparable dosimetry, for mainstream and sidestream smoke because there are so many factors that go in. It may be a topic for research, but at this point in our scientific knowledge it just is not a simple thing that you can do, and certainly cigarette equivalence doesn't really seem a reasonable way to approach the whole thing. There are numbers in this chapter in which the number is -- for example, there's a graph of a regression line for RSP versus nicotine measured in homes, which I think is some very important data. And the slope is given as 10.8. But there should have been some estimate of the uncertainty in that slope because they are scattered on the data, and I L" J c4m.Ea.i.can 4ZRe1i.otEess. 17L' /zo2j zq6-oz6i 'l/04 /703/ 644-7636 ~-4 Ln ko N

11-45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 think we need to recognize how large it is. That may not be in the paper, but certainly the author could provide that I think. And I think that's important to get in there. And there are some other numbers like that where , there are averages reported, but there's no confidence interval in it. So the quantitation and the sense of how large a magnitude of uncertainty you're talking about I think needs to be refined in the chapter. In terms of the exposure data, I think that was not as up-to-date as it could have been. And I'm a little bit puzzled because there's a new book out by Mike Guerin which has done a fairly comprehensive job of pulling together all the newer -- well, and the older -- measurement data. And for example in the measurements in office buildings, there are many more measurements in office buildings that are noted here. The one in particular that I'm aware of was one that was done at LBL, and it did not get into the peer-reviewed literature. It is an,LBL report, which means it's been internally reviewed by three scientists who were not authors. And that one is cited here, so I guess you can find it in the abstract. I think that one is important and interesting because it's 40 office buildings and multiple sites; smoking Ln W QT1rLE'tI.CQ.I2 GREfJ.o'itE'LS. 17C /zozf zqb-ozbi Cl04 J703J 644-7636

11-46 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 and nonsmoking, indoor and outdoor, and they measured respirable particles, and they measured polycyclic aromatic hydrocarbons in those books. I think one of the most interesting things is that the difference in respirable suspended particulate matter concentrations in -- the difference between smoking and nonsmoking areas in terms of the geometric mean concentration was 29 micrograms per cubic meter. It's identical to residences, which I think is fortuitous, but it's also certainly of the same order of magnitude. There is a statement made in there, and I think I have that in my detailed comments, the basis of it, that the concentrations in offices are lower than in residences. It may be that your exposure is lower, but there's no reason to think that the'concentrations are lower. I have not seen data to support that statement, so I think that has to be modified. I did bring,up the business of the cotinine, that there were two cotinine studies mentioned in Chapter 3, but the ones that were used for background corrections were not discussed in Chapter 3. And I think we need to modify that and look critically at the sampling and analysis aspects of those to make sure there are no problems with it. Amzi.can ::,Rs#ottets 2e tZozl 296-OZ6, q~a4 hosl 644-7636

II-47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Nowhere in the report was there any mention of the dietary exposures. And I think that, while maybe you'd like to sweep it under the table, there is some evidence that there are dietary exposures to nicotine, and that would influence the cotinine levels, and it probably influences the lower end of the tail and nonsmokers. And I think you simply have to consider that and discuss it and address it in some way. Those are my major comments. There are some specific comments as well. DR. LIPPMANN: Tim. DR. LARSON: Thank you. Well, I'd like to second what Dr. Daisey said'about the fact that I would have liked to have seen a more complete argument about the similarities and differences between the two different smokes that seem to be discussed a lot here. And so I would agree and would encourage that. It seems to,me that independent of that argument, you still are left with the question of simply demonstrating, I think, reasonably well whether or not people have been breathing air that passed recently through a cigarette. And in that regard, I think, again, I would agree that the correlations that were presented between respirable dgmExica.n eRe#ottzzs. L~e t2021 296-OZ6t wo4 00316aa-7636

11-48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 particulate matter and nicotine are very important ones and should be emphasized. To that end, I noticed that in the Table 3.1 that was presented in the draft, there's no nicotine in the vapor phase constituents, at least that I could find, but that in fact in sidestream smoke we're led to believe that that's the major chemical form, physical form of the compound. And so when one looks at the ratios of nicotine to 3-ZS particulate matter in Table 3.1 on page 3-21-5, you've got numbers that can reasonably be construed to be 10 to 1, such as are the slopes of figures that were mentioned at the end of the chapter. But that doesn't square with the fact that the statement is in there saying that sidestream smoke is 85 percent or more in the vapor phase. So somehow that has to be addressed. 'It seems to me that it's -- I don't know quite how that's resolved. But to that same point, when one looks at the data presented in Figures 1.5 and 3.6 on pages 33-3,6 and 33.7, again, in my own mind trying to see how consistent that ratio is, which is impossible to do without looking at all the data -- but when I just look at maximum values or median values in these datasets, in some of these environments it seems that something in the order of 10 to 1 micrograms of respirable ~ ~ Ln N Anezi.ca.n eRe~so~zs rn L9 ~ 2~e /2021 Z96-o26r 0% 'i/df l7osl 6o-760

) 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-49 particulate to microgram of nicotine seems to be consistent with that dataset. And in another cases, especially for maximum values, those ratios go up to 20 or 30 to 1 or higher. And I don't know to what extent it's important to establish a consistent set of ratios within and between locations. But I think that the argument is centered on the fact that these are measurements of air -- indicators that air has passed recently through a cigarette, rather than being a chemical statement about other properties or tracers or whatever -- trace properties, you know, compounds we don't really fully know about. I think that that's really the more important issue, and so some strengthening of that whole discussion of the relationship between those two apparent important markers would be, I think, appreciated. I have a number of specific comments, and I will include those in my written comments, and I will say that I agree again with Dr. Daisey; you may never be able to prove the equivalency on great detail, and so I think you're left with trying to establish some consistent correlations between specific tracers of interest. So I would hope that you could emphasize that part. And I would, again, to that end, say . d¢meti.ccsn cj-9,ehot&ts Ze /2021 Z96-o26, qUdf 0031 644-7636

11-50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 it's important to talk about the balance. In the summary on page 3- Z , I guess it's a somewhat picky point, except I just listened to some statisticians talk about nonparametric models recently at a seminar and they made the point that you assume linearity in your model and then you fit the data to it. So really, the statement that says that indoor levels of RSP and vapor-phased nicotine have been shown to vary in linear fashion with tobacco consumption -- I think it's a picky point, but it's probably more correct to say that assuming a linear relationship has been shown not to be inconsistent with that. You haven't really done a study to show what the functional relationship is. At least I don't think you have. I'll submit my detailed comments. Thank you. DR. LIPPMANN: Dr. Hammond? DR. HAMMOND: I think, again, that this chapter reflects what I mentioned yesterday about the problems of different authors of different chapters not being fully seen together. As the title suggests, the chapter focuses on the estimation of exposure to environmental tobacco smoke, and comparison, environmental tobacco smoke and mainstream is a relatively small part of the chapter. oO9nzeti.can cRe#ot&U ZG' /zo2f 296-0261 q'd{ /703` 644-7636

11-51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Overall, the rest of the document relies on this chapter predominately for that, to present the similarities of mainstream and environmental tobacco smoke. And so that leads to some problems that have been mentioned earlier. So, I think that the estimation of exposure has been done in a very interesting way, and there's been a lot done there. But I think the smaller emphasis will have to be expanded to support the rest of the document. I would suggest adding on page'33 in the last paragraph, among the differences between mainstream and sidestream, they're attributable also to the differences in the concentration of oxygen -- it's about 16 percent mainstream and 2 percent in sidestream -- so that the sidestream-generating environment is much more reduced in the mainstream. And that, of course, affects the distribution of the compounds that you get. I think that there are ways to strengthen the discussion of the chemical similarities, even if one may not be able to -- as Dr. Daisey said, it's very difficult to say: What does it mean? What is the test for chemical similarity? And it may be true that only a small fraction of the compounds in ETS have been identified, but it looks to me like a lot of emphasis has gone into those compounds which ~ ~ co ~ ~/ ~ ~ Ch ~ PTII2EZCCQ.12 Gl~E/20lfE'LS. , U j Ze tYo2l 296-oZ6, i k-A b q'6q /7031 644-7636

11-52 ) 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 have been identified in mainstream and identified as toxic or carcinogenic. So one would say that the energy has gone to what one would like to call the most important compounds, you might say that. So that could be discussed a little bit. For example you could say that -- I don't know the right number -- of the X identified carcinogens in mainstream, Y are also present in sidestream, and Z are emitted in greater quantities in sidestream than in mainstream. That type of discussion could be helpful I think. And if it's true -- and I didn't try to establish this -- but if it's true, one might be able to say that most, if not all, of the carcinogens that have been identified in mainstream are also present in sidestream, and what percentage are in higher concentrations. Within Table 3.2, virtually all of the five known human carcinogens, the nine probable human carcinogens, and three animal carcinogens, are listed -- are emitted in sidestream and then almost all of them are in higher levels, emitted in higher levels in sidestream. Now, I understand that that's diluted. That's another issue. But in terms of the similarities of smoke, I think that we need to focus on what are the critical similarities. ti (4 b QT1)2ELLCQI2 GRE#ORtEU Ze t20zi Y96-OP61 q~Vq 17031644-9636

II-53 Basically, that's what I would add to the other comments on the composition. in terms of looking at markers for environmental tobacco smoke -- this•is getting one of the charge questions as to how well those were addressed -- there are Meveral t 6 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 possible markers mentioned. Two of them, respiral~le particles and nicotine, are discussed most fully, probably because those have been used the most widely. Some of the others, the disadvantages are mentioned, and for ipthers they are not, and perhaps that might be included to extend that discussion. And I think that there maybe needs to 1rie a little more discussion of some of the uncertainties relit•d to the two markers that are discussed the most. For instance, the respirable particles; there are other sources in the environment. That the background RSP might vary*from environment to environment so that in low smoking levels RSP attributable to ETB is very difficult to determine directly. And estimates of ETS in these situations would be unstable. Conversely, ETS is the only source of rticotine in most environments, so that that's not the.probl.o*. The major limitations of vapor-phase nicotine that ore repQrtod in the literature include deposition onto surfaces more,quickly than ,=4.ncttccsft rJ~tuts ` 2K' /totf tq6.ot661 51592 6521 qU6l /wyi aqa-96so

11-54 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 other components of ETS, and also that it's re-emitted from these surfaces. On the other hand, the successful use of these two markers in field studies argues that these uncertainties are relatively small. And I think that the wide range of environmental tobacco smoke concentrations that are reported in human environments where smoking occurs and which are reported in the table are also very important. If we're looking at several orders of magnitude variation exposures, and typically there are two to three orders of magnitude that you've demonstrated here, then perhaps -- of these other uncertainties might be less important, but they should be described and quantitated, as Dr. Daisey mentioned, so that could help the interpretation and understanding of other criticisms of that. With regard to biomarkers, cotinine and nicotine on body fluids have been discussed most extensively because those, again, are the.most commonly used. I think that there -- well, actually, I had the same comment that you had about that the dietary sources do need to be mentioned because those are brought up often, and they may be important, especially if you have later in the document where the background corrections are discussed. And I'll be mentioning ,VqmEU:can re1iottezsA r M tzozl Y96-oZb, w6q /703/ 644-7b36

11-55 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 this later as well. But I think we need to have some understanding of to what extent dietary contributions might exist. Do they account for a large percentage or a small percentage of nonsmokers' urinary cotinine? Thank you. DR. LIPPMANN: Dr. Eatough, you have a comment? DR. EATOUGH: Yes. Many of them have already been mentioned, but let me just amplify on a few points. I think the chapter needs to really serve two purposes, and certainly right now it does a much better job of serving one of those purposes than the other. And that is, as it's tied to the rest of the document, it needs to establish the chemical composition. There have been a number of comments that have been made relative to that. One additional point that I thought I might make is that there is a lot of variability in the ratio of various components between mainstream and sidestream smoke, and yet the direction of that variability is very well defined. And at least for both the,chemical characterization and the toxicological data that I'm aware of, all the inferences are that sidestream smoke is going to be expected to be more potent per unit of exposure than mainstream smoke. That establishes a sign of what you expect to be the result of what's going to happen due to exposure to environmental og=zEcan -cQE#otEEZs !be tzozf 2q6-o26, NGq 1703/ 644-7636

11-56 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 tobacco smoke. And I think that there needs to be a point made of that. While you've got a lot of uncertainty in establishing the cigarette equivalent association, at least on the basis of the material to which you're exposed, there is a direction that you expect that relative toxicity to take based on the available data. There also is a suggestion, based on the wide variability in the ratios-that you see, that in fact the emissions are highly variable. And in fact, if you look at the data -- and I think Mike Guerin makes quite a strong point of that in the paper here, the book that Paul referred to -- much of the variability and the ratio of sidestream to mainstream components is controlled by the variability in mainstream smoke and not by the variability in sidestream smoke, and that in terms of there being a constant material, sidestream smoke is much more well defined than mainstream smoke for which we have a lot more epidemiological data. And of course, that results mainly from the fact that attempts have been made to alter the composition of mainstream smoke in the production of the cigarettes themselves. I think that a stronger point could be made of the fact that in fact in many ways sidestream smoke is better N ogmczi.can ::RafiottzRs ZC /zo2f zq6-oz6r NGq /7031644-7636

11-57 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 defined than mainstream smoke. It is a point that I think ought to appear somewhere. A couple of other comments. The second -- of course, that's the whole purpose of trying to establish a database that you can make reference to when you talk about the plausibility arguments. The second purpose that the chapter serves is helping to identify how well source apportionment can be made based on the use of specific markers. I have a couple of comments that I'll come back to again as we talked about the risk assessment, but there are a couple of areas where I think the chapter could be strengthened. I think there could be more quantitation that puts some perspective for what we expect to see as more typical exposures, the relative importance of environmental tobacco smoke respirable mass versus others. The difference is talked about, but that difference is not put in perspective to the va;iability of other sources. And I think that might be worth doing. I had some problems in knowing for sure that I knew how to use the figures that are given in the back for prediction of expected concentrations of mass from ETS because as I went through them to try and produce a number PrIf2EtCCQ.l2 --JtEj20'CtE41. Ze tZOZ1 296-0261 q'6q [7031 644-7636

11-58 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 that I thought I ought to expect to see, I ended up with figures that were factors of 5 to 10 higher than I expected to predict from the figures. I don't know if that reflects a lack of my understanding on how to use the figures, or some inherent problem. But that's something that may want to be looked at. Another area where I think the chapter could be expanded is that while the emissions from sidestream smoke, which make up for the majority of environmental tobacco smoke, are really fairly constant, there are very significant changes in the relative composition of environmental tobacco smoke because of some of the effects that have been talked about. Those are really quite well understood. There's a lot of data related to the relative rates of losses of nicotine and other components of tobacco smoke that affect how conservative or nonconservative it might be, and in fact, at least -- and I'll *ake some comments when we get to the other chapter -- my expectations are that most of those changes will lead us to an underprediction of exposure, rather than an overprediction of exposure. I think that at least to give the sign is predictable and something should be said about that. .; Ln CArtE4.l.CRI2 GRE#O'LfE4.s. %e tZ021 296-0261 q~04 /7031 6aa-7as6

11-59 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 I had another thought that I kind of lost as I was going in connection with the exposure estimate. I've lost that thought, so if I think of it again, I'll come back to that again. There is one other area that I think is maybe a small point, but one that we ought not to lose sight of, particularly as we begin to try another -- when you look at the background correction, you have to make some assumptions then about what environmental tobacco smoke looks like at lower concentrations. And we uniformally make the assumption that nicotine is dominantly present in the gas phase. In fact, there is really not any data related to that question. The great bulk of the data that show clearly that nicotine is in the gas phase all come from measurements of diluted sidestream smoke with relatively little alteration in that diluted sidestream smoke. There are very few measurements -- and then those assumptions have generally been used then to move from there into broader measurements. But there are really very few measurements that use techniques that unequivocally tell us what the distribution is that have been made in environments that are typical of low exposure, which of course is the background correction that we try to make. And the few data dgmEZ.i.ca.n cRE#otteas. Ze tYO21 296-0261 Nd{ /7o3f 644-7636

11-60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 that do exist suggest that there is much more of the nicotine that is not in the gas phase than one would expect based on the diluted sidestream measurements, and that, again, will affect our interpretation of the data. I think those are issues that we at least need to think about and address as you make estimates of what's going to happen to the conservancy or nonconservancy of nicotine as a marker. It certainly is the marker of choice for the data that we currently have available. Those were some of the thoughts that I had as I went through the chapter. DR. LIPPMANN: I guess the evidence that it's in vapor phase is that passive samples for nicotine have collected and they could only collect vapor. DR. HArIIMOND: Right. In fact, I've collected side- by-side active and passive samples in the field and seen similarities. It may not be true in all situations, but I've done it in several pl4ces and found that as long as the concentration is high enough that the passive sample in a short time and can respond -- I mean, that's often a problem -- but given that you've got your minimum levels, then I see good comparability. I'm not sure that's true in every situation, but in most of the situations I've tested. Cyqmazica.n ::,REhottets. 17C /zoz/ zq6-oz6Y W /703/ 644-7636

11-61 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. LIPPMANN: Dr. Leaderer, you've been dumped on a fair amount. We'll come back and get you, but we also need to recognize that we're looking at a first draft on this chapter, as compared to a second draft. So whereas the other chapters have been cleaned up a lot of the initial problems that one has, we are looking this time at the first effort. It's also I think clear to me that perhaps the charge to you in preparing this may have been less than ideal in that the critical dependency of discussions in other chapters on this may not have been apparent to you, and you may not have anticipated the uses to which this chapter would be put in the rest of the document. DR. LIOY: I think that's a very good point. I think that was my first point with Bryan, is the fact that this is the first draft, and having been through this process with other pollutants and being the author of a chapter similar to this, the same issue comes up every time, you don't have all the data and it's clear that it's a massive effort. I would make a major suggestion that this book here by Guerin is something you should have and use in dealing with your revisions because I think it's quite valuable, it has quite a lot of information, and it can fill in some of OT h2E 2CCQ.12 GRE/20 RtE R.S. Ze tY021 296-0261 wo4 003164a-7636

11-62 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 0T~2CCQ/2 GRE,fiA'LEEt!i M t202i 296-o26, W Ix3/ baa-76_q6 the gaps that have been suggested by colleagues on the Committee. I'm a little taken aback by Dr. Daisey's need to see that there is an equivalency established, or even considered, between mainstream and sidestream smoke, and I don't think that's the major point. DR. DAISEY: No, no. I'm saying -- DR. LIOY: Let me finish my point. DR. DAISEY: I'm saying it hasn't been demonstrated. DR. LIOY: You didn't qualify that, and I'd like to qualify that. And I think that Dr. Eatough made a very important point that in terms of the positive direction of sign -- which is my concern at this point -- it's clear by just looking over the data in Guerin's book over dinner the other evening and reviewing Table 1, it's clear that the ratios are all much higher; in some cases dramatically higher for some of these compounds. And that is an area that you have to consider in your revisions and make some deliberate comments in that regard in terms of the nature of some of the carcinogens and what it has in terms of meaningfulness for support or rejection of some of the hypotheses in Chapter 4. C4 co ti o) y cu m U,

11-63 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Y think an important point that has to-ki:e made is to look at the introduction of 4.1. It's the firect paragraph. The statement that's madv in the midQle of the paragraph that you have to either support, accept; reject or modify that statement in some way, shape or form 8o that we can move forward in Chapter 4 with discussion -- ¢11 right? -- because right now I think it's ambiguous, and I think the supporting arguments in Chapter 3 right now are etdbiguous. Again, it has to do with the fact that there's some interpretation that has to provide us with a database. I think another issue that was brought up this morning -- and I'm not sure if I agree with my colleague from Massachusetts down there that it's irrelevant -- it's the issue of where particles deposit in the lung -- I think is very important in whether or not we histologically know the exact location of the cancer. It's still clear to me that when you have differences in particle size distribution and where they deposit, it!s important to consider it and at least put it out on the table as a scientific erqument, and the plausibility of that argument, and prepare the discussions of the nature of the health effects. We can debate about how people read siids and pathology, but I think it's clear from a lot of work that's oqrne.t.lean ~e~otEsu 2~C f sosJ sqd-pprb~ w~R 170sl 44-16s6

11-64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 L'G'~.n~1n»S done in Dr. bittvn's laboratory and other laboratories around that world that there are differences in where particles deposit, and size distribution has a significant effect. And depending upon the chemical composition of those particles, there may in fact be an enhancement or a decrease in the carcinogenic potency of those particles. I think, Bryan, that's something you have to deal with rather directly in this chapter because I think it's not a side argument, and I think that you didn't deal with it directly. And I think you have to deal with that rather directly in your discussions. I think my final point, although I have a lot of little ones, and I think it's a major point, is that I'm a little surprised that you still -- that you made the mistake of misdealing with exposure and concentration. Concentration and exposure, as you well know from all your work, it's concentration time. And I think some of the figures are misleading in the sens,e that it shows concentration distributions, but it doesn't show exposure distributions. I think it would be well worth some effort, especially in terms of some of the statements that are made in the other chapters about the importance of the spousal smoking, that issues of exposure are dealt with, as well as. 44m.eti,ecsn eReho%teu Ze IZOZI Y96-OY61 wo4 (7031 644-7636

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-65 concentrations, because I think they're important, it shows the magnitude of the possible dose or at least the potential dose to the lung, and I think that issue has to be laid out clearly in this chapter so that it can be used again in arguments that are directed toward Chapter 5 and Chapter 4. Again, it has to do with completion, and I think a lot of information is here, but I think it has to be more crisply stated and stated in such terms that are clear to me what exposure is versus what concentrations are. I mean, I don't like to see them like mixed up in this particular chapter. I think those are my main comments. DR. LIPPMANN: Dr. Wesolowski? DR. WOODS: It's Jim Woods. DR. LIPPMANN: Oh, I'm sorry. Jim Woods. DR. WOODS: Thank you. First of all, I am very pleased to see the Chapter 3 in here from the first draft to this draft. It helps.a lot, I think, to be able to have a chapter that's devoted to the characterization of environmental tobacco smoke. I guess there are a couple of things that bothered me in the chapter, though. One is -- and this is probably true for most of the chapters -- I keep looking for an Ln L" %D N c4I)i.E'LLCQ/2 GRE/20Y.tE'L4. M tYC21 296-0261 qV 0031644-7636

11-66 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 objective of the chapter. And in this particular chapter, I agree, I think, with Dr. Eatough or Dr. Lioy that indicated that the purpose of the chapter needs to at least link to the other chapters. It needs to give a rationale for why certain methods of analysis were used in the other chapters. And I would look for that really in the first or second paragraph. I'm still bothered that there is some confounding or confusion. I'd like to see it brought more forward in distinguishing between environmental tobacco smoke, sidestream and mainstream. And then we get the term "passive smoking." This whole thing needs to be clarified, I think, and this is the chapter that that type of definition needs to be brought forward. The environmental influences of environmental tobacco smoke I don't believe are characterized to the extent that they should be. For example, the influences of temperature and humidity, the impact that that may have on the particle size distribution of sidestream smoke may be a factor, especially when you begin to deal with the aging factor of the ETS. So those types of characteristics I think need to be brought forward and really clarified in this chapter. The other is, I think that a direct linkage of the Lq ~ ~ ~ o, w .+/ Ln c~{m~zi.can ~e~so~z~ l~e t202J Y96-026, ClUd{ 17031644-7636

11-67 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 information that is being presented in this chapter to where you'll find it in the other chapters ought to be developed. For example, I think it was on one of the biomarkers -- I'm not sure that the 4-ABP is really used very much in the other chapters. DR. DAISEY: But it's important. DR. WOODS: Yes. You bring it up as a very important biomarker. So if that's important, we need to get this cross-linked, I think, to the chapters. I've got a couple of things I wanted to bring up, but the other major point is in modeling. In the section on modeling you introduce two types of modeling; the statistical modeling, which I believe basically is used in the other chapters; but the rational modeling, or the physical modeling, you're presenting, but it's not used very much in the other chapters. And I think that there's a real power that we've lost in not being able to use the modeling effects. , Also the information that's in here is broad-based, I think, from the standpoint of different types of environment settings: restaurants, offices, residences, et cetera. And I think, again, that needs to be more clearly linked. But from an assessment standpoint, I don't believe . P~'LCCQ.12 GRS/2o'CtE4S. ZC (2op/ 296-oz6, qGq 0031644-7636 ~ tD r Ln 01 Cn W ~

11-68 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 that that linkage is made as well as it should be. Almost everything that we're dealing with, I believe, in the subsequent chapters focused on residential exposure. And we need to think in terms of other types of exposure. Now, for the preschool children, that obviously is not a related factor, but for other populations, I believe it's a factor. Now, with regard to modeling, the problem I have with this -- and I was really surprised at how weak that section is with regard to what's known in modeling -- I think a serious critical review of the models that are available in the literature need to be brought forward here. And some of those are very comprehensive models. Some, Brian, I know you've dealt with; the modeling that was done, for example, in aircraft cabins, et cetera. There's some rather sophisticated modeling that can be brought to bear, and that needs to be critically reviewed. I think the.one equation that's in here is an empirical equation that's not validated, or not well validated, and yet it's not used anywhere else either. So I think, again, in this modeling concept that that aspect -- either drop the modeling, or the rational modeling, because it's not going to be used. or if it's going to be used, then C*KE'LLCQ/2 GRE#O'LtE'CS ZC IZo2l 296-OZ6, qGq 170316"-7636 ko cn Ln 1-

11-69 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 use it. I guess the nit that I wanted to bring up is on page 3-8. I may be just misreading this, but I can't make the text in Table 3-3 and Figure 3.1 agree. DR. HAMMOND: I spent some time on that. I think the numbers are reversed. If you check, they do agree. DR. LEADERER: Something was lost in the transcribing. DR. WOODS: Okay. That was the first problem. The second problem is I can't get the value of the increment of 28 micrograms per cubic meter out of Table 3.2. I can't find how you get from here to there. DR. LEADERER: That was reported in the paper, in publication itself, as the estimate, without giving the details of how that estimate was arrived at. DR. WOODS: But what that's basically -- if I look at Table 3.3, page 328, you're showing these ETS as 28 micrograms per cubic meter. DR. LEADERER: No, that's another subject. DR. WOODS: I'm looking at the right-hand column. The second entry down. Now, is that 28 as compared to 0 or nonexposed? PARTICIPANT: That's 28 over and above the OTI'12E'LLCQ.I2 GRB#A'CfEYS. Ze tZ02i 296-OZ6, we4 /7osl 6aa-a636

11-70 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 background -- DR. EATOUGH: You get that number, I believe, from Figure 3.1, the distributions is where the number 29 comes from. It's the 50 percent cumulative percentage, the difference between nonexposed and exposed. DR. WOOD: Okay. But the problem is that in Table 3.3 that looks like an absolute value. It looks like it's just a concentration and not a difference. DR. LARSON: Is it the difference of the mean, or the mean difference? (Simultaneous conversation.) DR. WOODS: I guess it just needs some work, I think, to clarify that because compared to the other values that are in here, I'm not sure of those differences. There's something that's not quite right, I think. I guess the other thing, on questionnaires, I felt that the literature review on that was rather weak, and I think there's additio0al work that can be done on what questionnaires are out there. I know there are a few that were not cited in here, and especially some critical review with regard to the validity of reliability of the questionnaires that are being used I think needs to be in this chapter. It's the only place that I think that we get. Ln Ln W N CAma-ti.can z~?c/zotfzu 2~C /zoz/ zq6-oz61 Ndq /7031 644-763b

11-71 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 stuck up front that is going to link back to the chapters, and I think it's important. PARTICIPANT: Oh, absolutely. There's no question. DR. LIPPMANN: I guess Mike wants to add his comments. I'd be surprised if he didn't. DR. LEBOWITZ: Well, I was going to say that based on what I know of the high quality of Dr. Leaderer's work, that I had to assume that he didn't have enough time to put into the chapter. I think that in addition to the comments already made that -- and I would certainly start my comments with the notion about questionnaires and the relation between questions asked and either RSP, PM10 nicotine measures, et cetera, that you do need to do more work, and I would start with work that you've participated in, including the standard and environmental inventory questionnaire that has documented the validity of certain questions asked about number of cigarettes smoked in ppecific environments and RSPs, for instance. And also continue then with some other work that you have laying around, or Dr. Stolwijk does, from WHL Euro effort in which we plotted a lot more for the RSP cigarette, the number of cigarettes smoked in specific environments, et oqm=¢1ca.n Z~?E#ottata. Ze IZ0Z1 Y96-oPbl wo4 /7o316aa-76s6

11-72 1 2 3 4 5 6 7 8 . 9 10 11 12 13 14 19 20 21 22 23 cetera, and use some of that additional data to flesh some of the relations. out Now, in addition, if you can find the time, I'd like to see you use some of the RSP dosimetry work in here well -- for instance, Heller's work, for which you've had references in the past -- you can find them, or I can find as them for you -- shows about 11 percent sidestream deposition of particles, RSP we'll call them. They're actually smaller, but that doesn't matter. And that would go along with the 10 to 20 percent figures as versus the nicotine figure in terms of smoking equivalency. Also, if you have problems fitting additional stuff in, you can get rid of redundancies that exist in many places and give yourself more room to do them. But I think that, again, you haven't had the opportunity to see the other chapters as much, and so there is still lack of correspondence in terms of reference to misclassification dat$ in studies and the effects of misclassification that occur in this chapter compared to what has occurred elsewhere or was discussed yesterday. And I'm sure the staff can fill you in on some of that that you missed. 3-17 The modeling in the modeling section, page 3~3, Ln 1~ Ln W N ~ M oginazi.ecs necsn ~a#O~ts ~ m `Ll e t2021 Y96-OP61 wz4 hOsl 64a-76sb

1 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 11-73 ignore certain things, like if there's no change in smoking and no change in residence, then in fact one can model past exposure as well as current exposure, and that current exposures are relevant for a shorter term; effects, acute effects that might be related to those exposures. In fact, in some estimates that I have made, for instance, in 1/6 of all the childhood leukemia cases we're studying and others our colleagues have studied, 1/6 of them over a 20-year period, their current exposure is their past exposure as well. In 1/4 of chronic respiratory diseases over a 20-year period, their current exposure is their past exposure as well, and if that 1/4 represents 1,000 or more cases and controls, then obviously you can use current exposure to estimate them. So there are a number of issues here that I think can be pursued further now that you've seen the overall document and may have more time to look at some of the previous work you've 4one and work others have done to flesh it out. But I certainly do appreciate the fact that the chapter is in here, and I think it can provide some basic numbers, arguments, et cetera, for the other chapters. And so I think that that's what we're hoping for, and I'm sure Ln N Ln W ) oTinExican GRctt.ottcts 2~C /zozJ 296-0261 q'6q /7031644-7636

11-74 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 the improvements will do that. Thank you. DR. SAMET: I'd like to turn back to this end of the table because I've think we've asked Brian to make a major expansion of what was a relatively minor part of the chapter. And hearing Delbert and*Paul and John and Tim discuss the issue of the comparabilities of three complex mixtures -- mainstream smoke, sidestream smoke, and ETS, which have some definition, but ETS is a time-variable mixture -- what principles can the Committee give to Brian for this assessment of comparability? I don't think he can go to work and come back and we can't then say, this isn't right. So I think we need to evolve some principles. And I don't think this is easy. So I think we ought to have some discussion of that issue. DR. LIPPMANN: I think Jan wants to address that. DR. STOLWIJK: I couldn't agree more with you. The difficulty that we're dealing with here is basically what is practically a chaotic,condition, which is tobacco smoke, which has been described in various ways and in various times. We can ask ourselves whether further description, more details, and more qualification, in fact, even if we could get it, would help a lot in this particular issue. It 44nezi,can Z~?F0otizts !bC Iiozf 2qG-o26, W /7031644-7636

11-75 1 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 might not. I think what we are looking for is the drawing of conclusions from past reports that say something about relationships between these various types that hold true matter whether you have a different kind of cigarette or whether you have this or whether you have that, because we are going to flounder, I think, in terms of trying to capture the detail because the detail has too much variability in it to be a meaningful utility. What is useful is to find the order that is still there. And the order, for instance, says that mainstream smoke is much more varied, depending on the source and so forth, than sidestream smoke is. There are some things that have been recognized, I think, and mentioned by several others here where certain categories are on a mass-percentage basis much more prevalent in sidestream smoke than they are in mainstream smoke. That's a very important thing to extract from the information that is in the literature. What you find in the literature is descriptive stuff. Maybe there are some relationships that might not even have been noted that you can extract from what the literature says in several of the articles. <z4If2E'LLCQ.12 GRq].OtEE'CS. 2Se r2021 296-o26, q,6q /7031 644-7636

11-76 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 I think that is the direction in which you go in which you actually discover the order that actually exists in the chaos that is the ultimate description of what if going on at any particular moment or any particular time because the smoke, for instance, in a smoke-filled room tends to be more aged than the smoke in a nonsmoke-filled room. Now, all of these things had to go into it. If you try to account for all of that simultaneously in detail in terms of concentrations, you will probably get stuck. If, on the other hand, you look for the order of this, then you are going to make progress. You also, by the way, will also make progress then in the definition of your models because your models actually reflect that order that is there. They also reflect the input variables that are highly variable. But there is an order in how the model should behave and reality behaves. And the extraction of order from the detailed description of lots of chemicals ovez lots of time in lots of different places, is really the task. The task is not to show the -- and I think this is where I really would question Joan because the finding out precisely what all the concentrations are of everything in their total account is not a useful approach. Ln ~ Ln to J (n ~ ~TI~2Eil.CR12 GRE~.o'tEE'LS '~a Ze tzo21 P96-oz6, W /7031 644-7636

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-77 DR. DAISEY: I'll comment afterwards. I think perhaps I was misunderstood. DR. LIPPMANN: Do you want to comment now? DR. DAISEY: My comment is that the data in that chapter simply do not demonstrate scientifically that they are similar. There simply are not enough data. And I agree with you that you're not going to have that data, and even if you did, you'd have to decide on criteria for what constitutes similarity and what does not constitute similarity. Does that mean you can't do anything with it? No. I think Jan is exactly correct, and I think some of the points made by my colleagues on the table are the ones to make. First of all, the comment that Kathy made that the carcinogens have been the main focus of the characterization, and that the major carcinogens that have been found in mainstream smoke have also been found in sidestream and environmental tobacco smoke. In rooms where there is heavy smoking, they are found in elevated concentrations. There's no question about that. Delbert's point about as a complex mixture, you have animal data for the complex mixture as a whole, animal o~{rne~i.can ~e~so~~ 1~. e [2021 296_C26, qUo4 /703J 644-7636 N to Ln m ~ ra-

11-78 ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 bioassay data, and you also have the business of the ratio of the carcinogens in these mixtures. All of those indicate that, if anything, sidestream smoke would be expected to be more potent than mainstream. I think that's an important point. The point I tried to make here but perhaps didn't make too clearly, is that all of the combustion mixtures, the major ones we know about, to which people are exposed that have been tested, and that includes -- let me go back a step. All the ones that have been tested extensively in animal bioassays have been found to be carcinogenic. And many of those mixtures have been shown to be carcinogenic in humans exposed in industrial situations. So if you think about it, it's certainly reasonable to expect another combustion mixture, and even though the conditions of combustion are somewhat different for sidestream and mainstream, you would expect it to sort of follow the same principles. You've got carcinogens; it's reasonable to expect it to be carcinogenic in humans as well as animals. So I think those are some major things, and I think there may be some other ones. DR. LIOY: That's a good point. I think the idea O7I12E~l.CQ12 GRE~2Cz.LC'i~ Ze 12021 296-0261 q/04 /7osl 6aa-7636 VI N Ln W n)

11-79 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 that there are carcinogenics has been shown -- other similar mixtures have been shown to be carcinogenic have been reported. And the equivalency is not necessarily in terms of point-by-point relationship between the ratios of the individual compounds in the mixture. That makes much clearer what you put on the board, Joan, and -- DR. DAISEY: My only point here was: Not proven. DR. LIOY: Now I understand what you were driving at. I think those three points are critical. But then the linkage to the other chapters, Brian, in terms of exposure are very crucial because the other chapters do in fact use epidemiologic/exposure assessment principles to establish exposure response relationships. And I think the linkage is there. It has to be firmly established in your chapter so that it can in fact be demonstrated that this is not done by taking something out of the air. Let me try that again. (Laughter.) , DR. LIOY: The point is that I think that you have to establish that relationship and try to link your information more closely to Chapters 4 and 5. Not force it, but where there are plausible arguments, link it; and where there are inconsistencies or confusion, at least provide that 47q~2f.CQ12 'rR'e#O'LtLLS, 2~C t2021 296-026! q~o4 t703] 644-7636

11-80 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 as a degree of uncertainty that one has to deal with when one tries to establish a causal response or quasi-causal response of total response between exposure and some kind of effect. And I think those are the four or five most important points. If anyone on this side wants to add anything -- I think those are the most important issues. Not detail. Adding 35 more tables is not going to help. DR. LIPPMANN: One last Committee comment at this phase from Dr. Woods. DR. WOODS: I guess I'll just make a plea. I hear what Joan is saying with regard to the carcinogenic constituents in environmental tobacco smoke. But I think this chapter needs to cover the whole book or the whole volume here. And to the extent that we're dealing with other types of diseases or other types of effects, I think the same intensity of presentation of the material needs to be dealt with as you would deal,with the carcinogenic substances. DR. LIPPMANN: All right. I think everything up to now has been more or less related to carcinogenic properties, but clearly that's the more trivial of the public health impacts of environmental tobacco smoke in my view, and that the effects in kids are really what we need to make sure we Ln Ln o W N ~ ,vIn.EtI.cQI2 G)?E#OtEEZS Ze tzo21 z9&o.-by wc4 I7v3/ 644-7636 OD

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 properly address as well. Let me give you an option and you decide whether you want to respond to these comments now or after the break. We're overdue for a break. We've taken much more time than we probably had planned for,on this chapter, but we needed to. You may want to discuss with Dr. Steve briefly next to you -- (Laughter.) DR. LIPPMANN: -- what your responses to these comments and suggestions are. Would you like to do it now would you like to probably do it after the break? or DR. LEADERER: I'd like to say a word now, and then probably we'll have more to say after the break. I gave final exams in one of my classes this May, and one of the students came back with the question and didn't do very well at all. He had said to me, "Well, I answered the question." And I said to him, "No, you didn't." And he said, "Yes, I d.id." And I know how that student feels. (Laughter.) DR. LEADERER: Let me say I think to some extent I've been laboring under a considerable misconception. The chapter was meant to cover some of the central points in ~ ~ ~ ~ N 01 OTYrtEZLCG.12 GRE#A'LtE4.S. Ze tzozl z96_ozbj6 qVcq /703j 644-7636 ~ tD .6

11-82 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 exposure assessment. That is, one could easily do a book, certainly a paper, on chemical composition; one would easily do a book or a chapter on the individual constituents, their measurement, the ability to measure them, the concentrations in different environments. Certainly with the questionnaires, that's a substantial thing in its own right, in looking at how the questionnaires have been used and how they're validated. The models themselves, again, relates to what easily could be considered a chapter. I think the problem is that there are several different areas covered in this presentation, and they're covered in a survey form. They're meant to highlight some of the major points in those areas. Now, the chapter was written in isolation in ignorance of the other chapters. I can honestly say I have not read the other chapters because I've just received this copy. The integration clearly needs to be done. I think areas can be strengthened. And I appreciate, believe it or not, the comments received here. It gives me a sense of direction in terms of what you would like to see. Some of the comments I agree with: some of the comments I do not agree with. 0'/lT2E4.l.CCi/2 G~40'LtE2S. Ze I202I Y()6-OY6/ q~cf{ 0031644-7636

} J 1 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-83 I think integration does need to be done. I think that will take some major restructuring. In fact, one might want to consider even additional chapters or the breaking up of this chapter into a couple of different subsections. If there has to be a greater emphasis on the chemical composition of the sidestream versus mainstream, and then even comparisons with other complex mixtures that are considered carcinogenic, well, then maybe that needs to be broken off and dealt with-in a more detailed form. I share Jan's concern that one easily could get lost in tables and numbers. And, in fact, I should say too that this was written before Mike Guerin's book was out, so that that data was not available and a lot of the recent data was not available at the time the chapter was put together. And Mike's book is a good example of how one can spend a book dealing with a couple of the issues. And I think it's a serious problem. I was under.the working hypothesis that just comparisons of some summary statements needed to be made, and I will very carefully take your suggestions and look at it again and talk to my colleagues at EPA to see how we might restructure things to provide better integration with this exposure chapter with the other chapters and provide you with c4rmti.can eRE#otEezs Ze t2021 Z96-026, wdl 00316aa-176s6

11-84 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 the information that you deem necessary. DR. LIPPMANN: Okay. I think at this point, let's -- DR. BAYARD: I feel that a lot of the information was put in the other chapters on exposure as it pertains to the specific materials that we were getting into. In Chapter 4 we talked about the connections between mainstream and sidestream smoke, vis-a-vis mutagenic and mutaconicity tests, cigarette smoke condensates, and animal inhalations. That's when we dealt with active smoking. In Chapter 6 when we did the quantitative risk assessment we dealt with cotinine levels in adults related to questionnaires. In Chapter 8 we talked about cotinine levels in children related to questionnaires and then we talked about cotinine levels vis-a-vis misclassification of smokers. So we felt that it would be redundant to put it both places. DR. LIPPMANN: Clearly, at least, there needs to be more cross-referencing in Chapter 3 that this will be discussed further here or there, or whatever, and vice-versa. And there were some inconsistencies noted which were inevitable. In fact, this author didn't know what was in the c4mEZica.n cv'?E#oitEts 2$e t2021 296-026I q'6q (7031 644-7636 L" N

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-85 other section, and probably vice-versa. But in any case, these issues I'm sure can be dealt with, and as the overall organizer of the volume, these comments are directed to you, as much as Brian, to make sure that he knows what he needs to cover in revising this chapter and that you make sure that there's appropriate cross- referencing in the various chapters to make sure that there's no misunderstanding. We might want to discuss this a little further after the break, but I think why don't we all take a break now and think about whether we're finished with this or not and move on. (Recess.) DR. LIPPMANN: Before we move on to the discussions of Chapters 6 and 8, let's make sure that we've properly closed out on our review and discussion of Chapter 3. You indicated, Dr. Leaderer, that you wanted to say something before the 4reak and might want to say something after as well. Do you had any further response that you'd like to make? DR. LEADERER: I had several comments on some of the specific comments that were made, but I don't think that I really need to get into those now. L" ) 44m.EtEca.n =Refiodev.s. Ze tzozj P96-or& wdf t7o-ql 6aa-7636 W W

11-86 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 What I would ask is that the guidance that the Committee gives me be as specific as possible so that I might, in going back and revisiting this chapter, come out the second time with something that's more functional in terms of the structure of the document overall. It's my understanding from what I heard here this morning and talking to people at breaks that the basic conclusions of the chapter are not a problem. It's a question of the material brought together, and how it was brought together and used, that could be strengthened. That is, more support for the conclusions and some identification where the weaknesses might be. But that there is not -- at least, I didn't sense any idea that the general conclusions of the chapter were a problem. Is that fair? DR. LIPPMANN: Well, of course, we can't tell you what the conclusions are when you expand your discussion and come to other conclusions. But clearly, this chapter is not there for standalone Burposes, but to provide a firm underpinning for what other chapters conclude. So its conclusions are statements of knowledge, rather than judgments of toxicity, and I think the more important thing is to look at it as a basis on which the final conclusions from the other chapters are based. OTI)2E2GCQ./Z GRE#OTtE4,1. SNC t2021 296-02b1 q/64 /7031644-7636 ~ ~ ~ 4P

, ~ 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 • 20 21 22 23 ,... . DR. LEADERER: Well, olaaxly, thdt ooo0i,nation would be on the chapters as needed.: I think fjfait ;Ust would be oertainly more --- but I do look CorVard to r*ootvihq . • . ;s r . . . detailed oonmOnts that wiil guide me iri the riqht":4irootion. . :: DR. LIPPMANNs I think lrou awt1,:lvave *0~~ -or those ~.., . almost immedhitely because most ot our Coynmitt**:.#'#mbsrs have written these 'comments, and Bob .FleOk, ft be Ao41* 't have . .r. extra copies, can qet copies to you within days.':zThose that . ..:; haven't writtsn out their key commehts will csPt#ttrily do so immediately and give them to Bob or direatly to yau.ns well. Dr..8ayard, do you want -to sdy anythinoftare about :_.~ .. Chapter 3? . , (No auaible response.) DR. : LIPPMANNa Let 'a nmove. on then to: -040tsir 6, Population Ridk oI Lung Ctnnoer from Passiva 8mq:c0t., Liby. Dr. DR. . LIOY: I'1•1 be with you 3,nt u mas3*Adj ; DR. •,LIPPMUN! Do you want tdi tb start•~'ii'th someone •~ , . , ..; : .. DR, • LxOY tPo, no. 1just want a si,OiOtA . ao Qet myself orqanixed. . . - . ,. DR. -LYPPId1WNi Okay. surb. ' ~ CHAPTRR 6- pOPUIATION RIdK OF INNG -CAN.OPR ZC j:o~J sqa-o~o, wd! 170:% Awosb

11-88 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 FROM PASSIVE SMOKING DR. LIOY: Okay. Chapter 6 and I had a long intimate discussion in the last two and a half weeks. I found a lot of good information there. However, I found it to be quite dense in,terms of presentations. And this is an overview. I think in some ways, because of the denseness in presentation, one gets a little lost as to where you're going with some of your analyses. And one of the initial assumptions and limitations that you are working under, in terms of the application of certain equations -- and I think that has to be clarified for myself and I think the audience that's going to read this document because clearly to me there are things that -- I'll give you a good example. 6-Z) At the end of the chapter, page 621, when you start talking about sensitivity analysis, you do the right thing. You start talking about what are the variables in the chapter that one has to concern oneself with in terms of how they may affect the risk, and their use in terms of quantifying the risk to the population. But that's the time that I saw them all in one place. All right? When I go back to the beginning, when I start looking at the equations, I get bits and pieces of that w H Ln ~ 01 ~~~ ~J1 G./ / I'12E4.LCQ/2 ~E~70 iLCZS 01 ~e t202I 296-026, wd~ 17031 64a-7636

11-89 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 information, but I have to start piecing together the whole argument. And it's not satisfying. Again, it has to do partly with my background, and also in trying to present things in a systematic way to the public -- and that's not just the general public; I think the scientific community in terms of what you're doing. 6-) 0 I think, again on page 6'10, there's another example, which is striking. You put in parentheses 6-10 additional -- page 644, 5th line from the bottom -- you say, "additional parameters" and then put parentheses (R,,) et cetera, et cetera. And they're not clearly defined. Again, I have to go to the next page and dig out the fact that there's an equation over here which has to deal with some of those things. And even then I struggled to figure out the definition, and in some cases, the only way I was able to figure out the definition was by working through the arguments. And, again, that left me with a degree of concern. I think it's very important for the delivery of your message that you have to lay this out clearly. I think the most serious deficiency I have with this chapter -- the resolution of it did not arise until I went into the sensitivity analysis. At that point I began to 07I)2E'i.I.CQ.1'L GR*A'LfE'l1 T~C /zoz/ zq6-oz6y qGq /703/ 644-7636

11-90 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 recognize the fact that you were working under a limited set of assumptions, one of which -- and if I'm wrong, please state clearly -- was the fact that you were assuming spousal smoking was constant in these analyses. Is that a true assumption? DR. BAYARD: The current level was represented (inaudible). DR. LIOY: That the current spousal smoking remained constant in your analysis. That is not stated clearly at all throughout this entire section, and to me that's a fundamental assumption that you're making, because if spousal smoke increases, if you make it as an increasing variable or decreasing variable in the sensitivity analysis, the equations have to be looked at in another view in terms of a limit as Z approaches infinity, so that you don't come up with the illogical conclusion that with spousal smoking the risk decreases. You understand my point there? DR. BAYARD:, Yes. DR. LIOY: It's clearly a limited case that you're working with. And I think for the benefit of the readership, this has to be clearly laid out. DR. BAYARD: We did that in Chapter 8. I can see where it has to be done in Chapter 6. Ln dgmeti,can d?e#ottzts. Ze tYOZ1 296-DY6i W /7031 644-7636

11-91 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. LIOY: Yes. Well, that was my point. When I read through Chapter 8, I began to recognize what you were dealing with in Chapter 6. Again, the threads of the argument were put together logically in your mind, but they were not very apparent, or readily apparent, to me, as the reader of this document, and looking at it from the standpoint of a fundamental mathematical formula, which is trying to explain a relationship that one is predicting is going to satisfy that relationship. And I think that this clearly has to be laid out in the assumption that you make. And also, the generalizable case has to be presented also because there are going to be situations where you may find that spousal smoking may increase rather than decrease -- increase or decrease as a function of time. I think some of the information presented yesterday by Dr. Sears, which was quite eye-opening to me, suggests that we may still have more spousal smoking than we anticipate because his. Z-factors were going up rather high in some of his studies; in fact, over 10. Maybe it's due to the fact that we have some spousal smoking microenvironments that heretofore have been underestimated. So I think clearly the way you present this issue weighs heavily on how one views the Z-factor. And I think aqmEqf.ccsn cRehottzsrs. 2~e t2O2/ Z96-026, qUdf t7031 644-7636

11-92 1 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 it's very clear that you have to present this in a logical way. My suggestions are, first, to define Z exactly as to what it means and how it relates to the generalized equation, and then what assumptions you make about the generalized equations to deal with the specific case at hand. In this case you assumed in the document constant spousal smoking. Then, in another section, explaining how in the limit, as Z goes to infinity, the relationship has to go to the much more fundamental relationship of Z -- of RR02 going to RR2. I believe that's the nature of it; when you go into limited Z, going into infinity, which would be the case of spousal smoking. And so I think those points have to be very clear. I think the definition of all terms have to be right up front to make this a less dense document for people to understand, and so that the arguments are clearly defined. DR. BAYARD:. Do you think that putting it into that section right after 6.3, consisting of the model of the parameters we're using in the -- DR. LIOY: I think the idea of sensitivity now was a very logical thing to do. In fact, it reduced my degree of concern about the chapter because I began to understand where c4nZE'L(.CQ/2 GRE/204tEt1 7e /zoz/ zqb-czb, qoq t7031 644-7636 U1 Ln Lo N

;,.., 2 3 4 5 7 8 the chapter was going and what your' assumptions 44%+s,. even • . . though you didn It give Us a degree pf oOrlYideftai •`*r the bslinninq as to knowing wh8re in thi hsok I was fjoinq with this whole argument tieoause I kept takittiq ratios :4ied x-kept . . , anealysis, it was clear to me that thinqs -wsrs..Uu!% sore getting confused. But once I dealt~ with'the neiihitivitK copacetic than I initielly thouqht.: Now, a =inal point m DR. BROWNs 8xdube me. Do you think IV.:q`d~lld be 11 • 12 13 14 15 16 17 18 19 20 21 22 23 helpful to put these terms in a little '.teibla in ,"n4 of a glossary? : DR. : LIOYi As lon.q as I filavs a dsliti~t~~ ~f ~vhat •:: the terms mean. Okay? The terms in a:qlossary it;th, a derinition of what the terms are would be tini, PARTICIPANT: Well, that's whpt a qlop*is~ is. DR..LZOY: Yes. I quess you'.re riqW.:il'm getting tired. I thir,tk that#a a good point. I think another issue was the sshsiti.y#y .iRna•iysis, and I think this is a threat that we have to avkn*Wledqe. some of the s0oakors yesterday, I notioscl, theit t4ay •cauqht it. before I did. And I think it ' s been pretty taken that in some caseR, not only do we havi to deal 1jith sensitivity anaiysis, but in some oases Y think a-;dsqres of ~I1t.~t.(C'Ari ~~~CtKSt 2aC .f:o~l ~-oss, tv

xI-94 1 2 3 4 6 6 7 8 9 lo 11 12 13 14 15 16 17 18 19 20 ' 21 22 23 , concern with the unaartainty in some of t.hos* ipeaRUraments should be at least provided Ln the text whero.pomlribls. I know that's very, very 4d!!libsi7,t ti~v.t*ae in *any . :., inbtanoes uncortainti.s are not reported Q1ea4y:l It's only been in the last couple of years thet people b11q&h..t.o really consider unoertaintios in -their dis`,ouasions. Vilt'.1 think that maybe in' an eppehdix or it maybe in an ax410# you could . .. :~ . run through tOe hypothesis of certa°in types ot 'uli?qoi,tainty, evsn it it's a theoretical analyais:. Z! I havi~ -$us or minus J . : 2o:percent on'two or three parameters of Oorcern, iaaed upon the total weiqht of evidence for those .paramet*=li11.what would , it` do to my ptediotiftac? i tbitsk that 344ht bi: .'helpfal. ' 8ut; again, making sure that &t'e fooui.p.-.Qn the . , , . . . ~. ..•; . .• particular tlalw of th: equation that you're deslifiq with. All right? Mdke suro you're very ekplioit. . ~ x hdpe I've made aryse3.f o;'L!ar. I've U-**q: to .. r rethink how to state this so you at-leagt undexptiqtid.whoro my , . : .. points are coming from. I think that's 1hy fiain 44-13it. OR. :LiP1~4Ar1Nc Dr. Rookat~4. ` + OR. ;ROCKM$s I egz*e,....cbrtA*,oly; vLt.h-'taolOe of ~ho - clarity I thitik from the example -- the aame exatVls Ipioked up that's on page t0 where you introduoe some pgiy~aeters you will introduce and th*n neither Qafined thsm or .~~ us what • - : ~. . .

11-95 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 they were. It kept leaving me hanging where I go to the later sections to try to find out what they were. So I think it could be made certainly clearer. I think the sensitivity analysis is an important part of this, and I think as much as possible, again, that should be put in the tables. I found your one table very useful in regard to the sensitivity analysis. And I think that that's particularly important, given the difficulty you would have in placing any-type of conventional confidence in this thing. And I think, you know, that without knowing what the co-variances are of all these different parameters from one another, I quite frankly don't know how you would put confidence in the convention in the sense that people talk about when they have a well-defined sample space. And I think the sensitivity analysis is an important part of that. I guess one of the things that came up, I found you know, are the assymptions reasonable? I found that the assumptions that you made were put in there, and I felt that they seemed to be pretty reasonable. One that bothered me just intuitively -- and I have no way to know -- I guess it was just one as a lay person that would stand out in my mind as being surprising and there, (07 y OTIf2dtl.CRYl GREf7A4tE'Li Ze 12021 Z96_C2b, Nd; t7031 644-7636

r 11-96 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 didn't seem to be offered much of an explanation -- is why the unmarried never-smokers should have such a higher level -- DR. BAYARD: Why is their level comparable to the females married? DR. ROCKETTE: Well, you made that assumption, but I guess your level is in the -- it was based on the level that was in the blood -- DR. BROWN: It comes from data. Yes. They seem to get more exposure elsewhere. That's just the way the data suggested, that the unmarried women are more likely to be exposed than married women and unexposed married women. DR. ROCKETTE: I guess it just was, again -- DR. BROWN: It needs to be clarified. DR. ROCKETTE: Well, I don't know if you can clarify it. It's just when one -- DR. BROWN: Where that came from. DR. ROCKETTE: Yes, I see where it was made. It just was a little surprising without understanding exactly why that would be. But maybe there's nothing to be done about it. The other issue, of course, was the one of the males, when you applied the adjustment that you had to that, c4112E'LI.CQ.l2 -_RE#04.fEia Ze t2021 296-0261 q~vq 00316qq-7636

11-97 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 of course, the numbers didn't work out. Presumably this can happen with small numbers. One of the things that comes up since you are using your adjustment procedures in Section B, which I guess was brought out before as being non-refereed at this point in time, would it also have turned out that way had you used some of the published methods of adjustment? DR. BAYARD: It wasn't really the method that was different, as we show in Appendix B. It was the parameters of what went into the method. DR. RocxMBs Well, okay. Again, that4a what I suspected, but I think I just wanted to clarify that it's not something that your partioular method of adjustment would have produce. Specifically, I guess, the inaccuracy or the small numbers used for these population -- for these frequency rates that are probably causing the probXem. But, in other words, it's not specific to your method of adjustment. DR. WELLS: It might be specific only to.the extent of the survey's olassification factors and on hindsight. They don't apply to males. We don't have enough never- smokers to measure. And you get screwy results. And I think the indication is you really need to survey from c4mettaan r.Rt#osEcu ze /roe/ sq6-otdi CV.4 /7a9/ 644-mo

r 11-98 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 classification (inaudible). We don't have any (inaudible) but -- so you don't know what (inaudible). DR. ROCKETTE: The other issue, I just reacted a little bit to this, I guess, second sentence in Chapter 6; 6.1. If an effect is large enough to detect in epidemiologic studies investigating the consequences of ETS exposure at common exposure levels, the individual risks associated with exposure is considered to be high compared to most environmental contaminants assessed. I think I know what you're trying to say by that, but it seems to me that when you talk about an effect being large enough to detect, that that also is, of course, intimately tied in with the sample size as well as the magnitude of the effect. And also, when you're talking about it, it is considered high enough to be -- the exposure is considered to be high compared to most environmental contaminants. Of course, that ignores the whole issue of potential bias. I just found that statement to be kind of over- simplistic. In other words, just the fact that I can identify something does not necessarily mean that I can make that statement. There are these other issues such as power, and everything else that comes in. So I just found that o~ Lq o) o, OTI,2EtCCQI2 fRE#04.tE'tS. Ze t2021 296-0261 qo4 /7031 644-7i636

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 _17 18 19 20 21 22 23 11-99 statement to -- I actually said that it's either vague. Or, you know, the way I interpreted it, I'd say it's probably incorrect. I mean, I think I could shoot holes in it being an exactly correct statement. I think that about covers the comments that I had. DR. LIPPMANN: We had two other people who have prepared detailed comments that I know about. Dr. Bill Blot from the National Cancer Institute is going to be with us. He, of course, provides a very nice 2-page summary of what the chapter says, and I won't read that because you've all read the chapter. I'll read his critique. "While the overall estimate of approximately 3,000 lung cancer deaths due to ETS exposure annually in the United States is based on reasonable assumptions, the chapter fails to convey the uncertainty in the estimate. The citation of a range of 2,500 to 3,300 ETS-related LCDs based on varying only one of the parameters involved in the estimation is misleading and implies a greater degree of precision in the estimation than is warranted. "The chapter tends to ignore the sampling variation inherent in the estimate of the relative risk of lung cancer among nonsmoking women married to smokers versus nonsmokers. Taking this variation alone into account would result in a QTIYlG2l.CQ.I2 GRE`J.O'LuL'L 2~e tZOZi P96-oz6, q,o4 /7031 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 II-100 confidence interval based range for the annual ETS-induced LCDs of as low as a few hundred or less, to as high as 5,000 or more. "Other sources of uncertainty associated with each assumption employed will tend to widen the potential range even further. "Thus I would recommend that for the summary conclusions, the 2,500 to 3,300 range of LCDs not be quoted. I would also suggest that the statement 'confidence in these estimates is judged to be medium to high' be deleted. Instead, it could simply be stated that the actual number of LCDs per year in the United States that may be caused by ETS is uncertain, but the best estimate is about 3,000. "In the text prior to these conclusions I would more fully describe the uncertainty in the overall estimate. The use only of data from American studies to estimate LCDs in the U.S. seems justifiable. Data from other countries are relevant to and bolste,r the conclusion that ETS exposure can increase risk of lung cancer in humans (that is, that ETS is a Class A carcinogen) but the magnitude of the effect in the United States seems better estimable using U.S. data alone. "The use of data among nonsmoking women to estimate LCDs due to ETS among nonsmoking men offers the advantage of ~ cn ti ) d7inezl,can 1:::R1r#otfev. 2$e /2021 z96-o26, q~04 /7031 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 the much greater stability of the estimated effect of ETS among nonsmoking women. "The U.S. studies included over 1,000 nonsmoking women with lung cancer, but the number of nonsmoking male cases studied appears to be nearly an order of magnitude less and probably too few to base adequate national projections on ETS-related LCDs. "The chapter cites problems in adjustment for misclassification as the rationale for not using the male data, but a stronger case might be made on grounds of limited information among males and the considerably greater precision of the female data. "Nevertheless, the text should more fully discuss the claim that misclassification bias may be greater for male than female nonsmokers and acknowledge the possibility that the number of ETS-related LCDs among male nonsmokers may be less than, in addition to greater than, the estimate of 500. "The chapteg states that confidence in the estimate of LTDs due to ETS is lowest for former smokers. It makes the assumption that the risk of lung cancer from active smoking largely diminishes after about five years from quitting; an assumption that, depending on how one interprets 'largely diminishes,' is not necessarily true. CO ~m.asi.can ~e~sottczs Ze tZOZ1 zq&oz6, wo4 17031 64a-7636

11-102 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 "Large epidemiologic studies of smokers show that some significant excess risk generally persists well beyond five years of smoking cessation. Although ETS may cause some lung cancer among former smokers, the excess risk from ETS is likely to be less, perhaps much less, than the excess risk from their past active smoking. "Because of the document's stated low confidence in the estimate and the difficulty in estimating directly or precisely the risks of lung cancer from ETS among former smokers, it may be preferable to restrict estimates of ETS- related lung cancer deaths to never-smokers, excluding former smokers, just as active smokers have been excluded from these analyses. "The decision to not use cigarette-equivalent low dose interpolation techniques to estimate the number of LCDs due to ETS, seems appropriate. It is not clear which one or several of the many carcinogens in cigarette smoke is responsible for the ir1creased risk of lung cancer in active smokers, nor how variation in bodily cotinine concentration relates to variation in lung exposure to these carcinogens. Thus, there would be considerable uncertainty in using cotinine or other biomarker data to translate ETS exposure by nonsmokers into cigarette-equivalent exposures. Ln Ln co ogrna-Li.can =Rc#otEeU 1~e 1202l 296-0261 q,6q t7031 644-7636

11-103 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 "The biomarker data clearly show that the nonsmokers inhale and metabolize components of cigarette smokes and help establish that exposure to ETS can increase the risk of cancer. But they are less useful in estimating the number of cancers induced by ETS." He offered a specific response to Question 5. That is, does any of the new information alter the SAB conclusion regarding the categorization of ETS as an EPA Group A carcinogen? His response is, "The conclusion that ETS should be categorized as a Group A carcinogen made in the initial EPA draft document and by the 1990 SAB, remains valid with thee addition of the new information in the revised draft. "The totality of epidemiologic evidence indicates that nonsmoking women married to smokers have experienced an increased risk of lung cancer. Increased risks have been observed in most of the investigations in various parts of the world, including %he United States. "Biomarker studies have shown that nonsmokers exposed to ETS inhale and metabolize compounds in cigarette smoke, leaving little doubt that some of the various classes of carcinogens in tobacco smoke reach bodily tissues. "While low concentrations of tobacco smoke L" W ti dgm.ati.ca.n --R8#os#au M (zozf z96-oz6i r q~04 t7031 644-76.36

11-104 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 components are much lower in ETS than in mainstream cigarette smoke, ETS and mainstream smoke share many qualitative chemical similarities. "Because cigarette smoking is such a powerful risk factor for lung cancer, it is biologically plausible that prolonged inhalation of ETS may result in some increase in risk of lung cancer among nonsmokers. Thus, despite the uncertainty about the magnitude of the risk of lung cancer due to passive smoking, the overall evidence is sufficient to declare that prolonged exposure to ETS is ideologically related to lung cancer, and that ETS should be regarded as an EPA Class A carcinogen." He has further responses to Questions 6 and 7. By and large he is agreeing with the position in criteria document. And 8, he finds again that procedures used are reasonable. And 9 is a restatement of what he said earlier. Other comments have come in from Dr. Burns. His written comments are solely in respect to the questions, rather than a descriptive, but I'll read them to you. His response to question 6, "Is the approach used to derive estimates of U.S. female never-smoker lung cancer risk is scientifically defensible?" his answer is, "Yes. The estimates are based on both human epidemiologic and dosimetry OTI='LCCQI! GRE,fio'CfE'LS `1k' (zozJ zq6-oz6, q'6q /wsl 644-T636 ti

11-105 2 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 data, the database upon which these estimates are based is more complete than is usually available for assessment of population risk for environmental carcinogens. "The approach used of using U.S. data results in a more conservative estimate of the magnitude of the population risks, but the document adequately describes the reasons for using this conservative estimate and the probabilities that the estimate may indeed be an underestimate of the risk to the U.S. population." Question 7: "Is the approach used to extrapolate lung cancer risk from female never-smokers to male never- smokers and former smokers of both sexes scientifically defensible?" His answer is, "Yes. The data that are available on active cigarette smoking do not show substantial differences in the dose-response relationships between males and females for lung cancer. "There are likely differences in the amount in the ETS exposure between males and females in the U.S. population, but the uncertainties involved in these variations are discussed in the document and adequately dealt with in the calculation of population risk. "Once again, the risk estimate is probably more conservative than is appropriate for this issue. However, Ln N ogrneti.can eRe#o:.ttu Ze t202J Y96-m& qGq /7031 644-7636

11-106 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 the use of higher relative risks would generate only modestly larger numbers, and it is a defensible position to adopt the most conservative approach to risks calculation." Eight: "Are the assumptions used to derive these lung cancer population estimates and the uncertainties involved characterized adequately?" His answer is, "Yes. There is an extensive and detailed discussion of the uncertainties and limitations involved in characterizing the lung cancer risk for the U.S. population. The estimates that are derived use the more conservative approach in these calculations, and the limitations and uncertainties of the approach are well described. "Once again, it is important to note that the database upon which these estimates are made is substantially more complete than that available from any other Class A carcinogen, and therefore the risk estimates are able to be delivered with more confidence." Nine: "Is the degree of confidence in these estimates as stated appropriately characterized?" His is, "Yes. The confidence in the female estimate being answer moderate to high is again conservative. The only limitation is really the extrapolation of dose and exposure data to the entire population, and there is now an adequate sample of the d{rnEZi.can ~a~otEets a, ~ L~e tY02J Y96-OZ6, V ~ wdf t7osJ baa-7s36

r 11-107 1 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 various environments and various populations to have great confidence in the estimation of exposure for the U.S. population. "The estimates for male are less certain because of the more limited database on relative risk. However, the description of confidence of that estimate is adequately described." Are there any other members of the panel? DR. STOLWIJK: I would like to bring up a question that I think was already brought up a earlier, and that is the sensitivity of the various parameters on the error of bounds around the parameters, and there are -- for each of these parameters it is essential or a single estimate and there is a range and/or a confidence limit. Many of these are co-related, but some are really quite independent of one another. My question is, are these uncertainties with respect to one another, adequately propagated in all of the analyses that you're doing? Was there an attempt made to do an analysis of the combined effect of several of these parameters? Because my guess would be that that might give you a better idea of the combined effect of uncertainties than you now are able to present. <z;~'LGCQ.12 GRE`2O'LtES.1 Ze t2021 296-0261 q4~q /703/ 644-7636

IZ-108 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 16 19 20 21 22 23 DR. BROWN: We vreated some various exaYaples, and very wide-ranged, because we were essentially treating the variables independently. I'd like to make a comment that these extremes were likely not be immense because of the correlation between the variables. DR. STOLWIJK: Yes. Well, to the extent that they correlated, you are saved by that. But to the extent that they're not, or to the extent that they're partially correlated, you will have to account for the possibility that they combine in unpleasant ways and spread the tota7l uncertainty considerably wider. The question that I had also for myself is one that I think is also expressed by Dr. Blot, and that is that I have a feeling that they are not adequately propagated and that in fact your ranges would widen considerably it you did that. And I was wondering why that hasn't been done because it's a relatively simple thing to try out. DR. BROWN: :I think it goes the other way is what I'm saying, the propagation. That this arrangement we come up with would shrink rather than go up. On just the sensitivity, now. There are other areas of uncertainty that are not quantified, and they would have an effect. And we tried to explain that. c~mcsLcan rJQc~otEcts`.Z~C' /tnrl spb-vsb~ ~~d{ f7royl 644•76.0

II-109 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. LIPPMANN: I think you do it in words and say that if we do this and if we do that -- but we are not made privy to what the combined effects of these would be if they're both -- or several of them are applied at the same time. DR. BROWN: When can we do then to make that clear? DR. LIPPMANN: Dr. Kabat. DR. KABAT: In re-reading Chapter 6 todqy, I was struck by how it really does pile assumption on top of assumption. And one assumption which at the very least needs more justification than it gets is our ability to partition lung cancer deaths into those not at all related to tobacco smoke and into those four different compartments. On page 617 and also on page 625 the same -- this relates to the use of the Z-factor which is computed for women at 1.75. It relates to the application of that Z- factor to men. Now, you may be right that this would only produce a conservative estimate, but you state on page 617 that males have higher background ETS than femalss. But you ought to at least add that based on what we know, males would have lower spousal exposure to ETS, which is the other component of the Z-factor. And that happens again on page 625, just to give 04rrutlca.n G~eriotuts (tot f rq6-os61 . l•.,-a A,r.-YRn-1

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 the whole story. And then I think you should look at whether using the female Z-score for men couldn't produce an overestimate. I'm not certain. DR. BROWN: Are you saying do men actually have lower exposure if they are married to women, or are you just saying that there are fewer men that are exposed? DR. KABAT: Historically, since men over the past 30 or 40 years have had a higher prevalence of smoking and smoke more cigarettes per day, we see in our data, and I think it's reasonable to assume, that women with smoking husbands have more ETS exposure than the reverse. DR. BAYARD: What you're saying is that the Z-value of the smoke -- DR. KABAT: Would be smaller. DR. LIPPMANN: Dr. Samet. Oh, I'm sorry -- DR. KABAT: I was going to bring up the point that Dr. Blot made about quitters, and I feel uncomfortable about the extension to quitters. I really endorse his point, and that came up in 1990. The conventional dogma is it takes 15 years for quitters to -- for their risk of lung cancer to return to that of someone who has never smoked. But in fact, it probably takes even longer. We see that even after 15 years, a quitter has a risk of C4p2EZl.CQ/2 GRE#O'LfE43. Ze tZ0Z1 P96-oZ6, 4~vq 003/ 644-7636

J 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 approximately 2, and most of these estimates don't adjust for the fact that lung cancer cases smoked more amounts than other determinants. So certainly, after five years you don't want to consider an ex-smoker as someone who never smoked. That's it. DR. LIPPMANN: John. DR. SAMET: Well, actually my point was really the same one. You quote the 1990 Surgeon General's Report, which exhaustively reviews this information, and comes to the conclusion that no one -- I mean, in fact, even the question of how you conceptualize what the risk of returning to is a very difficult and nontrivial question which is trivialized in the language here. And I would suggest that the material on former smokers is best excluded, as Bill Blot suggested. Certainly the assumption that the risk is somehow absent after five years is not defensible. DR. LIPPMANN: Mike. DR. LEBOWITZ: Yes. On that same comment, it's quite obvious from a number of studies that former smokers often quit because of respiratory health reasons which are highly correlated with the probability of lung Ca. And that persists throughout their life, as far as we can tell in at eO9rne¢i.can ZRe#o%tZas Ze tZ0Z1 296_026, qoq /7031644-7636 Cn ~ co N

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 least 18 years of our study; that they have a much higher death rate from lung Ca than the never-smoking group. And that has come out in two other studies; one of which I was just reading. I don't have the reference. But in any case, I think the overwhelming evidence is that the ex-smokers are very different. Now, in women it's still one-third to one-half are quitting because of respiratory reasons, and their lung cancer death rates also correlate with that. So, in both men and women, there is too much correlation there to -- too much excess risks throughout their life to consider them ever -- I think, the same as a complete never-smoker. DR. BROWN: We have a little internal disagreement on this issue. My concern was that former smokers may have an ETA (inaudible.) DR. LIPPMANN: Any further Committee input? (No audible response.) DR. LIPPMANIJ: It seems in this case, unlike Chapter 3, that we're talking about some fine-tuning of the chapter, taking into consideration the comments that you've, received. But I don't think the Committee consensus is clearly that it needs only fine-tuning and that the justification that there are some at excess risk will still o4m.Fti.can d?r0otteas ZC r2021 296-0261 q/04 17031 6q4-7i636

) 1 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 stand, unless I stand corrected by anybody on the Committee. DR. EATOUGH: I do have a couple of comments that are somewhat related to points that were made while we were talking about Chapter 3 that probably should have been brought up in connection with this chapter. One of them is one that Kathy made, and that is, that it is certainly true that smoking patterns outside the home have changed significantly in the United States during the last decade, and one ought to be very careful about taking values of Z that are being determined now that are probably much better measures of what's happening now than the former measures were of what was happening then, but nonetheless are not applicable to what was happening then. I think that's an important point to keep in mind as you try to interpret the new data that come to the front as that situation has changed very much. The other comment that I've made in my earlier comments are related t,o the cotinine data, which in a sense really are a measure of cigarette exposure equivalency, even though we've regarded using that directly of the various population groups. And you may want to give some thought to the impact that using the best tracer that we currently have, using a tracer that at least in the conventional CMB analysis OT/f2Etl.CLL/2 lz~?E#AltE21 2~e tYOPl 296-026, q~oq /7031 644-7636

11-114 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 meaning is really not conservative, what that means with respect to data. For example, in fresh sidestream smoke the trail gets really produced, the ratio of mass to nicotine is about 2 to 3. The average ration of 10 is indeed what is seen as an average ratio in homes. And that ratio changes because of the selective removal of nicotine. What that's going to do to the data is that, as you go to lower concentrations and that as you go to more aged material, you're going to tend to have a larger difference between what you infer to be the exposure and what really will be the exposure with the direction being underestimation. What that's going to mean is that you probably have, if you like, a misclassification between those that are spousally-exposed plus background-exposed, and those that are only background-exposed that will tend to drive your calculations to be an,underestimation of the -- rather, an overestimation of the difference between the two and an underestimation of Z. Which means that using nicotine as a marker is, again, I think going to tend to underestimate the overall impact because you're going to tend to calculate values as much actually larger than they really were because OT!)2E4.GCQ.K :rRE12OtEtt1 2~e tZO21 296-o26, q,64 003J 644-7636

l 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-115 of the fact that nicotine really isn't conservative. I think there are data that will at least let you put some kind of balance on the error that that may introduce into the calculation. But the direction, at least as I have thought about it, certainly ought to be that you will tend to underpredict rather than to overpredict the risk. The only thing that will happen then -- the data was presented yesterday, using what I at least considered to be a more conservative, if certainly not a conservative, tracer compared to measurements of nicotine and the relative ratio to Z-values that were obtained were exactly opposite of* what I would have predicted them to have been. But I haven't seen any details of that study yet, so I don't quite know what to make of that. But I think you may want to give some consideration to the effects of the changing composition of environmental tobacco smoke and what it means in trying to calculate Z- factors using a traceZ that's the best that we have, but is not really conservative. DR. LIPPMANN: Okay. I think, as I said, what you have to do here is pretty straightforward. We've taken more time this morning than the schedule indicated. I think it's not reasonable to try and do Chapter 8 before lunch. Chapter a'gme-tEcan cREfio%tz%s Ze 12021 Yqb-oz6, ql6q l7osl baa-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 8 may be an interesting chapter because it's a new chapter. There's nothing like it in the earlier draft. We asked for it. You've done it. And how happy we are remains to be seen. It may in fact be the most important chapter in the document because it comes to some judgments about the more important -- I think -- impacts of the environmental tobacco smoke. So please, let's try to be back here by 1:00 p.m. We have enough time, I think, to bring the discussion to a close on time, but not necessarily a lot. We'll adjourn for this morning and try to resume at 1:00 p.m. or a few minutes after. (Whereupon, at 12:07 p.m., the conference in the above-entitled matter was recessed, to reconvene this same date at 1:00 p:m.) OTI)2S21.CQ/2 GRE#0'LfE4s. 2~e t202l Y96-026i q'04 (703f 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 AfTF $ II 4 0 H S E9914 N ( 1:15 p.m.) DR. LIPPMANN: Mike, we would like to start our discussion of Chapter 8, and I had you as the lead discussant. CHAPTER 8 - ASSESSMENT OF INCREASED RISK FOR RESPIRATORY ILLNESSES IN CHILDREN FROM ENVIRONMENTAL TOBACCO SMOKE DR. LEBOWITZ: Some of my comments have to be taken in the light of the fact that I believe that this may be the most important chapter in terms of bottom line in the document. So I want to see a little bit more thought and some careful rewriting; a little bit more inductive logic than deductive logic. I think that the emphasis on trying to remove the effects of confounding has been a little bit arduous, and in the process of trying to remove some of these considerations the authors have gone a little bit too far. I mean, there really isn't reason to. deny that sometimes others exposures for instance, do have a greater impact than ETS and that sometimes the effect of ETS, as an independent factor at least, may disappear in terms of statistically independence significant contribution. I've tried to document some places where this has OTIIiE4.l.CQ.I2 GRE f1AttE%S. 'L1 C /zoz f zq6-o26I q404 /703l 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 i 13 14 15 16 17 18. 19 20 21 22 23 occurred; where this has been shown, and a couple have been mentioned yesterday as well. Without belaboring it, I don't think that one -- well, let me put it another way. I don't think that that's either undeniable or even when you have enough power that that's necessarily a strange phenomena in terms of the level of contribution to certain outcomes that you're looking at. Sometimes the other exposures can just overwhelm the effect of ETS as an exposure. However, there's one aspect to this that hasn't really been brought up and it should, and I was considering and mentioning it at lunch. That is that sometimes, in fact, the effect of ETS may be an interactive either additive or multiplicitive effect, whereas in one of the studies of ours quoted, in here we saw ETS contribute to formaldehyde exposures in childhood in terms of being strongly related to the prevalence of asthma. This in fact, is still a contribution of ETS, as pointed out, but the effect there maybe is either a promoter or just as an additive effect. In that case, the independent effect of ETS on asthma in that specific study population wasn't found. Obviously, it's much more difficult to calculate d¢me-ti.can ~a~oo~tcts ZC' (zo2f 2q6-oYbi q/04 00316qq-T636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 the risk of ETS as a contributing factor, say, rather than as an independent factor in those cases. But I think you'll find in the literature a number of reports that show that and would strengthen the argument that ETS is important in asthma incidence and prevalence in the same way that it would strengthen the argument that ETS is related say to, say, bronchial responsiveness and cough in kids. And both of those taken together are pretty well indicative of asthma in kids, whether diagnosed or not. You know, not only biological plausible, but as key indicators of the role and the mechanism of ETS in asthma. Now, this takes a little bit more thought. The other areas where something similar happens, where ETS contributes to, say, greater decrements in lung function in childhood or reduced growth of lung function in childhood, i that although it occurs independently, it also occurs somewhat more strongly in those who have persistent wheeze. Persistent wheeze and,asthma aren't absolutely different. It's a question of whether the kid has been diagnosed and under treatment and know whether they make the other clinical criteria for asthma. But in any case, if the effect is there for both, I mean, these things do go together. Therefore, in fact, they OTI'l2E'L(.CQ.12 GREfiA2EE'L'i ae tYOY1 Y96-oY6, q04 0031 644-7636 -j

l 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-120 both add to the biological plausibility of it but also add to the risk, both to the risk of reduced function and to the risk of asthma. And I think it's appropriate to look at that. As mentioned yesterday, in some places some of us think you have to also modify the statements relating to in utero exposure to in fact increase the importance of it. It's not just an effect modifier of ETS and the volume is now called passive smoking in part because we thought the last time that in utero exposure was important and is a form of passive smoking. It's not just what they're exposed to in the environment after birth. So there are a number of approaches that are going to require, I think, more thought to bring out the full relevance of passive smoking and the ETS component of it. There are other statements that I think again reflect an attempt to minimize potential bias. The notion that ETS effects incrgase with age. Well, if the intensity and duration of ETS exposure increases with age, I don't know why biologically you would assume that ETS effects wouldn't increase with age. In fact, lung function is an excellent example in kids where the effect does increase with age. I think what you want to say there is that there ogrneti.ca.n cRe~.os~tu 2)-e tZ0ZI Z96-m6, wo4 00316aa-a60

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 are a lot of effects that occur in early childhood and other effects in fact do increase with age. So again, it's a way of thinking and looking at it and stating it properly. In connection with other parts of this monograph, specifically Chapter 3, in what I envision as its new form, and in Appendix B, you will find that the likelihood of ETS exposure outside the home increases with age too. And time activity studies of children show that although the initial few months usually are predominantly in the home setting, that this starts changing rapidly with working mothers and fathers and with the activity of the child even up to the age of 5. And then you start seeing the increases of the outside exposure factors. One of the areas I think should be more emphasized, in fact, is the exposure in the day care setting. I know that Dr. Martinez and the group with whom he's working on a study of children in fact have shown the effect of ETS in the day care setting on sqme of these factors. This didn't come out and it needs to come out. So that in that aspect, the exposure definitely increased and the effects, therefore, will increase as well within the narrow age range. Getting back to something I said yesterday vis-a- vis Appendix B, the biases in fact do sometimes work toward. Ln ~/~zEtGCQl2 GRE~2otEEt1 ZC (zozf 296-0261 CUd4 /703/ 644-7636

11-122 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 the alternate hypothesis in terms of classification of exposure. And so one has to be careful with that. On the other hand, they're not balanced by the increase in exposure that hasn't been measured here and certainly some of the endpoints in which you concentrate on a very,narrow age -- in fact, as you increase range of when these effects arise, you mention it, but in fact the work has to be put in to show the effects up to 3 years of age, up to 5 years of age, and so forth, to get appropriate attributable risk and appropriate population impact numbers for the endpoints. As it were, in each of the age ranges that you look at you have to consider the'time activity of the child and you're going to have a different Z-factor. In other words, based on that time activity and some possible misreportirig bias. But certainly, the exposure outside the home will overwhelm any misreporting bias of just the mother smoking. I think that would in,fact lead to statements about much greater effects, and I've documented a fair amount of this. The one thing you haven't done in this chapter that has been done in some of the cancer chapters, including Chapter 6, is to try to determine the importance of certain studies and the lack of importance of others. I think you c;0jniEti.cal2 CRzli.ottets Ze 1Y02/ 296-CZ6, q'6q 003/ 6"-7636

11-123 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 should do this a little bit more. You have at your hands in the references already, plus some additional ones, enough good studies so that you can ignore others that appear to be either outlyers or for which there appear to be many doubts. The Striken studies, I can tell you in evaluation for other reasons, do not appear to be good. It was quite appropriate also to leave out the Sally Young study since later data show that they were probably wrong in their conclusions as expressed by them. On the other hand, the studies that Dr. Weiss and his group have done and Dr. Martinez has done, I would say were much stronger and should be relied on a little more. Now, there's one down side to that and that comes when you start estimating the likely number of incident cases of asthma due to ETS. I might add, in•the calculations of that I did get lost a few times as documented; including how you went from prevalence rates to incidence rates and how you determined proportion.exposed to 10 or more cigarettes a day and so forth. Looking at some of what I would accept as Z-factors and the number of asthmatics and prevalence rates and so forth, and looking at the observed risk ratios and what they would be adjusted, and then adjusting at the upper end, such Clqmcti.ca.n Ze t2021 Y96-OY61 q4~q /703/ 6q4-7636 cn o) W

11-124 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 as the 2.5, 2.55 odds ratio that Dr. Martinez found, but then adjusting for mother's education, where he also showed that the incidence rate really -- that the high odds ratio really had occurred only in those with lower education, which represents only 50 percent of the population -- every time I see that, in fact I have seen that, it shows about the same thing. So if you correct for parental reporting bias and education, then my lower limit of the number of new cases of asthma per year is about half of what you show. On the other hand, by changing the Z-factors, you may increase it again. And by changing the lower -- well, removing some of the higher Z, you probably will see a higher rate, especially at the upper end of your estimates. These are the kind of reasonings that I would like you to go through. Let me talk also specifically to the issue raised yesterday as to what are the important facts and, as someone said, the first one is: Were there real causal relationships or where can you say that the relation is very likely to be causal based on Bradford Hill's criteria or someone else's? Well, certainly the LRI, certainly the lower lung function I think is much closer based on what I've suggested d¢»zeti.can cReftotfZts 1.~C /zo2/ 2q6-oz6i w64 00316aa-abs6

11-125 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 today; the asthma. And definitely the exacerbations of asthma. In the exacerbations of asthma, I don't think you spend enough time on estimating likely proportion exposed and how many of those exposed are likely to be exacerbated. My answer is that almost all of those exposed are likely to be exacerbated, whether it's small, infrequent asthma attacks in the mild cases when there aren't any other contributing factors, or large numbers, as found by Evans and the one other study -- O'Connell and Logan. So even with the corrections based on the populations evaluated and the mildness of cases and social status, I get a rate where the minimum estimate is that 37 percent of asthmatics are going to have asthma attacks due to ETS. My higher estimate is closer to 67 percent. I reference my own studies, also, in which I show that if you follow the asthmatics long enough -- all the asthmatic kids that I've looked at -- with good monitoring and get the daily diary of symptoms and peak flow, then in fact, 100 percent of those exposed are going to show exacerbation from ETS at one point or another. And that this occurs at least once a year. DR. BAYARD: Once a what? DR. LEBOWITZ: At least once a year. Not that d¢rnzti.can rRE#ottzss. L~G' (zoz/ 296-0261 4aq (7031- 644-7636

11-126 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 hospitalization does, but that a definite asthma attack does occur, requiring increased medication; if they're not under self-management programs, visits to physicians and so on. I give you reference to those. So for the asthmatic exacerbations and the LRI,s for the reasons I mentioned, my estimates are much greater than yours. And you don't actually calculate the true effect of the decrements in lung function -- and I correct that. So it's larger. For the actual incidence rate of asthma, I think the argument is stronger but the numbers may be different, depending on what you finally end up as a Z-factor and the two adjustments that I mentioned from the studies that I mentioned. I think that the impacts we're talking about, which are not only thousands, but hundreds of thousands in terms of LRIs, are so great and so much more important in my eyes than some of the other things we've talked about. I really think the chapter deserves the added attention that you and your authors could bring to it to clarify it and do a better job of your risk assessment. Thank you. DR. LIPPMANN: Thanks, Mike. Dr. Rockette. c4inEti.ccsn ZRa#o%tZts l.~e 1ZC?1 296-0261 Nvq /7031644-763b

r 11-127 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. ROCKETTE: I guess I had some concerns, as was mentioned before, over some of the way the confounding was discussed, particularly on page 8.2. You have "In summary, there is no single or combined confounding factors which explain the observed respiratory effects of passive smoking in children." Again, if you look at the previous points that you made, many of them were setting up potential sources of confounding and give reasons why they're not great enough. Although, Number 4, where you talk about the effects of passive smoking -- it's not numbered but I guess the one that's the next to the last -- in my mind presents some potential sources of things that could be confounders and I see nothing there that rules those out. That's almost a non- sequitur to have that implying the confounding factors away. Also, I don't think -- and this has come up before a little bit -- I don't know that anything here really addresses this issue qf combinations of confounding factors. I mean to me, if you looked at some of these issues relative to them being potential confounders, to say there's no combined confounding factors that could explain it away, I think if you put in some estimates there, you might get -- I mean, you may argue that you don't think that they're likely PTmEtCCQI2 GREl1.otECt1 ZC /zozI zq6-o26i w6q 1m] 6aa-a6s6 ti (-q 44 N O1 ~

11-128 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 to explain it, but to rule it out I think is a rather strong statement. Also, again on page 8.4, there's a statement relative to the -- I guess it's the 6th line from the bottom, or 7th -- that the evidence available clearly indicates that this is very unlikely. I think you should document that a little better. You give one example about it not explaining the under-reporting at the older age group but, again, I don't know if what you've presented here merits that kind of statement. I think it needs further documentation. I also had difficulty, as was mentioned before, going to these estimates too. This is page 8.10, where you actually get to your 8,000 to 26,000. Again, I was a little bit thrown off track about all the prevalence figures being given and then you jump to incidence. So maybe I'm missing something but the previous reviewer also noted that it's not clear I think. I think most readers will have trouble getting to that 8 to 2L6, particularly when you're setting the stage with prevalence and then jumping to incidence. You know, that transition is -- I think you need to clarify how that number was attained and how you got this distinction between prevalence and incidence. Also on that same page, there's one type of 07 ti ti c4meti.ecsn :rRe#otEeu ZC /zozJ 296-oz6r NGq 17o3J 644-7636

I 11-129 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 statement that actually it's not just in this chapter but it's in others. I think it was pointed out once before. I really don't know how much it adds to these kind of statements where we say confidence in the estimate is medium or confidence in the statement is high. I don't know how others feel. My personal preference is I feel that's a very subjective thing and I think it takes away from just stating what the conclusions are and putting some kind of estimates of error on. Again, that's a very subjective type of opinion and I think it's here and it's also later on where in some of your other estimates the confidence is high. I guess your last statement I find interesting and it caused me to go back to Chapter 7. It's been pointed out earlier that some of these are interrelated. You make the statement that the estimates that ETS also leads to increased prevalence of middle ear effusion. I think that might be still somewhat of a controversial issue as to whether that's true. I went back to your Chapter 7 and looked at those studies you referred to. I think there's some studies, first of all that you're missing, and I certainly didn't do a literature review of that area. Secondly, there is no dgmetican cRefiottcss W /2ozJ z96-oz6i qar{ /703J 644-7636

11-130 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 distinction made, or there seems to be in your writing of some of these like chronic otitis media, otitis media with effusion -- are kind of used interchangeably. In some of those studies, as you pointed out, you did make the distinction of acute otitis media. But then you have the issue of otitis media with effusion, which, of course, in early infants, 80 percent of the kids probably have at least one bout of otitis media with effusion by one year of age, as opposed to chronic otitis media with effusion. I'm not so sure if you sat down and you took all of the studies and did them as rigorously as you did your epidemiological review for the ETS and, as I said, made an effort to get all the studies, I'm not so sure you would be able to demonstrate that they are consistent in maintaining that prevalence or increase in prevalence. And again, when we talk about that I think it's probably important to,distinguish the chronic otitis media from the otitis media itself because otitis media, the fact the kid has one episode, may or may or may not have much ramification and may be more difficult to identify. So I actually found from what you said -- again, I don't think that what was presented is convincing to me that cA='LCCQ.IZ 4d40'itL'L:S. Ze t202f 296-0261 qo4 0031644-7636

1 2 3 4 5, 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 smoking increases the prevalence of otitis media with effusion. Finally, I guess the other thing that we might mention is if one goes back and does the simulations, which was suggested by others on Chapter 6, to get some combined estimates of error for the cancer, then probably a similar type of thing should be done here. I think that's probably all I have. DR. LIPPMANN: We'11 then we move on to Dr. Woods. DR. WOODS: Thanks, Mort. I think the first thing I'd like to do is congratulate you for tackling this. Trying to put an assessment on the various types of non-cancer disease is no easy task and I think that even though this is a first shot, I think that it's a brave effort and you should be commended in attempting to do this. I think you're moving the science forward in the process. In going through the chapter, one of the issues I was curious about is the confounders. My feeling is that the confounders in this multitude of non-cancer type illness is going to almost have to be looked at, and as you go through each disease you're going to have to think about the confounders are. For example, in the asthma, or in the lower (4 c0 ti o0qlrlE'U.CQ.12 GRE#0'LtE'LS 1~. e [2021 P96-OY61 qd{ /7031644-7636

11-132 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 respiratory illness, effects such as temperature and humidity or other types of contaminants, especially microbials and the interactions that might occur, it seems to me could cause you to underestimate the impact of ETS on the effect of non- cancer illnesses. And that has not been addressed. I'm not sure it's quantifiable at this point but I think a caveat needs to be made that the model that you're using, the assessment procedure that you're using now, does not really take into effect some of the interactions that might occur, I think specifically in terms of temperature, humidity and microbial concentrations that probably are interactive with this. Again, on the comment about that no single or combined confounding factor explains the observed respiratory effects, I'm not sure that they have been studied to the point that you can say that with certainty at this point. So there's going to have to be some kind of a hedging with regard to the confounders I think even more so. And then, I think you need to clearly address the potential effect of confounders as opposed to interactions. I'd like to make that specific distinction characteristic between confounders. DR. BROWN: And can it affect modifiers? DR. WOODS: And can it effect modifiers, right. O&EYCCQI2 GRE#0'LtL'CS l~e tZOZ1 Z96-OZ6, q'6q 00316aa-a636

11-133 1 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 think you've done a good job in attempting to deal with misclassification where you could. And again, this is difficult. I like very much the idea that you worked with both a threshold and non-threshold model in trying to deal with the problem and then select which of the two seemed to make more sense under a particular situation. That added power, I think, to the assessment that you're working with. In trying to relate Chapter 7 and Chapter 8, there's actually several areas that you were not able, apparently from the information available to you, to make estimates. But there's not really much of a statement made to the effect as to why you couldn't do it. I think it's important for completeness of the document to indicate why it was not attempted or why you were -- maybe it was attempted and you found out that the data were just not available to be able to give you the confidence to make that estimate, But I think a statement to that effect is important. I think that probably is the extent of the comments I have. I think it's a good first attempt and it's very difficult to do. DR. LIPPMANN: Thank you, Dr. Woods. Dr. Hammond.. LN oO9~nEtl.CQ/2 CRE f2O'LfE'CS S)e t202i 296-OZ6, qo04 1703) 644-7636

11-134 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. HAMMOND: I just have a couple of comments. I think that one does need to be careful to distinguish confounders and effect modifiers and when we deal with panels we need to remember that. I'm not sure that one necessarily adds up confounders, as has been mentioned, and I'm not sure that one expects those to all add up in one direction; to be directional. That's just a comment. In terms of background exposures, I'm a little concerned -- I don't really know where truth lies here; it's just conjecture. But to see such a large percentage of children of non-smokers having exposure surprises me. And I wonder -- it's not clear in the presentation of whether there was any cutoff of cotinine or if any cotinine was translated as exposure to'ETS. If we're talking about very low levels, as some of the techniques now allow us to measure, maybe we need to take into account again the dietary sources that have been mentioned. I don't necessarily think it's a lot and I don't think it is necessary is a major at high levels but at the very low levels. I guess I would like a little bit of that to be addressed so that there is some sort of cutoff. I believe that's all I have to say. OTI=tI.CG.!'L GRB#o'CtE'C& Ze (2021 296-0261 N~ (7031 644-7636 N W ~ cn iLn

11-135 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. LIPPMANN: I have some written comments from Dr Burns to share with the Committee. I think I'll just restrict my reading of it to Questions 13 and 14. On the previous which relate to 7 and 8, he simply says, yes, he agrees with you, so I won't go into that. , Question 13: Is the evidence with respect to maternal smoking and Sudden Infant Death Syndrome properly characterized? Should this evidence be included in this report? His response is: "The evidence is appropriately characterized in terms of the strong association and the suggestion that environmental tobacco smoke may play a role. The conclusion that it is impossible, given the existing data, to determine whether this is an in utero effect or an effect of environmental smoke exposure is appropriately characterized and I believe the data do not yet clearly define the ability to calculate a risk from environmental tobacco smoke exposure. I believe that it is important to leave these data within the context of the report because they address a major concern relative to the exposure to ETS and one upon which public policy may well be made. The fact that there is ample data to raise concern y (~1 P*I2L'LI.CQl2 GRE#o'LtE'CS ZC /202j 296-0261 qd{ 0031644-7i636

11-136 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 on this issue should be brought forward to those responsible for making public policy, even though the data to derive a causal ideologic association are not yet complete." Question 14: Is the presented population impact of ETS on lower respiratory infections and asthma in children scientifically defensible? Dr. Burns' response is: "The major limitations in these data are the limitation and the understanding of the dose and ETS that children, and particularly children with asthma, are exposed to and the prevalence of these problems in U.S. society. Both of these limitations are common in the assessment of any respiratory impact. The numbers that are used for the U.S. population are defensible and similar to the numbers that are used for other assessments and, therefore, it is appropriate to use them for the impact of ETS. The dose and exposure data for environmental tobacco smoke in childxen is again probably better than for most respiratory irritants." Question 15: Are the assumptions and certainties and degree of confidence in the ranges of population impact estimates adequately characterized? His answer is: "Yes. This is the most complete t0 N OTIi2E.'CI.CQ.1't GR-0otEC'LS. 17C /zozJ z96-oz6i q64 /7031644-7636

I 11-137 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 description of the assumptions and uncertainties in the data set that I have ever seen. This is an enormous improvement from the first draft of this document. It provides a solid scientifically defensible base upon which to draw the conclusions drawn by the document." His final words are, "Once again, I regret that I will be unable to attend the meeting but would like to congratulate the EPA staff on what has been a monumental effort that has produced a truly superb document." You have at least one enthusiastic fan. Anyone else? John. DR. SAMET: I'd just like to return to the issue of confounding because I feel that a misimpression has been left, perhaps by the comments of Howard and Jim, concerning the rigor with'which that issue has been addressed in published studies, particularly as related to lower respiratory tract lung function. I think which confounding think if you look example for lower illness and to respiratory symptoms and Chapter 7 did try to address the degree to had been controlled in these studies. I across the array of studies -- and for respiratory tract illness, in my last count related to ETS under age 5, I found 30 -- confounders have Ln to ogrnzzi.can eRa#ott&ts ZC /zozf z96-o261 q~~q 17031 644-7636

11-138 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 been controlled. Perhaps not every study has controlled for every potential confounder, but the major confounders have been controlled and the pattern persists through these studies of an effect. So I'm not sure that there should be concern that somehow there's some confounder that remains across a whole body of studies from different countries that is introducing an exposure response relationship in the case of LRI, for example, that is misleading us. I'd like to hear your response. I understand that there's always the possibility of uncontrolled confounding. This has been, in fact, one of the favorite criticisms lodged against epidemiologic studies from many parties for many years. But it's really sort of a know-nothing kind of approach because it's not based on a biological argument. So I think the issue here is what confounders are you concerned about. And in this literature, the literature specifically on LRI artd on respiratory symptoms and lung function, should we be concerned. Like you, I always share the concern that there's a missing confounder that I haven't controlled for but I need to know what it is and I need to know why I'm worrying about it. I think we need to discuss this because I think C-q PT1rtE'LLCQI2 GRE,h0'C&t1 !be tZO21 296-0261 q/dQ /7031 644-763b I o~ m Cr1

11-139 1 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 Chapter 7 delved into these matters and there's a sentence sort of dangling in Chapter 8 that's not well supported by what comes before it. But in regard to confounding, and I think where we have some large data sets that have accumulated over a long period of time, what are we missing? DR. WOODS: Well, maybe I can take the first shot at it. I guess the though process I was going through is that most of the studies I believe that are reported with regard to children deal with the baseline exposure of residences. So I think we've got to think about what's going on from the environmental standpoint in residences. Now, if you look at other types of studies that have gone on with regard to inhalation, the two types of interactive effects I think that are fairly well documented would be asbestos and radon, and the interactive effects that can occur because you've got the presence of some other factor in there making more than the individual component. So I would look at that as a modifier rather than a confounder. What I'd like to see is somebody address the potential modifiers of what could happen and distinguish that from the confounder, as to something that you can't tell . ~{rrr.~zi.can ~~~.ottcu 2e [2021 296-026, wo4 /7osl 644-7636

11-140 1 2 3 4 5 ~ 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 whether it's this or that. DR. SAMET: Jim, I think these issues are pretty complex both biologically and conceptually in epidemiology. And the example that you raised is one having to do with carcinogenesis in fact. I think what we have to really deal with is with the set of non-malignant outcomes. I mean, I agree that we should be concerned that the combined effects of factors may be different from what we would anticipate from their individual effects. That's what we were really referring to, is the effect modification here. But to the extent that using the available techniques and attempting to control for these variables, either by considering that they may have confounding effects or that they may have modifying effects, and still in effect of the environmental tobacco smoke exposure variable emerges, I'm not sure we're missing anything within the context of present technology. DR. ROCKETTE: I'd like to respond also. First of all, my comments related solely to what was written here and not an evaluation of how they were addressed in the literature. If you look, again, there were five dots given here relative to statements and then in summary of those Ln I-A cn W N Ch 01 ~ // ~ PTmL'LCCQ/2 GRE#0'CtCt1 !be t2O21 Y96-026, m GD q'oq 0031 644-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 , 17 18 19 20 21 22 23 statements, they're saying that there is no single or combined confounding factors that can explain. If I were just reading this, I would think that those five things were all providing evidence to me that I can make that statement. The important one, which is the first comment I. made, in no way seems to me, if what you said is true, that they've been ruled out in all of these studies. Then perhaps the fourth one should say that in addition these factors have already been addressed or controlled for adequately in previous studies and, therefore, cannot explain what is being observed. But it doesn't say that. It just poses that they're potential confounders. I also did not see any of these five points that related to a combination being explained away. So in my mind, without going back and coming up with it and seeing what Chapter 7 says, in my mind I think that that statement being made is still somewhat of a non-sequitur from particularly the fourth point being made in that it does not apply to the combination. And if in fact you're right and there's been adequate studies done that handle all of these, then perhaps those points could simply be strengthened and then the statement could be made. If that's not the case, then the L" cn N ~{rnezi.can t~e~sotta~s. 2~C /2oz/ 296-0261 (V114 17031644-7636

11-142 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 statement should be weakened. DR. SAMET: Well, since this discussion seems to be perhaps misplaced and more appropriately belongs in Chapter 7 and is in fact largely included in Chapter 7, because it has to with the judgment as to the causality of the association and not to the risk assessment itself, it might be better if in fact this were removed and this dangling statement that is troubling you, and your interpretation of it is troubling me, this could be taken out and then perhaps that would solve it. I mean, I agree it's sitting there quite lonesome with a bunch of dots above it that don't have much to do with it. DR. LIPPMANN: Mike. DR. LEBOWITZ: Let me address Jonathan's concerns from a know-something attitude. I agree that it should be removed to put in Chapter 7 for the reasons in the way he expressed. And I certainly agree that in terms of LRIs that the "confounders, even the effect modifiers, et cetera, et cetera," have been handled very well, and possibly only in the case of LRIs. In the case of lung function and in the case or asthma I would say it's not the confounders that haven't been handled but it's the fact that there's new evidence that effect modifiers increase the effect of the ETS, as it were, OT»zeti.can GRc,&ottcts !be tZOZI 29f)-026, qc4 /7031 644-7636

11-143 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 or that the effect of ETS is greater in certain subgroups and that that part belongs in Chapter 8. That is, what is the additional risk to certain subgroups of the population from ETS in terms of decrements in lung function during childhood; in increased asthma rates, both incidents and exacerbations in childhood, et cetera? Part of what I understood from Jim Woods and part of what I was saying is that this chapter should include the risk assessment that includes the increased risk in subgroups of the population, and it doesn't address it. DR. SAMET: But again, I think it may be that you can draw out that certain subgroups are at greater or perhaps lesser risk from ETS because of effect modification. That is relevant if you want to draw the inferences for those segments of the population, not necessarily relevant to drawing the overall assessment of hazards of what they are. DR. WOODS: John, I agree with that. I wasn't trying to say that what is in here is an overestimation of the problem. But it may be an underestimation of the problem. DR. STOLWIJK: Well, I think that the first part of 8.1 of the text before the bullets, and the sentence after the bullets, could be pulled together and they can and should L" d~qi.can --RF-1i.oztZts l~e 12OZ] 296-o26/ ~Ud4 1wsJ 6aa-a636

. 1 ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-144 be dealt with in Chapter 7 where they belong. Then the confusion goes away because here you set off a bunch of modifiers and confounders, or potential confounders, that appears you're not have been dealt with but in fact they were already dealt with in Chapter 7. I think you would clarify the situation a great deal by removing the various strawmen that you have behind the bullets here. Then the confusion that now arises in this discussion wouldn't have been there. DR. LIPPMANN: Dr. Weiss, you weren't particularly assigned to this, but I know this is an area you have strong interest. Do you have anything to add to this? DR. WEISS: Well, I was asleep here a little bit because I thought we had this discussion yesterday. (Laughter.) DR. WEISS: Paul made these points yesterday. I mean, we said we were going to bring it up again today, but I think we're beating a,dead horse here. You know, let it come out and deal with it in the other chapter. It's not relevant here. Just to belabor the obvious here, it seems to me that even some of the public testimony from Dr. Witorsch yesterday made it very clear that confounders cannot explain OTH2E'LCCLtI2 GR~20iLEl& 2e 12021 296-0261 q'64 t7O-qj 644-7636 t0 KJ 01 01 N

r 11-145 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 at all what's going on with LRI. I think that that just reemphasizes the point that John made. I think that there is new information that would suggest that the in utero exposure has some important biologic effects here that is difficult to perhaps from the postnatal exposure, and that may be the most relevant of all of these points. And again, we dealt with this yesterday and Dr. Laties made the comment and I seconded it. So I really think that probably if you just remove it, you won't have the problem. DR. LIPPMANN: What about the rest of Chapter 8? Do you have any problem with the quantitation of the extent of effect? DR. WEISS: Maybe this is just my own problem, but I think that both here and in the cancer area there are so many leaps from the biology, the stuff that we were talking about with regard to dose and deposition -- one of the other areas which I think iq critical is the timing of exposure in phase of human development with regard to active smoking, for example, and it's relationship to lung cancer. I think that those of us that do clinical medicine have seen cases clearly, where relatively small amounts of active smoking at a particularly vulnerable phase in Ln I- Ln oginEti.can eRe#otlxts ZC (zvz/ zq6-o2& q'6q 17031644-7636

11-146 1 2 3 4 5. 6 7. 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 development, adolescence for example, maybe uniquely predisposes an individual to the development of malignancy at a much later date in life. These issues are so complex from a biological point of view that I think it's probably not worthwhile to nit-pick about these estimates. I mean, the numbers here, from a public health standpoint, are quite substantial and most of the issues I think would probably tend to bias the estimates toward a higher level rather than a lower. So I think that within the limits of the risk assessment, and given that you're pretty far removed from the biology, I think that I don't have big problems with the numbers here. DR. LIPPMANN: All right. We're obligated to review the basis and use of science in coming to quantitative risk assessment, which some of us don't normally do. But still, the arithmetic is science -- DR. WEISS: Right. DR. LIPPMANN: -- and we have to say whether EPA is being reasonable. I gather what you were saying was that you feel they are being reasonable. DR. WEISS: Yes. I would say these are probably conservative estimates based on what I presume to understand O7qZEtl.CQn GREfiO'ZEG'C:f. Ze 12021 Y96-C26, Ndq /7o3l 644-7636 ~la

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-147 about the biology. DR. LIPPMANN: Anybody else? Paul? DR. LIOY: In terms of the methodology I think in contrast to Chapter 6, this is much more clearly laid out. I had no intellectual difficulties with going from one place to another. Obviously, you made a number of assumptions, but they were reasonable and credible. Again, I'11 go back to my point from yesterday that the emphasis in Chapter 8 should be reflected by that summary presented in Chapter 7, based upon the emphasis of what the effects concern and the magnitude. And I think clearly that the lower respiratory effects are really, I think, eye openers to me and I think that this has to be seriously considered. Again, I'll get back to the issue of SIDS as being something that I really think that we have to walk very gently on. And I think on page -- I'1l bring it up from yesterday just for the. sake of completion -- I guess it's in Chapter 8, page 13, I guess paragraph 5 -- I really think you should reconsider the thrust of that statement because I don't think that the data supports it. In other areas, I think there's very strong supporting evidence, but I think in terms of SIDS right now I AMt"n L~e 12021 296-0261 w6q t7o.91 644-760 W

11-148 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 think the data is not that supportive, unless someone has other information that would lead me to another conclusion. You made a lot of qualifiers in Chapter 7. I think it's an important point to bring out and it's an observation and something that should be considered and seriously delved into in the future. But I don't think it warrants being in the conclusions at this particular point in time. In fact, I think it detracts from some of the more supportable evidence that you have in here and I don't think it makes any difference to me whether or not that's in the conclusions. It's an important point; something that has to be dealt with; but I don't think it warrants it in the conclusions in the way that you state it. DR. LIPPMANN: Well, that's Question 13 on our charge. Is the evidence with respect to maternal smoking and sudden infant death syndrome properly characterized and should the evidence be included in this report? Is what you!re saying that you believe it should be in Chapter 7 and not in Chapter 8 at all? Is that what you're saying? DR. LIOY: That's my opinion but I'll defer to more experienced colleagues. DR. LIPPMANN: Clarify what you're saying. You're <vlrtELCCQIZ GRF-f204.fE?1 l.~e I2021 296-oY6, q/6q /7osl 6aa-T6-0

11-149 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 simply saying delete the paragraph? DR. LIOY: Right. DR. LIPPMANN: That's your message. DR. LIOY: That's my message; especially in the conclusions. DR. LIPPMANN: That is one of the items of our charge. We have to give an answer. How do other people feel about suggesting that it not be cited in Chapter 8 at all? (Chorus of ayes..) DR. STOLWIJK: I would agree with that. I think it probably should also be removed from the summary. DR. LIPPMANN: Anybody wish to speak for including it? DR. WEISS: I think it's important to distinguish between ETS and passive smoking here. You know, I have not reviewed this literature recently, and I think it's actually written quite well in Chapter 7. But if you look at the number of studies that have looked at in utero cigarette smoke exposure and its relationship to SIDS, the evidence is substantial, I think, that there is an association. Now, going back to the point that Dr. Lippmann made earlier with regard to some of the inhalation toxicology and 44f2Etl.CLiI2 GRE,ho'LfE'LS. ZC /zozf 296-0261 q'6r{ /7031644-7636

11-150 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 its relationship to cell type with regard to lung cancer, I think that there is at least some plausible biological mechanisms as to how in utero exposure might be related to the development of the respiratory controller and how that might ultimately influence SIDS occurrence. I think that it's really stretching it a bit to say that this is ETS-related. But in considering the issue of passive smoking, I think that the relationship is substantial. That's my only point. I don't care whether it's included or not particularly, but I think that I just wanted to speak to the issue of the science as to whether in utero cigarette exposure is related to SIDS, and I think that there is clearly data to support that. DR. LIPPMANN: The Agency clearly has not taken the position that this is an encyclopedic review and quantitation of ETS-related health effects. In fact, it's explicitly excluded cardiovascular effects, which may or may not be substantial. Since we haven't reviewed it, we can't say. So there's no reason why we can't have other exclusions where the data are not at a point where judgments can be made. Mike? DR. LEBOWITZ: I would remove it from 8 and keep it 0 JI)2EZ.LCQ.12 GRE#O'LfE'is 'L1C /zo2J z96-o26i QUo4 /7osl 6aa-7636

1 2 3 4 5 6 7 8. 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 in 7 only because you're not making new risk assessment on SIDS in 8 and I would keep 8 preserved for the things that we feel are strongly related and on which you're doing some risk assessment calculations. DR. LIPPMANN: I think we have a consensus then that -- DR. WOODS: I'd like to talk to this other side for just a minute. I think if you take it out and leave it in 7, there are at least 5 other subjects that were mentioned in 7 that are not assessed in 8. So rather than make an assessment in 8, or if it's going to be left in Chapter 7, what I'd prefer to see is some kind of a statement to say that there's not sufficient information to be used to make an independent assessment. DR. LIPPMANN: On these 6 things, rather than just on the one. DR. WOODS: On the six, right, that's fine. DR. LIPPMANN: There seems to be an easy consensus for us to come to here. Any other discussion of Chapter 8? Jerry? on any aspect DR. WESOLOWSKI: Two points. First of all, I want to congratulate EPA in taking what is a rather a bold but . o¢meti.can ~`a~ot~ts Ze [202J 2t)6-OY6, W /703] 644-7636

11-152 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 very reasonable step that I think will help a lot of risk assessors who can't seem to do a risk assessment unless they have mortality data. So I was pleased with this chapter. There is one error, I think on page A-12, the last sentence. It says "when new cases of asthma in continuously symptomatic children under 18 years of age, we estimate that" I think it should be an asymptomatic. The other thing, as I mentioned to you, Dr. Bayard, yesterday, was on page A-12 where we get LRIs, you stated that the population at risk is 5.5 million and you should use the same algorithm that was used to get asthma cases. I naively did that and got a much different number and you pointed out to me how it should be done. That took about six separate steps because the population at risk is not 5.5 million children but a much smaller amount which are the people who have had attacks. So because this is such an important step from the ART to the number of children, I think that a very simple -- and it is simple -- but detailed discussion of how you to go from ART of .1 to 150,000 would be very rewarding for the readers. DR. BAYARD: Just for the record, though, you're not arguing with the calculation or the way it was (4 L7 14 ti ogmcsi.can v0?,e#o%Eets !be t202i 296-0261 W 00316qq-7636

11-153 J 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 calculated? You just want it more explicit? DR. WESOLOWSKI: No. The calculation is correct; 150,000 is correct. I get that answer now but I didn't get that answer when I used the words that were in there. . DR. LIPPMANN: Any other suggestions or comments on Chapter 8? (No audible response.) DR. LIPPMANN: Okay. Well, this was a new chapter which we didn't have any need to go back to completely rewrite but simply a little more fine-tuning than on the previous chapter. So in terms of the bottom line conclusions about health impacts of environmental tobacco smoke, it appears that there's not likely to be any major changes in numerical estimates or judgment as to which are deserving of quantitation. ' Clearly, there have been suggestions and input to you which will cause you to reexamine many specific points and come to perhaps some different judgments or numerical predictions on some of them. Clearly, the statements in the draft remain those of draft nature and are subject to final,,i calculation. J COMMITTEE REVIEW OF CHAPTERS 1 AND 2 DR. LIPPMANN: Having come this far -- and I'm glad OTl12EtGCQ.I1 GRE f2o'LtE'L1 ~LIC (2021 z96-oz6i q'6q /9031644-7636

11-154 ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 we're at this point at this time, we can turn our Committee attention to the review of Chapters 1 and 2, as appropriate, and our decisions and commitments for our own committee business of writing our report to the Administrator and finishing our task in a reasonable time. The Science Advisory Board, of course, determines its own agenda and schedule as it sees fit. But clearly we do take into account the requests of the Administrator, whom we serve when we're here. We will make every effort that we're likely to be able, to come up with our report on a relatively short time schedule as SAB reports go. I know the Administrator has expressed the desire that we have our report ready for the Executive Committee review in late October, and I see no reason why we can't do that. I don't see any lack of consensus on most issues. We have had a hard working committee who has for the most part already prepared written commentary, which will require relatively little modification, some addition, and I imagine the committee members will be able to get their additional material and/or their modified material to Mr. Flaak within the next week or two and he can work with me on pulling it together into a draft report. I think the order of our business is ogmezi.can ez~stau ZC /zoZf Pq6-oz61 q~vq /7osl 644-?636

11-155 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 straightforward. our charge is very specific in terms of the 15 items which have been laid out. Most of the panel has already addressed them explicitly and from what I've seen we don't have very many inconsistencies to work out. Those will be our task to work out to the extent that we do. I think we directly or indirectly addressed each of these 15 items as we've gone. And if you have any questions about whether you've gotten the message on our views on these items, perhaps you should ask us-now. We then have a fairly clear understanding of where we've come from and where we've been the last few days and of where we're going. It's clear that we have some difficult issues to wrestle with still, as a committee, in coming to consensus judgments. It's apparent from the consensus judgment of our last meeting that there is some excess of cancer associated with spousal smoking remains. The number again can change, in particular with the previous draft. We do encourage you, when you do some of the additional calculations, to not only come up with your best central estimate but the range of uncertainty. I think, clearly, since our collective judgment is 23 1 different from that of a good number of the public ~ ~ N O7I'IZEYI.CQ.12 ~E~70'LtCt1 ZC /2021 2q6-o26i wc4 /7031 644-7636

11-156 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 commentors, I think perhaps I should try to reiterate where I think this panel is coming from, ask for any comment from other members of the panel and some reaction from you. I think a lot of the confidence on which we come to this conclusion that there is an excess is largely on a basis of the coherence of the data and the total weight of the evidence. That is, we clearly are not happy with the judgment in Chapter 4 that just because mainstream smoke is a carcinogen, and there are some of the same materials in sidestream smoke, that therefore ETS is, on that simple basis, a human carcinogen. Clearly, the coherence demonstrated in the discussion of Chapter 4 lends weight to our belief that the relative risks in Chapter 5 are plausible. Together they give us that degree of confidence. I must say that this is a very different issue in my view from the other Class A carcinogens, which are largely based on human epidemiology with larger relative risks but of much smaller populations. This is in the class issue, it seems to me, that we deal in the criteria pollutants -- ozone, the TMSOX and lead -- of which there are no studies that look at one material alone or even one material that's dominant but are looking at a situation with relatively low relative risks and relatively y N U? co PTIf2E'Lf.CQ.l2 GRE#A'LtE'Ci. Ze t2C21 296-0261 q/04 003) 644-7i636

r 11-157 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 _17 18 19 20 21 22 23 high numbers of exposed people. The weight of the evidence really determines the extent of belief that there is an association or not. And so, one can say that as in those pollutants, as well expressed by David Bates, I believe, that when you get to coherent evidence from many sources, the weight of the burden of proof should logically shift from those who say you must be absolutely sure to those who want to deny the weight of the evidence and say because you can't be absolutely sure in the calculation based on many uncertain inputs, that, therefore, you can't do anything by concluding that the risk is there. At least in my personal view, this is a situation where the coherence of the overall evidence is so nearly complete and the potential impact is so high that the burden of proof should be on the other side. Therefore, not only must there be bias, but it must be a differential bias which is hard, in my view, to establish a basis for. So I would say that one could interpret the criteria for causality in the guidelines to be that they're met; that there is no identified bias that could explain the association. There I emphasize the word identify. ti Ln 44 Amcttcan eIQc,&ottcu Ze (202l 296-o26, q~vq 0031644-7636

11-158 1 2 3 4 5 6 8 9 10 11 12 13 14 19 20 21 22 23 You have indicated the biases you can identify and have given, in my view, a reasonable account of them and whether they're likely to be differential and I accept your conclusion that there aren't. The possibility of,confounding has been considered and ruled out. It's hard to really say they've been ruled out, but on the other hand again, I see no evidence that confounding is likely to overwhelm the judgment. Clearly the third item, that the association is unlikely to be due to chance, is an extremely compelling reason for making the judgments that you have. Some may argue that we have not come to a sufficiently compelling case for these three criteria being made, but that's for a matter of individual judgment to decide. This is, as I say, not the normal carcinogen risk assessment but one based on a coherence of many, many studies, each of relatively weak power but where the most likely confounders have been studied in greatest detail, as in Fontham, it doesn't seem to shift the results from those where there's been less attention. I invite other members of the committee to comment on my comments and to disagree if they choose and see if we need to modify the consensus I've tried to state. L" H ~T ninattcan GRE#os.tzU Ze t2o2i Y96-OY6, q04 0031644-7036

11-159 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 DR. STOLWIJK: Well, I would most probably endorse to the comments you made, Mort. I think the only reservation that still remains with me is the calculation of the likely population impact and developing the narrow ranges as I've indicated here. I have some discomfort with that, although I cannot point to a particular reason why they are there. But I would like to make sure that these uncertainty ranges are carefully explored before they get put into a final document. They look to me narrower than they have a right to be. That is based on instinct as well as other opinions of clarifying simultaneous uncertainties. DR. LEBOWITZ: Jan, is that strictly for cancer, lung cancer? DR. STOLWIJK: Yes. DR. LIOY: Just cancer? DR. STOLWIJK: Yes. I think that the situation is most serious with cancer, although it propagates to some extent into the other diseases also. But I think that the likelihood that there are a great many things simultaneous tracking the other things, it's just not as high. DR. LIPPMANN: Anyone else? (No audible response.) P-A ko N ogm.zucan z~?ehottEu ZC (zo21 296-0261 Q'aq /7031644-7d36

11-160 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 DR. LIPPMANN: While we still are in committee then, let's turn to some other committee business. We're early enough to engage in that. Does anybody want to offer specific comments now? DR. STOLWIJK: I have a comment on Chapters 1 and 2 that has perhaps some effect throughout the report. It is a matter of language and specifically language that could be identified and interpreted as being expressing the opinions of the authors, or the hopes of the authors, or the fears of the authors, and they deal with an excessive use of strong evidence instead of just evidence. The words "strong evidence" appear -- the association with strong evidence appears in a large number of cases and there is a tendency in general to put adjectives there that reflect a value judgment as opposed to just a presentation. I think it is not helpful in documents of this type to have an excessive use by the authors of imposing their judgments or values in statements that are being made in the document. It invites the suggestion that there was a bias in the writing, that there was a hoped-for result or a planned-for result. I trust such biases did not exist when the writing was done, and I think the appearance of such biases or wishes PTIYLE'tI.CLLI2 G~E~20'C~CY~ ZC tzvZf zq6-oz6, q,6q /7031644-7636

! 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 existing within the minds of the authors should be avoided. And it requires if you do a search through the text where you look for "strong" and "highly" and everything else, you'll find that are a very large number of occasions where you can omit the word. DR. BAYARD: I'm having trouble with this, Jan. DR. STOLWIJK: It is there in such magnitude that I though there might be trouble with it. But I still suggest that -- DR. BAYARD: It was obviously a conscious effort to try to characterize the evidence as we felt, as we tried to establish in the report, so you wouldn't just get a (inaudible) stating evidence. DR. LIPPMANN: But you don't need the adjectives. If you end up with the numbers, they speak for themselves. DR. BAYARD: That's right. I won't argue that. DR. LIPPMANN: And there are uncertainties, of course. DR. BAYARD: That's right. I'm referring to the extent that we made a statement of the cases of asthma and such. We didn't feel that the evidence for asthmatic (inaudible) exacerbations. So just stating the numbers without characterizations of how we felt about that would be OT!'12E'LI.CQ.I2 GJ`E#OttEtS Ze 1202/ 296-o261 wGq 0031644-7636

11-162 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 inappropriate. DR. STOLWIJK: But if you then say one line of evidence is stronger that the other, I would be happy. But if you say in one case it's strong and in the other case it's even stronger, that comparison doesn't upset me. But when you have summary and conclusions and you find a substantial number of these occasions, which I will hand in so you can look at them -- but I'm giving it to you as a general concern, probably grew from being a conservative person who does not put too many exclamation marks in his writing, I think the disappearance of many of these words would not hurt the document and it might actually make the document somewhat more evenhanded in the way it sounds. DR. LIPPMANN: I think we already got this message in part in reference to words in other chapters. As was mentioned yesterday, your summary and conclusions chapter is the one that's most likely to be read by large numbers of .people and it's especially important there not to go beyond the evidence. DR. STOLWIJK: I have one other plea that applies throughout the document and that is the use of exposure response terminology. Somehow it offends me a little bit to think of an exposure response in a population being there's y PTIfLELCCQIZ <Z1?L#O'CfE'ks. Ze 12021 Y9&OY6l qGq /7osl 6aa-ae36

) 1 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-163 an excessive number of people who die of lung cancer. That's not the kind of response, I think -- that's not the kind of terminology that I would like to see. Exposure and effect would perhaps be better. I don't know. But that again, a searching could probably take care of almost all of that without too much difficulty. I don't know how the rest of the panel feels, but exposure and effect gives me more solace than exposure response. DR. LIPPMANN: Paul. DR. LIOY: I think the way that the judgment -- you know, stronger or strong, weaker -- I think should be based upon some kind of supporting evidence that you put in. If it's going to be put in the conclusions, you'd better back it up to say supporting evidence-- you know, don't just say the words. I understand where Jan's coming from, that sometimes you make a statement about "strong" but I don't see the rationale clearly,presented. I think that in an executive summary of that sort, if you're going to make that kind of statement, that it should be supported in that statement. It doesn't mean it has to be 14 paragraphs but there has to be a concise statement behind it saying exactly why. And I think that will, again, make your use of the word p~t[.CCLI2 GREl2o'LtE'L,ti ZC /2vzf zq6-oz6, 'V1=4 003) 644-7636 .

11-164 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 "strong" more prudent and more effective. DR. LIPPMANN: Kathy. DR. HANIIMOND: Along a different line. I actually am a strong advocate in general, not just because of yesterday, in having an executive summary which is accessible to the layman. I think that the suggestion was very good, that we have a science writer or someone who can help. Now I'm speaking for myself and not the panel at this point, although someone else may want to add to that. But I myself think that, especially when there is public interest in something, it's good to have a document that is written the people who are the authors of the document that people can read and understand as distinct from someone else translating it. DR. LIPPMANN: Jim. DR. WOODS: I guess I have two comments and I would sure second that point. In the summary and conclusions, I had some trouble trying to separate the summary from the conclusions. I'm not sure you wanted to do that but it seems to me that this document is massive enough that you need to be able to summarize and maybe have either a section in the summary and conclusions that would be the conclusion section or that you o4nzezi.can lz~?&fsottzts `1X' tzo2f zq6-oz6, q'a4 /7osl 6aa-760 ,

11-165 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 ,17 18 19 20 21 22 23 have a separate chapter of conclusions that might even -- I don't know that we're going to have recommendations, but at least separate out the conclusions so that they can be highlighted. The other point is that in the introductory chapters, Chapter 1, Summary and Conclusions, and in the Chapter 2, Introduction, I think an overall objective statement is really needed. What is this document about? That should be clearly stated. The only place I could find that is to kind of wade through the preface a little bit and you see it there. But I think it needs to be in the body of the material; you know, specifically what is it that's going on here. I guess the last point I'd like to make is in the summary and conclusions section and in the introduction, I think you really need to try again to more clearly distinguish between passive smoking and sidestream and environmental tobacco.smoke. I read that and I've read it over and over again to see if I'm missing something. And now this afternoon we heard a little bit more about the in utero effects, which can be a passive smoking phenomenon. I think that we really need to distinguish between these term and get some concise cn ~ ogrne-ti.ean :.-RE,&ottets 1'~G' /zoz) zq6-oz6, N64 /7031644-7636

11-166 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 definitions into the introductory paragraphs, both in the summary and conclusion and in the introduction chapter. DR. LIPPMANN: One other thing. I think the only chapter we're anticipating a really major rewrite is Chapter 3. And I don't think that justifies us suggesting that you need to have another public review at all. But on the other hand, there's sufficient amount of work to be done that you may want some SAB committee input into it. I suggest that you decide how much of that you want and we can try and arrange for some further guidance from a subcommittee of this group on that chapter. DR. BAYARD: Thank you, very much, Dr. Lippmann, and thank you very much all the members of the committee for your hard work and great effort, and the sound advice you've given us. On behalf of my office in ORD and OAR I want to thank you. DR. LIPPMANN: Thank you. Bob, do you have any leftover housekeeping? MR. FLAAK: No. DR. WESOLOWSKI: I want to make one last statement. We earlier talked about the chapters and we wanted -- on the health effects on adults and health effects on children. I want to make the suggestion, but there seems to a better way PTr12ER.l.CA.K GRE~AYECRS L~e 12021 296-0261 q/c4 t703/ 6q4-T63b m W ~ra

1 2 3 4 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-167 to classify and also take away some of (inaudible) of lung cancer. DR. BAYARD: I think so. DR. WESOLOWSKI: If you do that, then I presume that the title page should also change to Health Effects of Environmental Tobacco Smoke on Adults and Children. DR. KOPPIKAR: Respiratory health effects. DR. WESOLOWSKI: Yes. Thank you. OTHER ISSUES AND FUTURE BUSINESS DR. LIPPMANN: Turning to other business, Dr. Stolwijk has drafted a statement entitled Reduction of Population Exposure to Air Toxics. Everybody on the Committee, if they haven't looked at it so far, at least has it in front of them. For those in the audience who chose to stay, let me briefly tell you what we're discussing. As many of you know, the Clean Air Act Amendments of 1990 had a major s9ction on air toxics. It identified 189 chemicals for which the EPA is required to apply best control technology initially and eventually write criteria documents on each of those entities -- chemical, or a chemical mixture or a group of chemicals -- and to come up with residual risk calculations. cn J colbnutican ~e~sotf.Et sEts Ze tzo21 Q96-o26, q~a4 0oAI baa-~~

) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 r 11-168 This is a tremendous burden. I don't know if Bill Farland will last long enough to ever see the end of that exercise, if he has to in fact do it. But it's a logical absurdity in that to control the outdoor air of most of these chemicals when the dominant exposure to these chemicals of people is indoors and in fact those of us exposed indoors become sources perhaps requiring control when we ventilate outdoors. As a committee on indoor air exposure and total human exposure, clearly this causes us some concern about whether EPA is going to be devoting its scarce resources in an intelligent way as a contributor to the reducing risk report. It clearly in my mind would not be. And Dr. Stolwijk proposes that we examine this draft statement and, if we can endorse it, it could form the basis of a letter to the Administrator telling him what we think. A slightly analogous situation arose recently in SAB where the Radiatiqn Advisory Committee_sort of rebelled against helping to construct drinking water standards for radon because under the provisions of the regulations for drinking water standards they were obliged to come up with drinking water content of radon, which added a less than 1 percent incremental risk to that which is associated with the 07/12EZl.C6LI2 GREf].02ELt6 Ze 11021 Z96-C26, w64 00316aa-a636 ~ ~ ~ N

11-169 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 4 peak per liter guideline for indoor air radon. So this was an absurdity which needed to be pointed out, and it would cost perhaps billions of dollars in cumulative cost to water treatment authorities who would be obligated to remove radon from drinking water to remove a negligible incremental risk of radon. The Administrator has been very anxious to be advised on some of the absurdities that go on in the Agency because he is committed to the philosophy of reducing total risk. And if it requires legislation, at least he can ask for legislation from the scientific base, and give his priorities to the most important things. So we don't have time to discuss this in committee since you have all just gotten it, but I ask each member of the committee to look at it in the reasonably short-term and to provide any comment to us. If you endorse it or endorse it with only modest changes, we can send that letter to the Administrator and ask,for a discussion with him or his representatives on that issue. As the standing members of the Committee already know, we have sent sometime ago a letter to the Administrator on asbestos and substitute fibers and are awaiting a response as to how we can help deal with some of the'absurdities in . o4mett.can _-~ostzts Ze tZ0Y1 Z96-o26, N6q /7csl 644-7636 N

I 11-170 1 2 3 4 5 6 7 8 9 10 11 12 13 14 19 20 21 22 23 the asbestos program in terms of providing scientific advice on overall reduction of risk in a cost-effective manner. My hope is that we will be able to get sufficient expression of interest from the Administrator to meet with him or his designated deputy at a meeting in September at, which time we will perhaps also meet in committee to polish up our report from this review to get it to the Executive Committee by the first week in October. So we don't have any committee date established yet, but keep in mind that we may want to get together on these issues, depending on the one hand on whether we need to meet in committee to polish up our report and, second, whether the Administrator, or any of his people, are willing to meet with us to discuss what role we might play in meeting the Agency needs on these two issues. DR. DAISEY: Can I just raise a question? The Clean Air Act, of course, came out of Congress and it has a little bit of the flav,or of Congress legislating science to some extent and a number of scientific organizations spoke out against that particular recommendation. I guess I'm asking whether the Administrator is the appropriate audience for these comments or whether it should be addressed to the appropriate congressional committee. In ~ N Lq %0 ~ N O1 d¢meti.ean V1Qc#osteu O1 w !be t202l 2qfl-0261 ~ q/04 /7031 fiqq-7636

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 a way, the Administrator has his hands tied; Congress has told him to do this. DR. LIPPMANN: But Congress is aware and interested in the Reducing Risk Report which has come out since that Act and I think might well be responsive,in this era of diminishing resources and to perhaps go through the Administrator in getting to Congress. At least that's my perception. DR. DAISEY: I just raise that issue as a procedural one, not one in substance. DR. STOLWIJK: We normally report to the Administrator and not to a congressional committee. DR. DAISEY: That's true. MR. FLAAK: Although I might point out that the SAB committees in the past have had mandates to respond directly to Congress, including this Committee in its initial report on the implementation of the Indoor Air Program. So there is precedent for and, in fact, often requests from Congress, particularly several of the committees, for the committees of the Board to provide direct responses to them. So this has happened in the past, although generally committees don't do this unsolicited. DR. LIPPMANN: Because it has a separate charter. QTlfZEtI.CQ.I2. GRE#A4tCL1 2~L' /zo2J zq6-o261 wo4 /7031 644-7636 i

11-172 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Jerry? DR. WESOLOWSKI: Well, I read this quickly. It's a nice letter, Jan. The question is a little bit vague as to what we're asking the Administrator to do, except take into account the total exposure. We can probably call on people and say what did they mean and some of those people might say, well, go after funding the NEXAS project. Is that the bottom line here? DR. STOLWIJK: No. I think the bottom line, Jerry, would be that there are avenues within the Agency that can in fact make a direct contribution to the reduction of these exposures by going through the Indoor Air Office or other similar arrangement. It doesn't have to be through regulation that this is accomplished. As a matter of fact, it's probably not possible to accomplish it through regulation. PARTICIPANT: But there are non-regulatory ways in which the EPA can act. DR. WESOLOWSKI: But you do ask in here for an effective scheme for monitoring the population exposure and this goes back to a recommendation that I made a number of years ago, and others as well. To take some of the some of the money from your fixed station monitoring network and ogmeti.c.a.n 1ZRE/i.ottet& 2~e t202] 296-o26I q'6q 003] 6aa-a636 m

11-173 I 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 close some stations and instead do maybe a biyearly random sample of the population in several cities in order to determine what their exposure is in various toxic and bacterial pollutants to determine if you strategies are working. DR. STOLWIJK: But I think that introduces an element of being able to be held accountable for the effectiveness of what you're doing. DR. WESOLOWSKI:• From the point of view of exposure and hence, health effects? DR. STOLWIJK: From the point of total population exposure reduction, right. DR. LIOY: Well, Jan I think that what Jerry's leading towards is this in a sense is a supportive statement of the ideas of NEXAS, the National Human Exposure Assessment Program. DR. STOLWIJK: It would give a task to NEXAS, yes. DR. LIOY: The question I have is right now I think most of you are aware, or some of you are aware at least, that the NRC is looking at residual risk portion of these 189 air pollutants and how would we interface that? Because it's very important, I think, that SAB perspective be brought on this issue, and I think Jan's statement is very reasonable OT1rtE4GCQQn d?SfiOttE'LS. Ze t202l 296-0261 Qod{ /703/ 644-7636

r 11-174 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 and clear. DR. WESOLOWSKI: NEXAS I think at the moment is still a concept looking for actualization in -- (Interruption by audio interference.) DR. WESOLOWSKI: -- including the one who said that environmental tobacco smoke should be one of the pollutants of interest, which makes sense, as we've all heard today. Now, if we write this letter, then in a sense we're saying without having formally met to talk about the NEXAS Program that we are endorsing it. DR. STOLWIJK: No. We're saying there's a task to be done and if NEXAS is applied to it, then that would be one way of doing it. There are other ways of doing it. You could have another team. DR. LIPPMANN: Consider this. Tell us what you think; if you want to draft and alternate statement. I think the point is well taken that indoor air and total human exposure, which is why. we're a committee -- and we think those are both important things -- are inconsistent with this regulatory approach and we'd like to offer our services to the Administrator to help him come up with an Agency position. 23 1 I don't see where that directly impacts the cn W N ,vq1'lt8'ii.can GRFOo'LtE4s ZC t202J 296-0261 Nd4 00316q4-7i63G

J 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-175 National Academy Committee who can listen to us or not listen to us as they see fit. DR. LIOY: Oh, yes. I don't think that's a major -- but that's a point of information that's going on at the present time and just be aware of it. MR. FLAAK: I might point out that the Agency expects to present the exposure survey to the Committee sometime in the December or January timeframe at a formal review. DR. LIOY: I'll be in a workshop tomorrow and Friday on that topic. It'll be interesting. DR. DAISEY: Could we get a list of the 189 toxics? DR. LIPPMANN: Do you really want it. DR. DAISEY: Yes, because not all VOCs have been measured in the homes and I want to make sure that -- DR. LIPPMANN: They're not all VOCs. DR. DAISEY: Yes, part are called toxics. I just want to make sure that -- sort of the refinement of the letter, not the substance. The substance of the letter I agree with. DR. STOLWIJK: The letter is a 2-page thing which could probably be expanded to 4 or 5 pages or so. DR. DAISEY: It doesn't even have to be. og»seti.can cRe#osfet1 ZC /zozf 296-oz61 q~vq 17031 6aa-7636

1 1 2 3 4 5 6 8 10 11 12 13 14 15 16 17 18 19 20 21 22 23 11-176 DR. STOLWIJK: And it could probably take care of the various (inaudible) to present the insights that Committee has. It is just that such letters do not get spontaneously developed unless there is a proposal -- DR. DAISEY: No, I agree with the thrust of the letter. - DR. WESOLOWSKI: When is this Committee going to meet on NEXAS, in September or October? MR. FLAAK: The Program Office has advised us it won't be ready until around the end of the calendar year. DR. WESOLOWSKI: So when we meet in September it will be on asbestos? DR. LIPPMANN: We may or may not have a meeting, but if we have a meeting, it would be hopefully on both asbestos fibers and as a last chance to polish our report on this review. MR. FLAAK: When Mort talks about a meeting, he's talking about the standing committee at this point, the meeting in September. We'll evaluate the availability of that information of whether or not we can in fact hold a meeting, in the next week or so and check back with you for dates if in fact we're going to do this. If not, the next meeting of the Committee PTH2Ctl.CQ/2 ::RE/20'LtEti Ze tZOZ1 Y96-0Z6, Nd{ hosl 644-7636

11-177 1 2 3 4 5 6 7 8 will probably be in December or thereabouts, depending on when these other issue become available. Perhaps sooner. DR. LIPPMANN: Thank you all for your very good contributions to a productive review and discussion. It was a pleasure working with you. I guess we are adjourned. (Whereupon, at 2:50 p.m., the hearing in the above- entitled matter was adjourned.) OTrl2E'LLCQI2 'Vi?E`20'itC'i1 Ze t2021 296-0261 q~vq t703/ 644-7636