vvEPA
United States
Environmental Protection
Agency '
Office of Health and '
Environmental Assessment
Office of Atmospheric and
Indoor Air Programs
Washington DC 20460
Research and Development/Air and Radiation
Health Effects of
Passive Smoking:
Assessment of
Lung Cancer in
Adults and
Respiratory
Disorders in
Children
EPA/600/6-90/006A
May 1990
External Review Draft
Review
Draft
(Do Not
Cite or Quote)
Notice
This document is a preliminary draft. It has not been formally
released by EPA and should not at this stage be construed to
represent Agency policy. It is being circulated for comment on its
technical accuracy and policy implications.
-------
TO: OTS LIBRARY STAFF jyj^j 2 5 |QQn
FROM: TIM
OTE THIS F.Y.I.:
EPA AND PASSIVE SMOKING
ACTION
On June 25, 1990, EPA transmitted to its Science Advisory Board (SAB) for review
and comment two draft documents on Environmental Tobacco Smoke (ETS) entitled
Health Effects of Passive Smoking: Assessment of Lung Cancer in Adults and Respiratory
Disorders in Children (EPAI600I6-90/006A) ana Environmental Tobacco Smoke: A Guide
to Workplace Smoking Policies (EPA/400/6-90J004.) Companion notices in the June 25,
1990 Federal Register also announce a simultaneous public review process, commencing
June 25,1990 and ending August 31,1990, in w ich the public is invited to comment on the
draft documents.
STATUS
Both documents are public review drafts. They have been released by the Environmental
Protection Agency only to solicit scientific and public input on their contents and, therefore,
do not represent Agency policy. Consequently, it is inappropriate to quote or cite
information from these documents until they f^e released in final form by the Agency.
DESCRIPTION OF DOCUMENTS
The first document — Health Effects of Passive Smoking: Assessment of Lung Cancer
in Adults, and Respiratory Disorders in Children —proposes to classify ETS according to
EPA's carcinogen risk assessment guidelines, tn estimate the excess lung cancer deaths
attributable to ETS exposure, and to assess the association between passive smoking and
respiratory effects. The draft risk assessment was prepared by the Human Health >
Assessment Group of the Office of Health and Environmental Assessment of the Office of
Research and Development at the request of tli3 Indoor Air Division of the Office of
Atmospheric and Indoor Air Programs in the (Office of Air and Radiation.
The second document - Environmental Tobacco Smoke: A Guide to ' Workplace
Smoking Policies - is intended to provide go /eminent and private sector decision-makers
with information on the technical basis for cc itrolling involuntary exposure to
environmental tobacco smoke and to describe a variety of technical and policy options for
instituting effective smoking restrictions. The guide to workplace smoking policies is based
on the overall body of literature on passive smoking, including the 1986 reports of the
Surgeon General and the National Research Council. Its review has been timed to coincide
with the review of the risk assessment in order to ensure that up-to-date information from
the risk assessment would be incorporated into the guide to workplace smoking policies.
The draft workplace policy guide was prepared by the Indoor Air Division.
A third document under development — but not yet available for public review — is
a Technical Compendium of Information on Environmental Tobacco Smoke. This
document was jointly conceived and funded by several agencies of the Department of
Health and Humanjjervices in addition to EPA, including the Office on Smoking and
Health (Centers fofDisease Control), the Office of Disease Prevention and Health
Promotion ^Public Health Service) the Heart, Lung and Blood Institute, and the National
Cancer Institute (National Institutes of Health). The compendium consists of individual
-------
chapters by several authors summarizing existing information in a number of areas
relevant to ETS, including exposure, health effects, risks, public attitudes, and smokuig
policies. The technical compendium has already been subjected to a limited external peer
review and its chapters are now being revised to incorporate changes. When those
revisions are completed, EPA's Science Advisory Board will also be given an opportunity to
review this document before it is finalized as a multi-agency publication.
OT oFthese documents are being devloped under the authority of Title IV of the
Superfund Amendments and Reauthorization Act of 1986, which provides EPA with broad
authority to conduct research and disseminate information on all aspects of indoor air
' ''
REVIEW
The Health' Effects of Passive Smoking: Assessment of Lung Cancer in Adults and
Respiratory rQi§o/d£t? in.? Children and:The Guide to Workplace Smoking Policies will be
the subject of an SAB* review meeting. A separate Federal Register notice announcing the
, dates o|Jhej;SAl^jeTOj(8w.-rfleeting will be published^shortly.
HOW TO OBTAIN COPIES OFTHE FUBIIOitEViEW DRAFTS
A"Jimit«±riuft\l>er'eFcot^e3"are available and may be obtained by contacting:,
Office, CERI-FRN
f. U.S. Environmental Protection Agency
"" 26 West Martin Luther King Drive
, ... . Cincinnati, Ohio 45268
(513) 569-7562; FTS 684-7562
COMMENT PERIOD^
ta r> rrJ: bs
Comments on both documehts^just be submitted jeparately, in writing, and must be
postmarked NO LATER THAN AUC-UST-31rl990rAddresses for submission of comments
are provided witb4haijublic review drafts. •.,:
2.S sial' .v*no ro£ - -
PLEASfeKOTE THAT EPA IS UNABLE TO PROVIDE POLICY OR TECHNICAL COMMENT
^THE«CG|EF;ENT Of ^THER DOCUMENT AT THIS TIME SINCE THEY ARE
AMTIPTroTTrutr\mjwc AMT»
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
>* . (FRL • - ) - '- ~:^VJ' . ,
= • . ; , ' <•' '( Oi'J. £.
Health Effects of Passive Smoking: Assessment of^Lung Cancer in
Adults and Respiratory Disorders in Children,; n'External Review
Draft , , .
'".-., .f,i..,-., -^i
AGENCY: U.S. Environmental Protecti8n;Agendy \ ' . ,
; '. ; r. '.c ; 0(o ': j
ACTION: Notice of availability"of external'review-draft and
request for public comments. 7 -: •• " -' " •" -T' '
SUMMARY: This notice announces the availability of 'the external
review draft of the Health Effect* of Passive Smoking:
Assessment of Lung Cancer in Adults and Respiratory Disorders in
Children, EPA/600/6-90/006A. This document will be the subject
of a Science Advisory Board (SAB) meeting. Notice of the date
and place of the SAB meeting will be published in a subsequent
• •. ' i' . i !Vi ij..'
Federal Register notice. . / •.!-::vj2f.j
. •,.,»->, --.•-'
DATES: The Agency will make' the draft document available for
public review and comment on or about Monday, June,25, 1990.
Comments must be postmarked by Friday, August' 3iv 1990.
••"•-. '"«-'.•."? • :
ADDRESSES: To obtain a single copy of the draft document,
interested parties should contact the ORD Publications Office,
CERI-FRN, U.S. Environmental Protection Agency, 26 West Martin
Luther King Drive, Cincinnati, OH 45268, (513)569-7562 or
FTS/684-7562. Please provide your name and mailing address and
-------
•-sfceques'ti this .eternal review draft by title and" E^A*1 number..
!jt .document also will be available for 'p'tiblic
T • " •- • „.
rsand- .copy ing in the Public Information Reference Unit
fecary,, U.S. Environmental Protection Agency
'headquarters j Waterside Mall, 401 M Street, SW, Washington, DC
-30460^0^*; vsc^rJcrs,..
53- *=i s2fiaratenli€J3fc,jarfe; requested to submit their comments in writing
-»3f6pjo SPrc^jeclJ:vP|j|i9er,,for Environmental Tobacco Smoke, Technical
: Staff- , -Off ice of Health and Environmental Assessment
^ **-..'./, e_, ,*. £» 7£ "2 O *
yrBfi'S. /.Environmental Protection Agency, 401 M Street, SW,
~iH?FORMATIOM CONTACT: Steven Bayard, (202)382-5722 or
*" " *" ~ *• -•-• ~S -'•' • •—"'
SU^PEEMEKTARY I»|a^|TION: Because Environmental Tobacco Smoke
(ETS) is a widespread pollutant of indoor air and because it has
~" ' ". '^* ••* - ^- ** «->\
•*-"- be£en previously -found, to be associated with several respiratory
¥•£ tracts diseasesr and^- .disorders, EPA's Office of Indoor Air and
requested that the Human Health Assessment
!Odlfijq%jOJ^.HjMjlthJTand Environmental Assessment, prepare a
"*fci*v*»^ c^ C « . C? H. ' •.
^as^sjmenrl^^cument^for ETS. This project is carried out
s Harder th« sutho5ityn9|, Title IV of Superfund (The Radon Gas and
t'- Ihdoor/Air^Quality Research Act of 1988).
^ " — - •>••-. ^ , _'CJc ' '
r., vcrr Tha-h^alrthtjass^sment document addresses the scientific,
:s -•m%>st:l!y~epJidemio3spgi.c, evidence on the possible respiratory
^•5-e¥feets,< including. Ijung. cancer, of passive smoking or ETS. This
» - •- .. ^ .,» .. t.rfj~ _n c»^";j5 2^~
issue was examined previously in two 1986 reports,' one by the
£
2
-------
U.S. Surgeon, General, titled The Health Consequences of
Involuntary Smoking; the other by the National-Refciarehs Council,
titled Environmental Tobacco Smoke, Measuring ^feKposurjeTs and
Assessing Health Effects. The EPA draft '
analyses of those reports to include sub
evidence ;on the potential association betwieH
cancer in nonsmoking adults, and 2) respiratory disorders u in
children. With respect to lung cancer^ ilf %d£l^"isttMK>araft report
concludes that 1) ETS is causally a~s~socf at&f 3/£tte Jn&sg qajvcer in
nonsmoking adults and that, according to
carcinogen risk assessment, ETS"ls~ini3roup-A-
carcinogen; and 2) that approximately 3 86 6 l£hg. oancerr1 deaths per
year among nonsmokers (never-smoker"1ind'>fliriiei: i*mS3terB)T(p-f both
sexes in the United States are attributable to" -ETS .; $ Thi-s. figure
is for total ETS exposure, with no ~sHji&ra&» 6re'akdd.wir*f-(0js
domestic vs. occupational or sociii e'xposufi^ '»' £ '. : .;"!'£;
With respect to respiratory effects- in children, rthe draft
report concludes that ETS from parental smoking)- especially
during infancy, is associated with ^n^lreased1 pr^ew»i^nQ^fif acute
lower-respiratory tract infect ions"w05^nch^1^te2a&ia ^»ei®pnia) ,
•"• ' ' "*" ^ ^ T '_f*' r "^ ^
respiratory symptoms of irritation fcougSf ^p^fttum, -Whre.eae) , and
middle ear effusions (a sign "of " chr8nic*m£ddi;6 ear dise&se) . It
also concludes that ETS is* associate^ H€6'!redttc*d~-lung • function
and with a small reduction in the1 re^-of "pu*lm6nary growth and
development in children of mothers who ^smoker -during: ddxedr early
•• ' »e " "e-j •• *•• • ^^ • .
childhood. No conclusions are 'drawn regarding: a ^potential
-------
association of parental smoking with increased acute upper-
respiratory tract illnesses (colds and sore throats), an
increased prevalence of asthma, or exacerbation of symptoms in
asthmatic children.
JUN 1 9 '990
L ^Carl/ft. Gerb.er. Acting
(Date) yKSsistatft Administrator for
// Research and Development
-------
EPA
United States
Environmental Protection
Agency
Office of Health and
Environmental Assessment
Office of Atmospherici,afid~-
Indoor Air Programs
Washington DC 20460
EPA/600/6-90/006A
May 1990
External-flfeview Draft
Research and Development/Air and Radiation
Health Effects of
Passive Smoking:
Assessment of
Lung Cancer in
Adults and
Respiratory
Disorders in
Children
Bev&aw
(Do Not
dte or Quote)
Notice
This document is a preliminary draft. It has not been formally
released by EPA and should not at this stage be construed to
represent Agency policy. It is being circulated for comment on its
technical accuracy and policy implications.
-------
DRAFT EPA/600/6-90/006A
DO NOT QUOTE OR CITE May 1990
External Review Draft
\
HEALTH EFFECTS OF PASSIVE SMOKING:
»
ASSESSMENT OF LUNG CANCER IN ADULTS
AND RESPIRATORY DISORDERS IN CHILDREN
NOTICE
THIS DOCUMENT IS A PRELIMINARY DRAFT. It has not been formally released by the
U.S. Environmental Protection Agency and should not at this stage be construed to represent
Agency policy. It is being circulated for comment on its technical accuracy and policy
implications.
Office of Health and Environmental Assessment
Office of Research and Development
and
Indoor Air Division
Office of Atmospheric and Indoor Air Programs
Office of Air and Radiation
U.S. Environmental Protection Agency
Washington, D.C.
-------
DRAFT-DO NOT QUOTE OR CITE
DISCLAIMER
This document is an external draft for review purposes only and does not constitute
Agency policy. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
ii 05/17/90'
-------
DRAFT-DO NOT QUOTE OR CITE
CONTENTS
Tables vi
Figures ix
Preface x
List of Abbreviations xii
Authors, Contributors, and Reviewers xiv
1. EXECUTIVE SUMMARY 1-1
1.1. ETS AND LUNG CANCER 1-2
1.2. ETS AND RESPIRATORY DISORDERS IN CHILDREN 1-6
2. INTRODUCTION 2-1
3. EPIDEMIOLOGIC EVIDENCE OF LUNG CANCER
FROM ETS 3-1
3.1. INTRODUCTION 3-1
3.2. META-ANALYSIS OF CASE-CONTROL STUDIES FROM
RAW DATA 3-14
3.3. META-ANALYSIS OF CASE-CONTROL STUDIES THAT
INCLUDE AN ADJUSTED STATISTICAL ANALYSIS 3-21
3.4. EVIDENCE OF DOSE-RESPONSE IN CASE-CONTROL
STUDIES WITH MORE THAN ONE EXPOSURE LEVEL 3-25
3.5. BASIC ISSUES IN POTENTIAL BIAS FROM
MISCLASSIFICATION IN CASE-CONTROL STUDIES 3-31
3.5.1. Background 3-31
3.5.2. Sources of Bias 3-31
3.6. COHORT STUDIES: BACKGROUND 3-34
3.7. SOME COMPARATIVE ASPECTS OF THE TWO
MAJOR COHORT STUDIES: HIRA(Coh)
ANDGARF(Coh) 3-35
3.7.1. Overview 3-35
3.7.2. Comparative Review and Discussion of the
Cohort Studies 3-36
3.7.3. Comparative Data Analysis of the Cohort
Studies 3-40
iii 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
CONTENTS (continued)
3.8. SUMMARY AND CONCLUSIONS 3-41
4. ASSESSMENT OF LUNG CANCER RISK FROM ETS 4-1
4.1. INTRODUCTION 4-1
4.2. PREVIOUS ESTIMATES OF RELATIVE RISK FROM
EPIDEMIOLOGIC DATA 4-3
4.2.1. The NRC Report and Wald et al. (1986) 4-3
4.2.2. Other Risk Assessments Based on
Epidemiologic Data 4-8
4.3. APPROACHES TO RISK ASSESSMENT BASED ON
CIGARETTE-EQUIVALENTS 4-15
4.4. CURRENT ASSESSMENT OF LUNG CANCER RISK 4-22
4.4.1. Combining Evidence Across Studies 4-23
4.4.2. Adjustment to Relative Risk for Smoker
Misclassification 4-25
4.4.3. Parameter Sensitivity 4-27
4.4.4. Adjustment to Relative Risk for
Background Exposure 4-27
4.4.5. Population-Attributable Risk and Excess
Lung Cancer Deaths 4-29
4.4.6. Adjusted Relative Risk and Population-
Attributable Risk by Individual Study 4-31
4.5. SUMMARY AND CONCLUSIONS 4-37
5. ENVIRONMENTAL TOBACCO SMOKE AND RESPIRATORY
DISORDERS IN CHILDREN 5-1
5.1. INTRODUCTION 5-1
5.2. EXPOSURE OF CHILDREN 5-3
5.3. RECENT EPIDEMIOLOGIC EVIDENCE 5-5
5.4. RESPIRATORY SYMPTOMS 5-9
5.4.1. The U.S. Surgeon General's Report on
Respiratory Symptoms 5-9
5.4.2. The National Research Council Report on
Respiratory Symptoms 5-11
iv 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
CONTENTS (continued)
5.4.3. Recent Studies on Respiratory Symptoms 5-12
5.4.4. Summary and Discussion of Respiratory
Symptoms 5-16
5.5. ACUTE RESPIRATORY ILLNESS 5-18
5.5.1. The U.S. Surgeon General's Report on Acute
Respiratory Illness 5-20
5.5.2. The National Research Council Report on
Acute Respiratory Illness 5-20
5.5.3. Recent Studies on Acute Respiratory Illness 5-21
5.5.4. Summary and Discussion of Respiratory Illness 5-23
5.6. PULMONARY FUNCTION 5-25
5.6.1. The U.S. Surgeon General's Report on Pulmonary
Function 5-27
5.6.2. The National Research Council's Report on
Pulmonary Function 5-28
5.6.3. Recent Studies on Pulmonary Function 5-29
5.6.4. Summary and Discussion of Pulmonary Function 5-31
5.7. RELATED RESULTS 5-33
5.7.1. Middle Ear Effusion 5-33
5.7.2. Acute Upper-Respiratory-Tract Illness 5-35
5.7.3. Asthma 5-35
5.7.4. Symptoms in Asthmatics 5-37
7.7.5. Non-Specific Ailments 5-39
APPENDIX A: SUMMARY DESCRIPTIONS OF ELEVEN
CASE-CONTROL STUDIES A-l
APPENDIX B: MATHEMATICAL FORMULAS
AND RELATIONSHIPS B-l
APPENDIX C: DOSIMETRY OF ETS C-l
APPENDIX D: ALTERNATIVE APPROACHES FOR
ESTIMATING THE YEARLY NUMBER OF
LUNG CANCER DEATHS IN NON-SMOKERS
DUE TO ETS BASED ON DOSE-RESPONSE
MODELING D-l
REFERENCES AND RELATED BIBLIOGRAPHY R-l
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLES
3-1 Case-control Studies of ETS: Characteristics 3-2
3-2 Case-control Studies of ETS: Characteristics 3-5
3-3 Case-control Studies of ETS: Sources 3-8
3-4 Case-control Studies of ETS: Measure of Exposure
to ETS and Other Substances 3-10
3-5 Case-control Studies: "Unexposed" vs. "Exposed"
from Raw Data 3-15
3-6 Case-control Studies: "Unexposed" vs. "Exposed"
Females from Adjusted Statistical Analyses 3-23
3-7 Case-control Studies: Exposure Response Trends
for Females 3-26
3-8 Two Cohort Studies: Female Lung Cancer Data for
Similar Age and Exposure Groups 3-42
4-1 Epidemiologic Studies Included in Overall Relative
Risk in this Report (Females Only) and Several
Other Sources 4-11
4-2 Adjusted Relative Risks and Population-attributable
Risk of Individual Studies (Females) 4-35
4-3 Population-attributable Risk by Study (Females) 4-36
5-1 Evidence of Respiratory Disorders Related to ETS
Exposure from Selected Studies Subsequent to the
U.S. SG and NRC Reports of 1986 5-6
5-2 Studies on Respiratory Symptoms Referenced in the
U.S. SG and NRC Reports of 1986 5-10
5-3 Studies on Respiratory Illness Referenced in the
U.S. SG and NRC Reports of 1986 5-19
5-4 Studies on Pulmonary Function References in the
U.S. SG and NRC Reports of 1986 5-26
vi 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLES (continued)
B-l Definition of Terms in Equation Bl Relating
Observed RR (RRO) and its Value Adjusted for
Misclassification (RRM) B-2
B-2 Parameters and Their Alternative Specifications
Required for Equation Bl B-3
B-3 Conversion of Parameters Between Reported and
Correct Classifications B-5
C-l Steady State Ratio of Concentrations of Nicotine
in Body Tissues or Organs C-16
C-2 Approximate Composition of Mainstream Smoke and
Diluted Sidestream Smoke from One Non-filter
Cigarette C-19
C-3 Summary of Concentrations and Daily Intakes for
Constituents of Cigarette Smoke, Assuming Fresh SS C-22
C-4 Measurements of ETS Constituents in Environmental
Settings C-23
C-5 Daily Integral Organ Burdens for Particulate Phase
Chemicals as Calculated in this Report C-32
C-6 Summary of Dose Measures Calculated in this Report C-38
C-7 Summary of Ratio of Measures (ETS/MS) Calculated
in this Report C-43
D-l Selected Sources of Information Potentially Useful
for Deriving a Dose-Response Relationship for ETS D-5
D-2 Data That Can be Used to Obtain Relative Potency
Estimates for ETS Constituents D-7
D-3 Historical Lung Tumor Control Data for
Osborne-Mendel Rats D-7
D-4 Relative Potency Estimates of ETS Constituents D-8
D-5 Example of Animal Inhalation Dose-Response Model
Syrian Golden Hamsters Exposed to B[a]P Via NaCl
Aerosol D-10
vii 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLES (continued)
D-6 Number of Lung Cancer Deaths and Person-Years of •*
Observation for British Male Physicians D-17
D-7 Lung Cancer Death Rates Per 100,000 Person-Years and
Observed and Predicted Number of Lung Cancer Deaths
Among Men and Women Who Never Smoked Regularly D-21
D-8 Relative Potency Estimates of Complex Mixtures
of Incomplete Combustion Products of Hydrocarbons
Compared to B[a]P D-24
D-9 Relative Potency Estimates of Agents Compared
toB[a]P D-25
D-10 Lung Cancer Mortality in Japanese Women by
Husband's Age Group and Smoking Habits
(Patient Herself a Non-Smoker) D-28
viii 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
FIGURES
3-1 Percentage of controls exposed to ETS by study 3-13
3-2 Ordered values of the S statistic from raw data of
studies in Table 3-5 3-20
3-3 Ordered values of the S statistic from
adjusted analyses of studies in Table 3-6 3-24
3-4 Plots of relative risk against exposure for studies
in Table 3-7 3-28
C-l The general anatomy of the lung from the trachea
down to the distal bronchioles C-7
C-2 A compartmental model of the human body, displaying
organs, tissues, fluids, and their interconnections C-13
C-3 A metabolic model for the conversion of nicotine
and the excretion C-35
D-l Goodness-of-fit of two-stage model to non-smokers--
age-dependent lung cancer data D-22
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
PREFACE
This assessment of the health effects associated with
passive smoking has been prepared by the Human Health Assessment
Group, Office of Health and Environmental Assessment, Office of
Research and Development, which is responsible for its scientific
accuracy and conclusions. The assessment was prepared at the
request of the Indoor Air Division, Office of Indoor Air and
Atmospheric Programs, Office of Air and Radiation, which defined
its scope and provided funding.
The document has been developed under the authority of Title
IV of Superfund (The Radon Gas and Indoor Air Quality Research
Act of 1986) to provide information and guidance on the potential
hazards of indoor air pollutants.
A comprehensive search of the scientific literature for this
document is complete through September 1989. In addition, a few
studies published since September of 1989 have been included in
response to recommendations made by reviewers.
Due to both resource and time constraints, the scope of this
review has been limited to an analysis of respiratory effects,
with emphasis on the epidemiologic data. Further, because two
thorough reviews on passive smoking were completed in 1986 (by
the U.S. Surgeon General and the National Research Council), this
document provides a summary of those reports with a more
comprehensive analysis of the literature appearing subsequent to
those reports and an integration of all the results.
With respect to quantitation of lung cancer risk, the
document has used the actual epidemiologic data and vital
statistics to estimate the number of nonsmokers affected. It
does not use high- to low-dose extrapolation models since
exposures in the epidemiology studies were at true environmental
levels. However, measures of exposure-response and modeling are
examined in two appendices. One appendix analyzes the two main,
currently-used proxies for environmental tobacco smoke (ETS)
exposure, respirable suspended particulates and body cotinine
levels. The other appendix examines methodologies and models
treating ETS as a complex mixture of carcinogens with both
initiating and promoting properties. It also outlines several
possible approaches for exposure-response assessment.
Two other issues that have not been addressed in this draft
but which have drawn comments from reviewers of earlier drafts
are (1) the possible synergistic lung cancer effect of ETS and
radon, and (2) the relative lung cancer hazards and risks of home
and the workplace. These issues will be more fully covered in a
revised version of this document.
It is the Agency's intent to revise and update this document
following the public comment period and review by the Agency's
Science Advisory Board.
x 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
examined in two appendices. One appendix analyzes the two main, currently-used proxies for
environmental tobacco smoke (ETS) exposure, respirable suspended particulates and body
cotinine levels. The other appendix examines methodologies and models treating ETS as a
complex mixture of carcinogens with both initiating and promoting properties. It also outlines
several possible approaches for exposure-response assessment.
Two other issues that have not been addressed in this draft but which have drawn
comments from reviewers of earlier drafts are (1) the possible synergistic lung cancer effect of
ETS and radon, and (2) the relative lung cancer hazards and risks of home and the workplace.
These issues will be more fully covered in a revised version of this document.
It is the Agency's intent to revise and update this document following the public
comment period and review by the Agency's Science Advisory Board.
xi 05/17/90
-------
ACS
ADC
AMW
B[a]P
BR
CEA
C.I.
Coh
COPD
CS1
CSC
DA
ES1
ETS
FEF.
25-75
FS1
FVC
GI
GSD
IARC
ICRP
LCD
DRAFT-DO NOT QUOTE OR CITE
LIST OF ABBREVIATIONS
American Cancer Society
adenocarcinoma
American war veterans
benzo[a]pyrene
bronchial responsiveness
cigarette-equivalents approach
confidence interval
cohort study
chronic obstructive pulmonary disease
current smoker
cigarette smoke condensate
Direct Approach
ever-smoker
environmental tobacco smoke
forced expiratory flow rate, mid-expiratory phase
forced expiratory volume at one second
former smoker
forced vital capacity
gastro-intestinal
geometric standard deviation
International Agency for Research on Cancer
International Commission on Radiological Protection
lung cancer deaths
1 used for both singular and plural forms
xii
05/17/90
-------
MAP
MMAD
MS
NCI
NHIS
NP
NRC
NS1
OR
P
PAH
PAR
RPA
RR
RSP
RSV
SCC
SDA
SS
TB
U.S. DHHS
U.S. DOT
U.S. SG
WHO
DRAFT-DO NOT QUOTE OR CITE
ABBREVIATIONS (continued)
marriage aggregation factor
mass median aerodynamic diameter
mainstream smoke
National Cancer Institute
National Health Interview Survey
nasopharyngeal
National Research Council
never-smoker
observed risk
pulmonary
polycyclic aromatic hydrocarbons
population-attributable risk
Relative Potency Approach
relative risk
respirable suspended particulates
respiratory syncytial virus
small cell carcinoma
Seventh-Day Adventists
sidestream smoke
tracheobronchial
U.S. Department of Health and Human Services
U.S. Department of Transportation
U.S. Surgeon General
World Health Organization
used for both singular and plural forms
xui
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
AUTHORS, CONTRIBUTORS, AND REVIEWERS
This document was prepared as a joint collaboration between the Office of Research and
Development, Office of Health and Environmental Assessment and the Office of Air and
Radiation, Office of Atmospheric and Indoor Air Programs, Indoor Air Staff. The project
manager with overall responsibility for this report and its conclusions is
Steven Bayard, Ph.D.
Human Health Assessment Group
Office of Health and Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC 20460
The primary author of this report is
Kenneth G. Brown, Ph.D.
Kenneth G. Brown, Inc.
P. O. Box 16608
Chapel Hill, NC 27516-6608
who wrote all chapters and Appendix B.
Three other authors are
Appendix A: Charles G. Humble
Department of Epidemiology
School of Public Health
University of North Carolina
Chapel Hill, NC 27514
Appendix C: Douglas Crawford-Brown, Ph.D.
Associate Professor
Department of Environmental Sciences and Engineering
School of Public Health
University of North Carolina
Chapel Hill, NC 27514
Appendix D: Todd W. Thorslund, Sc.D.
Vice President
Clement Associates, Inc.
9300 Lee Highway
Fairfax, VA 22031
xiv
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
A contributor to Chapter 5 was
Sandra A. Peters, Ph.D.
800 Powell Street
Chapel Hill, NC 27514
Technical Assistance to Chapters 2 and 3 was provided by
Chu-Chih Chen
Department of Biostatistics
School of Public Health
University of North Carolina
Chapel Hill, NC 27514
Earlier drafts, in whole or in part, were reviewed by
Michael Berry, Ph.D.
David Burns, M.D.
Rebecca Calderon, Ph.D.
Chao Chen, Ph.D.
Thomas Evans
Lawrence Glass, Ph.D.
Richard Hertzberg, Ph.D.
Aparna Koppikar, Ph.D.
Karen Milne
James Repace
Charles Ris, P.E.
Sherry Selevan, Ph.D. "
Marvin Schneiderman, Ph.D.
Richard Walentowicz
A. Judson Wells, Ph.D.
Jeannette Wiltse, Ph.D., J.D.
In addition, the following people have provided helpful discussions or supplied copies of
recent manuscripts on other material relevant to this report: R.C. Brownson, A. Charlton, A.
Collies, K.M. Cummings, M.P. Ericksen, D. Fennell, I.E. Garfinkel, R. Greenberg, F.W.
Henderson, D. Hoffman, W. Hoffman, M.J. Jarvis, L.C. Koo, M. Layard, W.K. Lam, M.D.
Lebowitz, J. Lewtas, T. Martonen, A.B. Murray, B.D. Ostro, J. Repace, E. Riboli, M.A.
Richardson, J.M. Samet, D. Sandier, and S.R. Tannenbaum. Overall document preparation and
typing was done by Stacy Henkle.
xv 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
1. EXECUTIVE SUMMARY
In 1986, the National Research Council (NRC) and the U.S. Surgeon General (U.S. SG)
assessed the health effects of exposure to environmental tobacco smoke (ETS). Both of the
1986 reports conclude that ETS exposure is causally associated with lung cancer and that
children of parents who smoke have increased frequency of respiratory symptoms and acute
respiratory illnesses and evidence of reduced lung function. The two reports were developed
and edited by different processes, which strengthens the validity of the conclusions common to
both reports. The NRC report is the product of a committee of experts; the U.S. SG report is a
composite of contributions from individual experts that were edited, based on the review of
other knowledgeable individuals, and then cleared through the U.S. Public Health Service.
This document extends the analyses of those reports to include subsequent epidemiologic
evidence on the potential association between ETS and (1) lung cancer in nonsmoking adults,
and (2) respiratory disease and pulmonary effects in children. It concludes that passive
smoking is causally associated with lung cancer in adults and that exposure of young children to
ETS from parental smoking, particularly during infancy, is associated with increased prevalence
of acute lower-respiratory-tract infections (bronchitis and pneumonia), respiratory symptoms of
irritation (cough, sputum, wheeze), and middle ear effusions (a sign of chronic middle ear
disease). Passive smoking in early childhood is also associated with reduced lung function in
children of mothers who smoke and with a small reduction in the child's rate of pulmonary
growth and development. No conclusions are drawn regarding a potential association of
parental smoking with the child's increased acute upper-respiratory-tract illnesses (colds and
sore throats), an increased prevalence of asthma, or exacerbation of symptoms in asthmatic
children.
1-1 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
This report also estimates that approximately 3800 lung cancer deaths per year among
nonsmokers (never-smokers and former smokers) of both sexes are attributable to ETS in the
United States. This figure is an extension of the estimate of 1750 (95% C.I. 910, 2660) for
female never-smokers alone, an overall value calculated from all the epidemiologic studies on
lung cancer and ETS. The 1750 value and its confidence interval includes reasonable estimates
of exposure and risk for single female never-smokers. Projection of the 1750 estimate to 3800
for all nonsmokers of both sexes is based on reasonable estimates of exposure and risk for all
never-smoking men and for former smokers of both sexes. No further quantitative estimates of
ETS-related health effects in adults or children are made.
1.1. ETS AND LUNG CANCER
The U.S. SG (1989) estimates that smoking is responsible for more than one of every six
deaths in the United States and that it accounted for 87% of the lung cancer deaths in males
and 75% in females in 1985. Smokers, however, are not the only ones exposed to tobacco
smoke. The sidestream smoke (SS) emitted from a smoldering cigarette between puffs (the
main component of ETS) has been documented to contain many of the same carcinogenic
compounds (known and suspected human and animal carcinogens) that have been identified in
the mainstream smoke (MS) inhaled by smokers. Exposure concentrations of these carcinogens
to passive smokers are variable but much lower than for active smokers. An excess cancer risk
from passive smoking, however, is still biologically plausible. (U.S. EPA guidelines [Fed. Reg.,
1986] assume that unless there is evidence to the contrary, any level of exposure to a carcinogen
carries a potential risk of cancer.)
Based on the firmly established causal relationship of lung cancer with active smoking, the
lung is considered to be the site most likely affected by passive smoking. The ubiquity of ETS
and its absorption by members of the general population have been well documented by air
1-2 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
sampling and by bioassays for nicotine and cotinine. This raises the question of whether any
direct evidence exists for the relationship between ETS exposure and lung cancer in the general
population and what its implications may be for public health. This document addresses that
question by reviewing and analyzing the cumulative evidence from epidemiologic studies. These
studies compare individuals with higher ETS exposures to those with lower exposures. Typically
the study subjects are married women who have never smoked but are married to a smoker
(higher exposure) and those married to a nonsmoker (lower exposure). Following the
nomenclature of the literature, the higher and lower exposed persons are referred to as
"exposed" and "unexposed." Of course there is exposure to ETS from sources other than spousal
smoking, collectively designated as "background" exposure, which applies to the so-called
unexposed as well as the exposed. Background exposure is taken into account in characterization
of population risk (Chapter 4), but it is not required for the statistical assessment of the
evidence of excess lung cancer risk from spousal smoking (Chapter 3).
The epidemiologic evidence of a lung cancer hazard is statistically assessed by methods of
meta-analysis to obtain overall results. The data and study results included apply to female
married never-smokers. Several studies include male subjects, but the percentage of male never-
smokers is relatively small and the data are scant by comparison. In some instances former
smokers are included with never-smokers. All the ETS exposures are considered to be at true
environmental levels.
Based on these analyses and following the U.S. EPA guidelines for carcinogen risk
assessment (Fed. Reg., 1986), EPA concludes that environmental tobacco smoke is a Group A
(known human) carcinogen. This conclusion is based on a total weight of evidence, principally:
• Biological plausibility. ETS is taken up by the lungs and distributed throughout the
body. The similarity of carcinogens identified in SS and MS along with the established
causal relationship between lung cancer and smoking make it reasonable to suspect that
ETS is also a lung carcinogen.
1-3 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
• Consistency of response. The two completed cohort studies and sixteen of the 21 case-
control studies observed a higher risk of lung cancer among the female never-smokers
classified as exposed to ETS. Evaluation of the total study evidence from several
perspectives leads to the conclusion that the observed association between ETS
exposure and increased lung cancer occurrence is not attributable to chance.
• Upward trend in dose-response. Of the two major cohort studies, the Japanese study
(Hirayama) demonstrates a strong association between passive smoking and lung
cancer, including an upward trend in dose-response. The upward trend is well
supported by the preponderance of evidence in the 13 case-control studies that
classified data by exposure level. The Hirayama study has undergone extensive critical
review that led to some corrections and revisions but failed to discredit the findings.
Differences in life-style and culture may be a factor in the Japanese study reporting a
stronger association between ETS and lung cancer than the American study (American
Cancer Society).
• Detectable association at environmental exposure levels. Within the population of
women who are lifelong nonsmokers, the excess lung cancer risk of those married to a
smoker is large enough to be observed. Carcinogenic responses are usually detectable
only in high exposure circumstances, such as occupational settings or in highly dosed
experimental animals.
• Broad-based evidence. The 21 case-control and three prospective studies provide data
from eight different countries and from a wide variety of study designs and protocols
conducted by many different research teams. No alternative explanatory variables for
the observed association between ETS and lung cancer have been indicated that would
be broadly applicable across studies.
• Effects remain after adjustment for potential bias. Current and ex-smokers may be
misreported as never-smokers, thus inflating the apparent cancer risk from ETS
exposure. The evidence remains statistically conclusive, however, after adjustments for
smoker misclassification. The summary estimate of relative risk from raw data of both
the case-control and cohort studies is 1.41 (95% C.I. 1.26, 1.57) before adjustment for
misclassification and 1.28 (95% C.I. 1.12, 1.45) afterward (p < 0.01).
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 the ubiquity of ETS. Current smokers comprise approximately 30% of the adult
U.S. population and consume over one-half trillion cigarettes annually (1.5 packs per day, on
average), causing nearly universal exposure to ETS. Cotinine, a metabolite of the tobacco-
specific compound nicotine, is detectable in the blood, saliva, and urine of persons recently
exposed to tobacco smoke. Cotinine has typically been detected in 50% to 75% of reported
1-4 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
nonsmokers tested (50% equates to 62 million U.S. nonsmokers of age 18 or above). The
estimate of 3800 lung cancer deaths per year in nonsmokers attributable to ETS is based on
data from epidemiologic studies at actual environmental exposure levels. Some mathematical
modeling is required to adjust for expected bias from self-reported misclassification of smoking
status and to account for ETS exposure from sources other than spousal smoking. The approach,
however, does not rely on a mathematical model of dose-response or low dose extrapolation of
observations obtained at extraordinarily high exposure levels.
The components of the 3800 figure include approximately 1750 female never-smokers, 800
male never-smokers, and 1250 former smokers. The 800 value for males is probably low based
on information in the limited epidemiologic data for male never-smokers. Little is known about
the lung cancer risk of ETS to former smokers. The estimate of 1250 former smokers is based
on the assumption that the risk to former smokers is the same as the risk to never-smokers,
which may be conservative from a biological perspective. If the estimate of 3800 total lung
cancer deaths per year is recalculated using the lower (upper) confidence limit from study data
on female never-smokers and the lower (upper) plausible value regarding population exposure
to ETS, then a value of 1800 (6100) is obtained. It is unlikely that the number of lung cancer
deaths per year attributable to passive smoking by nonsmokers is below 1800 or above 6100.
Other quantitative approaches to characterize population risk could have been used. Dose-
response assessments based on the cigarette-equivalents approach to relate the risk of passive
smoking to active smoking are reviewed. Published variations of this general approach have
typically used either cotinine concentrations or respirable suspended particulates as a surrogate
measure of exposure to ETS. They typically ignore the epidemiologic data on ETS and lung
cancer and follow the paradigm of low-dose extrapolation from a dose-response curve (a dose-
response curve for the lung cancer risk of active smokers in this case). Examples of the
cigarette-equivalents approach provide additional estimates of lung cancer risk that, with the
1-5 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
exception of one low value, range from a few hundred to over 4000 lung cancer deaths per year.
Although this report prefers the more direct approach based on epidemiologic data to
characterize the lung cancer risk, there are also potential benefits from a dose-response
assessment. For example, a dose-response model would be useful for evaluating the
effectiveness of abatement procedures and the risk from varied environmental conditions.
Methods based on the cigarette-equivalents approach could benefit from improved
understanding of the biokinetic similarities and differences between passive and active smoking
that may affect extrapolation of risk from active smoking to risk of passive smoking. A
mathematical model relevant to that objective is developed as a basis for future study
(Appendix C). Three additional quantitative methods for dose-response assessment of passive
smoking, along the lines of the general cigarette-equivalents approach, are outlined with
solicitation for comments and advice (Appendix D).
1.2 ETS AND RESPIRATORY DISORDERS IN CHILDREN
Exposure to ETS from parental smoking has been previously linked with increased
respiratory disorders in children, particularly infants. Several studies have confirmed the
exposure and uptake of ETS in children by assaying saliva, serum, or urine for cotinine. A
recent study of 433 healthy neonates in central North Carolina found that 64% of them lived in
households with a smoker and that 75% of smoking mothers smoked near their infants.
Cotinine concentrations were correlated with smoking (especially by the mother) in the infant's
presence. Nine to twelve million American children under five years of age may be exposed to
cigarette smoke in the home (American Academy of Pediatrics, 1986).
With regard to the respiratory effects of passive smoking in children, this report focuses
on the epidemiologic evidence appearing since the two major reports of 1986 (NRC and U.S.
SG) that bears on the potential association of parental smoking with detrimental health effects
1-6 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
in their children. These include symptoms of respiratory irritation (cough, sputum, or wheeze);
acute diseases of the lower respiratory tract (pneumonia and bronchitis); indications of chronic
middle ear infections (predominantly middle ear effusions); reduced lung function (from forced
expiratory volume and flow-rate measurements); prevalence of asthma; exacerbation of
symptoms in asthmatics; and acute upper-respiratory-tract infections (colds and sore throats).
The thirty or so recently published studies essentially corroborate the previous conclusions of
the NRC and U.S. SG regarding respiratory symptoms, respiratory illnesses, and pulmonary
function; strengthen support for those conclusions by the additional weight of evidence; and
extend research in some directions. In particular, recent studies on middle ear effusion
strengthens previous evidence to warrant the conclusion in this report of an association with
parental smoking. Additional research also supports the hypothesis that early respiratory illness
is associated with long-term pulmonary effects (reduced lung function and possibly increased risk
of chronic obstructive lung disease).
The NRC and U.S. SG reports conclude that both the prevalence of respiratory symptoms
and the incidence of respiratory infections are higher in the children of smoking parents.
Estimates of the increased risk of wheezing vary from zero to over sixfold. In the seven studies
of respiratory symptoms subsequent to the two reports, increased cough was observed in a range
of ages from birth to mid-teens. Recent studies also supplement the evidence for increased
wheeze. Overall, the cumulative evidence supports the previous conclusion of the NRC and U.S.
SG. Six of the studies subsequent to those reports have addressed the topic of parental smoking
and respiratory illness in children, and all have reported statistically significant results. The
cumulative evidence indicates strongly that parental, especially the mother's, smoking is
associated with increased incidence of respiratory illnesses in the first one-to-two years of life,
particularly for bronchitis and pneumonia. Recent studies also solidify the evidence of a link
between parental smoking and increased middle ear disease in young children.
1-7 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The U.S. SG and NRC reports both conclude that children of parents who smoke have
small decreases in tests of pulmonary output function of both the larger and smaller air passages
when compared with the children of nonsmokers. This conclusion is statistically supported, but
the topic itself is difficult to study and to evaluate quantitatively because of the relatively large
inter-individual variability in temporal patterns of lung growth and development. Family history
may be an important factor as well. As noted in the NRC report, if ETS exposure is the cause
of the observed decrease in lung function, the effect could be due to the direct action of agents
in ETS or an indirect consequence of increased occurrence of acute respiratory illness related to
ETS.
Study results on ETS and lung function in children that have appeared since those reports
add some additional evidence supporting an association of ETS exposure with decreased lung
function. Furthermore, this evidence adds support to the supposition that acute respiratory
illness during childhood has a long-term effect on lung growth and development (suggesting an
indirect association with ETS exposure, by virtue of its association with increased pneumonia
and bronchitis in infants). Overall, the weight of evidence indicates that ETS exposure is
associated with decreased lung function in childhood and with a small reduction in their rate of
pulmonary growth and development.
This report concludes that the detrimental respiratory effects described in children are
associated with exposure to ETS, but a causal association has not been established. Causation is
biologically plausible, but other factors that cannot be fully assessed may be influencing the
observed study data. One confounding variable, for example, is direct transmission of
respiratory infections from smoking parents, who tend to have more infections than nonsmokers.
Also, parental recall and the increased incidence of respiratory symptoms in smoking parents
may be contributing an upward bias to the response attributable to ETS exposure. Studies have
generally not controlled for in utero exposure to agents in tobacco smoke (Chen et al., 1988, is a
1-8 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
notable exception). Some studies that include children above the age of seven may also be
upwardly biased by subjects' unreported experimentation with cigarette smoking.
It is improbable, however, that sources of bias or confounding factors account for the
totality of study findings on increased occurrence of detrimental respiratory effects in children of
parents who smoke. The overall evidence of a health risk is based on numerous investigations
that vary broadly in design and location, source of data, objective, protocol, and methods of
analysis. Most studies have controlled for bias and potential confounding factors to the extent
possible. The upward dose-response trends exhibited in several studies suggest that an
alternative explanatory factor would have to be highly correlated with the level of ETS exposure,
e.g., the number of cigarettes smoked per day by the mother. In view of these considerations,
the substantial epidemiologic evidence, the serious health consequences of some of the observed
effects, and the large number of children potentially at risk, it is prudent and reasonable to treat
passive smoking as a risk factor for acute respiratory diseases and chronic obstructive pulmonary
disorders in infants and young children.
1-9 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
2. INTRODUCTION
Over 300,000 deaths per year from all causes of disease in the United States are
attributable to cigarette smoking. Some 100,000 of these smoking-related deaths are from lung
cancer, which is almost 90% of lung cancer cases from all causes (figures for 1985, U.S. SG,
1989). Forty-three known or suspected carcinogens have been identified in tobacco smoke, most
of which are in both mainstream smoke (MS, produced during "puffs") and sidestream smoke
(SS, from the cigarette tip between puffs). The relative distribution of carcinogens and other
toxins differs between SS and MS, however, often with much larger total amounts (by weight) in
the SS from a cigarette. More of the cigarette tobacco is burned in the generation of SS than
MS, on average, but a more significant factor is the less complete combustion of tobacco at the
lower temperatures producing SS. Environmental tobacco smoke (ETS) to which a passive
smoker is exposed principally consists of SS, usually in greatly diluted concentrations depending
on the proximity to the source and related environmental conditions, e.g., ventilation. Aging
also affects the composition of chemicals in ETS and their relative distribution between the
vapor and particulate phases.
Passive and active smokers are exposed to many of the same carcinogens, however, and
active smoking has been firmly established as causally related to lung cancer. It is biologically
plausible that passive smoking is also causally related to lung cancer. Consequently, the
epidemiologic studies available on lung cancer and ETS exposure are examined for detectable
evidence of increased lung cancer risk from passive smoking. It should be noted that it is
extremely unusual when considering a low dose exposure to an agent known to be carcinogenic
at higher doses to have as complete or as extensive a set of measures of exposure to the
population at environmental levels; it is almost unprecedented to have epidemiologic evidence in
as many different populations as is present for ETS. Nevertheless, statistical analysis of the
2-1 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
combined study results could be inconclusive. Statistical significance is evidence that an effect is
at a sufficiently high level to be detected with the data available; lack of significance only
supports the conclusion that it is below a level that the data have adequately high power to
detect with assurance.
A first step is to gain some idea of the magnitude of exposure to ETS in the general population.
Surrogate measures of ambient concentrations of ETS, such as respirable suspended particulates
(RSP), have confirmed passive smokers' exposure in real and simulated environments.
Concentrations of cotinine, a tobacco-specific metabolite of nicotine measurable in blood, saliva,
and urine, confirm the uptake and systemic distribution of nicotine from ETS in passive
smokers. Positive cotinine concentrations in 50% to 75% of self-reported nonsmokers, including
persons reporting no exposure to tobacco smoke in the detectable period (up to a few days,
depending on the body fluid tested), demonstrate the ubiquity of ETS. Conservatively, over 100
million U.S. adult nonsmokers are exposed to ETS at levels detectable in urinary cotinine assays.
The first epidemiologic results associating passive smoking with lung cancer appeared in
the early 1980s. The epidemiologic studies amassed quickly, along with interest in the results
and controversy over issues and conclusions. At the request of two Federal Agencies, the U.S.
Environmental Protection Agency (Office of Air and Radiation) and the Department of Health
and Human Services (Office of Smoking and Health), the National Research Council (NRC)
formed a committee on passive smoking to evaluate the methods for assessing exposure to ETS
and to review the literature on the health consequences of exposure. The committee's report
(NRC, 1986) addresses the issue of lung cancer risk in considerable detail, including summary
analyses of the evidence from ten case-control and three cohort (prospective) studies. It
concludes that "Considering the evidence as a whole, exposure to ETS increases the incidence of
lung cancer in nonsmokers."
2-2 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The NRC committee was particularly concerned with potential bias in study results from
current and former smokers incorrectly self-reported as lifelong nonsmokers (never-smokers).
Under plausible assumptions for misreported smoking habits, it concludes that the true relative
risk lies between 1.15 and 1.35, with 1.25 the most likely value. When these relative risks are
also corrected for background exposure to ETS to make the risk relative to a baseline of zero
ETS exposure, the resultant estimate is 1.42, with a "plausible" range of 1.24 to 1.61. (Technical
Note: "Relative risk," RR, is used somewhat generically throughout this report for technical ease
and because the name is descriptive of the intended measure. For 2-by-2 contingency tables of
raw data from classifying exposed/unexposed against cancer/non-cancer, as analyzed in the
NRC report, the sample odds ratio is estimated in place of RR. These terms are discussed in
standard texts of statistical methods for epidemiology. A good technical discussion with
examples may be found in Agresti, 1990, Section 2.2.)
Two other major reports on passive smoking have appeared: the U.S. Surgeon General's
report on the health consequences of passive smoking (U.S. SG, 1986), and the report on
methods of analysis and exposure measurement related to passive smoking by the International
Agency for Research on Cancer (IARC, 1987). The U.S. SG's report concludes:
The absence of a threshold for respiratory carcinogenesis in active smoking, the presence
of the same carcinogens in mainstream and sidestream smoke, the demonstrated uptake of
tobacco smoke constituents by involuntary smokers, and the demonstration of an increased
lung cancer risk in some populations with exposures to ETS leads to the conclusion that
involuntary smoking is a cause of lung cancer.
The IARC committee emphasized issues related to the physicochemical properties of ETS,
the toxicological basis for lung cancer, and methods of assessing and monitoring exposure to
ETS. The report quotes the summary statement on passive smoking of a previous IARC
working group that found the epidemiologic evidence available at that time (1985) compatible
with either the presence or absence of lung cancer risk. Based on other considerations related
2-3 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
to biological plausibility, however, it concludes that passive smoking gives rise to some risk of
cancer. Specifically, the report states:
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, however, leads to the
conclusion that passive smoking gives rise to some risk of cancer.
The summary analysis across epidemiologic studies conducted by the NRC is extended in
this report to include nine additional case-control studies for which the raw data are available
(for 19 from a total of 21 case-control studies). The committee's approach to hazard
identification is further complemented by other statistical analysis across studies (meta-analysis)
of results adjusted for potential confounding variables. The two major cohort studies (from
Japan and the U.S.) are reviewed and compared. Their results are combined with those from
the case-control studies to give an overall estimate of relative risk (Chapter 3). The population
risk is then characterized by estimation of the annual number of lung cancer deaths among
nonsmokers attributable to passive smoking (Chapter 4).
Results from dose-response risk assessments of lung cancer and ETS are examined in this
report. The methods tend to be variations on the "cigarette-equivalents" approach, the basis of
which is extrapolation of lung cancer risk for passive smoking from a dose-response curve of
active smoking. Although this approach makes use of the dose-response information previously
obtained for active smoking, the information contained in the epidemiologic studies is not fully
utilized. The cigarette-equivalents approach is limited by incomplete knowledge regarding the
biological basis for comparing the carcinogenic potential of passive and active smoking. A
mathematical model has* been developed to aid the study of the biokinetic similarities and
differences between passive and active smoking. The model also serves to identify parameters
where information is currently lacking.
2-4 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The conclusions of this report do not require a dose-response risk assessment, but such a
construct could be useful for evaluation of the effectiveness of abatement procedures and for
confirmation of population risk estimates from another perspective. Three alternative
approaches are described in this report (Appendix D), with solicitation of comments and advice.
The three alternatives include: (1) a relative potency approach in which the dose-equivalent
potency of ETS relative to a positive control, such as benzo[a]pyrene (B[a]P), is established from
animal lung implant studies. That relationship is then included with the relative cancer potency
of various polycyclic aromatic hydrocarbons (PAHs) and an existing inhalation dose-response
model for B(a)P to establish a dose-response model for ETS; (2) a mixture approach based on a
modification of the method in (1) and the relative concentrations of the known lung carcinogens
in ETS; and (3) a direct approach using epidemiologic cohort studies on ETS.
Acute health effects in children from household exposure to ETS is a second health-
related concern examined in this report (Chapter 5). Epidemiologic evidence subsequent to the
major NRC and U.S. SG reports of 1986 is summarized and compared with the conclusions of
the two previous reports for respiratory symptoms, respiratory illness, and pulmonary function.
Recent studies on related health concerns in children are also reviewed. These studies
investigate the effects of parental smoking on prevalence of asthma, chronic middle ear diseases,
and upper-respiratory-tract infection, and on the severity of conditions in asthmatics. Potential
confounding factors and sources of bias that limit quantitative estimation of health hazards
attributable to ETS are addressed.
2-5 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
3. EPIDEMIOLOGIC EVIDENCE OF LUNG CANCER
FROM ENVIRONMENTAL TOBACCO SMOKE
3.1. INTRODUCTION
The 21 Case-control studies currently available are listed in Table 3-1. The studies are
denoted by the first few letters of the first author's name for easy reference. The NRC report
(1986) reviews and analyzes ten of the studies shown in Table 3-1: AKIB, BUFF, CHAN,
CORR, GARF, KABA, KOO, LEE, PERS, and TRIG. The study designated as WU in the
table is excluded because the raw data were not available. (Raw data consist of the number of
exposed and unexposed subjects among lung cancer cases and controls, where a subject is
typically classified as exposed to ETS if married to a smoker.) The NRC also excludes an
earlier version of the KOO study and the studies by Knoth et al. (1983), Miller (1984), and
Sandier et al. (1985) for various reasons (NRC, 1986). Aside from WU, these studies are also
omitted from this report.
The U.S. SG's report (1986) contains particularly good summary reviews of the studies
available at that time. The studies are selectively described or compared in several sources as
well (NRC, 1986; IARC, 1987; Baiter et al., 1986; Blot and Fraumeni, 1986; Correa, 1986;
Eriksen et al., 1988; Kuller et al., 1986; Repace and Lowrey, 1985; Riboli, 1987; Samet, 1988a,b;
Saracci and Riboli, 1989; Weiss, 1986; Wells, 1988b; Uberla, 1987; and Varela, 1987). Appendix
A contains summaries of the studies subsequent to the NRC report, two of which are
unpublished dissertations (LAMW and VARE). The other studies described in Appendix A
include BROW, GAO, GENG, HUMB, INOU, LAMT, SHIM, SVEN, and WU. Tables 3-2, 3-
3, and 3-4 display descriptive characteristics of all 21 case-control studies.
3-1 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-1. CASE-CONTROL STUDIES OF ETS: CHARACTERISTICS
Study
AKIB
(Akiba et
al., 1986)
BROW
(Brownson et
al., 1987)
BUFF
(Buffler et
al., 1984)
CHAN
(Chan and
Fung, 1982)
CORR3
(Correa et
al., 1983)
GAO
(Gao et al.,
1987)
GARF
(Garfinkel
et al., 1985)
GENG
(Geng et
al., 1986)
HUMB
(Humble et
al., 1987)
Location
Japan
(Hiroshima,
Nagasaki)
USA
(Colorado)
USA
(Texas)
Hong Kong
USA
(Louisiana)
China
(Shanghai)
USA
China
(Tianjin)
USA
(New Mexico)
Matched
variables
Age, sex,
residence,
RERF part-
icipant
Age, sex
Age, sex
Matched but
variables
unspecified
Age (±5),
sex, race
Age (± 5)
Age (± 5)
Age (± 2),
sex, race,
marital
status
Age (± 10),
sex, ethnicity
Final Includes an
sample adjusted
matched statistical
for ETS analysis7
Yes No
No2 Yes
No2 No
No2 No
No2 No
No2 Yes
Yes Yes
No2 No
No2 Yes
(continued on following page)
3-2
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-1. (continued)
Study
INOU
(Inoue and
Hirayama,
1988)
KABA
(Kabat
and Wynder,
(1984)
KOO
(Koo et al.,
1987)
LAMT
(Lam et al.,
1987)
LAMW8
(Lam, 1985)
LEE
(Lee et al.,
1986)
PERS
(Pershagen
et al., 1987)
SHIM
(Shimizu et
al., 1988)
Location
Japan
(Kanagawa,
Miura)
USA
(New York)
Hong Kong
Hong Kong
Hong Kong
England
Sweden
Japan
(Nagoya)
Matched Final Includes an
variables sample adjusted
matched statistical
for ETS analysis7
Age, year of No2 Yes
death (± 2.5),
district
Age (±5), Yes No
sex, race,
hospital
Age ( ± 5), No2 No
residence,
housing
Age (±5), No2 No
residence
Age, socio- No2 Yes
economic
status,
residence5
Age, sex, No2'4 Yes
hospital
location,
time of
interview
Age (± 1), Yes Yes
sex
Age (±1), Yes Yes
hospital, ad-
mission date
(continued on following page)
3-3
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-1. (continued)
Study
SVEN
(Svensson et
al., 1988)
TRIG
(Trichopoulos
et al., 1981)
VARE
(Varela,
1987)
WU
(Wu et al.,
1985)
Location
Sweden
(Stockholm)
Greece
(Athens)
USA
(New York)
USA
(Los Angeles)
Matched
variables
Age
Age, socio-
economic
status6
Age, sex,
county,
smoking
history
Age (±5),
sex, race
Final
sample
matched
for ETS
No2
No2
Yes
No2
Includes an
adjusted
statistical
analysis7
Yes
No
Yes
Yes
Adenocarcinoma only.
Not matched on smoking status (smoker/nonsmoker).
Bronchoalveolar cancer excluded.
Ongoing study modified for passive smoking with follow-up.
"Similar" in age, SES, and residence.
"Similar" in age and SES.
Generally refers to (conditional) logistic regression and to
stratification or standardization of variables in analysis.
W.K. Lam is the author of LAMW and co-author of LAMT, a separate
study.
3-4
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-2. CASE-CONTROL STUDIES OF ETS: CHARACTERISTICS
Percent
proxy
Study response1
Ca Co
AKIB 90 88
BROW 69 39
BUFF 82 76
CHAN
CORR * *
GAO 0 *
GARF * *
GENG * *
HUMB
INOU 100 100
KABA
KOO *
Female2
age
Ca Co
70.2 *
35-95 *
66.3 68.2
30-79 30-79
39-70 39-70
* *
35-69 35-69
MO >_40
^65 ^65
^85 ^85
* *
61.6 53.9
* *
Source
of
controls
Atomic
bomb
survivors
Cancer
cases4
Cancer
cases6
Orthopaedic
patients
Hospital
patients8
General
population
Cancer
cases9
*
General
population
Cerebro-
vascular
disease
(deaths)
Patients10
"Healthy"11
Number
female
controls
270
47
196
139
133
375
402
93
162
64
25
136
Percent
female
controls
"exposed"3
70
15s
84
47
46
74
61
44
56
*
60
49
(continued on following page)
3-5
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-2. (continued)
Percent
proxy
Study response1
Ca Co
LAMT * *
LAMW * *
LEE 3813 38
PERS *15 *
SHIM
SVEN 0 0
TRIG * *
VARE 3319 3319
WU *
Female2
age
Ca Co
* *
67.5 66
35-74 35-74
*16 *
59 58
35-81 35-81
66.3
62.8 62.3
67.P 68.119
<76 <76
Source
of
controls
"Healthy"12
Hospitalized
orthopedic
patients
Patients14
*17
Patients18
General
population
Hospitalized
orthopedic
patients
New York
State Dept.
of Motor
Vehicles
Neighbor-
hood12
Number
female
controls
335
144
66
347
*
174
190
21820
52
Percent
female
controls
"exposed"3
45
44
68
43
*
66
43
*
63
1 "Ca" and "Co" stand for "cases" and "controls", respectively.
2 Single values are the average or median. Paired values are the
range.
3 Definition of "exposed" differs between studies. See Table 3-4
and 3-5.
4 Persons with cancers of bone marrow or colon in Colorado Control
Cancer Registry.
5 "Exposed" to ETS 4 or more hours/day.
6 Population-based and decedent comparison subjects selected from
state and federal records.
8 Assorted ailments.
9 Colo-rectal cancer
3-6
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
10 Diseases not related to smoking.
11 Selected from a healthy population.
12 Living in neighborhood of matched case.
13 Applies only to the 143 patients in the follow-up study.
14 Excluding lung cancer, chronic bronchitis, ischemic heart
disease, and stroke.
15 No overall percentages given.
16 Two control groups: 15-65 and 35-85 for both cases and controls
in groups 1 and 2 respectively.
17 Two control groups were randomly chosen from the cohort under
study
18 Patients in the same or adjacent wards with other diseases.
19 Includes males and females and long-term ex-smokers.
20 Includes 69 ex-smokers.
* Data not available.
3-7 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-3. CASE-CONTROL STUDIES OF ETS: SOURCES
Spouse(s)
Study 1 > 1
AKIB X
BROW X
BUFF
CHAN1
CORR X
GAO X
GARF X
GENG X
HUMB X
INOU X
KABA X
KOO X
LAMT X
LAMW X
LEE X
PERS X
SHIM X
SVEN X
TRIG X
Adulthood
Others Away from
at home home
X X
X
X X
X X
X2 X
X X
X
X X
X3
X X
X X
Childhood
exposure from
mother/father
X
X
X
X
X
X
(continued on following page)
3-8
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-3. (continued)
Study
Spouse(s)
1 >1
Adulthood
Others
at home
Away from
home
Childhood
exposure from
mother/father
VARE
WU
X
X
X
X
X4
X
X
1 Not stated. See footnote 1 of Table 3-4.
2 Exposure to mother's or father's smoking is presumed to mean in
adulthood.
3 Separate for workplace, travel, leisure.
4 Separate for workplace and social circumstances.
3-9
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-4. CASE-CONTROL STUDIES OF ETS: MEASURE OF EXPOSURE
TO ETS AND OTHER SUBSTANCES
Study
AKIB
BROW
BUFF
CHAN
CORR
GAO
GARF
GENG
HUMB
INOU
KABA
KOO
LAMT
LAMW
LEE
PERS
SHIM
SVEN
ETS exposure measures
Cig./ Total Total
day years cigs. Other
X
hrs./day
X
X1
pack-yrs.
X
X2 hrs./day
X X
X X
X3
no units
XX X4
Xs
no units
X6
no units
X
X7
Related exposures
Cooking/ Work/
heating environ.
X
X
X
X X
X
X
X
X X
(continued of following page)
3-10
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-4. (continued)
Study
TRIG
VARE
WU
ETS
Cig./
day
X8
X
exposure measures
Total
years
X
Total
cigs.
X
X
Related
Cooking/
Other heating
person-yrs.
X9 X
exposures
Work/
environ.
X
Exposed/unexposed determined from a single question, "Are you
exposed to the tobacco smoke of others at home or at work?" (Lam
T.H. et al., 1987; Chan and Fung, 1979).
Cig./day smoked by husband at home.
Smoker at home defined as >_5 cig./day.
Others include total hours of exposure and mean hrs./day.
A woman was considered exposed to her husband's smoke if they had
lived together continuously for at least one year.
Exposure designated as 0 (unexposed), 1,2,3.
Exposure is "yes" or "no" for each source.
Exposed within the last 5 years.
Exposed if husband smoked.
3-11
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
These studies compare individuals with higher ETS exposures to those with lower exposures.
All of them have made observations on never-smoking married women. Those married to a smoker are
assumed to be at higher exposures than those not married to a smoker. These two groups are referred
to as "exposed" and "unexposed", respectively, following the established terminology. Those terms refer
to spousal smoking only, however, and obscure the presence of other sources of ETS, collectively
referred to as "background" sources. Background exposure is taken into account in the next chapter
under characterization of the population risk, where it becomes necessary to think of the unexposed
classification as referring to exposure from background sources only and the exposed classification as
including both background and spousal smoking. For the purpose of hazard identification in this
chapter, background exposure does not explicitly enter into the discussion. The relative risk comparison
of exposed to unexposed individuals, however, is implicitly a comparison of "exposed to both
background and spousal smoke" to "exposed to background only".
Several differences in the case-control studies may be noted. The unit of measure of ETS
exposure (e.g., cig./day or total years of exposure) varies between studies. A few studies include former
smokers as nonsmokers if they have abstained from tobacco usage for some minimum period, while
others do not. Classification of a subject as ETS-exposed depends on the questions asked, which differ
across studies. The proportion of controls classified as exposed is shown by study in Figure 3-1.
Exposure percentages cover a range from 43% to 84%, a two-fold difference, with the BROW study is
an outlier at 15%. The referent populations are defined by a number of parameters, such as whether
the subjects were alive or dead and, if alive, whether or not they were hospitalized. Other general study
characteristics that vary relate to study design, protocol, interpretation, and analysis of data, potential
confounding factors included in the matching and/or data analysis, and confirmation of primary lung
cancers. Study differences do not invalidate statistically testing the hypothesis that exposure to ETS is
unrelated to lung cancer occurrence. (Technical Note: The Breslow-Day test for homogeneity of the
odds ratio is not significant, p = 0.48. [Breslow and Day, 1980].) ETS studies have primarily
3-12 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
BROW
TRIG
PERS
GENG
LAMW
LAMT
CORR
CHAN
KOO
HUMB
KABA
GARF
WU
SVEN
LEE
AKIB
GAO
BUFF
*******
*********************
**********************
**********************
**********************
***********************
***********************
************************
************************
****************************
******************************
******************************
********************************
*********************************
**********************************
***********************************
*************************************
******************************************
20 40 60 80
PERCENT
15
43
43
44
44
45
46
47
49
56
60
61
63
66
68
70
74
84
PERCENT
FIGURE 3-1. PERCENTAGE OF CONTROLS EXPOSED TO ETS BY STUDY
3-13
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
considered exposure to married never-smoking women, generally classified as exposed or unexposed to
ETS based on their spouse's smoking status. Consequently, analysis in this report primarily focuses on
the data for females. Data on males is sparse by comparison, and there may be sex-related exposure
differences to ETS (Cummings et al., 1986; Friedman et al., 1983).
The case-control studies are tested for an association between ETS exposure and lung cancer
from three perspectives.
1) Raw data are available for 19 of the 21 studies in Table 3-5. Three statistical tests are
applied to these data to test for an association between ETS exposure and lung cancer
occurrence.
2) Eleven studies provide results of statistical analyses that adjust for covariables, e.g., age,
duration of spousal smoking, etc. Statistical modeling, such as logistic regression, or
stratification on variables are two general approaches. The estimated relative risks and
confidence intervals from these methods are used in two statistical tests of the combined
evidence from the eleven studies.
3) Fourteen case-control studies contain crude classifications of exposure to ETS suitable for
evaluating whether there is an upward trend in lung cancer occurrence as ETS exposure
increases. These "dose-response" data from the studies are plotted for comparison.
This statistical testing is in the next three sections (Sections 3.2-3.4). It is followed by a discussion of
potential sources of bias in the case-control studies (Section 3.5). The two major cohort studies are
then described and compared (Sections 3.6 and 3.7), followed by the summary and conclusions of this
chapter.
3.2. META-ANALYSIS OF CASE-CONTROL STUDIES FROM RAW DATA
The statistical power to detect a small but meaningful increase in lung cancer risk from a
single case-control study is often small, but it can be improved by analyzing the total evidence from all
studies simultaneously (meta-analysis). The NRC (1986) followed the lead of a committee member
(Wald, 1986) who estimated an overall relative risk across all studies by the statistical method in Yusuf
et al. (1985). Blot and Fraumeni (1986) and Wells (1988b) achieved the same objective using the
3-14 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-5. CASE-CONTROL STUDIES: "UNEXPOSED" VS. "EXPOSED"
FROM RAW DATA
Study
AKIB
Female
(cig./day)
Male
(cig./day)
BROW3
Female
Male
(hrs./day)
BUFF
Female
(tot.
yrs.)
Male
(tot.
yrs.)
CHAN
Female
CORR
Female6
Male
(pack-
yrs.)
Exposure
0
>.!
0
>.l
unexposed
exposed4
0-3
>_4
unexposed
exposed18
unexposed
exposed18
unexposed
exposed5
0
2.1
0
2.1
No.
cases
21
73
16
3
15
4
2
2
8
33
6
5
50
34
8
14
6
2
No.
controls
82
188
101
19
40
7
11
8
32
164
34
56
73
66
72
61
154
26
RR1
1.52
1.5
1.8
1.52
1.5
1.38
0.81
0.8
0.5
0.75
0.8
2.07
2.07
2.0
C.I.1
(0.88,2.63)
(1.0,2.5)'
(0.5,5.6)
(039,5.99)
(034,1.90)
(0.3,1.8)
(0.2,1.7)
(0.43,130)
(0.4,1.3)
(0.82,525)
(0.8,5.0)
S1 Fs
1.48 0.07
0.61 027
-0.49 0.69
-1.02 0.85
1.52 0.06
(continued on following page)
3-15
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-5. (continued)
Study
GAO
Female
(tot.
yrs.)
GARF
Female
(cig./day)
GENG
Female
(cig./day)
HUMB8
Female
(cig./day)
INOU9
Female
(cig./day)
KABA
Female
Male
KOO
Female
(cig./day)
Exposure
0-19
^20
0
^l7
0
>.!
0
^1
<4
^4
unexposed
exposed10
unexposed
exposed10
0
exposed11
No.
cases
57
189
44
90
20
34
5
15
4
18
11
13
7
5
35
51
No.
controls
99
276
157
245
52
41
71
91
17
30
10
15
7
5
70
66
RR1
1.19
131
1.31
2.16
2.16
234
1.8
2.55
0.79
1.00
1.55
1.55
C.I.1
(0.82,1.73)
(0.87,1.98)
(0.99,1.73)
(1.08,429)
(1.05,4.53)
(0.81,6.75)2
(0.6,5.4)
(0.74,8.78)
(0.25,2.45)
(0.20,4.90)
(0.90,2.67)
(0.94,3.08)
S1 Ps
0.91 0.18
129 0.10
2.19 0.01
1.57 0.06
1.50 0.07
-0.41 0.66
1.56 0.06
(continued on following page)
3-16
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-5. (continued)
Study
LAMT
Female
(cig./day)
LAMW
Female
LEE
Female
Male
PERS
Female
SVEN
Female
TRIG
Female
(cig./day)
WU17
Female
Exposure
0
>:!
unexposed
exposed11
unexposed12
exposed13
unexposed12
exposed13
unexposed15
exposed16
unexposed
exposed16
0
>.!
unexposed
exposed4
No.
cases
84
115
23
37
10
22
7
8
34
33
10
24
24
38
9
19
No.
controls
183
152
80
64
21
45
16
14
197
150
60
114
109
81
22
33
RR1
1.65
1.65
2.01
2.01
1.03
l.OO14
1.30
1.2814
1.2814
126
2.13
1.41
1.2
C.I.1
(1.16,2.35)
(1.16,2.35)
(1.09,3.71)
(0.41,2.55)
(0.37,2.71)14
(0.38,4.42)
(0.76,2.15)
(0.75,2.15)
(0.57,2.82)
(1.19,3.83)
(0.54,3.67)
S1 Ps
2.77 0.003
224 0.01
0.05 0.48
0.91 0.18
0.57 028
2.55 0.005
0.70 0.24
3-17
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Footnotes for Table 3-5.
1 Values of RR, C.I., S, and Ps on the first row of an entry (boldface) are our calculations
for Mantel-Haenszel odds ratio. Values in the second row are from the study. S is the
square root of the Mantel-Haenszel statistic with sign of (-) if R< 1 and ( + ) if R> 1. Ps
is one-tailed significance value from normal tables, and equals one-half the corresponding
two-sided p-value for the M-H chi- squared statistic. Confidence intervals are 95% unless
noted otherwise.
2 90% C.I.
3 Data communicated from R.C. Brownson.
4 Exposed if husband smoked.
5 Exposure based on single question, "Are you exposed to the tobacco smoke of others at
home or at work?" (Lam et al., 1987; Chan et al., 1979).
6 Data partially from Table 12-4, NRC (1986).
7 Cigar or pipe smoking by husband while at home is included in category of .>_! cig./day.
8 Data communicated from C.G. Humble. Eight (total) cases were observed in males, so a
separate odds ratio for males alone was not reported in the study.
9 Raw data were calculated from information given in the reference to INOU by A. Judson
Wells.
10 Based on spouse's current or past smoking habits.
11 Exposed if husband ever smoked in presence of spouse.
12 Only the controls in the follow-up study.
13 Exposed if husband ever smoked during marriage.
14 Standardized for age.
15 Data for controls from Saracci and Riboli (1989).
16 No measure of exposure given.
17 Raw data were calculated from information in Table 11 of the Surgeon General's report
(U.S. SG, 1986, p. 99) by A. Judson Wells.
18 Any current household member who smokes regularly.
3-18 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Mantel-Haenszel (M-H) method (Mantel and Haenszel, 1959; Mantel, 1963), a standard
approach for the combination of information from 2-by-2 contingency tables. (Wells included
adjusted analyses when available, using the raw data to calculate weights.) The two methods are
basically equivalent. In this report, the M-H method is applied to the raw data for females in
the case-control studies shown in Table 3-5.
The M-H estimate of relative risk (odds ratio) and its associated confidence interval are
shown in boldface type in Table 3-5 for females of each case-control study where the raw data
are available (SHIM and VARE are excluded). Each study's own relative risk estimate and
confidence interval are also displayed when available. The M-H estimate of the overall RR for
females is 1.42 (95% C.I. 1.24, 1.63) for 19 case-control studies. (All confidence intervals
hereafter are 95% if not indicated otherwise.) The corresponding value in the NRC report for
females of 10 case-control studies and 3 cohort studies is 1.32 (1.16, 1.51). The NRC reports
1.62 (0.99, 2.64) for males. Wells (1988b) obtained an overall relative risk estimate of 1.44 (1.26,
1.66) for females from 14 case-control studies and three cohort studies. Wells' higher estimate
may be due in part to his exclusion of the CHAN study, the first and probably least
sophisticated of the four Hong Kong studies (CHAN, LAMW, KOO, and LAMT). It also
produced the lowest relative risk estimate in Table 3-5, namely 0.75 (0.43, 1.33).
A statistic referred to as "S" is included in Table 3-5. (Technical Note: S is the
square-root of the M-H chi-squared statistic with the sign " + " if the odds ratio exceeds 1 and
the sign "-" if otherwise. Equivalently, it is the estimated log-odds ratio [ln(RR)] divided by its
estimated standard error. The estimator S is approximately normally distributed with mean zero
under the null hypothesis that the true relative risk equals one [Woolf, 1955].) The values of Ps
in Table 3-5 are the one-tailed significance levels of S for testing the null hypothesis that ETS
exposure is unrelated to lung cancer occurrence. The rank orderedcvalues of S are displayed in
Figure 3-2.
3-19 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
S Statistic
CHAN
BUFF
KABA
LEE
WU
SVEN
BROW
GAO
PERS
GARF
AKIB
CORR
KOO
HUMB
GENG
LAMW
TRIG
LAMT
_ ,*
**********
*****
****
..00 0.
-1.02
-0.49
-0.41
* 0.06
**** 0.39
****** n R7
»*****""" \J • *S /
****** 0.61
********* 0.91
********* 0.91
************* 1.29
*************** 1.48
*************** 1.52
**************** 1.56
**************** 1.57
********************** 2.19
********************** 2.22
************************* 2.53
**************************** 2. 77
00 1.00 2.00
S Statistic
FIGURE 3-2. ORDERED VALUES OF THE S STATISTIC FROM RAW
DATA OF STUDIES IN TABLE 3-5
3-20
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Three statistical tests are conducted of the null hypothesis. The extended M-H
chi-squared test statistic is significant at p < 0.001 based on the combined evidence from the
raw data shown in Table 3-5. The Wilcoxon sign-ranked test (Hollander and Wolfe, 1973)
applied to the values of S in Table 3-5 also is significant at p < 0.001. (Technical Note: The
signed-rank test requires the test statistic for each study to be symmetrically distributed about
zero under the null hypothesis. Since S is approximately normally distributed with mean zero,
that assumption is appropriate here.) If the null hypothesis is true, then the four studies with ps
< 0.05 are making a Type I error (false positive). The chance of that many Type I errors in 19
independent studies is 0.013. The predominance of small values of ps, in general, is informative.
Over half of the 19 ps values are 0.1 or less, an outcome that would occur with probability less
than 0.001 if the null hypothesis is true. All of the tests based on the raw data are statistically
significant.
3.3. META-ANALYSIS OF CASE-CONTROL STUDIES THAT INCLUDE
AN ADJUSTED STATISTICAL ANALYSIS
An adjusted analysis is generally preferable to an analysis of the raw data, even when the
data are matched (Schlesselman, 1982). The two studies for which the raw data are unavailable
do include results of an adjusted analysis (SHIM, VARE). Table 3-1 identifies the studies with
results adjusted for other variables. Some authors have not included complete details, so the
choice of studies for inclusion in this section may be subjective. In eleven reports, the relative
risk and confidence interval are given for two or more levels of exposure, e.g., 1 to 20, 21 to 40,
or 41+ cig./day smoked by the spouse. The RR and confidence interval at a high exposure in
these studies were generally selected for inclusion in Table 3-6, but with some exceptions. For
example, the highest exposure category in VARE is so extreme (80+ cig./day) that it contains
very little data. In this instance, the response for smoking 20 cig./day as predicted from the
3-21 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
logistic regression model fitted to all the data is given in Table 3-6. The table entry for GARF
is also from an adjusted analysis for a spouse smoking 20 cig./day. Two studies that provide a
relative risk without an associated confidence interval (LAMW and SHIM) are displayed in
Table 3-6 for completeness, although they cannot be included in the summary analysis.
The RR and confidence intervals in Table 3-6 are the study authors' conclusions that
depend on their methods of analysis. To combine the results across studies, the statistic S was
calculated from the RR and confidence interval for each study. (Technical Note: It is assumed
that ln(RR) is approximately normally distributed. S equals ln(RR) divided by its estimated
standard error, as calculated from the confidence interval reported.) The values of S are
displayed in Table 3-6 and plotted in Figure 3-3. Five of the studies had significant values of S
(p < 0.05). The probability of observing five or more Type I errors in 11 independent studies is
less than 0.001. Thus, it is highly unlikely that so many significant test results would be observed
if there were, in fact, no association between ETS exposure and lung cancer incidence.
(Technical Note: No multiple comparison adjustment is necessary because the choice of a single
exposure level is made without regard to statistical significance. Test results reported at
exposure levels other than the one used are not relevant and no adjustment for multiple
comparisons is needed.)
The Wilcoxon signed-rank test was also applied to the S statistics of Table 3-6, as
conducted previously with the raw data, to provide another statistical test of the null hypothesis.
The outcome is significant (p = 0.014). In this test the magnitude of the evidence from each
study is a factor. Both statistical tests indicate that the cumulative evidence that lung cancer is
related to ETS exposure would be very unlikely to occur by chance alone.
3-22 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-6. CASE-CONTROL STUDIES: "UNEXPOSED" VS. "EXPOSED"
FEMALES FROM ADJUSTED STATISTICAL ANALYSES
Study
BROW
GAO
GARF
HUMB6
INOU7
KOO
LAMW2-3
LEE5
PERS1
SHIM2
SVEN
VARE4
WU
Exposure
_<3 vs. 2.4 (hrs./day)
0-19 vs. MO (yrs. with
smoking husband)
0 vs. 20 (cig./day)
0 vs. 2.21 (cig./day)
<4 vs. 2. 20(cig./day)
0 vs. 2. 21 (cig./day)
Exposed by husband
Exposed by husband
0 vs. 2. 16(cig./day)
Exposed by husband
Exposed in both child-
hood and adulthood vs.
exposed in neither
0 vs. 20 (cig./day)
Exposed by husband
RR
1.68
1.7
1.70
1.2
3.09
1.19
2.64
1.00
2.40
1.1
1.9
0.94
1.2
95% C.I.
(0.39,2.97)
(1.0,2.9)
(0.98,2.94)
(0.26, 5.5)
(1.04,11.81)
(0.46,3.03)
*
(0.37,2.71)
(0.6,8.7)
*
(0.2,3.7)
(0.76, 1.17)
(0.6,2.5)
S
1.78
1.95
1.90
0.23
1.65
0.36
*
0.00
1.33
*
1.89
-0.54
0.49
PS
0.04
0.03
0.03
0.41
0.05
0.36
*
0.50
0.09
*
0.03
0.70
0.31
1 See footnotes 15-17 of Table 3-2.
2 Higher RR values associated with adult exposure to smoking by mother or by
father's husband. Insufficient information to calculate the S statistic.
3 No units of exposure. RR=2.64 with p = 0.02, and RR= 1.61 with p = 0.19, for
peripheral and central lung adenocarcinoma, respectively.
4 From Table 4 of Varela, 1987.
5 See footnotes 12-14 of Table 3-5.
6 Discussed in Humble et al. (1987) following Table 4, with 90% confidence
interval of (0.3,4.4).
7 The authors assume that husbands who smoke less than five cig./day do not smoke
at home.
3-23
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
VARE
LEE
HUMB
KOO
WU
PERS
INOU
BROW
SVEN
GARF
GAO
***********
1
S
*****
*******
**********
***************************
*********************************
************************************
**************************************
**************************************
**************************************
1 1 — _ I - - 1 inn 1 J 1 III n-
Statistic
-0.54
0.00
0.23
0.36
0.49
1.33
1.65
1.78
1.89
1.90
* 1.95
-0.40 0.00 0.40 0.80 1.20
S Statistic
1.60
FIGURE 3-3. ORDERED VALUES OF THE S STATISTIC FROM ADJUSTED
ANALYSES OF STUDIES IN TABLE 3-6
3-24
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
3.4. EVIDENCE OF DOSE-RESPONSE IN CASE-CONTROL STUDIES WITH MORE
THAN ONE EXPOSURE LEVEL
Data for studies that report relative risk (RR) by levels of exposure are given in Table 3-7,
along with the results of statistical tests for trends when available. The RRs are plotted against
exposure and shown in Figure 3-4. Both adjusted and unadjusted estimates of RR are
presented, as data permit. Some observations are apparent from the plots. For example, the
estimated RRs increase in seven studies: AKIB, CORR, GAO, GENG, INOU, PERS, and
TRIG; decrease slightly in one case: LEE; and are variable in the remaining five plots: GARF,
HUMB, KOO, LAMT, and VARE.
If RR is independent of ETS, then the predicted RR at the highest exposure level is less
than one with probability at least one-half. (Technical Note: If the distribution of the observed
RR under the null hypothesis is symmetric, then the value is one-half. The distribution depends
on the statistical method used in a study. All appear to be skewed to the right, judging from the
confidence intervals. For a right-skewed distribution, the median is less than the mean. Thus,
the probability of values greater than one exceeds one-half.) The observed RR at the highest
exposure level is less than one in only two of the 13 studies above. The probability of two or
fewer such occurrences by chance alone is approximately 0.012. It can be concluded, therefore,
that the plots for trend are consistent with conclusions of the previous statistical tests conducted
on the case-control studies, i.e., that the observed association between exposure to ETS and
increased occurrence of lung cancer deaths is statistically significant.
3-25 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-7. CASE-CONTROL STUDIES: EXPOSURE RESPONSE
TRENDS FOR FEMALES
Study
AKIB
(cig./day)
CORR
(pack-yrs.)
GAO
(tot.yrs.)3
GARF
(cig./day)
GENG
(cig./day)
HUMB
INOU
(cig./day)
KOO
(cig./day)4
Exposure
1-19
20-29
2.30
0
1-40
2.41
0-19
20-29
30-39
MO
0
1-9
10-19
2.20
0
1-9
10-19
2.20
0
1-20
2.21
0-4
5-19
>20
0
1-10
11-20
2.21
RR
1.0
1.3
1.5
2.1
1.0
1.18
3.52
1.0
1.1
1.3
1.7
1.0
1.15
1.08
2.11
1.0
1.40
1.97
2.76
1.0
1.8
1.2
1.0
2.58
3.09
1.0
2.33
1.74
1.19
C.I.1
(0.7,2.3)2
(Q.8,2.8)2
(0.7,2.5)2
*
*
(0.7,1.8)
(0.8,2.1)
(1.0,2.9)
(0.8,1.6)
(0.8,1.5)
(1.1,4.0)
(1.1,1.8)
(1.4,2.7)
(1.9,4.1)
(0.6,5.6)2
(Q.3,5.2)2
(0.4,5.7)2
(1.0.11-8)2
(0.9,5.9)
(0.8,3.8)
(0.5,3.0)
Analysis
P-Trend Unadjusted Adjusted
0.06 X
X
X
< 0.025 X
* X
* X
<0.05 X
* X
(continued on following page)
3-26
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-7. (continued)
Study
LAMP
(cig./day)
LEE6
PERS7
(cig./day)
TRIG9
(cig./day)
VARE10
(cig./day)
WU"
(yrs.
exposed
as adult)
Exposure
0
1-10
11-20
2.21
0
Low
High
0
1-15
2.16
0
1-20
>21
0
1-20
21-40
41-60
61-80
80+
0
1-30
2.3 1
RR
1.0
2.18
1.85
2.07
1.0
0.92
0.81
1.0
1.8
6.4
1.0
1.95
2.55
1.0
0.79
0.91
1.23
0.42
2.86
1.0
1.2
2.0
C.I.1
(1.14,4.15)
(1.19,2.87)
(1.07,4.03)
*
*
(0.6,5.3)
(1.1,34.7)
*
*
(0.6,1.1)
(0.6,1.3)
(0.6,2.4)
(0.1,2.3)
(0.3, 27.7)
*
*
Analysis
P-Trend Unadjusted Adjusted
<0.01 X
* X6
* X8
* X
* X
* X
9
10
11
Confidence intervals are 95% unless noted otherwise.
90% confidence interval.
Years lived with a smoking husband.
Cig./day smoked by husband.
All histologies.
Exposure at home only. Standardized for age, spouse smoking, and whether
currently married.
Small cell carcinoma only. Observed risk was lower for other histologies
combined.
Stratified analyses and conditional (logistic) regression produced consistent
results.
Data from Trichopoulos et al. (1983).
From Table 2 of Varela (1987) for spouse smoking, presumably including males.
Adenocarcinomas only.
3-27
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
AKIB
0 1-40 >-4l
CORK
o-ii jo-at jo-at >*4o
GAO
e 1-t it-it >">o
GARF
FIGURE 3-4. PLOTS OF RELATIVE RISK AGAINST EXPOSURE
FOR STUDIES IN TABLE 3-7
3-28
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
•**••
• •••»
*****
*****
•*•**
•«*«•
•«**•
•••••
*****
•**••
•*•••
**•*•
*****
GENG
HUMB
• *••
*•••
• *•*
**••
• *«*
• *•*
• *•*
«•**
INOU
KOO
FIGURE 3-4. (continued)
3-29
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
.1»
,!
• ****
*•***
• •*••
*•••• ***
1*10 11-20
LAMT
1-10 >-H
TRIG
LEE
PERS
VARE
FIGURE 3-4. (continued)
3-30
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
3.5. BASIC ISSUES IN POTENTIAL BIAS FROM MISCLASSIFICATION
IN CASE-CONTROL STUDIES
Bias, particularly misclassification of subjects by smoking status, is not limited to case-
control studies. Quantitative adjustment of the overall observed RR for possible misreporting of
smoking habits is addressed in detail in Section 4.4.2.
3.5.1. Background
Estimation bias is due to study design, protocol, or method of analysis that apriori makes
the expected outcome too large (positive bias) or too small (negative bias). Although sample
size and dispersion contribute to outcome variability, neither repeated sampling nor increasing
the sample size will affect bias. In either case, the estimate (on average) simply becomes
arbitrarily close to the unknown value of interest, i.e., the true RR plus the bias. In practice,
each case-control study has its own bias. If bias is largely random over a set of studies, some
averaging effect toward zero would be expected. If there is a consistent source of bias in studies,
however, sometimes referred to as "systematic bias," then it cannot be expected to disappear as
the number of studies increases.
3.5.2. Sources of Bias
Ever-smokers (former and current smokers) are more likely to incorrectly report
themselves as never-smokers (NS) than the reverse. A smoker is more likely than a nonsmoker
to marry a smoker, so these misclassified ever-smokers (ES) are more likely to be exposed to
ETS than true NS. Among lung cancer cases in a study (all of whom are reported NS), those
who are actually former smokers (FS) or current smokers (CS) are disproportionately classified
as exposed to ETS (assuming that the smoking habit of the spouse is not misreported as well).
These cancer cases classified as exposed to ETS may have been at higher risk for lung cancer
3-31 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
because of a history of smoking, not just because of exposure to ETS. Hence, a lung cancer
effect from active smoking may be contaminating the evaluation of the risk of passive smoking.
Misreporting among controls tends to overstate the percentage classified as exposed to ETS, due
to the concordance of smoking habits in married couples. This artificially elevates the exposed
percentage in controls relative to cases, so it contributes to underestimation of risk (or bias in
the negative direction).
Subjects classified as unexposed are rarely "truly unexposed"--as supported by data on
measurements of cotinine. The NRC report takes this "background" exposure into account in
adjusting an overall relative risk for a "net" bias. Recent survey results of Cummings et al.
(1989b) provide additional evidence of background exposure in NS. Detectable levels of
cotinine were found in 132 of 162 (81%) of the nonsmokers who reported no exposure in the
four days preceding the interview. A mean urinary cotinine level of 8.8 ng/mL was found among
nonsmokers. Although the study is based on self-selected volunteers, the authors note that the
results are consistent with reports from other studies. Cummings and colleagues conclude that
exposure to ETS is extremely prevalent, even among those not living with a smoker.
The overall estimate of RR in the NRC report places the excess risk of lung cancer
associated with spousal smoking at about 34%. An adjustment for possible misclassification of
the never-smoker status reduces the value to 25%. A second adjustment to make the risk
relative to a truly unexposed subject, i.e., to take into account a background level of exposure,
raises the increased risk to 42%. Consequently, the net adjustment for bias is upwards. The
reports of the U.S. SG (1986) and IARC (1987) do not adjust their overall risk estimate for
possible bias.
Potential sources of bias have been given considerable attention in the literature, often
with an emphasis on the potential for positive bias (over-estimation of relative risk in this case).
The following discussion is not complete, but it raises some of the more prominent issues on this
3-32 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
topic. Many of the potential sources of bias that have been raised in the literature could
conceivably create negative bias as well. Of greatest concern, of course, has been the potential
for bias from misreported classification of smoking status, i.e., CS or FS reported as NS.
The diagnosis of lung cancer in cases may be a source of bias, e.g., a cancer that
originated at another primary site and then metastatised in the lung may be incorrectly
diagnosed as a primary cancer of the lung (Samet, 1988b). As an example, Garfinkel et al.
(1985) report that about 12% of lung cancer patients identified through hospital records were
reclassified after histological review. Some studies addressed this issue by including only
pathologically confirmed lung cancers or by considering histological cell type in their analyses
(e.g., CORR, GARF, PERS, and others).
Bias due to a proxy respondent in place of the subject has also been raised as an issue by
Mantel (1987b) and by Kilpatrick (1987), with evidence from two studies. As reported in Eriksen
et al. (1988), respondent bias can be a source of bias in either direction (Sackett, 1979). In
general the information provided by surrogates has been comparable to that provided by the
individuals themselves (Blot et al., 1985). The recent study by Cummings et al. (1989a) of the
passive smoking histories of 380 NS further supports that conclusion. They report substantial
agreement between subjects and surrogates on most exposure measures.
Vandenbroucke (1988) and Mantel (1987a) have questioned whether there may be a
publication bias, i.e., whether studies with non-significant results are less likely to be published.
Vandenbroucke constructed a quantitative approach but found publication bias only for the
studies on men. Wells (1988a) reviewed the subject and found it unlikely that publication bias
has any substantial effect on the RRs that have been calculated from published reports for
passive smoking for either men or women.
3-33 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
3.6. COHORT STUDIES: BACKGROUND
The three cohort studies that have been conducted are: Garfinkel (1981); Gillis et al.
(1984), which was recently updated by Hole et al., (1989); and Hirayama (1981a, 1984);
abbreviated as GARF(Coh), GILL(Coh), and HIRA(Coh), respectively. (The "Coh" in
parentheses means a cohort study.) The three studies are included in most of the references
cited for summary descriptions and comparisons of case-control studies in Section 3.1. The U.S.
SG's report (1986) sketches the basic features of the cohort studies and the salient topics of
controversy and discussion that appeared in the literature. The Scottish study, GILL(Coh),
which observed only a very small number of lung cancer deaths (6 men and 8 women), is
included in the risk assessment in the next chapter but is not discussed further in this section.
Unlike the case-control studies, several of which have appeared since the NRC, IARC,
and U.S. SG reports of 1986 and 1987, the two major cohort studies, GARF(Coh) and
HIRA(Coh), first appeared in 1981. Consequently, most of the issues regarding these two
studies and their somewhat dissimilar results surfaced well before the three major reports were
prepared. Critical scrutiny of the Hirayama study had already appeared and had been
adequately addressed by Hirayama, as described in the U.S. SG and NRC reports. Judging from
the roundtable discussion at the symposium "Medical Perspectives on Passive Smoking"
(Lehnert, 1984), previous challenges to Hirayama's work regarding data analysis and other
issues appear to have been resolved, aside perhaps from the issue of misreported smoking
habits. Even one of the strongest critics of epidemiologic findings (P.N. Lee) offered a qualified
acceptance of the strength of the statistical evidence in the Hirayama study: "It is ... clear in Dr.
Hirayama's data that if one takes the age of the husband or wife into account and does the
analysis correctly, there is a statistically significant association in lung cancer risk, but the
significance is not nearly as marked as in the incorrect analysis."
3-34 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
In contrast, the study by Garfinkel and colleagues at the American Cancer Society (ACS)
has undergone much less questioning and critical examination, although the problems
experienced in conducting the study and the potential for error in the results have not gone
unnoticed. The difference in outcomes in HIRA(Coh) and GARF(Coh) has been a source of
concern to many, but to our knowledge no one has conducted a statistical review of GARF(Coh)
or compared the statistical methodology in the two studies. Those topics are addressed in the
following section.
3.7. SOME COMPARATIVE ASPECTS OF THE TWO MAJOR COHORT STUDIES:
HIRA(Coh) AND GARF(Coh)
3.7.1. Overview
An increase in risk of lung cancer from ETS was observed in both cohort studies, with
statistical significance (p < 0.05) achieved in HIRA(Coh) but not in GARF(Coh). In the former
study, the observed risk increases as spousal smoking increases (a "dose-response" relationship
that would be expected if passive smoking is causally related to lung cancer). Data from the
American study, however, estimate a higher risk at the lower of two exposure categories (spouse
smokes < 20 cig./day) than at the higher one (spouse smokes 20+ cig./day). Some researchers
have interpreted this outcome as evidence that there is not a "dose-response" relationship in the
American study, or more strongly, that the results demonstrate that there is no increased risk of
lung cancer from ETS exposure. The statistical evidence supporting an association between
lung cancer incidence and ETS exposure in GARF(coh) is inconclusive-it is consistent with
either the presence or absence of a true dose-response relationship. This conclusion follows
from the 95% confidence intervals for the lung cancer mortality ratio at the low (< 20 cig./day)
and high (20-f cig./day) exposures, equal to (0.85, 1.89) and (0.77, 1.61), respectively. These
confidence intervals are consistent with a wide range of possibilities. For example, 1.0
3-35 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
(corresponding to no increase in lung cancer mortality) is in both confidence intervals, but so
are values corresponding to a substantial dose-response relationship, e.g., 1.25 and 1.50 at the
low and high exposures, respectively.
In the following section, the Japanese study (HIRA(Coh)) and the American study
(GARF(Coh)) are reviewed, with an emphasis on the cultural differences in the populations
sampled and the differences in study design, execution, and analysis of data that may help to
compare outcomes of the two studies. In the final section, data comparisons are made for the
two studies to evaluate if there are widespread differences across all age-exposure group
combinations, or just specific ones.
3.7.2. Comparative Review and Discussion of the Cohort Studies
HIRA(Coh) is a census-population based study of adults aged 40 or above, begun in 1965
in 29 Health Center Districts in Japan. A total of 200 cases of lung cancer occurred among the
91,540 nonsmoking married women who were followed. A total of 265,118 subjects were
enrolled for the entire study (122,261 males and 142,857 females, including unmarried women)
accounting for 94.8% of the total census in the study area. Subjects were tracked by establishing
a record linkage system between the data/interview records and death certificates (Hirayama,
1983b, 1984). Interviewers were blind to the smoking status of subjects (NRC, 1986).
In the Japanese study, relative risks of 1.42, 1.58, and 1.91 were observed for nonsmoking
wives with husbands who smoked 1 to 14, 15 to 19, and 20+ cigarettes per day, respectively. The
corresponding value for women whose husbands were former smokers is 1.36, which falls
between the values for nonsmoking and light smoking husbands (Hirayama, 1984). The observed
increase in risk across the exposure categories, with former smokers classified between
nonsmokers and the 1 to 14 cig./day group, is statistically significant by the Mantel-Haenszel
test (one-tailed p < 0.002). Also, RR for women married to smokers increased with age and
3-36 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
with duration of exposure to spousal smoking (age and duration of exposure are likely
correlated, so these should not be construed to be independent results).
GARF(Coh), the ACS's Cancer Prevention Study (CPS-I), began in 1959 when 68,000
volunteers in 25 states enrolled more than one million men and women for long-term follow-up.
Volunteers were instructed to recruit people they knew well. Subject participation was fairly
evenly divided across large cities, small cities and suburbs, small towns, and rural areas. Overall,
about 3% of the population over the age of 45 in 1121 counties was recruited. Enrollment
included all family members of age 30 or above, provided at least one member of the household
was at least 45.
Each year, for six years, the volunteers were asked to report the vital status (alive or
dead) of the persons contacted. For subjects who had died, death certificates were obtained
from state departments of health to determine the cause of death. Additionally, physicians who
certified the cancer deaths were contacted and asked to supply information to verify the primary
sites of the cancers. In the first six years, information was received confirming the primary sites
of cancer in 78% of the cases, and microscopic confirmation was obtained in 69% of the cases.
Death certificates overstated the lung cancer rates by 11.8% (Garfinkel, 1981, 1984, 1985). The
study was essentially terminated after six years, as originally planned in 1965, until it was
decided to conduct a second follow-up beginning in 1971. Follow-up was achieved for 98.4% of
the subjects. The follow-up was terminated, however, because tracing became increasingly
difficult due to death or movement of the volunteers and their substitutes (Garfinkel, 1985).
Apparently death certificates did not continue to be followed up by a medical report after the
first six years. For lung cancer cases in all women, married or not, 203 out of a total of 564
(36%) reported by death certificates were accompanied by a medical report.
The American study does not provide conclusive results regarding a possible association of
lung cancer with ETS exposure. The ratio of observed to expected lung cancer deaths, referred
3-37 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
to as the mortality ratio (Garfinkel, 1981), is 1.27 and 1.10 for nonsmoking women with
husbands who smoked < 20 cig./day and 20+ cig./day, respectively. Neither value is statistically
significant. When data from the Hirayama study are grouped according to the same exposure
levels (< 20 cig./day and 20+ cig./day), the observed relative risks are 1.45 and 1.91, with
one-tailed p-values of 0.03 and 0.001, respectively (Hirayama, 1984).
The American cohort study appears to contain more statistical uncertainty than the
Japanese study. Some of the general factors contributing to uncertainty in study data are
related to sample size, variability in the population sampled, sample design and protocol,
treatment of missing or incomplete data, accuracy and reliability of collecting and reporting
data, and methods of statistical analysis. When the data produce a clear pattern, such as
HIRA(Coh), with a consistent upward trend across exposure categories and age groups that
cannot be ascribed to chance alone, one has some assurance that the sources of variability are
not obscuring a dose-response relationship. Apparent differences between outcomes of the two
studies could be due to one or more sources: (1) a real difference in risk in the populations
studied (perhaps due to higher exposure or uptake of ETS); (2) differences in the way the
studies were designed, conducted, or interpreted; or (3) chance occurrence alone. There is
suggestive evidence for the first two alternatives. Subjects in the American study were followed
for 12 years compared to 16 years in the Japanese study (Hirayama, 1984), so the proportion of
subjects with lung cancer would be expected to be lower in the American study. As reviewed in
the U.S. SG's report (1986), the relatively high risks observed for nonsmokers whose husbands
smoked led to speculation that Japanese women may report themselves as nonsmokers when
they actually smoke (also see Lehnert, 1984). However, some reassurance of the validity of
self-reported information from Japanese women is provided by the AKIB study, which found
strong concordance between self-reported smoking status and the reports from the next-of-kin.
3-38 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Hirayama has emphasized the importance of properly defining passive smoking. He
classifies direct passive smoking as exposure from within approximately 1 to 1.5 meters of the
source and indirect passive smoking as exposure from a greater distance (Hirayama, 1984;
Lehnert, 1984). Direct passive smoking is of much greater concern than indirect passive smoking
(Lehnert, 1984). Japanese wives may experience more direct passive smoking if they tend to be
in closer proximity to their smoking husbands than American wives. Related factors that may
contribute to a net increase in exposure for Japanese wives relative to their American
counterparts include house sizes, the number of smokers per volume of air, proximity of
nonsmoking spouse's sleeping area to spouse's smoking area, and the amount of time a
nonsmoking spouse is in the home. Hirayama (1981b) notes additional differences between
Japan and America that may influence exposure, such as a higher percentage of office workers
among females in the United States than in Japan and a higher divorce rate in the United
States. Japanese wives may be much less exposed to ETS from sources other than spousal
smoking in the home than U.S. wives, i.e., background exposure to ETS may be lower for
Japanese wives. (Technical Note: An increase in exposure to household ETS and a decrease in
exposure from other sources for Japanese women relative to U.S. women would both contribute
toward a higher RR for the Japanese if passive smoking is causally related to lung cancer.) Two
additional factors that may affect a comparison between the United States and Japan include
total cigarette consumption and lung cancer rates due to causes not related to tobacco smoke.
These differences between the U.S. and Japan weigh heavily in favor of Japan as the more
fertile sociological environment for observing an excess risk of lung cancer from passive smoking
by means of an epidemiologic study based on spousal smoking. Exposure to household ETS
appears to be higher in general, with more direct passive smoking, and exposure from other
sources (background exposure) appears to be lower. These factors contribute to a larger relative
exposure to ETS between the so-called exposed and unexposed groups. If passive smoking is a
3-39 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
risk factor for lung cancer, then the observed relative risk should be higher when the relative
exposure to ETS between the comparison groups is higher (mitigating factors not withstanding).
There are also some differences in the methods of analyzing and interpreting data in the two
cohort studies, as described in the next section.
3.7.3. Comparative Data Analysis of the Cohort Studies
The measures of risk reported in HIRA(Coh) and GARF(Coh), the odds ratio and the
mortality ratio, respectively, are not identical statistics. Neither are the statistical methods to
control for age the same. The method applied by Hirayama (Hirayama, 1984) is the
Mantel-Haenszel procedure, commonly used to standardize for age and other factors that may
have an influence. To control for age by this method, for example, study observations are
grouped by time intervals. Comparisons between exposure groups are made at each time
interval, and then the results are combined across intervals to test for a difference between
exposure groups (the extended M-H procedure). The method of analysis used by Garfinkel
(1981) is somewhat different (described more fully in Hammond et al., 1975, 1976). Adjustment
for age is handled by assigning weights according to person-years with a smoking husband. The
results are analyzed as quantal response data.
The method previously applied to non-ETS data in the American study is used to
statistically adjust for potential confounding variables (Hammond et al., 1975, 1976). Groups are
formed from the data matched on age, race, highest educational status of the husband or wife,
residence, and whether or not the husband is occupationally exposed to dust, fumes, or vapor
(Garfinkel, 1981). The ratios of the number of adjusted lung cancer deaths in the low (< 20
cig./day) and high (20+ cig./day) exposure categories to the corresponding number in the
control group, i.e., the nonsmoking women with nonsmoking husbands, are reported to be 1.37
and 1.04, respectively, neither of which is statistically significant. Using data from the American
3-40 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
study that includes age at time of death, duration on study, and whether death was due to lung
cancer or another cause (supplied by L. Garfinkel), the Mantel-Haenszel method was applied
controlling for age and duration on study. Controlling for these two factors simultaneously,
however, did not produce statistical significance or otherwise alter previous conclusions.
For further comparison, the descriptive data from GARF(Coh) corresponding to the age
groups and exposure classifications of data published for HIRA(Coh) (Hirayama, 1984, Table 1)
were placed side-by-side for visual comparison (Table 3-8). Relative to the general pattern of
response in the Japanese study, the American data appear to be at greatest variance from what
might be expected in the two subgroups at highest exposure (20+ cig./day) in the age
classifications 40 to 49 and 50 to 59. Further review of those data for completeness, possible
sources of bias, or unanticipated anomalies may be illuminating.
3.8. SUMMARY AND CONCLUSIONS
The primary focus of this chapter has been on hazard identification, in this case a
statistical assessment of the combined evidence from 21 case-control studies and three cohort
studies of an association between ETS exposure of never-smoking women and lung cancer. The
case-control studies vary with regard to inclusion of raw study data, an adjusted statistical
analysis (adjusting, or controlling, for covariables in the statistical analysis), and dose-response
information to test for upward trend (lung cancer occurrence reported by the amount spouse
smokes). The statistical assessment was conducted from these three perspectives. Overall,
analysis of unadjusted odds ratios (19 studies) indicated a significant lung cancer relationship (p
<_ 0.001 for both statistical tests used. The two statistical tests based on authors' adjusted
statistical procedures in 11 studies were also significant (p <_ 0.001 and p = 0.014). To check
for an upward trend in response, observed relative risk was plotted against exposure (e.g.,
number of cig./day smoked by the husband) for the 13 case-control studies reporting these data.
3-41 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 3-8. TWO COHORT STUDIES: FEMALE LUNG CANCER
DATA FOR SIMILAR AGE AND EXPOSURE GROUPS1
Husband's smokine habit2
Age3
40-49
50-59
60-69
Study4
G
H
G
H
G
H
Nonsmoker
9/23,743
(3.8)
4/6,229
(6.4)
31/25,108
(12.3)
10/7,791
(12.8)
23/15,138
(15.2)
18/7,120
(25.3)
1-19
6/11,791
(5.1)
14/13,779
(10.2)
25/13,528
(18.4)
28/13,720
(20.4)
16/6,884
(23.2)
37/9,756
(37.9)
20+
12/26,918
(4.5)
16/10,764
(14.9)
21/24,184
(8.7)
24/9,820
(24.4)
20/7,299
(27.4)
23/4,651
(49.4)
1 Entries are (number of lung cancer deaths)/(number at risk), stated as a percentage (x 100)
in parentheses. Data for age 70-79 are omitted because of small sample sizes and small
number of lung cancers observed. Data for "G" were supplied by L.
Garfinkel. Data for "H" are in Hirayama (1984).
2 Cigarettes/day.
3 Women's age for G; Husband's age for H.
4 G:GARF(Coh) H: HIRA(Coh).
3-42 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Although outcomes vary, an upward trend was observed in more instances than could be
attributed to chance alone. Tabled study characteristics did not suggest common features among
the studies that might explain their findings. The statistical results solidly support the
conclusion that the observed association between lung cancer and ETS exposure in case-control
studies is not attributable to chance occurrence.
The cohort studies were addressed separately from the case-control studies for several
reasons. The size of the two major studies, the U.S. study by the American Cancer Society and
the Japanese study by Hirayama, might dominate the outcome of some comparisons with case-
control studies. Also, the cohort studies differ from the case-control studies in design, execution,
and numerous other characteristics. The Japanese cohort study alone provides compelling
evidence of a lung cancer risk associated with ETS exposure. Although some corrections to the
initial calculations were required, it has withstood extensive critical examination since its
appearance in 1981 (see NRC, 1986). Results of the American cohort study are less conclusive.
Differences in culture and life-style between the U.S. and Japan suggest that ETS exposure from
spousal smoking may be higher, and exposure from background sources lower in the U.S. than
in Japan. In view of other study evidence of an upward trend in response, the more pronounced
outcome observed in the Japanese study might be anticipated.
Although the American cohort study weakly indicates an increased lung cancer risk from
ETS exposure, the data have an observed inversion in dose-response, i.e., lower response at high
exposure to spousal smoking than at moderate exposure. Further study of the data to see if the
inversion can be explained may be warranted, especially since this study is the largest in the U.S.
and is the only U.S. cohort study. A statistical analysis of the data adjusted for survival was not
helpful. Consequently, the well-patterned response of lung cancer occurrence by subjects' age
and husband's smoking status in the Japanese experience were used as a model to illuminate
departures in the American survey. Data in the American study deviate most from what might
3-43 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
be expected in the two subgroups at highest exposure (over 20 cig./day) in age groups 40 to 49
and 50 to 59. Further examination of this subset of the data may be warranted.
Based on the statistical results of this chapter, this report concludes that passive smoking
is associated with an increased risk of lung cancer. The stronger conclusion of a causal
association, however, is not warranted from these statistical tests alone. Other factors must be
considered as well, including the likelihood that the observed association is attributable to
systematic bias or the presence of a confounding variable. Further analysis relevant to whether
the stronger conclusion of causal association is relevant is given in the next chapter.
3-44 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
4. ASSESSMENT OF LUNG CANCER
RISK FROM ETS
4.1. INTRODUCTION
The preceding chapter addressed the topic of hazard identification and concluded that
ETS exposure is associated with lung cancer. Statistical tests alone, however, do not generally
warrant the conclusion that two variables are causally related, i.e., that one of the variables is a
contributing cause of the other. The total weight of evidence needs to be considered. In
particular, the likelihood that the observed association is attributable to a confounding variable
or a systematic source of bias needs to be considered. If it is concluded that ETS is causally
associated with lung cancer, then the next step is to characterize the magnitude of the
population risk.
Review and analyses of the epidemiologic studies described in Chapter 3 and
supplemented by Appendix A have not indicated a correlate of ETS that may explain the
observed association between ETS and lung cancer. Among the potential sources of bias
discussed, however, misclassification of smoker status needs to be examined as a possible
explanation of the observed effect. Its significance is determined from the remainder of the
estimate of overall relative risk after subtraction of an amount attributable to smoker
misclassification. A model is implemented to estimate the overstatement of relative risk due to
smoker misclassification, the technical details of which are included in Appendix B. The overall
relative risk of lung cancer remains significant after adjustment for smoker misclassification.
Based on this outcome and other evidence, it is concluded that ETS is causally associated with
lung cancer. After numerically adjusting for background ETS (sources other than spousal
smoking), the lung cancer risk of ETS from all sources to the U.S. population of nonsmokers
4-1 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
(never-smokers and former smokers of both sexes) is characterized in terms of the number of
lung cancer deaths (LCDs) attributable to ETS (estimated at 3800).
Several other authors have estimated the population risk of lung cancer from exposure to
ETS also. Two approaches have been used almost exclusively. One analyzes the overall
epidemiologic evidence available from case-control and cohort studies, as done in this report.
The second approach estimates a dose-response relationship for ETS exposure based on
"cigarette-equivalents" determined from a surrogate measure of exposure common to passive
and active smoking. Cotinine concentrations in body fluids (urine, blood, or saliva) and tobacco
smoke particulates in SS and MS have commonly been used for this purpose. The lung c; "U :
risk of ETS is assumed to be equal to the risk of actively smoking at the rate determined by the
cigarette-equivalents.
The NRC report is a good example of the first approach. An overall estimate of relative
risk (RR) for never-smokers exposed to spousal smoking is obtained by statistical analysis across
all available studies (as in Chapter 3 of this report). Two adjustments are made then to the
estimate of RR. The first adjustment accounts for expected bias from former smokers (FS) and
current smokers (CS) who may be misclassified as never-smokers (NS) and it results in a
decrease in the RR estimate. The second adjustment, an upward correction, takes into account
the risk from background exposure to ETS (experienced by a NS whether married to a smoker
or not). Population risk can be characterized then by estimating the annual number of LCDs
among NS attributable to all sources of ETS exposure (spousal smoking and background). This
calculation requires the final-adjusted estimate of relative risk, the annual number of LCDs from
all causes in the population assessed (e.g., NS of age 35 or above), and the proportion of that
population exposed to spousal smoking. The entire population is assumed to be exposed to a
background level of ETS.
4-2 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The cigarette-equivalents approach calculates an estimate of the lifetime excess risk of
lung cancer from exposure to ETS, extrapolated from a dose-response curve constructed for
dose-response of active smoking. Multiplying this estimate by the size of the population exposed
to ETS characterizes the population risk in terms of the excess number of lifetime lung cancers.
It is important to note that the population excess risk in the first approach is stated in terms of
an annual number of lung cancer cases; in the second approach, population risk is stated in
terms of the lifetime excess number of lung cancers. There are variations on the two basic
approaches described, but the published literature largely falls into the two camps
described—inference from the epidemiologic studies or extrapolation from a dose-response
assessment for active smoking. A recent review of risk assessment methodologies in passive
smoking may be found in Repace and Lowrey (1990).
Examples from the published literature illustrating both approaches and their results are
reviewed in Sections 4.2 and 4.3. Section 4.4 describes the risk assessment of lung cancer from
ETS exposure for this report, based on epidemiologic data (the first approach described above).
The overall relative risk to female NS married to a smoker is estimated from the epidemiologic
studies with 95% confidence bounds and then an adjustment for smoker misclassification bias is
calculated from an extension of the NRC/Wald formula. Background ETS is taken into account
in a second adjustment to RR (implicitly redefining RR at that point to be "relative" to the risk
at zero-ETS instead of at an average background level of ETS). The population attributable risk
then is estimated to obtain an annual number of U.S. LCDs in never-smoking women
attributable to total ETS exposure. The predicted number of LCDs due to ETS is further
extended to include male NS and then to include former smokers of both sexes.
4-3 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
4.2. PREVIOUS ESTIMATES OF RELATIVE RISK FROM EPIDEMIOLOGIC DATA
4.2.1. The NRC Report and Wald et al. (1986)
The NRC report follows the construct of Wald and coworkers to adjust for potential
misclassification bias. The technical details of the adjustment are contained in Wald et al.
(1986) and to a lesser degree in the NRC report. An illustrative diagram for the implicit true
relative risk of lung cancer from exposure to ETS in women from spousal smoking is shown in
Figure 2 of Wald et al. A similar example is in Table 12-5 of the NRC report. The formula for
an extended version of these examples is given in Appendix B (Equation Bl), with the required
parameters described in Tables B-l through B-3. The summary relative risks from observed
data reported by the NRC are 1.32 (95% C.I. 1.16, 1.51) for females and 1.62 (0.99, 2.64) for
males. Both RR estimates apply to NS married to smokers, i.e., the estimates apply to lung
cancer risk in exposed NS (actually exposed to spousal smoke and background sources) relative
to the risk in unexposed NS (actually exposed to background sources only). The terminology
adopted from epidemiologic studies comparing an exposed group with an unexposed group can
be misleading when sources of ETS other than spousal smoking are taken into account (i.e.,
background ETS). Both groups experience background exposure to ETS, so one group is at a
higher exposure level (from spousal smoking and background) and the other one is at a lower
exposure level (backgound alone).
After adjusting for expected negative (downward) bias in the RR estimates due to smoker
misclassification, the NRC concludes that the relative risk for both females and males is likely to
be 1.25, and probably lies between 1.15 and 1.35. The "relative" in RR, however, still means
relative to the risk from background exposure alone. To estimate the number of LCDs in NS
attributable to ETS, risk estimates need to be relative to the risk of lung cancer at zero-ETS
exposure instead of at background exposure. This is accomplished in the NRC report by using
data on cotinine concentrations to compare exposure to ETS at the higher level (background
4-4 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
plus spousal smoking) with exposure to the lower level (background only). This report uses the
same method; it will be discussed in more detail. The resultant estimates of RR apply to
exposed persons and to unexposed persons, but are now relative to risk at zero-ETS instead of
risk at the background level. This adjustment for background sources changes the NRC
estimate of RR for an exposed person to 1.42 ("ranging" from 1.24 to 1.61); the change is due
only to implicit redefinition of RR to mean risk relative to zero-ETS, however, instead of
relative to a background level of ETS. Similarly, the RR estimate from background ETS
becomes a positive value, since RR now means relative to zero-ETS. The estimates given by the
NRC are for both sexes, with the qualification that the adjustment for background and the
estimation of LCDs attributable to passive smoking (to be described next) are "crude".
The NRC report estimates that about 21% of the lung cancers in nonsmoking women and
20% in nonsmoking men may be attributable to exposure to ETS (NRC, 1986, Appendix C).
When applied to the ACS's estimate of 6500 (3000) LCDs among NS women (men) in 1988, the
number attributable to ETS exposure is 1365 (600), a total of about 2000. To obtain these
figures for annual LCDs attributable to ETS in NS requires the RR figure for exposed and
unexposed persons obtained from data on married NS. Not all NS are married, of course, and
the NRC estimates that 17% of all NS women (married or not) and 12% of NS men (married
or not) are exposed to ETS at the higher exposure level (equivalent to background plus spousal
smoking); the remaining percentages are assumed to be exposed at the lower exposure level
(background ETS only). These exposure percentages are based on a sample of cotinine
concentrations. In effect, the NRC is estimating the RR for the population of NS by taking a
weighted average of the RR at the higher exposure level and the RR at the lower exposure
level, weighted by the proportions of the population at the higher and lower exposures. The
population is not bipolarly distributed at two exposure levels, but a judicious assignment of
values to the proportions assumed will produce a weighted average that approximates RR for
4-5 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
exposure near the population average. An implicit assumption is that RR is linearly related to
ETS exposure between the lower and higher exposure levels.
The remainder of this section discusses details of the NRC and Wald et al. (1986) analyses
to adjust the overall RR for bias from smoker misclassification (the first step in the procedure
above). In particular, NRC parameter values used for the adjustment are compared with those
to be used in Section 4.4 of this report. Both the NRC and Wald et al. assume that 50% of
women are NS. Wald et al. assumes that the remaining 50% consists of 35% CS and 15% FS,
based on a survey conducted in Britain. Estimates for the U.S. for 1985 from the National
Health Interview Survey (NHIS), as reported in the U.S. SG (1989) report, place the percentage
of U.S. females who are CS and FS at 27.8% and 16.9%, respectively, for a total of 44.7% ever-
smokers (ES) in the U.S. Corresponding values for 1982 from the initial phase of the ACS's
Cancer Prevention Study II (CPS-II) are 22.1% and 22.6%, respectively, for the same total
percentage (44.7%). Consequently, the 50% value for ES is probably a little high for the U.S.,
and the composition of female ES in the U.S. between 1982 and 1985 is probably closer to 22%
to 28% CS and 17% to 23% FS. In Section 4.4, this report assumes 45% female ES, consisting
of 25% CS and 20% FS.
The NRC report estimates that the lung cancer risk of smokers relative to nonsmokers
may be as high as 8.0; based on an ACS study in 1966, and the British Physician's Study in 1980,
and then allowing for a reasonable increase by the mid-1980s. The initial phase of CPS-II
indicates that the relative risk for female CS aged 35 and older has soared from 2.9 between
1959 and 1965 (CPS-I) to 11.9 (95% C.I. 10.0, 14.3) between 1982 and 1986 (CPS-II) (S.G.,
1989, p. 153). The large change is related to an increase in the number of women who began
smoking two to three decades ago, a sufficient period for lung cancers to appear in the mid-
1980s. The relative risk reported for female FS in the general population from CPS-II is 4.7
(95% C.I. 3.9, 5.7) U.S. SG, 1989). ES (CS and FS) misreported as NS are assumed to have a
4-6 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
lower risk than correctly reported ES. The NRC report suggests an overall relative risk of two
for misclassified ES (but notes that it could be as high as four). Wald et al. adjusts the smoker's
relative risk for misreported CS and FS (from eight to two), assuming the misclassification
percentage of ES is an aggregate of 2.1% attributable to CS and 4.9% due to FS.
In this report, calculations for the adjustment to relative risk for misclassification are
made directly from the specific values assumed for CS, FS, and NS, without combining CS and
FS into a common category (ES). This approach is somewhat more general and should provide
additional flexibility and accuracy. The extended formula is given in Appendix B. The
calculations in Wald et al. and the NRC report are a special case. Wells has undertaken to
model the impact of misclassification in further detail by using parameter estimates that are
characteristic of the time and place of each study for which a bias is estimated, e.g., in estimates
of relative risk for CS and FS. His initial results indicate that the overall relative risk estimate
increases with this modification (personal communication from A.J. Wells).
The NRC report and Wald et al. make a correction to the overall relative risk estimate,
using identical procedures, to account for background exposure to ETS. Urinary cotinine is
used as a surrogate for recent exposure to tobacco smoke in NS. In the study by Wald and
Ritchie (1984), the urinary cotinine levels among NS exposed to smoking spouses were three
times those of NS married to NS. To make the lung cancer risk to exposure to ETS relative to
a zero-exposure group, it is assumed that the excess lung cancer risk from ETS exposure is
proportional to urinary cotinine concentration. Note that this assumption does not imply that
cotinine (or its precursor nicotine) determines the carcinogenic potential of ETS. The formula
to calculate the adjustment is in Equation B2 in Appendix B.
The NRC report calculates the population-attributable risk, assuming that 17% of
nonsmoking women and 12% of nonsmoking men are exposed to ETS and that the remaining
percentages are exposed to a background level of ETS (NRC, 1986, Appendix C). The exposure
4-7 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
percentages for men (12%) and women (17%) are from a sample of cotinine measurements in
men reported in Wald and Ritchie (1984). The 17% is a very low figure for female NS when
compared to the percentages exposed in the case-control studies (see Figure 3-1). The 121
married males tested by Wald and Ritchie are not representative of the U.S. population. Wald
et al. (1986) assume an exposure percentage of 59% in their example calculation for female NS
and the NRC report (p.236) uses an odds ratio of 1.3/3.3 (= 39%) for the exposure of female
nonsmokers (NS in this case) in its example. In the analysis of this report in Section 4.4, 60% is
assumed, with a plausible range from 45% to 75%. These values are reasonable for the case-
control percentages in Figure 3-1.
4.2.2. Other Risk Assessments Based on Epidemiologic Data
Wells (1988) provides a quantitative risk assessment that includes several epidemiologic
studies subsequent to the NRC and U.S. SG reports. Like the NRC report, the epidemiologic
data for both women and men are considered, for which separate estimates of overall risk and
attributable risk are provided. The three cohort studies addressed in the NRC report are also
included (no additional ones have appeared). Fourteen case-control studies are analyzed, three
of which have appeared in the literature subsequent to the NRC report (Brownson et al., 1987;
Humble et al., 1987; Lam et al., 1981; denoted in Table 3-1 as BROW, HUMB, and LAMT,
respectively). Other differences compared to the NRC report are: WU (Wu et al., 1985) and
the study by Sandier et al. (1985) are included; part of the data of BUFF and KABA are
excluded; and CHAN is excluded. The reader is referred to Wells (1988) for discussion of
criteria for study inclusion.
Wells calculates an overall relative risk of 1.44 (95% C.I. 1.26, 1.66) for females and 2.1
(1.3, 3.2) for males. Following the general approach of Wald et al. (1986), the misclassification
percentage for ES is assumed to be 5% (compared to 7% for Wald et al.). Rates were adjusted
4-8 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
for background exposure to ETS except in studies from Greece, Japan, and Hong Kong, where
the older nonsmoking women are assumed to experience very little exposure to ETS outside the
home. A refinement in the estimation of population-attributable risk is provided by adjusting
for age at death (which also appears in the calculations of Robins [NRC, Appendix D]). The
relative risk is calculated under two assumptions regarding risk-constant with age and declining
with age, but there is little difference in the two outcomes. A figure of 76% is used for the
fraction of female nonsmokers (never-smokers) exposed to ETS, assuming that 60% are exposed
to spousal smoking and another 16% are exposed at a comparable level otherwise. The total
percentage for males is 61%. The calculation of population-attributable risk applies to FS as
well as NS, which is a departure from Wald et al. and the NRC report. The annual number of
excess LCDs in the U.S. is estimated to be 1,232 (females) and 2,499 (males) for a total of
3,731. About 3,000, however, is thought to be a best current estimate.
Robins (NRC, 1986 Appendix D) explores three approaches to assessment of lung cancer
risk from exposure to ETS, each with attendant assumptions clearly stated. Method 1 is based
solely on evaluation of the epidemiologic data applying two assumptions: 1) adjustment of
relative risk for background exposure to ETS independent of age, and 2) the excess relative risk
in a nonsmoker is proportional to the lifetime dose of ETS. The validity of both assumptions
are questioned by the author in later remarks. The age-adjusted population-attributable risk is
estimated for females and males separately. (The reader is referred to Robins et al., 1989).
The age-specific fraction of LCDs due to ETS exposure is also required. Data from the
controls of one of the case-control studies in the U.S. (Garfinkel et al., 1985) are used for that
purpose. The age-specific LCD rates for women are also used in calculating population-
attributable risk for males in lieu of no data specific to males. Robins assumes that a relative
risk of 1.3 is associated with ETS exposure (1.14, the summary for U.S. studies alone, is also
considered). He omits any adjustment for misclassification, but does adjust for background
4-9 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
exposure. In view of the NRC report's emphasis on the potential for misclassification, this
omission is surprising. The author estimates lifetime risk of LCD attributable to ETS, but
epidemiologic data alone are not sufficient. Further assumptions implying some similar
characteristics in lung cancer risk from active and passive smoking are introduced for that
purpose.
Blot and Fraumeni (1986) published a review and discussion of the available epidemiologic
studies about the same time as Wald et al. (1986), the NRC, and the U.S. SG reports appeared.
The set of studies considered by Blot and Fraumeni are almost identical to those included in the
NRC report (see Table 4-1), except for omission of one cohort study (Gillis et al., 1984), and
inclusion of WU, the case-control study excluded by the NRC because the raw data were
unpublished. An overall relative risk estimate calculated from the raw data for females yields
1.3 (95% C.I. 1.1, 1.5). When the results are combined for high exposure categories, the overall
relative risk estimate is 1.7 (1.4, 2.1).
Wigle et al. (1987) apply the epidemiologic evidence to obtain estimates of the number of
LCDs in NS due to ETS in the population of Canada. A total of 50 to 60 LCDs per year is
attributed to spousal smoking alone, with 90% of them in women. Overall, involuntary exposure
to tobacco smoke at home, work, and elsewhere may cause about 330 LCDs annually. The
percentage of LCDs in NS used in the calculation is 1.6% for males and 12.4% for females,
values obtained by pooling results from U.S. and Canadian reports. The fraction of NS with a
smoking spouse is estimated from the control groups in two U.S. studies (Dalager et al. [1986]
and the case-control study of the ACS reported in Garfinkel et al. [1985]). A pooled prevalence
rate of 40% from these two studies is assumed. The estimated number of deaths from lung
cancer attributable to passive smoking is calculated separately for males and females, using age-
specific population figures for Canada and age-specific rates of death from lung cancer
attributable to ETS (Repace and Lowrey, 1985).
4-10 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 4-1. EPIDEMIOLOGIC STUDIES INCLUDED IN OVERALL RELATIVE RISK IN
THIS REPORT (FEMALES ONLY) AND SEVERAL OTHER SOURCES
Study
AKIB
BROW
BUFF
CHAN
CORR
GAO
GARF
GENG
HUMB
INOU
KABA
KOO
LAMT
Sex
F
M
F
M
F
M
F
F
M
F
F
F
F
F
F
M
F
F
Observed
relative risk1 1
1.52(0.88,2.64) F
1.80(0.50,5.60)
1.52(0.39,5.99)
1.38( - , - )
0.81(0.34,1.90)
0.50(0.20,1.70)
0.75(0.43,1.30) F
2.07(0.82,5.20) F
2.00( - , - )
1.19(0.82,1.73)
1.31(0.87,1.98) F
2.16(1.09,4.28)
2.34(0.83,6.61)
2.55(0.74,8.78)
0.79(0.25,2.48) F
1.00(0.20,4.90)
1.55(0.90,2.67) F
1.65(1.16,2.35)
2
F
M
F
M
F
F
M
F
F
M
F
Sources2
3
F
M
F
M
F
M
F
M
F
F
F
M
F
F
4
F
M
F
M
F
F
M
F
F
F
M
F
5
F
F
F
F
F
F
F
F
F
F
F
F
F
(continued on following page)
4-11
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 4-1. (continued)
Study
LAMW
LEE
PERS
SVEN
TRIG
WU
GARF
(Coh)
GILL
(Coh)
HIRA
(Coh)
Sex
F
F
M
F
F
F
F
F
F
M
F
M
Observed
relative risk1 1
2.01(1.09,3.71)
1.03(0.41,2.56) F
1.30(0.38,4.42)
1.28(0.76,2.15) F
1.26(0.57,2.81)
2.13(1.19,3.81) F
1.41(0.54,3.67) F
1.18(0.90,1.54)
1.00(0.20,4.91)
3.25(0.60,17.65)
1.63(1.25,2.11)
2.25(1.04,4.85)
2
F
M
F
F
F
F
M
F
M
Sources2
3
F
M
F
F
F
F
F
M
F
M
4 5
F
F F
M
F F
F
F F
F
F F
F F
M
F F
M
1 Figures for case-control studies are as recorded in Table 2-5,
with the values calculated in this report from raw data used
for females. Figures for cohort studies are taken from NRC
(1986). Parentheses contain 95% confidence intervals.
2 Sources are:
1. Blot and Fraumeni (1986).
2. NRC (1986) and Wald et al. (1986).
3. Wells (1988): includes study by Sandier et al. (1985)
on women, and the data for males in HUMB (Humble et
al., (1987), both of which contain very few cases of
lung cancer.
4. Saracci and Riboli (1989).
5. This report.
4-12
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Unlike the previous examples discussed, Wigle et al. use the relative risk estimates
obtained from a study comparing Seventh-Day-Adventists (SDAs) (Phillips et al., 1980a, 1980b)
with a matched group of non-SDAs who are also NS, as reported in Repace and Lowrey (1985).
The SDA/non-SDA comparison is used as a basis for assessing lung cancer risk from ETS in a
broader environment, particularly outside the home, than the case-control and cohort studies. It
provides an independent source of data and an alternative approach for comparison, to be
described further in the review to follow.
Repace and Lowrey (1985) suggest two methods to quantify lung cancer risk associated
with ETS. The one based on epidemiologic data estimates the relative risk of LCD from all
sources of exposure to ETS, i.e., in the home, at work and elsewhere, in what they describe as a
"phenomenologic" approach. A comparison of LCDs in the study by Phillips et al. (1980a,
1980b) referred to above, wherein SDA NS and a demographically/educationally matched cohort
of non-SDA NS provides the basic data. Information regarding the number of age-specific
LCDs and person-years at risk for the two cohorts is obtained from the study. The comparison
of two groups of NS is based on the premise that the non-SDA cohort is more likely to be
exposed to ETS than the SDA groups due to differences in life-style. Relatively few SDAs
smoke, so an SDA NS is probably less likely to be exposed at home by a smoking spouse, or in
the workplace, or elsewhere if associations are predominantly with other SDAs. One of the
virtues of this novel approach is that it contributes to the variety of evidence for evaluation and
provides a new perspective on the topic.
Phillips et al. reported that the non-SDA cohort experienced an average lung cancer
mortality rate equal to 2.4 times that of the SDA cohort. Using 1974 U.S. Life Tables, Repace
and Lowrey calculate the difference in lung cancer mortality rates for the two cohorts by 5-year
age intervals and then apply this value to an estimated 62 million NS in the U.S. in 1979, to
obtain a number of LCDs attributable to ETS annually. The result, 4665, corresponds to a risk-
4-13 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
rate of about 7.4 LCDs per 100,000 person-years. In an average lifespan of 75 years, that value
equates to 5.5 deaths per 1000 people exposed.
A recent article by Vainio and Partanen (1989) assumes that the observed relative risk of
1.3 from Wald et al. (1986) represents a causal effect. No adjustment is made for possible
misclassification of subjects. The same correction method for background exposure used by the
NRC and Wald et al. is applied to the observed relative risk to yield 1.53. Two calculations are
made for population-attributable risk applying an excess risk of 0.53 to the exposed fraction of
the population of NS and a value of 0.18 for excess risk from background exposure to the
remaining fraction. The two calculations are identical aside from the different values of the
fraction exposed: 12% (men) and 17% (women) in one case; 28% (men) and 56% (women) in
the other. The first pair of values is identical to the fractions used in the NRC report, as
discussed previously in this section. The second set is from one of the case-control studies
(Humble et al., 1987). The value for population-attributable risk calculated from the exposure
percentages 12%, 17%, 28%, and 56%, are 18%, 19%, 22%, and 27%, respectively, which
illustrates only moderate sensitivity of the calculations to the different values assumed for the
fraction exposed. Vainio and Partanen use the same excess relative risk for both men and
women in their calculations. The authors conclude that the proportion of lung cancer cases
among nonsmokers that could reasonably be attributed to ETS is 20% to 30%. This range is
consistent with population exposure percentages of 20% to 75%. A plausible range for exposure
percentages is about 45% to 75% (Section 4.4.2). The approximate range for PAR concluded by
Vainio and Partanen is close to what the calculations in this report would be without a
downward adjustment of the RR estimate for smoker misclassification bias.
Saracci and Riboli (1989), of the International Agency for Research on Cancer, review the
evidence from the three cohort studies and 11 of the case-control studies (Table 4-1). The
authors follow the example of the NRC and Wald et al. in the studies to exclude, and add only
4-14 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
one additional case-control study (Humble et al., 1987). The overall observed relative risk for
the studies, 1.35 (1.20, 1.53), is about the same as reported by the NRC, 1.34 (1.18, 1.53). It is
not reported how the overall relative risk was calculated.
4.3. APPROACHES TO RISK ASSESSMENT BASED ON CIGARETTE-EQUIVALENTS
The cigarette-equivalents approach assumes that the dose-response curve for lung cancer
risk from active smoking also applies to passive smoking, after conversion of exposure to ETS
into an "equivalent" exposure from active smoking. For example, suppose the average cotinine
concentration in exposed NS is 1% of the average value found in people who smoke 30 cig./day.
The lung cancer risk for a smoker of (.01)30 = 0.3 cig./day is estimated by low-dose
extrapolation from a dose-response curve for active smoking, and that value is used to describe
the lung cancer risk for ETS exposure. This general explanation describes the nature of the
approach; however, authors vary in their constructed solutions and level of detail. The basic
assumption of cigarette-equivalents procedures is that the lung cancer risks in passive and active
smokers are equivalently indexed by the common measure of exposure to tobacco smoke, i.e., a
common value of the surrogate measure of exposure in an active and a passive smoker would
imply the same lung cancer risk in both.
A difficultly in assessing this approach lies in evaluating the assumption that apparent
differences between passive and active smoking are negligible or have cross-effects that cancel.
For example, MS and SS differ in the relative composition of carcinogens identified in tobacco
smoke and in their physicochemical properties in general. The lung and systemic distribution of
chemical agents common to MS and SS are affected by their relative distribution between the
vapor and particle phases, which differs between MS and SS and changes with SS as it ages.
Passive and active smoking also differ in characteristics of intake-intermittent (possibly deep)
puffing in contrast to normal (shallow) inhalation. To help illuminate relationships and identify
4-15 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
parameters where additional information would be helpful on this topic, a mathematical model
for comparison of dosimetry of passive and active smoking was constructed as a basis for
further study (Appendix C).
Several authors have taken issue with the validity of the cigarette-equivalents approach.
For example, Hoffmann et al. (1989), in discussing the longer clearance times of cotinine from
passive smokers than from active smokers, concludes "The differences in the elimination time of
cotinine from urine preclude a direct extrapolation of cigarette-equivalents to smoke uptake by
involuntary smokers." A recent consensus report of an IARC panel of experts (Saracci, 1989,
p.3) states that "Lacking knowledge of which substances are responsible for the well established
carcinogenic effect of MS, it is impossible to accurately gauge the degree of its similarity to ETS
in respect to carcinogenic potential." The U.S. SG report devotes a three page section to the
concept of cigarette-equivalents, quantitatively demonstrating how they can vary as a measure of
exposure (U.S. SG, 1986). It concludes with "These limitations make extrapolation from
atmospheric measures to cigarette-equivalents units of disease risk a complex and potentially
meaningless process." On a lesser note, it has generally been assumed that the dose-response
relationship for active smokers is reasonably well understood. Recent literature raises some
questions on this issue (Moolgavkar et al., 1989; Gaffney and Altshuler, 1988; Freedman and
Navidi, 1987a, 1987b; Whittemore, 1988).
The cigarette-equivalents approach has some important limitations, due in large part to
limited knowledge regarding similarities and differences between passive and active smoking and
how to adjust for them in a risk assessment. Legitimate reservations not withstanding, virtually
all analytic approaches bear some assumptions and weaknesses, and most contribute something
to our understanding. Further development of the cigarette-equivalents approach and the
knowledge base surrounding it may be worthwhile. Three new methods akin to cigarette-
equivalents approach are described in Appendix D, with comments and advice solicited. Several
4-16 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
published examples of the cigarette-equivalents approach follow. Although a risk assessment
based on the epidemiologic data is preferred in this report, it is worthwhile to consider the
spectrum of methods and approaches that have been tried.
Vutuc (1984) estimates that exposure of passive smokers to cigarette smoke is equivalent
to 0.1 to 1.0 cig./day actively smoked. This relationship follows Repace and Lowrey (1980),
except that Vutuc adjusts their figures to apply to a cigarette with tar content of 16 mg instead
of 0.55 mg, as assumed by Repace and Lowrey. For the smoking situations indicated by Repace
and Lowrey, who found passive smoking equivalent to actively smoking 5 to 27 cig./day, Vutuc
obtains an equivalent of actively smoking 0.2 to 1.0 cig./day.
Citing cigarette-equivalents calculated in other sources, Vutuc assumes a range of 0.1 to
1.0 cig./day for ETS exposure. Relative risks for nonsmokers are calculated for 10-year age
intervals (40 to 80) based on the reported relationships of dose, time, and lung cancer incidence
in Doll and Peto (1978). Relative risks for smokers of 0.1 to 1.0 cig./day give a range in relative
risk from 1.03 to 1.36. The author concludes that "As it applies to passive smokers, this range of
exposures may be neglected because it has no major effect on lung cancer incidence." As
observed by the author, however, the influence of ETS on lung cancer incidence becomes more
marked in the higher age groups, where the carcinogenic effect of tobacco smoke is strongly
influenced by the duration of exposure. From Vutuc's Table 1, the increase in incidence (per
million) in lung cancer for a smoker of one cig./day is 270 at age 79, 130 at age 70, 40 at age 60,
etc. These values of risk slightly exceed the acceptable levels typically used by the EPA and
other regulatory agencies in setting standards for pollutants under their authority.
Vutuc assumes that his figures apply to both males and females. If an exposure fraction
of 75% is assumed for both males and females, the range of relative risks given correspond to a
range for population-attributable risk. The number of LCDs among NS in the U.S. in 1988 is
about 6500 females and 3000 males (personal communication from the ACS). The number of
4-17 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
LCDs in NS attributable to ETS is estimated to range from 240 to 2020 (140 to 1380 for females
alone). So Vutuc's figures are consistent with several hundred excess LCDs among NS in the
U.S. These figures are from our extension of Vutuc's analysis, however, and are not the claim
of the author.
Methods 2 and 3 of Robins (NRC, Appendix D) are constructs of the cigarette-equivalents
approach. In both methods, a range of values is reported corresponding to a range of unknown
parameter values. Method 2 uses an overall relative risk value based on epidemiologic data, but
also makes some assumptions to appeal to results of Day and Brown (1980) and Brown and Chu
(1987) on lung cancer risk in active smokers. The author estimates the number of excess LCDs
due to ETS, assuming 7000 and 5200 annual LCDs in female and male NS, respectively.
Adjusting his results to 6500 females and 3000 males (for comparison purposes), the range of
excess LCDs attributable to ETS is 1650 to 2990 for females and 420 to 1120 for males.
Robins' Method 3 ignores the epidemiologic data on passive smoking entirely and
extrapolates from data on active smoking, along with several assumptions. Applying his results
to 6500 females and 3000 males, the range of excess LCDs due to ETS is 550 to 2940 for
females and 153 to 1090 for males.
Arundel et al. (1987) attribute only five LCDs among female NS to ETS exposure. The
corresponding figure for males is seven (both figures are adjusted to 6500 females and 3000
males). The expected lung cancer risk for NS is estimated by downward extrapolation of the
lung cancer risk/mg of particulate ETS exposure for CS. Their premise is that lung
carcinogenicity of ETS is entirely attributable to the particulate phase of ETS, and the
consequent risk in passive smoking is comparable to active smoking on a per mg basis of
particulate ETS retained in the lung. If the vapor phase of ETS were also considered, the
number of LCDs attributable to ETS would likely increase.
4-18 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Russell and coworkers (1986) use data on urinary nicotine concentrations in smokers and
nonsmokers to estimate exposure and risk from passive smoking. The risk of premature death
from passive smoking is presumed to be in the same ratio to premature death in active smokers
as the ratio of concentrations of urinary nicotine in passive to active smokers (about 0.007).
Calculations are made using vital statistics for Great Britain and then extrapolated to the United
States. The latter estimate, 4000+ deaths/year due to passive smoking, is for all causes of
death, not just LCDs.
Repace and Lowrey (1985) describe a cigarette-equivalents approach as well as the
procedure described previously. One objective is to provide an assessment of exposure to ETS
from all sources that is more inclusive and quantitative than might be available from studies
based on spousal smoking. They consider exposure to ETS both at home and in the workplace,
using a probability-weighted average of exposure to respirable suspended particulates (RSP) in
the two environments. Exposure values are derived from their basic equilibrium model relating
ambient concentration of particulates to the number of burning cigarettes per unit volume of air
space and to the air change rate. From 1982 statistics of lung cancer mortality rates among
smokers and their own previous estimates of daily tar intake by smokers, the authors calculate a
lung cancer risk for active smokers of 5.8 X 10-6 LCDs/year per mg tar/day per smoker of lung
cancer age. The essential assumption Unking lung cancer risk in passive and active smokers is
that tobacco tar inhaled poses the same risk to either on a per unit basis. Extrapolation of risk
from exposure levels for active smokers to values calculated for passive smokers is accomplished
by assuming that dose-response follows the one-hit model for carcinogenesis. An estimated 555
LCDs per year in U.S. nonsmokers (NS and FS) is attributed to ETS exposure (for 1980). The
ratio of total LCDs in 1988 to 1980 is approximately 1.37 (Repace, 1989). With that population
adjustment factor, the approximate number of LCDs attributable to ETS among nonsmokers is
closer to 760 for 1988 (including FS).
4-19 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The potential for bias due to misreported smoking habits was apparently first noted by
Lee (see discussion in Lehnert, 1984), and has been emphasized by him in several articles, e.g.
Lee (1986, 1987a, 1987b). In Lee (1987b), it is argued that smoker misclassification may explain
the entire excess lung cancer risk observed in self-reported NS in epidemiologic studies. A
hypothetical example is first provided to the reader to illustrate that if 5% of reported
nonsmokers are actually smokers, and the relative risk of lung cancer of smokers to nonsmokers
is 20, then a relative risk as high as 1.75 could be observed for ETS exposure to spousal
smoking. The example is a little misleading in view of the discussion that follows in the article
on the results of three separate studies aimed at measuring the accuracy of reported current
smoking (a cotinine study), the accuracy of reported lifetime smoking (a 1980/1985 follow-up
study), and concordance of smoking habits in married couples (a 1985 consumer study).
("Marriage aggregation factor" in NRC [1986] and Wald et al. [1986] is a measure of
concordance). All three studies were conducted on British or UK subjects ages 1.6 and above.
Following review of these studies, the author assumes more refined parameter values.
The relative risk for smoking is assumed to be 10 instead of 20. The evidence suggests that
about 1.4/2.5 (56%) of misclassified CS may be regarded as "regular smokers" and 1.1/2.5 ( =
44%) as only "occasional smokers." The relative risk of the latter is assumed to be 2.5 instead of
10. Based on cotinine measurements, 20 of the 689 self-reported NS (2.9%) are treated as CS, 9
as occasional and 11 as regular smokers. (In the notation of Table B-l, this equates to RR(E/C)
= 6.7). The relative risk of misclassified FS is assumed to be 2.0 (giving RR(E/F) = 2.0 in
Table B-l). The percentage of reported NS assumed to be FS is 10, equivalent to 21% of the
(true) FS. No supporting evidence was found for the 10% figure for FS, but the follow-up study
does provide evidence that the percentage may be lower for women than men. Applying an
adjustment for women to the 10% and 21% figures above gives 2.8% and 7%, respectively, for
4-20 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
women. (This correction for estimation of risk to women was of little consequence to our
calculations.)
The remaining parameter values from Lee (1987b) needed to apply the formula for
adjustment to misreporting bias in Equation Bl are as follows, where the variable identifiers are
described in Table B-2: (Vl.0.336), (V4.0.025; 0.011 for regular smokers plus 0.014 for
occasional smokers), (V5.0.483), (V8.0.181), (Vll.0.10), (V16.6.7; from a RR of 10 for CS and
2.5 for FS), and (V20.2.0). The observed relative risk for a true value of one from Equation Bl
is 1.18. With the correction for women described, 1.18 would be reduced to 1.16. These values
are close to what the NRC report and Wald et al. calculated for an expected value of the
observed relative risk when the true value is one. The excess above one is the anticipated bias
for smoker misclassification.
Assuming that the parameter values specified above accurately reflect the author's
description, the method of adjusting for misclassification detailed in Appendix B (adapted in
principal from Wald et al. [1986] and the NRC [1986]) does not support the claim that self-
reported misclassification would fully account for the excess risk of lung cancer observed in the
epidemiologic data. The lower 95% confidence limit on the overall summary observed relative
risk of 1.40 is 1.25, and the 99% lower limit is 1.21, still above the range potentially explained by
the misreporting of smoking habits alone.
The study results discussed by Lee for setting parameter values need to be included in the
larger pool of related information and study evidence available, particularly for inference on
U.S. women of age 35 + . It appears likely that there are sex-related differences in rate of
misreporting, and possibly in other parameters. Age may be a factor. It has not been indicated
whether persons of ages 16 to 30, too young to be included in epidemiologic studies, are
representative of the older age groups in terms of rates of misreporting and other factors. When
4-21 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
only 20 out of 808 reported nonusers of tobacco are reclassified as CS (11 as regular users at a
relative risk of 10 or so), a few subjects may have a large impact on the outcome.
Several further observations on this article may be noted. It is concluded that the overall
observed relative risk from spousal smoking to ETS can be explained as bias due to probable
overstatement of the number of reported NS. It would follow that the true excess risk is zero.
Extrapolation of risk from active smoking using average cotinine concentration in smokers and
NS married to a smoker, however, is recommended in place of analysis of the epidemiologic
data. It is concluded that there is a positive excess risk to ETS exposed NS (1.02 for women
and 1.07 for men). As discussed previously, values in this range cannot be regarded as
negligible.
4.4. CURRENT ASSESSMENT OF LUNG CANCER RISK
The data from epidemiologic studies currently available are evaluated for evidence of an
elevated occurrence of lung cancer associated with ETS exposure. The methodological approach
of Wald et al. (1986) and the NRC report are adapted with two minor modifications. The
adjustment for relative risk of ES is calculated by its separate components (CS and FS). The
second modification is to distinguish between parameter values for "reported" and "correct"
classifications. Some parameter estimates depart from values used by the NRC, largely
reflecting more current evidence and information available in some areas (Section 4.4.2). The
most significant difference in parameter values is probably in the relative risk of active smoking.
The value of eight assumed by the NRC is replaced by 12 (from the ACS's Study CPS-II, as
reported in U.S. SG [1989]). Following the adjustment to an overall relative risk estimate for
misclassification, a further adjustment is made to account for background exposure, i.e., for
exposure to NS aside from spousal smoking, the dominant measure in the epidemiologic studies
to distinguish between exposed and unexposed subjects.
4-22 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The twice-adjusted RR estimate is combined with the percentage of the population
exposed to ETS to obtain an estimate of the population-attributable risk (PAR), the proportion
of LCDs in female NS of age 35 and older attributable to ETS exposure. Upper and lower
confidence limits on the PAR are determined. These limits are conditional on the population
exposure percentages assumed. Multiplying the PAR by the number of LCDs in never-smoking
women in 1988 estimates the excess number of LCDs attributable to ETS.
In the final section of this chapter the population-attributable risk is calculated separately
for each of the 19 case-control studies with data available and the three cohort studies, using the
exposed fraction of controls from each study in the calculations. The ordered outcomes provide
a basis for reviewing similarities/dissimilarities between studies, such as the country of origin
and other study characteristics shown in Chapter 3. Although it is likely that the true
differential in ETS exposure for subjects classified as exposed or unexposed is relatively higher
in some sampled human environments than others (probably higher in Japan, for example, as
discussed by Hirayama and others), comparison of outcomes with study characteristics did not
reveal any apparent patterns or study characteristics associated with the findings.
4.4.1. Combining Evidence Across Studies
The overall relative risk estimate for case-control studies is determined by the extended
Mantel-Haenzsel procedure. The M-H method was applied to case-control studies in Chapter 3
to test for an association between ETS and lung cancer. This same method is now applied to
the raw data of case-control and cohort studies to obtain an overall estimate of RR and
confidence interval. The procedure used in the NRC report (1986, Appendix B) is basically this
same method (Yusuf et al., 1985). The method of combining cohort studies, and then obtaining
an overall value for case-control and cohort studies together, follows the NRC procedure.
4-23 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The case-control studies included in the analysis and their observed relative risks from raw
data are in Table 4-1 (also see Table 3-5). Two studies, SHIM and VARE are excluded because
the raw data were not available. The overall estimate of relative risk for the 19 case-control
studies and three cohort studies in Table 4-1 is 1.41 (95% C.I. 1.26, 1.57). This value may be
compared to 1.32 (1.16, 1.51) in the NRC report for females (three cohort studies and ten case-
control studies), and to 1.50 (1.3, 1.8) determined by Wells (1986) from three cohort studies and
14 case-control studies. The shortest confidence interval is for the analysis of this report,
indicating that the additional studies since the NRC report have reduced the statistical
uncertainty in the estimated RR as would be anticipated. The higher estimate of relative risk in
Wells' analysis is largely due to the choice of studies. The study designated as CHAN, reporting
a relative risk value of only 0.75, was excluded by Wells but was included in the NRC report and
in our analysis.
The summary RR of 1.41(1.26, 1.57) in this report is from the combined values of 1.42
(1.24, 1.63) for 19 case-control studies and 1.39 (1.15, 1.67) for three cohort studies. These
figures are almost identical, so the results of one type of study reinforce the outcome from the
other type. The consequence of combining the two summary outcomes is essentially just to
shorten the confidence interval which is equivalent to reducing the standard error of the
estimate of RR. For U.S. studies alone, the RR is 1.25 (1.03, 1.52), the combined RR from
seven case-control studies is 1.34 (1.00, 1.79), and GARF(coh) is 1.18 (0.90, 1.54).
Adjustment for smoker misclassification, using the method and parameter values described
below, reduces the overall observed relative risk of 1.41 (1.26, 1.57) to 1.28 (1.12, 1.45).
Modification for background exposure to ETS raises it to 1.48 (1.21, 1.87). The NRC
committee, with eight fewer case-control studies, obtained an overall summary value of 1.34
(1.18, 1.53) for both men and women that was adjusted to 1.25 (with 1.15 to 1.35 possible) for
misclassification, and then to 1.42 ("ranging" from 1.24 to 1.61) for background exposure. The
4-24 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
values for women alone should be only slightly lower judging from the initial overall relative risk
of 1.32 for women.
4.4.2. Adjustment To Relative Risk for Smoker Misclassification
The reduction formula for misreporting current smokers (CS) and former smokers (FS) as
never-smokers (NS) is described in Appendix B. The general equation (Bl) is accompanied by
a description of parameters in Tables B-l and B-2. Alternative specification of parameters in
terms of "reported" and "correct" values is sometimes useful. The algebraic relationships for this
conversion are shown in Table B-3. The description of parameter values used to adjust for
smoker misclassification in this report follows, with variable identifiers in parentheses for use
with Table B-2.
The percentages of reported CS, FS, and NS are 21.3, 24.0, and 54.7, respectively (V12, V17,
V15). These values are obtained from the ACS's study CPS-II (Stellman and Garfinkel, 1986)
except that 3% of the reported "never-smoked regularly" (56.4%) were reclassified as FS, leaving
54.7% estimated NS (see footnote d, Table 2, of U.S. SG [1989]). The NRC report and Wald et
al. (1986) use 25%, 15%, and 50% for CS, FS, and NS, respectively, from a study of smoking
habits in the U.K.
The percentages of reported NS who are misclassified as CS and FS are 2 (1.5, 2.5) and 4
(2, 6), respectively (VI, V2), where the parenthetical values denote a reasonable range. Wald et
al. describe evidence to support values of 1.6% and 4.9% for these parameters respectively. Lee
(1987b) makes a case for 2.5% (1.1% for regular smokers and 1.4% for occasional smokers) and
assumes 10% for misclassified CS and FS, respectively (both males and females.) Adjusting the
10% value to reflect a lower rate of misreporting by females (Lee, 1987b) would make it 2.8%.
Cummings (1989b) found that six subjects of a total 669 (1%) reported NS and FS were
misreported CS (urinary cotinine above 90 mg/dL). Jarvis et al. (1984) found that 21 of 121
4-25 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
reported nonsmokers (17%) were CS based on biochemical markers. This very high rate was
found among elderly patients attending clinics for smoking-related disease, whose doctors had
frequently urged them to give up smoking.
The percentages of misclassified CS and FS who are exposed (married to a smoker) are
both 82 (71, 92) (V5, V6). These values correspond to an exposure percentage of 60 (45, 75)
(V4) for reported NS, and a marriage aggregation factor (MAF) of 3.07 for females (from Lee,
1987b). (The MAF is the ratio of cross-products, i.e., the odds ratio, in a 2-by-2 table of
smoking status of subject by smoking status of spouse. The MAF is assumed to be the same for
NS compared with either CS or FS). Table 12-7 of the NRC gives MAF = 3.1 for females, a
value communicated from Wald et al. The NRC report considers values of 2.5, 3.5, and 4.5,
with 3-4 likely.
Evidence suggests that CS and FS who are misclassified as NS are likely to be only light
smokers and that misclassified FS have typically stopped smoking several years previously (Wald
et al., 1986). Estimates of relative risk for CS and FS (females) are 11.94 (9.99, 14.26) and 4.69
(3.86, 5.70) for 1982-1986 (ACS's CPS-II Study, as reported in U.S. SG [1989]). Our report
assumes the relative risks of misclassified CS and FS are 5.95 (5.0, 7.15) and 2.97 (2.7, 3.86)
(V7, V8), respectively (discussed and compared with other sources below). If exposed, the
values are incremented by the excess risk from exposure to ETS (1-V(9)) to obtain variables
V10 and VI1. (Technical Note: Equation (Bl) is implemented by setting a value for V(9) along
with other parameters to obtain a corresponding value of RRO. To find the value of RRM that
corresponds to a specific value of RRO may require a few iterations from the starting point
chosen for RRM.)
Parameter values for the RR of CS and FS misreported as NS have been determined in
other sources from the information available for the relative risk of smokers and FS in general.
To compare our parameter values of the RR for CS and FS misclassified as NS with those of
4-26 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
other authors, their formulas for reducing the RR of CS and FS who are misreported NS have
been applied to the RR figures from the CPS-II survey given above. For misclassified CS, the
results are-NRC (1986): the range from 3.0 to 6.0; Wald et al. (1986): 3.8(3.3, 4.3); Wells
(1988): 4.6(4.0, 5.4); and Lee (1987b): 8.0(6.7, 9.6). All but the last entry are below the value
6.0(5.0, 7.2) in this report. For misclassified FS, the figures in these sources would be--NRC:
the range from 3.0 to 6.0; Wald et al.: 1.9(1.8, 2.1); Wells: 1.9(1.8, 2.1); Lee: 3.0(2.5, 3.6). The
last entry is comparable to 3.0(2.7, 3.9) of this report.
4.4.3. Parameter Sensitivity
For the parameter values described above for this report, a true relative risk of 1.00
corresponds to an observed relative risk of 1.14, a 14% inflation due to smoker
misclassification. For each of the parameters with a range of input values (shown in
parentheses), the lower and upper values in the parentheses were also applied, leaving the other
parameter values fixed. The marriage aggregation factor varied between 2.0 and 4.0 as well.
The observed relative risk did not exceed 1.19. The 95% lower confidence limit on the overall
observed relative risk from the epidemiologic data is 1.25, still well above the likely range
explained by misclassification. (The 99% lower confidence limit is approximately 1.21, still
above the explainable range). Only one parameter was set at an extreme value at one time in
the sensitivity testing, and it is not improbable that some combination of parameters chosen this
way would produce a value exceeding 1.25. But the observed risks from the epidemiologic data
appear unlikely to be explained by misclassification alone, and no single parameter within a
plausible range alters that conclusion.
4-27 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
4.4.4. Adjustment to Relative Risk for Background Exposure
The relative risk of ETS in epidemiologic studies is relative to the baseline risk of a
female NS married to a nonsmoker, who still has some background level of exposure to ETS.
Some assumptions are required to approximate the lung cancer risk due to background exposure
to ETS. A means of estimating the proportion of total ETS exposure that is due to background
in an exposed individual (actually exposed to background and to spousal smoke) is needed. The
NRC report compared average cotinine concentrations in exposed and unexposed persons.
Assuming that lung cancer risk from passive smoking is linearly related to cotinine
concentrations at these low doses, lung cancer risk of passive smoking can be estimated at the
higher exposure level (background plus spousal smoking, applicable to an exposed person) and
at the lower exposure level (background only, applicable to an unexposed person), with both
estimates relative to the risk of lung cancer risk from zero exposure to ETS.
Background ETS appears to constitute about one-third of the total ETS exposure of a NS
married to a smoker, if cotinine concentrations are used as an index of total exposure. The
ratio of average cotinine concentrations in exposed to unexposed married NS is assumed to be
three in the NRC report, based on evidence from Wald and Ritchie (1984). The Wald and
Ritchie study applies to men, but Lee (1987b) reports a ratio of 1/.3 (= 3.3) in women and
Coultas et al. (1986) report a ratio of 3.41/1.45 (= 2.35) from saliva cotinine levels in a
population-based survey of Hispanic subjects in New Mexico. Three is used in this report (X =
3 in Equation B2), along with the assumption that cotinine is a constant multiple of the
carcinogenic potency of ETS at low doses. Applying the method used by the NRC (Appendix B
of this report) to take background exposure into account changes the overall RR estimate
adjusted for misclassification from 1.28 (1.12, 1.45) to 1.48 (1.21, 1.87) for female NS. It should
be noted that the meaning of RR is changed with this adjustment-from meaning relative to the
risk from background ETS to meaning relative to the risk at zero-ETS exposure. (Note: There
4-28 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
is an important distinction between the use of cotinine as a surrogate dose for ETS to estimate
lung cancer risk from background exposure and its use in the cigarette-equivalents approach
(Section 4.3). In the latter, the contention centers around the assumption that cotinine (or
anything else, such as respirable suspended particles) is an equivalent dose surrogate for both
passive and active smoking, i.e., that equivalent uptake in passive and active smoking implies
equivalent carcinogenic risk.)
4.4.5. Population-Attributable Risk and Excess Lung Cancer Deaths
The number of LCDs in U.S. female NS in 1988 is estimated to be 6500 (3000 for male
NS). The 6500 figure includes both married and unmarried NS. The ACS's CPS-II Study
(reported in Stellman and Garfinkel, 1986) percentages for marital status of all women surveyed
(not just NS) are: married, 75.3; divorced 5.1; widowed, 14.6; separated, 0.8; and single, 4.2.
Our estimates of risk apply to married female NS, about 75% of female NS, so it is necessary to
consider exposure to ETS in the remaining 25% of unmarried NS.
Cummings (1989b) obtained urinary cotinine levels on a total of 663 self-reported NS and
FS. The cotinine levels were only slightly higher in males than females (9.6 and 8.2 ng/mL,
respectively), and slightly more than half of the subjects were females. The average cotinine
level (in ng/mL) was 10.7 for married subjects if the spouse smoked and 7.6 otherwise (all units
in ng/mL). Interestingly, the average cotinine levels reported by marital status are: married,
8.3; never married, 10.3; separated, 11.8; widowed, 10.4; and divorced, 9.2. The study, which
includes 7% of age 18 to 29, and 47% of age 60 to 84, does not claim to be representative.
Nevertheless, the results suggest that in terms of ETS exposure, an unmarried NS is probably
closer, on average, to a NS married to a smoker (an exposed person) than to a NS married to a
ratodmoker (an unexposed person). This observation is also consistent with the findings of
Friedman et al. (1983).
Oe\Vl\£0 4-29 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The percentage of married female NS who are married to smokers is assumed to be 60,
with a plausible range of 45 to 75 (Section 4.4.2.). The choice of exposure percentages is based
on the distribution of values observed in the epidemiologic studies (Table 3-2 and Figure 3-1).
From the discussion of exposure to unmarried female NS above, it is reasonable to assume that
exposure to ETS, on average, is at least as large as the sum of 60% of the higher exposure level
(from spousal smoking plus background) and 40% of the lower exposure level (background
alone) experienced by a married female NS. For the calculations needed from these figures, this
assumption is equivalent to treating unmarried and married female NS alike, in terms of
exposure to ETS (60% exposed at a level equivalent to spousal smoking plus background and
40% exposed at the background level). Alternatively, average exposure in the population of
female NS (including marrieds and unmarrieds) is assumed to be at the background level plus
60% of the exposure from spousal smoking. The percentages assumed by others in the
literature have varied: 82 (Lee, 1987b), 76 (Wells, 1988), 59 (Wald et al., 1986), and 17 (NRC,
1986). The NRC percentage was taken from a sample of urinary cotinine bioassays appearing in
Wald and Ritchie (1984). The 17% figure is likely much too low to be representative of the
population of interest.
The population-attributable risk (PAR) for women is the proportion of LCDs in female
NS per year associated with exposure to ETS. Multiplying it by the total number of LCDs in
female NS gives the number of excess LCDs per year from ETS exposure, i.e., the number
attributable to ETS exposure. For calculation of PAR in this report, the percentage of NS
exposed to ETS is assumed to be the same for marrieds and unmarrieds (60% as discussed
above). The PAR for the 60% assumption in women is 0.27 (95% C.I. 0.14, 0.41). (Technical
Note: The calculations are from Equation B3 with P(E/N) = 0.6, RRM = 1.28 and RRB =ob
1.48. The confidence interval is calculated by using the upper and lower confidence boumdsifor
RRM and RRB in Equation B3. The confidence interval is conditional on the exposunarnbahH
4-30 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
percentage of 60%). Multiplying the PAR by 6500, the number of LCDs in female NS in 1988,
gives 1750 (910, 2660) estimated LCDs in female NS in 1988. The number of LCDs in male NS
is probably about one-half the number of females, although the evidence for males is scant in
comparison to females. The available data indicate that the risk to male NS is not likely to be
smaller than for female NS. Applying the PAR value of 0.27 to the total number of LCDs in
male NS (3000), gives an estimate of 810 LCDs per year in male NS due to passive smoking.
For both sexes combined, the annual number of LCDs in NS attributable to ETS exposure is
about 2500, with a range of 1300 to 4000.
A figure for FS also needs to be included in the estimate of LCDs attributable to ETS
since they constitute a large segment of the population. Repace and Lowrey (1985) and Wells
(1988b) estimated the risk to FS. Sandier et al. (1985) and Geng et al. (1988) found,
respectively, an increased total cancer risk and an increased lung cancer risk in active smokers
as a result of passive smoking. Sandier et al. found that relative to active smokers who had no
other smokers at home, smokers who had one, two, or three or more smokers at home had
increased risks of (total) cancers of 40%, 120%, and 160%, respectively. To calculate the annual
number of LCDs in FS attributable to ETS, RR is assumed to be the same for FS and NS. The
figure calculated for FS is 1260 (540 women and 720 men), making the yearly total of LCDs
attributable to ETS approximately 3800. Data on the number of FS in the U.S. population and
the calculational details are included in Appendix B.
The 3800 figure is based on a population exposure percentage of 60, the mid-point in a
plausible range of 45% to 75%. The results, however, are not very sensitive to the exposure
percentage. If 45% (75%) is assumed instead of 60% in the preceding calculations, then the
yearly total of LCDs attributable to ETS is approximately 3500 (4100) instead of 3800. This
range, 3500 to 4100, corresponds to the plausible range of the exposure percentage, 45% to
75%, for the estimate of RR after adjustment for misclassification (RRM = 1.28) and correction
4-31 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
for background exposure to ETS (RRB = 1.48). The RRM and RRB figures are based on the
epidemiologic studies of female NS married to smokers and the simple linear model using
cotinine concentrations to account for background exposure to ETS. Ninety-five percent
confidence intervals are available for the population values of RRM and RRB, given by (1.12,
1.45) and (1.21, 1.87), respectively. Calculating PAR from the low (high) value of the plausible
range of exposure percentages, 45% (75%), and low (high) values of the 95% confidence
intervals for RRM, 1.12 (1.21), and RRB, 1.21 (1.87), provides estimates that are probably too
low (high). Using these values to recalculate the population risk provides numerical markers for
low and high extremes of the estimated number of LCDs due to ETS. These markers are
approximately 1800 and 6100.
In summary, the estimated total number of LCDs per year due to ETS exposure is 3800.
It is approximately the sum of estimates for NS females (1750), NS males (810), FS females
(540), and FS males (720), assuming a 60% exposure rate. Using the lower (upper) confidence
limits for RRM and RRB and the lower (upper) end of the plausible range for exposure
percentage gives approximately 1800 (6100) total LCDs per year due to ETS. These low (high)
totals consist of NS females 820 (2800), NS males 380 (1290), FS females 250 (860), and FS
males 330 (1150). Characterization of the population risk is based on data for female NS from
22 epidemiologic studies of varied design and protocol conducted in numerous locations under
ordinary environmental conditions. Extension of these results to the population of all U.S.
nonsmokers is not without uncertainty, but is probably conservative. It is unlikely that the true
number of LCDs per year in U.S. nonsmokers lies outside the interval defined by the two
extreme values, 1800 and 6100.
4.4.6. Adjusted Relative Risk and Population-Attributable Risk by Individual Study
The estimates of RR for lung cancer from ETS exposure have been statistically combined
4-32 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
to give an overall estimate and confidence interval. The overall observed value (RRO) was
adjusted for possible misclassification (to be called RRM) and for background exposure to ETS
(to be called RRB). The latter "correction" was to make the risk estimates relative to zero-
exposure.
Combining results from all studies increases the statistical power to detect an exposure
effect on lung cancer. Aside from weighing each study's results according to a measure of
statistical uncertainty (influenced but not solely determined by sample size), studies are treated
as if they were qualitatively equivalent and as if the "true" RR were the same in all
environments studied. Qualitative differences exist in all studies, but there is little basis for
quantifying them. The true values of RR being estimated depend on both the study design and
protocol. Culture, environment, and life-style would influence inter-study differences. In
particular, one might expect these factors to contribute to inter-country variability in the
epidemiologic data. To extract this source of variability statistically would require multiple
studies, similar in design and execution, from each country. Although there is more than one
study from several countries among those analyzed (China [2], Hong Kong [4], Japan [2],
Sweden [2], U.S. [8]), the studies are not sufficiently similar within countries to test variability.
In particular, there is considerable dissimilarity between the U.S. studies, and this probably
contributes to their wide ranging results.
The observed relative risk from each study is adjusted for misclassification and
background rate and the PAR is calculated to evaluate the dissimilarity of results. The
percentage of exposed NS for a study is taken from the observed percentage among controls of
the study (Table 3-2 or Figure 3-1). The exposure percentages for CS and FS are then
calculated assuming a marriage-aggregation factor of 3. The same exposure parameters are
used in both the calculations of misclassified smokers and the population-attributable risk.
By using 95% confidence limits on the RR estimate for each study in the calculations,
4-33 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
limits are obtained for the PAR. Interpretation of these limits is conditional on the percent
exposed, which is a random variable. Aside from the exposure percentages described above, the
values and method of calculation are the same as used previously in Section 4.4. The results of
adjusting each study's observed relative risk for misclassification are shown in Table 4-2.
(Technical Note: The overall observed RR, denoted by RRO, is corrected for misclassification
(RRM) and then for background exposure (RRB). The excess RR for background exposure is
approximately RRB - RRM, which applies to all NS. The additional excess risk to the
proportion of NS exposed to ETS equals RRM-1.) A minus sign (-) in Table 4-2 indicates that
the observed excess risk after adjustment for misclassification is negative, i.e., the adjusted RR
< 1. The adjustment for misclassification decreases as the observed relative risk increases and
is treated as negligible for values above 2.5. The observed value that would correspond to a true
RR of one was also calculated for each study. These range from 1.12 (BUFF) to 1.22 (BROW),
and cluster from 1.13 to 1.16. The excess risk from background exposure to ETS shown in Table
4-2 was arbitrarily limited to 0.2. When the excess risk after adjustment for misclassification is
zero, a correction for background is not appropriate. The values for PAR have associated
confidence intervals predominantly with a lower limit of zero (the minimum) and an upper limit
in the range 34% to 75%. Exceptions include the case-control studies CHAN and TRIG, and
the cohort study H1RA, with percentages of (0,13), (3,58), and (14.49), respectively.
The estimates of PAR and their upper confidence limits are rank-ordered in Table 4-3.
Probable values are shown for three studies in which the calculation could not be made due to
insufficient data. The studies with a zero estimate of PAR were reviewed to see if some
common characteristic might be apparent. Remarks about some of those specific studies follow.
The studies with an estimate of zero PAR will serve to indicate some possible sources of
disparity in results across other studies as well.
WU includes only patients with adenoma (ADC) or small cell carcinoma (SCC) of the
4-34 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
lung in their case groups. Their results overall are somewhat ambiguous regarding ETS
exposure, which the authors attribute to the lesser etiologic role of ETS for ADC compared to
SCC. The number of SCC observed, however, was too small to warrant quantitative
comparisons. Of 29 ADC cases, 12 are bronchoalveolar cell carcinomas, which Correa et al.
(1983) found to have only a weak association with passive smoking. GILL(Coh) reports that
insufficient time had elapsed since completion of the recruitment phase of their cohort study to
observe a sufficient number of cases to allow firm conclusions. Only six instances of lung cancer
in exposed females had occurred. In BUFF, exposure refers to having ever lived with a
household member who smoked regularly. Exposure is not limited to spousal smoking, nor to
any relative time-frame for duration of exposure. This broad definition of exposure possibly
includes subjects who experienced little total exposure from ETS over the past 20 to 30 years.
The high percentage of female controls exposed to ETS (84%) leaves a relatively small
percentage of unexposed subjects.
KABA is one of the smaller studies in this report (24 cases and 25 controls, total).
Exposure refers to current or past smoking of a spouse. SHIM, listed in Table 4-3 as a
"probable O" found no association between risk of lung cancer and smoking by husbands,
fathers, siblings, or coworkers. A high correlation was observed, however, with smoking by the
father-in-law (p < 0.005), which the authors describe as plausible in the Japanese society (see
Appendix A). The unpublished study by Varela (1987), denoted as VARE in this report, is
quite large and warrants further attention. The author detects no effect from exposure to
spousal smoking, but does find a significant increase in lung cancer incidence at a very high total
exposure to ETS (spouse has smoked 150 person-years, in the author's terminology) (see Figure
3-4). Unfortunately, the pertinent data from this study are not included in the source for this
report. Attempts to obtain these data from the authors have been, thus far, unsuccessful.
4-35 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 4-2. ADJUSTED RELATIVE RISKS AND POPULATION-ATTRIBUTABLE
RISK OF INDIVIDUAL STUDIES (FEMALES)
Study
AKIB
BROW
BUFF
CHAN
CORR
GAO
GARF
GENG
HUMB
KABA
KOO
LAMT
LAMW
LEE
PERS
SVEN
TRIG
WU
H1RA (Coh)
GILL (Coh)
GARF (Coh)
Percent
controls
exposed1
70
15
84
47
46
74
61
44
56
60
49
45
44
68
43
66
43
63
74
72
72
RR adjusted
for misclass.1'2
(RRM)
1.42(-,2.62)
1.35(-,5.99)
- (-.1-83)
- (-,1-15)
2.00(-,5.20)
1.06(-,1.64)
1.19(-,1.91)
2.10(-,4.28)
2.31(-,6.61)
- (-,2.45)
1.44(-,2.67)
1.54(1,2.35)
2.01(-,3.71)
- (-,2.56)
1.13(0,2.09)
1.14(-,2.81)
2.07(1.03,3.81)
- (-,3.24)
1.53(1.13,2.05)
-(-,4.91)
1.05(-,1.44)
Excess RR from
background13
(RRB-RRM)
0.2 (0,0.2)
0.2 (0,0.2)
0 (0,0.2)
0 (0,0.08)
0.2 (0,0.2)
0.03(0,0.2)
0.11(0,0.2)
0.2 (0,0.2)
0.2 (0,0.2)
0 (0,0.2)
0.2 (0,0.2)
0.2 (0,0.2)
0.2 (0,0.2)
0 (0,0.2)
0.07(0,0.2)
0.08(0,0.2)
0.2 (0.02,0.2)
0 (0,0.2)
0.2 (0.07,0.2)
0 (0,0.2)
0.03(-,0.2)
Population
attributable
risk1'4
33(0,57)
20(0,49)
0(0,47)
0(0,13)
40(0,68)
7(0,40)
18(0,43)
41(0,62)
48(0,77)
0(0,52)
29(0,50)
31(0,45)
39(0,58)
0(0,56)
6(0,40)
15(0,58)
40(3,58)
0(0,62)
37(14,49)
0(0,75)
7(0,34)
1 Values in parentheses are calculated from confidence bounds.
2 Adjustments considered negligible for values of 2.5 or greater.
Minus signs indicate a negative excess risk.
' Values truncated at 0.2, assumed to be a reasonable upper limit on excess risk for this exercise.
4 See Appendix B.
4-36
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 4-3. POPULATION-ATTRIBUTABLE RISK BY STUDY (FEMALES)
Country
Hong Kong
U.S.
U.S.
U.K.
U.S.
Scot.
Sweden
U.S.
China
Sweden
U.S.
U.S.
Hong Kong
Hong Kong
Japan
Japan
Hong Kong
Greece
U.S.
China
U.S.
U.S.
Japan
Japan
Study
CHAN
BUFF
KABA
LEE
WU
GILL(Coh)
PERS
GARF(Coh)
GAO
SVEN
GARF
BROW
KOO
LAMT
AKIB
HIRA(Coh)
LAMW
TRIG
CORR
GENG
HUMB
VARE1
SHIM1
INOU1
Estimate
0
0
0
0
0
0
6
7
7
15
18
20
29
31
33
37
39
40
40
41
48
0
0
70
Confidence limit
13
47
52
56
62
75
40
34
40
58
43
49
50
45
57
49
48
58
68
62
77
1 Insufficient information to calculate. Estimate shown is a guess.
4-37
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Our remarks above pertain to studies that estimate zero attributable risk. It is easier to
depict study characteristics that reduce the likelihood of detecting an effect than to speculate
why a study may have concluded falsely that there is an effect. A false positive conclusion could
result from an undetected causal variable that is correlated with ETS exposure. The likelihood
of a false negative conclusion is enhanced if the sample size is small, or the study differentiates
poorly between exposed and unexposed subjects.
The cohort study in Japan (HIRA) provides strong evidence of an increased lung cancer
hazard associated with ETS exposure. Consideration of plausible confounding factors and
covariables has not produced an alternative explanation, and implication of ETS as a causal
factor is biologically plausible. The mixed results of epidemiologic studies in the U.S. may be
partly due to statistical chance and study differences that affect the power to detect a lung
cancer effect from exposure to ETS. There is also evidence to suggest that exposure
differentials from spousal smoking may be larger in Japan, and possibly some other countries,
than in the U.S. This would make a cancer-related effect more difficult to detect in the U.S.
An estimate of U.S. PAR from evidence in Japan alone might lead to an overstatement.
Prediction based on the overall summary has the advantages of using all the study data while
mitigating but not ignoring the influence of the extreme outcomes.
4.5. SUMMARY AND CONCLUSIONS
The overall summary RR (before adjustments) for female NS obtained by the NRC (for
10 case-control and 3 cohort studies) is 1.32 (95% C.I. 1.16, 1.51). The corresponding value in
this report (for 19 case-control and 3 cohort studies) is 1.41 (1.26, 1.57). The effect of nine
additional case-control studies in the analysis of this report is to increase the RR from 1.3 to 1.4
and to reduce the width of the confidence interval. (Both changes increase the statistical
significance.) The overall RRs for the 19 case-control studies by themselves, and for the three
4-38 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
cohort studies by themselves, are nearly identical, even though the two types of epidemiologic
studies have their own strengths, weaknesses, and potential sources of bias. The evidence from
the U.S. studies, however, is weaker than for non-U.S. studies. The summary RR from raw data
for seven U.S. case-control studies is 1.34 (95% C.I. 1.00, 1.79). For the 12 non-U.S. studies
with raw data (all but two studies), the corresponding estimate is 1.45 (95% C.I. 1.24, 1.69). The
overall summary RR for all U.S. studies (from seven case-control studies and one cohort study)
is 1.25 (1.03, 1.52), which is statistically significant (p = 0.025 for a one-tailed test).
The overall RR for women NS is adjusted downward to 1.28 (95% C.I. 1.12, 1.45), using a
modeling approach similar to that of NRC/Wald to estimate bias from smoker misclassification.
Parameter values for the model were taken from recent sources. The adjusted estimate of RR
is still statistically significant (p < 0.01). The parameter values also were varied over a plausible
range (one at a time, but not jointly), leading to the conclusion that observed values of overall
RR of up to 1.19 are consistent with a true RR of 1.0, i.e., could arise from misreporting bias
alone. Although possibly substantial, misreporting bias is not sufficient to explain the entire
excess risk associated with ETS exposure.
Based on these analyses and following the U.S. EPA guidelines for carcinogen risk
assessment (Fed. Reg., 1986), EPA concludes that environmental tobacco smoke is a Group A
(known human) carcinogen. This conclusion is based on a total weight of evidence, principally:
• Biological plausibility. ETS is taken up by the lungs and distributed throughout the
body. The similarity of carcinogens identified in SS and MS along with the established
causal relationship between lung cancer and smoking make it reasonable to suspect that
ETS is also a lung carcinogen.
• Consistency of response. The two completed cohort studies and sixteen of the 21 case-
control studies observed a higher risk of lung cancer among the female never-smokers
classified as exposed to ETS. Evaluation of the total study evidence from several
perspectives leads to the conclusion that the observed association between ETS
exposure and increased lung cancer occurrence is not attributable to chance.
4-39 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
• Upward trend in dose-response. Of the two major cohort studies, the Japanese study
(Hirayama) demonstrates a strong association between passive smoking and lung
cancer, including an upward trend in dose-response. The upward trend is well
supported by the preponderance of evidence in the 13 case-control studies that
classified data by exposure level. The Hirayama study has undergone extensive critical
review that led to some corrections and revisions but failed to discredit the findings.
Differences in life-style and culture may be a factor in the Japanese study reporting a
stronger association between ETS and lung cancer than the American study (American
Cancer Society).
• Detectable association at environmental exposure levels. Within the population of
women who are lifelong nonsmokers, the excess lung cancer risk of those married to a
smoker is large enough to be observed. Carcinogenic responses are usually detectable
only in high exposure circumstances, such as occupational settings or in highly dosed
experimental animals.
• Broad-based evidence. The 21 case-control and three prospective studies provide data
from eight different countries and from a wide variety of study designs and protocols
conducted by many different research teams. No alternative explanatory variables for
the observed association between ETS and lung cancer have been indicated that would
be broadly applicable across studies.
• Effects remain after adjustment for potential bias. Current and ex-smokers may be
misreported as never-smokers, thus inflating the apparent cancer risk from ETS
exposure. The evidence remains statistically conclusive, however, after adjustments for
smoker misclassification. The summary estimate of relative risk from raw data of both
the case-control and cohort studies is 1.41 (95% C.I. 1.26, 1.57) before adjustment for
misclassification and 1.28 (95% C.I. 1.12, 1.45) afterward (p < 0.01).
To estimate the number of LCDs per year due to passive smoking, a further adjustment to
RR is made to correct for background exposure to ETS, i.e., to make the estimate relative to
zero ETS exposure. This additional adjustment leaves the overall RR estimate for female NS at
1.48 (1.21, 1.87). This same estimate is assumed for the relatively small number of male NS
with a spouse who smokes. The available data on males suggests that the risk is at least as high
as for females. It is assumed that 60% of the population of female NS of age 35 and over are
exposed to ETS at levels equivalent to being married to a smoker, and that the remaining 40%
are exposed to an average background exposure level (equivalently, average ETS exposure in the
population of female NS is 60% of the way from the lower exposure level (from background
ETS alone) toward the higher exposure level (from spousal smoking and background). It is
4-40 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
estimated that about 27% (95% C.I. 14%, 41%) of annual LCDs in never-smoking women are
attributable to ETS exposure. Applied to the ACS's estimate of 6500 such cases in 1988, this
percentage equates to 1750 (910, 2660) ETS-related LCDs in never-smoking women. The RR
for male NS is likely at least as high as for female NS (NRC, 1986; Wells, 1988b). Applying the
RR estimate for females to the ACS estimate of 3000 LCDs of male NS in 1988 gives an
estimate of 810 males, making the total estimate 2560 for both sexes.
Compared to the wealth of epidemiologic data on never-smokers, particularly for women
married to smokers, the information available for estimation of lung cancer risk in FS is not
substantial. The absolute risk is higher in FS than in never-smokers, but the incremental risk
from exposure to ETS is essentially unknown. For the purpose of including FS in the
calculation of LCDs attributable to ETS, they are assumed to have the same relative risk from
exposure to ETS as never-smokers. Based on this assumption and data regarding the number of
former smokers in the U.S., the estimated annual number of LCDs due to ETS is 1260 (540
women and 720 men, as calculated in Appendix B). Inclusion of FS brings the estimated total
number of LCDs per year from ETS to 3800. The component of this figure for married never-
smoking females is based on the large quantity of epidemiologic data that is available. It has
been statistically estimated with a 95% confidence interval included to indicate a range of
statistical uncertainty. The component terms for unmarried women, for men, and for FS of both
sexes are less well substantiated and subject to greater uncertainty.
4-41 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
5. ENVIRONMENTAL TOBACCO SMOKE AND
RESPIRATORY DISORDERS IN CHILDREN
5.1. INTRODUCTION
Medical, epidemiological, and experimental research of the past forty years has implicated
smoking as a causal factor in an increasing array of diseases and detrimental health conditions.
A recent workshop (Speizer, 1989) describes chronic obstructive pulmonary disease (COPD) as
a heterogeneous group of disorders with the common element of obstructive airways disease. It
concludes that over the last 25 years there has been a drastic and relatively sharp rise in COPD,
particularly in men, and that much of the increase is associated with temporal trends in cigarette
smoking. As knowledge unfolds an increasing breadth and severity of health effects related to
smoking, the potential implications for passive smokers expand as well.
Unlike studies testing for a link between ETS and lung cancer, only a few studies have
used adult subjects to test for respiratory symptoms. Infants and small children generally have
been preferred because of better study control for exposure to ETS and confounding substances
and because of the greater likelihood of an observable response due to their higher
susceptibility. Children are more susceptible to respiratory disorders because their immunologic
and respiratory systems are immature and pulmonary function is still developing (NRC, 1986;
WHO, 1986). Also, "doses" of ETS are larger relative to their body size. Very young children
experience prolonged exposure since they spend much of their time at home, and they are not
exposed socially or occupationally to ETS or potential confounding agents. On the negative
side, regular active smoking and experimentation become potential confounding factors in
children at about age seven that may be understated by either parental or self-reporting.
At early ages, particularly up to one or two years when study results of some health effects are
most consistent, in utero exposure to products of tobacco smoke is another potential
5-1 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
confounding variable. Finally, children are also relevant study subjects because of their
population size. The number of American children less than five years old living in homes with
at least one smoker probably exceeds nine million (American Academy of Pediatrics, 1986).
As observed in the report of the Canadian Pediatric Society (1986), the results of
epidemiologic surveys on the potential effects of parents' smoking on their children were initially
equivocal because it was often asked if either parent was a smoker. Only when the mother's
and father's smoking habits were considered separately did it become evident that the mother's
smoking habit was more important. It is not surprising that maternal smoking is more
significant than paternal smoking in studies involving infants and young children. The
predominant effect of maternal smoking, however, further complicates differentiating the
influence of maternal smoking during pregnancy and during the postnatal period. A large study
by Chen et al. (1988) has found increased respiratory illness in infants of nonsmoking mothers
but with ETS exposure from other household smokers after birth. This somewhat discounts the
role of the potential confounding effect of in utero exposure to tobacco products from the
mother's smoking during pregnancy.
This chapter focuses on the evidence for an association of ETS exposure with chronic
respiratory symptoms, acute lower respiratory illnesses, and impaired pulmonary function in
children, with emphasis on epidemiologic studies that have appeared since the reports of the
NRC and the U.S. SG in 1986. Studies on related topics are discussed in the last section,
including the exacerbating effects of household ETS on asthmatic children and the potential
association of parental smoking with the prevalence of additional respiratory disorders-asthma,
upper-respiratory-tract infections, and middle ear effusion (an indicator of chronic middle ear
disease).
5-2 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
5.2. EXPOSURE OF CHILDREN
Biochemical data indicating exposure to ETS are predominantly assay results of cotinine
concentrations in urine, saliva, or serum. Nearly all of the studies on adverse health effects,
however, have relied on verbal reports of parental smoking or some other measure of exposure.
Two exceptions have related health risks to cotinine concentrations. Strachan et al. (1989), who
found a positive association between passive smoking and middle ear effusion in seven-year-old
children, evaluated exposure to ETS by assaying salivary cotinine. The authors note that
cotinine concentrations were related to the number of smokers in the household. Schwartz-
Bickenbach et al. (1987) observed urinary cotinine concentrations in infants as a measure of
exposure to nicotine and cotinine in breast milk and to nicotine in ETS. As noted by Jarvis
(1989), the implicit assumption in utilizing cotinine concentrations as a measure of ETS
exposure, i.e., as a dose-surrogate, to detect detrimental health effects that may be associated
with exposure to ETS, is that cotinine concentrations are directly proportional to the uptake of
risk-relevant compounds in ETS. Which compounds are risk-relevant to a particular detrimental
health effect is generally unknown, however, and uptake may vary between individuals. Cotinine
concentration appears to be the most promising internal measure of recent exposure to ETS.
Studies that have addressed the validity or reliability of cotinine as an indicator of exposure to
ETS are described next.
Jarvis et al. (1985) found saliva cotinine concentrations in nonsmoking schoolchildren to
be related to smoking within the family. A clear increase in cotinine concentration was observed
across the categories (1) neither parent smokes, (2) father only smokes, (3) mother only smokes,
and (4) both parents smoke. Pattishall et al. (1985) measured serum cotinine in young children
(6 to 12 years of age) and found a direct correlation with the number of smokers in the home,
the amount smoked by the mother, and the amount smoked by others in the home. Coultas and
colleagues (1987) surveyed salivary cotinine levels in children and adults. Similar to the study by
5-3 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Pattishali and colleagues, the major determinants of a detectable level of cotinine in children
were the mother's smoking, father's smoking, and smoking of other household members.
Henderson et al. (1987) found urinary cotinine to be a more sensitive indicator than serum
cotinine among exposed children of day-care age. Both measures of cotinine, however, were
significantly correlated with ETS exposure in the home as determined by levels of nicotine
recovered from air samples and by the number of cigarette butts collected. Henderson et al.
(1989) notes that urinary cotinine/creatinine ratios were remarkably stable in preschool children
over a one-month period, and Jarvis et al. (1987) reports that cotinine measures were reasonably
stable over one year in nonsmoking adolescent girls. The cotinine/creatinine levels of nine
children (9 months to 3.5 years of age) exposed to 26.4 micro-grams/m3 of nicotine from SS
peaked at 4 hours and the elimination half-life was 29 hours (Goldstein et al., 1987). Greenberg
et al. (1989) studied a representative sample of 433 healthy neonates in central North Carolina.
Sixty-four percent lived in households with smokers or had contact with nonhousehold smokers.
Urinary cotinine was found in 60% of all study infants. Seventy-five percent of smoking mothers
smoked near their infants. The amount smoked in the infants' presence was the most significant
correlate of cotinine concentration.
The cotinine studies clearly demonstrate uptake, metabolism, and systemic distribution of
ETS in infants and implicate mothers' smoking as a principal source of exposure in their own
children. There are indications of substantial host-related differences in cotinine concentrations
among children, i.e., in experiments with controlled exposure to ETS, cotinine concentrations
vary between individuals experiencing the same airborne concentration of ETS. Benowitz and
Jacobs (1984) reported similar results for studies of cotinine in adults. To our knowledge, no
researchers have estimated the component of variability due to inter-individual differences of
absorption and metabolism in studies comparing cotinine levels with airborne concentrations of
SS.
5-4 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
5.3. RECENT EPIDEMIOLOGIC EVIDENCE
Studies that have appeared subsequent to the NRC and U.S. SG reports of 1986 are
displayed in Table 5-1. They have largely taken potential biasing factors into account in the
study design, protocol, or data analysis. It is often not feasible to measure and control for all
the variables of potential interest, however, and it is difficult to do so perfectly when attempted.
For example, active smokers among study subjects have generally been identified by self-
reporting or parental reporting. Although there is no evidence to suggest that this method has
been inadequate, the possibility of classification bias cannot be fully excluded. Other examples
of variables to be considered include parental social class (Chen et al., 1988; Somerville et al,
1988; Willatt, 1986); heating and cooking fuels (Chen et al., 1988); parental illness that could
either cross-infect their children or make the children genetically vulnerable (Chen et al., 1988;
Willatt, 1986); and illnesses in the children that could mimic the effects of ETS (Chen et al.,
1988; Geller-Bernstein et al., 1987). Typically these variables have been entered into a multiple
regression or logistic regression model for statistical analysis or a method of stratification has
been applied.
The epidemiologic studies emphasized in the U.S. SG and NRC reports on the health
hazards of ETS to children between birth and adolescence are shown in Tables 5-2 to 5-4 for
respiratory symptoms, respiratory illness, and pulmonary function. About 30 additional studies
have appeared since the two major reports of 1986. Recent studies' characteristics are listed in
Table 5-1, which includes a few entries for ailments in addition to respiratory symptoms,
respiratory illness, and pulmonary function. Those last three categories, however, are the
principal focus of this report and the topics of the next three sections. The general formats of
the sections are similar: summary conclusions and issues from the NRC and U.S. SG reports
are reviewed; selected recent studies are described that may bear on the weight of evidence and
relevant issues; implications of the recent studies for the results and conclusions of the NRC and
5-5 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 5-1. EVIDENCE OF RESPIRATORY
DISORDERS RELATED TO ETS EXPOSURE FROM SELECTED STUDIES
SUBSEQUENT TO THE U.S. SG AND NRC REPORTS OF 1986.
Study
Age of
subjects
Section
reference
Reaction to
ETS exposure
Chan et al.,
1989
Charlton and
Blair, 1989
Chen et al.,
1988
Chen, 1989
(same sample
data as entry
above)
Evans et al.,
1987
Fleming et al.,
1987
Geller-Bernstein
et al., 1987
Hinton, 1989
Kallail et al.,
1987
Children (7)
Adol. (12-13)
Infants
Infants
Chil./adol.
(4-17)
Infants/
chil. (0-5)
Infants/
chil. (0-5)
Infants/chil.
Children
5.4.3 and 5.4.4
5.4.1 and 5.7.5
5.1, 5.3, 5.5.3,
5.5.4, and 5.7.3
5.5.3 and 5.5.4
5.7.4
5.7.2
5.3 and 5.7.3
5.7.1
5.7.1
Wheeze in low-birth-
weight cohort; cough
in controls.
Increased school
absenteeism.
Increased hospitalization.
Synergism of passive
smoking and artificial
feeding on hospitaliza-
tion.
Increased emergency
room visits; no increase
in hospitalizations; no
effect on pulmonary
function.
Increased upper respira-
tory tract infection.
Persistent wheezing in
atopic children who were
bottle fed.
Higher chance of hospital
admission for grommet
insertion (middle ear).
No association with
middle ear problems.
(continued on following page)
5-6
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 5-1. (continued)
Study
Age of
subjects
Section
reference
Reaction to
ETS exposure
Kauffmann et
al, 1989
Children (6-10) 5.4.4, 5.6.3,
and 5.6.4
Lebowitz and
Holberg, 1987
Marks, 1988
Martinez et al.,
1988
Masi et al.,
1988
McConnochie and
Roghmann,
1986
Murray and
Morrison, 1989
Neuspiel et
al., 1989
Ogston et al.,
1987
Chil./adol./
adult (5-25)
Children (5)
Children (9)
Adol./adults
(15-35)
Chil./adol.
(1-17)
Inf./chil.
(0-10)
Infants
5.6.3 and 5.6.4
5.4.3 and 5.4.4
5.7.3
5.6.3 and 5.6.4
Children (6-10) 5.4.3 and 5.4.4
5.7.4
5.4.3, 5.4.4,
and 5.7.3
5.5.3 and 5.5.4
Maternal (but not
paternal) smoking
associated with decrease
in FEV! and FEF^,
but not in FVC.
Long-term effect on
children'spulmonary
function.
More likely to cough or
wheeze during physical
exercise.
Increased frequency of
bronchial responsiveness
and atopy in males but
not females.
Exposure during lung
growth period may affect
males permanently.
Predicts wheezing only if
family history is positive
for respiratory allergy.
More severe asthma
symptoms, especially in
males.
Increased incidence of
post-infancy wheezy
bronchitis.
Increased incidence of
alimentary and respira-
tory illnesses.
(continued on following page)
5-7
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 5-1. (continued)
Study
Age of
subjects
Section
reference
Reaction to
ETS exposure
Ostro, 1989
Park and Kim,
1986
Reed and Lutz,
1988
Somerville et
al., 1988
Stern et al.,
1987
Strachan et al.,
1989
Teculescu et al.,
1986
Chil./adults
(0-6)/(18-65)
Children
(0-14)
Children
Children
(5-11)
5.7.5
Children (7)
Adol. (10-16)
5.4.3 and 5.4.4
5.7.1
5.3, 5.4.3, 5.4.4,
5.5.3, 5.5.4, and
5.7.3
Infants (0-2) 5.4.3, 5.4.4, 5.5.3,
5.5.4, 5.6.3, 5.6.4,
and 5.7.3
5.2, 5.4.4, and
5.7.1
5.4.3, 5.6.3, and
5.7.2
Willatt, 1986
Chil./adol.
(2-15)
5.3 and 5.7.2
Increased respiratory
restricted days in adults,
and bed disability days in
young children.
Dose-response relation-
ship observed for cough,
except with family history
of cough or phlegm.
Dose-response relation-
ship observed.
Associated with wheeze,
day and night cough, and
bronchitis attacks.
Asthma and morning
cough increased, but not
statistically significant.
Increased hospitalization
for chest illness; cough,
phlegm and asthma more
frequent.
Relation between salivary
cotinine and middle ear
effusion.
More prevalent respira-
tory symptoms and upper
airway infections;
decreased forced expira-
tory flow. Effects more
marked in males.
Associated with sore
throats.
5-8
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
U.S. SG reports are discussed. The recent evidence largely corroborates and strengthens the
support for the similar conclusions of the two reports.
5.4. RESPIRATORY SYMPTOMS
The studies on respiratory symptoms cited in the 1986 reports of the U.S. SG and the
NRC are listed in Table 5-2.
5.4.1. The U.S. Surgeon General's Report on Respiratory Symptoms
Children whose parents smoke were found to have a 30% to 80% excess prevalence of
chronic cough or phlegm compared with children of nonsmoking parents. For wheezing, the
increase in risk varied from none to over sixfold among the studies reviewed. The results of
some of these studies may have been confounded by the child's own smoking habits (Colley et
al., 1974; Bland et al., 1978; Kasuga et al., 1979). The association with parental smoking was not
statistically significant for all symptoms in all studies (Lebowitz and Burrows, 1976; Schilling et
al., 1977; Schenker et al., 1983). However, the majority of studies showed an increase in
symptom prevalence with an increase in the number of smoking household members in the
home.
Although misclassification of smoking children as nonsmokers must be considered, many
studies showed a positive association between parental smoking and symptoms in children at
ages before significant experimentation with cigarettes is prevalent. In addition, several studies
(Bland et al., 1978; Weiss et al., 1980; Charlton, 1984; Schenker et al., 1983; Dodge, 1982;
Burchfiel et al., 1986) found significant effects of parental smoking after adjusting for active
smoking by the children. Acute respiratory symptoms represent an immediate health burden for
the child. However, the long-term significance of chronic respiratory symptoms for the health of
the child is unclear.
5-9 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 5-2. STUDIES ON RESPIRATORY SYMPTOMS REFERENCED
IN THE U.S. SO AND NRC REPORTS of 1986
Study
Age of
subjects
U.S. SG
NRC
Bland et al., 1978
Bland et al., 1958
Charlton, 1984
Colley, 1974
Dodge, 1982
Ekwo et al., 1983
Kasuga et al., 1979
Lebowitz and Burrows,
1976
Schenker et al., 1983
Schilling et al., 1977
Tager et al., 1979
Ware et al, 1984
Weiss et al., 1980
ChiJdren/adol. (12-13)
Children (12)
Children/adol. (8-19)
Children (6-14)
Children (8-10)
Children (6-12)
Children (6-11)
Children (< 16)
Children (5-14)
Children/adol. (< 16)
Children (5-19)
Children (6-13)
Children (5-9)
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
5-10
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
5.4.2. The National Research Council Report on Respiratory Symptoms
Almost all of the cross-sectional studies that have compared children of parents who
smoke with the children of parents who do not smoke have reported increased prevalence of
respiratory symptoms, usually cough, sputum, or wheezing, in the children of smoking parents.
Some studies, including some that have not found a statistically significant increase in the
prevalence of respiratory symptoms in ETS-exposed children, observed an increase in prevalence
of respiratory symptoms as the number of household smokers increases.
Three problems related to interpretation of results are particularly relevant to studies of
respiratory symptoms in children-underreported active smoking on the part of the children,
recall bias leading to overreporting of symptoms by parents, and the confounding variables of
infections in parents. All three may lead to overestimation of symptom prevalence among
children of smokers. It has been observed that parents, especially mothers who have a history of
severe respiratory illness, report higher rates of respiratory symptoms in their children
(Schenker et al., 1983; Ferris et al., 1985).
Lebowitz and Burrows (1976), reporting on children in the Tucson Epidemiologic Study of
Obstructive Lung Disease, emphasized the need for controlling for parental symptoms. Ferris et
al. (1985) have argued, however, that correcting for parental symptoms represents an
overcorrection for respiratory symptoms in children since it also corrects for the parents'
smoking habits. In the Harvard Air Pollution Respiratory Health Studies (Six-Cities Study) of
children ages 6 to 9 years, the variable indicating whether the parent had a history of bronchitis,
emphysema, or asthma was found to be a highly significant risk factor for cough and wheeze and
a history of respiratory illness among children.
In both the Lebowitz and Burrows and Ferris et al. studies, adjustment for parental
symptoms or respiratory illness decreased the strength of the apparent association between
exposure to ETS and respiratory symptoms but did not eliminate it. This finding leads to the
5-11 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
reasonable conclusion that the exposures typical of ETS are sufficient to cause respiratory
symptoms in some children. The increase in frequency of cough was 20% to 50%, and as high
as 90%, when there were smoking parents. The increased frequency of wheezing was more
variable, which may indicate the difficulty in assessing this symptom. Furthermore, there
appears to be a dose-response relationship between exposure and the likelihood of the child's
developing respiratory symptoms or a respiratory illness.
5.4.3. Recent Studies on Respiratory Symptoms
Studies in Table 5-1 that address respiratory symptoms vary by objective and
methodological approach. For discussion they are categorized roughly by age of the study
subjects. Age is an important factor for several reasons. Susceptibility and the manifestation of
symptoms may be related to duration of exposure or to developmental growth and maturation.
It is also of interest to identify respiratory symptoms in infancy that may be predictive of chronic
respiratory disorders or impaired lung function in later years. As noted previously, evidence of
smoking in the home is more clearly differentiated between subjects in the first few years of life
where there is little exposure to ETS and confounding substances outside the home. For this age
group, maternal smoking has historically been more strongly related to respiratory symptoms
and illnesses in early childhood than have paternal smoking or other more general measures of
household exposure. This outcome is consistent with what would be anticipated if ETS exposure
is causally linked to an increased incidence of symptoms.
A further reason for considering age relates to the potential sources of bias, mentioned
above and discussed further in the summary and discussion to follow (Section 5.4.4.). Since
smokers tend to have more respiratory symptoms than nonsmokers, it has been claimed that
parents may overstate (or understate) their own children's symptoms. It seems unlikely,
however, that parental perceptions of their children's health are sufficiently distorted or
5-12 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
influenced by their own health conditions to have a broad-based influence across numerous
studies. Perhaps of greater significance is the potential for unknowingly treating active smokers
as passive smokers. Children of parents who smoke are more likely to become smokers than
children of nonsmokers. Consequently, subjects who smoke but are misreported as nonsmokers
may be more likely to be included with those exposed to ETS at home. If a subject's own
smoking contributes to respiratory symptoms, then the adverse effects of passive smoking may
be overstated. As noted previously, children may become smokers as young as 10 years of age,
and experimentation with cigarettes may start younger.
Of the studies examined in this report, one is on infants; four are on children of primary
school age (5, 7, 5-11, and 6-10); one is on adolescents (10-16); two cover a wide age range from
birth (0-14 and 0-10). Virtually all of the investigators report some adverse outcome associated
with ETS exposure (except possibly Park and Kim, 1986, for ages 0 to 14). Statistical
significance is not always achieved, however, and it is not always clear if the findings reported
include all symptoms investigated. Nevertheless, the evidence is very substantial. If there were
no effects of ETS exposure on respiratory symptoms, then the outcomes would be equally likely
to produce an observed increase or decrease in relation to ETS exposure. This is clearly not the
case for the published studies.
Stern and colleagues (1987) evaluated the effects of infant exposure to maternal smoking
in a cohort of over 4000 Canadian schoolchildren. Cough and phlegm were more frequent
symptoms in children whose mothers smoked during the child's first two years. (Cough was also
more common in children whose mothers smoked during pregnancy, which is a potential
confounding variable.)
Increased occurrence of cough and wheeze was commonly reported to be associated with
ETS exposure in studies of children in the next higher age group (5, 7, 6-10, 5-11). Specifically,
Marks (1988) conducted a multiethnic study of inner-city preschool 5-year-olds. Children living
5-13 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
in smoking households were more likely to experience coughing or wheezing during physical
exercise, although cigarette smoke exposure did not appear to influence other respiratory
symptoms. The effect of passive smoking on respiratory symptoms of children ages 5 to 11
years was investigated in over 4000 English children and nearly 800 Scottish children
participating in the National Study of Health and Growth in 1982 (Somerville et al., 1988). A
number of statistically significant positive associations were found between respiratory
conditions, including wheeze and cough, in English children and the number of cigarettes
smoked per day at home by their parents. Only wheeze was significant for Scottish children.
Frequent "cough first thing in the morning" showed a positive but not statistically significant
association in English children.
McConnochie and Roghmann (1986) explored the effects of passive smoking and non-
breast-feeding on wheezing in children ages 6 to 10. Maternal smoking, lack of breast-feeding,
and two measures of genetic tendency (family history of respiratory allergy and male sex)
predicted wheezing in children with a mean age of 8.4 years. Passive smoking determined
wheezing only among children whose family history was positive for respiratory allergy. The
final study in this age category is by Chan and associates (1989) who found evidence that low
birth weight children of age seven experience wheezing when exposed to ETS. In a comparison
of 121 seven-year-old English children with a history of low birth weight (< 200 g) and an
unselected reference group of schoolchildren of the same age, the investigators reviewed hospital
charts at discharge after birth; administered questionnaires on family, social, and clinical history;
and performed tests for lung function, bronchial reactivity, and allergies. Multiple logistic
regression was used to control for socioeconomic status, neonatal oxygen scores, atopy, and
family history of asthma. In the reference group, daytime (but not nocturnal) cough was weakly
associated with maternal smoking but not with smoking by other household members. Maternal
smoking was associated with wheeze but not cough in the low birth weight cohort.
5-14 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The two studies of children in the age ranges 0-10 and 0-14 are by Neuspiel et al. (1989)
and Park and Kim (1986), respectively, both of whom utilized large samples. The former
studied the effect of parental smoking on wheezing in 9670 British children. Children of
smoking mothers had a significantly increased cumulative incidence of post-infancy wheezing to
ten years of age, but it was confined to an increase in wheezy bronchitis. (The authors note that
some investigators have suggested that wheezy bronchitis is clinically and pathologically
indistinguishable from asthma, so this study is referenced in Section 5.7.3. as well.) There was a
14% increase in childhood wheezy bronchitis when mothers smoked over four cigarettes per day
and a 49% increase at 14 cigarettes per day after adjustment for covariables. The covariables
controlled in the analysis (multiple logistic regression) include paternal smoking, social status,
sex, history of family allergy, crowding, breast-feeding, gas cooking and heating, and bedroom
dampness. Some of the observed effect was explainable by maternal respiratory symptoms and
maternal depression, but not by neonatal problems, the child's allergic symptoms, or paternal
respiratory symptoms. Passive smoking was related only to wheezy bronchitis and not to parent-
reported asthma or wheezing for other reasons.
The Park and Kim (1986) survey of 3651 Korean children from 0 to 14 years of age was
conducted to ascertain if coughing is related to ETS exposure. A rural area was selected for the
survey where adult smokers showed a similar life-style and spent much of their time at home.
The prevalence and frequency of coughing was found to increase with the number of adult
household smokers and with the number of cigarettes smoked in an analysis unadjusted for
covariables. Smoking by children is unlikely to be a confounding factor as smoking is seldom
seen in children below 15 years of age in Korea. Data on 21 extraneous variables were collected
in the study. When included in an adjusted statistical analysis, the presence of coughers in the
family was found to be an explanatory variable of some importance. The association of ETS
exposure with increased coughing may be due to the indirect effect of family smoking through
5-15 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
coughers in the family, or a direct consequence of exposure to ETS. The mechanism by which
adult household smoking is related to increased coughing in children was not investigated.
Teculescu et al. (1986, in French) conducted a relatively small study that compared 46
nonsmoking children ages 10 to 16 years whose parents smoke with an identical number of
children matched for sex, age, and height whose parents were nonsmokers. Passive exposure to
parental tobacco smoke was associated with a higher prevalence of respiratory symptoms.
5.4.4. Summary and Discussion of Respiratory Symptoms
Studies on the relationship of ETS exposure and respiratory symptoms appearing
subsequent to the U.S. SG and NRC reports of 1986 provide additional support to those reports'
conclusions. Increased cough has been observed in a range of ages, including 0 to 2, 5 to 11 and
0 to 14 (Stern et al., 1987; Somerville et al., 1988; and Park and Kim, 1986, respectively). The
last reference, however, found that the substantial dose-response observed could be largely (but
not totally) explained by a family history of cough or phlegm. Similarly, McConnochie and
Roghmann (1986) found that ETS exposure predicts wheezing only in families with a history of
respiratory allergy. The cumulative incidence of post-infancy wheezy bronchitis through 10 years
of age increased with the amount mothers smoked, in the large study by Neuspiel et al. (1989).
They found that some of the effects could be explained by maternal respiratory symptoms.
These results are consistent with the explanatory effect of parental symptoms found in the
studies by Lebowitz and Burrows (1976) and Ferris et al. (1985) (Section 5.4.2). A weak
association of wheezing was reported in seven-year-olds of low birth weight (Chan et al., 1989).
Marks (1988) observed increased wheezing and coughing in exposed children of age five
following physical exercise.
Family history of respiratory symptoms and disease appears to be a confounding variable
for the interpretation of data. As noted in the remarks from the NRC report (Section 5.4.2),
5-16 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
bias may be introduced by parents who have a history of respiratory illness for several reasons.
They may be overstating their children's symptoms, or their children may actually have more
respiratory symptoms and illness. The latter possibility could be the result of intra-family
correlation of susceptibility (referred to as familial resemblance in Kauffmann et al., 1989) or
may be attributable to contagion between members of the same household. The conclusion is
simply that family history has been shown to be a confounding variable in some studies showing
an association of respiratory symptoms in children with parental smoking. This means that
family history, as well as any other factors correlated with parental smoking, is a candidate for
causally contributing to the occurrence of respiratory symptoms in children. When data have
been collected on family history one has the option of attempting to adjust for its effect
statistically. How meaningful the results are, however, depends on how confounded family
history is with parental smoking. Correlation between family history (as reported in the data,
whether biased or not) and household exposure to ETS will tend to lead to an understatement
of the statistical significance of ETS after adjustment for family history and conversely,
regardless of which (if either) is causally related to the observed increase in respiratory
symptoms. This difficulty has been noted in the literature (Section 5.4.2.).
Some additional precautions to those already discussed are in order. Controlling for
active smoking becomes an issue for subjects at about age ten (some evidence suggests seven or
eight years of age) depending on the culture, even though it may be light smoking. Most
researchers have been aware of the potential confounding effect of smoking and have attempted
to control for it. Some studies have excluded persons who smoke, and others have made them a
separate group for comparison. A greater difficulty probably lies in the potential for
misreported smoking habits. Young persons may be reluctant to admit to smoking cigarettes,
especially if they have been experiencing respiratory maladies. Data are often obtained from
parents, who may not be aware of a child's smoking. Future studies may include cotinine tests
5-17 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
to confirm the reported smoking status (as in Strachan et al., 1989). Of course, misreported
smoking status would not explain the observed relationship between ETS exposure and
respiratory symptoms in infants and children up to the middle years of primary school.
In conclusion, there is no apparent single source of systematic bias that might explain the
substantial evidence accumulating across the multiple age groups, investigative approaches, and
environmental and cultural conditions studied. Family history may cause overstatement of
conclusions when it is not taken into account in the analysis of data, but it may only indicate
that ETS is related to increased respiratory symptoms indirectly, or both directly and indirectly.
It is reasonable to conclude that parental smoking increases the incidence of respiratory
symptoms from infancy well into primary school years and probably through the adolescent
years. Misreported smoking status may have some influence in studies of older children.
Results in the higher age group, however, are not inconsistent with findings at other ages. In
addition to the unlikelihood that bias significantly distorts the overall results, several other
factors are supportive: (1) evidence that maternal smoking tends to be more of a factor than
paternal smoking in the first one to two years of life; (2) an observed dose-response function in
numerous cases; and (3) the biological plausibility given the increased incidence of respiratory
symptoms in adult smokers.
5.5. ACUTE RESPIRATORY ILLNESS
References to the evidence on acute respiratory illness in the reports of the U.S. SG and
the NRC of 1986 may be found in Table 5-3. As in Section 5, conclusions of the U.S. SG and
NRC reports are summarized, subsequent evidence is addressed, and the overall implications for
exposure to ETS are assessed in a summary and discussion section.
5-18 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE,
TABLE 5-3. STUDIES ON RESPIRATORY ILLNESS REFERENCED
IN THE U.S. SG AND NRC REPORTS OF 1986
Study
Age of
subjects
U.S. SG
NRC
Cameron et al., 1969
Colley, 1971
Colley, 1974
Dutau et al., 1981
Fergusson et al., 1981
Harlap and Davies, 1974
Leeder et al., 1976b
(also see Colley,
1974, and Leeder
et al., 1976a)
Pedreira, 1985
Pullen and Hey, 1982
Rantakallio, 1978
Said et al., 1978
Sims et al., 1978
Speizer et al., 1980
Ware et al., 1984
Children (6-9)
Infants
Infants
Infants/children (0-6)
Infants
Infants
Infants
Infants
Children
Infants/children (0-5)
Children/adol. (10-20)
Children
Children (6-10)
Children (5-9)
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
5-19
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
5.5.1. The U.S. Surgeon General's Report on Acute Respiratory Illness
The results of these studies show excess acute respiratory illness in the children of parents
who smoke, particularly in children under two years of age. This pattern is evident in studies
conducted with different methodologies and in different locales. The increased risk of
hospitalization for severe bronchitis or pneumonia associated with parental smoking ranges from
20% to 40% during the first year of life. Young children appear to represent a more susceptible
population for the adverse effects of involuntary smoking than older children or adults. The
time-activity patterns of infants, which generally place them in proximity to their mothers, may
lead to particularly high exposure to ETS if the mother smokes.
The possibility of bias due to the respiratory status of the reporting parent(s) must be
considered for the studies that have used questionnaires to measure illness experience. In all
studies in which potential reporting bias was examined, control for parents' status reduced, but
did not eliminate, associations of involuntary smoking with health outcomes (Colley et al., 1974;
Leeder et al., 1976a,b; Schenker et al., 1983; Ware et al., 1984). Further, the consistency of
these studies, in spite of differing study populations and methods, weighs against bias as the sole
explanation. Acute respiratory illnesses during childhood may have long-term effects on lung
growth and development. It may possibly increase the susceptibility of the lung to the effects of
active smoking and to the development of chronic obstructive lung disease (Samet et al., 1983;
U.S. DHHS, 1984).
5.5.2. The National Research Council Report on Acute Respiratory Illness
There is now strong evidence that bronchitis, pneumonia, and other lower-respiratory-tract
illnesses occur more frequently (at least during the first year of life) in children who have one or
more parents who smoke. Bronchitis, pneumonia, and other lower-respiratory-tract illnesses
occur up to twice as often during the first year of life in children who have one or more parents
5-20 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
who smoke than in children of nonsmokers. All the studies that have examined the incidence of
respiratory illnesses in children under the age of one year have shown a positive association
between such illnesses and exposure to ETS. There is a dose-response relationship that relates
more to maternal smoking than to paternal smoking as the source of ETS exposure.
The association between ETS exposure and increased occurrence of respiratory illnesses in
children is very unlikely to have arisen by chance. It may represent a direct association between
ETS exposure and disease (a causal explanation) and/or an indirect one (noncausal) arising
because children living in homes of smokers are at risk of such diseases for other reasons.
Some of the studies have examined the possibility that the association is indirect by allowing for
confounding factors-such as social class, parental respiratory illnesses, and birth weight-and
have concluded that such factors do not explain the results. This argues, therefore, in favor of
the causal explanation. Such an explanation is also supported by the evidence of a dose-
response relationship specific for respiratory disease. Regardless of the mechanism, however,
the exposure of small children to smoking in the home appears to put them at risk of respiratory
illness.
5.5.3. Recent Studies on Acute Respiratory Illness
Four of the studies that have appeared subsequent to the U.S. SG and NRC reports deal
with the potential relationship between exposure to ETS and lower-respiratory-tract illness
(excluding the study on asthmatics for now and counting Chen, 1989 and Chen et al., 1988 as
one study). All four reports indicate an association between ETS and increased respiratory
illness, although the approaches and health-related endpoints are varied. Three studies pertain
to infants. The remaining one applies to the age range 5-11.
Chen and his colleagues (1988) investigated the relationship between passive smoke
exposure and hospitalization for bronchitis and pneumonia by 18 months of age in a study of
5-21 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
2227 Chinese children. The results indicated a significant dose-response relationship of
household smoking to hospitalization for respiratory illness during the child's first 18 months.
No confounding variables were discovered. Further analyses indicated that infants who were
less than six months of age, had low birth weight, or were artificially fed were relatively more
susceptible to the effects of tobacco smoke. Moreover, the cumulative incidence of bronchitis
and pneumonia increased significantly with increased smoking of family members. This result
persisted when sex, birth weight, nursery care, father's education, coal for cooking, and cases of
adult chronic respiratory disease were taken into account. An interesting and important aspect
of this study is that of 1746 smoking families, there were no mothers who smoked. Presuming
that the reporting of smoking habits is accurate, or nearly so, smoking during pregnancy is
unlikely to confound postnatal exposure to ETS. In a later publication based on the sample
described above, Chen (1989) reported that among artificially fed infants the frequency of
hospitalization for respiratory illness was 2.5 times greater when more than 20 cig./day were
consumed in the home. The frequency dropped to 1.7 times greater if 1 to 19 cig./day were
consumed. Chen also concluded that passive smoking and artificial feeding work synergistically,
producing a detrimental effect much greater than that produced by their separate actions. Thus,
infants may be at greater risk from ETS exposure if bottle fed.
Further evidence of a relationship between passive smoking and respiratory illness in the
child's first year is provided by Ogston et al. (1987). These investigators conducted a
prospective study of 1565 infants involved in the Tayside Morbidity and Mortality Study. Health
visitors interviewed parents to gather data on social class, age of parents, method of heating and
cooking, father's and mother's smoking habits, and the presence of upper- or lower-respiratory-
tract infections (later confirmed by medical diagnosis). Parents were classified simply as
smokers or nonsmokers. Multiple logistic regression indicated that respiratory illness during the
first year of life was predicted by parental smoking. Moreover, a trend indicated increasing
5-22 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
incidence from (a) nonsmoking families to (b) only father smoking, to (c) mother or both
parents smoking.
A cohort study of 4099 Canadian children, 7 to 12 years of age, corroborates the findings
described above. Stern and associates (1987) found that children whose mothers smoked during
the first two years of their life were significantly more likely to have been hospitalized at least
once before the age of two years for a respiratory illness than children of nonsmoking mothers.
Moreover, ETS-exposed children hospitalized in their first two years of life were six times more
likely to be hospitalized for chest illness than unexposed children. Similar results were reported
for children whose mothers smoked during pregnancy, however, so in utero and postnatal
exposure to products of tobacco smoke may be partially confounded.
A study demonstrating a dose-response relationship between passive smoking and
respiratory illness in children 5 to 11 years old was conducted by Somerville and associates
(1988). The sample consisted of 4000 English and 800 Scottish children from the National Study
of Health and Growth. Data on each child's respiratory symptoms, parental smoking, and family
background were obtained from a self-administered questionnaire completed by the mother.
Passive smoking was assessed by the total number of cigarettes smoked each day by the mother
and the father. Multiple regression analysis indicated that the number of cigarettes smoked by
English parents was significantly associated with bronchitis attacks within the last two months.
Neither household crowding, parents' education, father's occupation, nor age or sex of the child
confounded this result.
5.5.4. Summary and Discussion on Respiratory Illness
As in the discussion of respiratory symptoms (Section 5.4), the respiratory status of
parents is a factor to be considered in interpreting the results on respiratory illness. As the U.S.
SG report notes, there is the possibility of biased reporting from parents who smoke and may be
5-23 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
more subject to respiratory ailments than normal. All of the studies reviewed used
questionnaires or interviews. The questions dealing with respiratory illness tend to be rather
specific, requiring little subjective judgment. For example, the number of hospitalizations for
respiratory or chest illness during the first year or two of life, and the number of doctor-
diagnosed cases, were indexes of health (Chen et al., 1988; Chen, 1989; and Stern, 1987). Ogston
et al. (1987) confirmed reports of respiratory infections by medical diagnosis. Consequently, the
potential for biased reporting from parents who smoke may be less of an issue for evaluation of
respiratory illness than for respiratory symptoms.
The consistent association between respiratory illness and ETS exposure in the home over
all studies to date, especially in studies of infants, is unlikely to be attributable to a confounding
factor. Such a factor would have to be consistently operative over the broad spectrum of
countries, cultures, and age groups studied and of substantial influence on respiratory health.
Most researchers have attempted to control for potential confounding factors, although only one
study has controlled for in utero exposure (Chen et al., 1988). The dose-response relationships,
reported in recent studies (Chen et al., 1988; Somerville et al., 1988; Ogston et al., 1987) and in
previous evidence assessed by the NRC and U.S. SG committees, are consistent with a causal
association; a plausible explanation in terms of systematic bias or a confounding factor is less
apparent. These arguments and the biological plausibility that ETS exposure increases the
incidence of lower-respiratory-tract illnesses in the first one-to-two years of life provides some
support for a causal relationship. The influence of potential sources of bias and confounding
factors, however, cannot be adequately assessed to conclude a causal association between ETS
exposure and the increased incidence of respiratory illnesses in young children. The consistency
of the conclusions across the cumulative recent and previous studies, however, cannot be
statistically attributed to chance occurrence; it is only the explanation behind the association that
is uncertain.
5-24 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The implications of early childhood respiratory illnesses for health in later years is an
important related issue that is difficult to study. Stern et al. (1987) concluded that the strong
relationship between hospitalization before the age of two years for a chest illness and
subsequent respiratory symptoms and decreased pulmonary function later in childhood suggests
that there are definite carryover effects of early acute respiratory illness. In a study not directly
related to passive smoking, Barker and Osmond (1986) found a strong geographical relation
between death rates from chronic bronchitis and emphysema in 1959 to 1978 and infant
mortality from bronchitis and pneumonia during 1921 to 1925 for regions in England and Wales.
The authors concluded that this relation provides strong evidence of a direct causal link between
lower respiratory infection in early childhood and chronic bronchitis in adult life. Although
these studies contribute some evidence of the potential for long-term effects from childhood
exposure, further support is needed for a conclusion. Long-term implications have been
expressed as a major concern by numerous authors. Chronic respiratory ailment or permanently
impaired lung function has largely been speculated from the results of multiple studies over
various age groups. No major longitudinal study following subjects from infancy through
adolescence has been conducted.
5.6. PULMONARY FUNCTION
Studies on pulmonary function referenced in the 1986 reports of the U.S. SG and the
NRC are displayed in Table 5-4. Conclusions of the U.S. SG and NRC reports are summarized,
and studies appearing since the U.S. SG and NRC reports are reviewed. The current state of
evidence on the potential association of ETS exposure with pulmonary function are summarized
and discussed.
5-25 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 5-4. STUDIES ON PULMONARY FUNCTION REFERENCED
IN THE U.S. SG AND NRC REPORTS OF 1986
Study
Berkey et al., 1986
Brunekreef et al.,
1985
Burchfiel et al., 1986
Chen and Li, 1986
Comstock et al., 1981
Dodge, 1982
Elcwo et al., 1982
Ferris et al., 1985
Hasselblad et al.,
1981
Kauffmann et al.,
1983
Kentner et al.,
1984
Lebowitz, 1984
Lebowitz and
Burrows, 1976
Pimm et al., 1978
Schilling et al.,
1977
Shepard et al., 1979
Age of
subjects
Children (6-10)
Adults
Infants/children (0-10)
Children/adol. (8-16)
Adults
Children (8-10)
Children (6-12)
Children/adol.
Children (5-13)
Adults
Adults
Families
Children/adol. ( < 16)
Adults
Children/adol. (< 16)
Adults
U.S. SG
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
NRC
X
X
X
X
X
X
X
(continued on following page)
5-26
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE 5-4. (continued)
Study
Sims et al., 1978
Tager et al., 1979
Tager, 1983
Tashkin
Vedal et al., 1984
Ware et al., 1984
Weiss et al., 1980
White and Froeb, 1980
Age of
subjects
Children
Children (5-19)
Children (5-9)
Children
Children (6-13)
Children (6-13)
Children (5-9)
Adults
U.S. SG
X
X
X
X
X
X
NRC
X
X
X
X
5.6.1. The U.S. Surgeon General's Report on Pulmonary Function
Cross-sectional studies have demonstrated lower values on tests of pulmonary function
(FEV 75%, FEVj, FEF^.vs, and flows at low lung volumes) in children of mothers who smoked
compared with children of nonsmoking mothers. Longitudinal studies confirm the cross-
sectional results and provide some insight into the implications of the cross-sectional data.
Dose-response relationships have been found in both cross-sectional and longitudinal studies
(Tager et al., 1979; Weiss et al., 1980; Ware et al., 1984; Berkey et al., 1986); the level of
function decreases with an increasing number of smokers in the home. As would be anticipated
from the mother's greater contact time with the child, maternal smoking tends to have a greater
impact than paternal smoking. Younger children seem to experience greater effects than older
children (Tager et al., 1979; Weiss et al., 1980), and in older children the effects of personal
5-27
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
smoking may be additive with those of involuntary smoking (Tager et al., 1979, 1985). Some
studies have reported greater effects on flows at lower lung volumes in girls than in boys
(Burchfiel et al., 1986; Tashkin et al, 1984; Yarnell and St. Leger, 1979; Vedal et al., 1984).
Flows at higher lung volumes seem to be more affected in boys (Burchfiel et al, 1986; Yarnell
and St. Leger 1979; Berkey et al, 1986; Tashkin et al, 1984). It is unclear whether these sex
effects represent differences in exposure, differences in susceptibility to environmental cigarette
smoke, or differences in growth and development. The observed reduction in lung function of
children associated with maternal smoking is small, on the average. Some children may be
affected to a greater extent, however, and even small differences may be important for children
who become active smokers in adulthood.
5.6.2. The National Research Council's Report on Pulmonary Function
It is often difficult (but not impossible) to measure lung function in young children, and it
is also hard to dissect out the relative contribution of ETS and that of natural variation and the
effect of respiratory infections to pulmonary damage. The most important contributors to
variation in lung function among children are size-related factors such as sex, age, and height.
These account for about 50% to 60% of the variation (Comroe et al, 1962). Nevertheless, a
majority of the studies has shown a small decrease (up to 0.5% of FEVj per year) in rate of
increase in lung function associated with normal growth in children living with one or more
parents who smoke compared with those living with nonsmoking parents. These differences
have usually been statistically significant. Although the mean effect is small, there are
individuals in each study who have large decrements in growth of lung function. Some studies
have found a dose-response relationship with the number of smokers in the home or the amount
smoked (Hasselblad et al, 1981). In most studies, only the maternal effect was statistically
significant. It is not possible to determine whether ETS is directly causing the decreased lung
5-28 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
function observed in children of smoking parents or if an increased infection rate in these
children is responsible for the decrease.
The annual small decrease in FEV,, which is related to exposure to ETS, is unlikely to be
clinically significant. However, the effect may be important in two respects. First, the existence
of statistically significant differences related to parental smoking leads to the conclusion that
there are pathophysiologic effects of exposure to ETS in the lungs of the growing child. It may
be an in utero effect, an effect on the growing and remodeling of the lung, or both. Second, it
raises the question of whether the child who is adversely affected by parental smoking may be at
increased risk for the development of chronic airflow obstruction in adult life. An accelerated
decline in lung function could increase the risk of chronic pulmonary disease (Samet et al.,
1983). An important unanswered question is whether exposure to ETS affects the way the lungs
grow and develop during childhood.
5.6.3. Recent Studies on Pulmonary Function
Kauffmann et al. (1989) asked what factors related to lung function may be correlated
between parents and their offspring between 6 and 10 years of age. A total of 1160 children
were included in the study whose parents had been examined in the 1975 French PAARC
Cooperative Study. Positive correlations between parent and child of FVC, FEVt, and FEF^s
were observed and exhibited an increasing temporal trend with increasing age of the children.
Comparisons between siblings of the opposite sex suggest that different growth patterns between
boys and girls may be a factor in lung function. Maternal, but not paternal, smoking was
associated with a significant decrease in FEVj and FEF^j, but not with FVC.
Masi et al. (1988) suggest that passive smoking during the growth period of the lungs in
early life permanently affects their mechanical properties in young men. They collected mail-in
questionnaire data on lifetime exposure to ETS. The exposure estimates were compared with
5-29 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
pulmonary data previously collected as part of a cross-sectional study of the evaluation of lung
function from adolescence to early adulthood. FEFjj.^ was inversely related to ETS exposure
estimates before age 17 in males, but not in females.
A longitudinal study of pulmonary function between the ages of 5 and 25 supports the
hypothesis that childhood respiratory illnesses have implications for pulmonary development
(Lebowitz and Holberg, 1987; also, Lebowitz et al, 1987). A total of 1502 observations were
made between 1972 and 1983 on 362 subjects 6 to 15 years of age when initially tested. Follow-
up averaged about 9 years. Measures of pulmonary function included FVC, FEV1; flow at 50%
of FVC(Vmax 50%), and flow at 75% of expired FVC(Vmax 75%). The study includes active
smokers as well as nonsmokers, which complicates analysis and interpretation of the data. The
statistical models assumed describe the data well, but the technical details are sketchy. The
authors conclude that respiratory illnesses and smoking had the biggest negative impact on
growth of lung function, using FVC, FEVj, Vmax 50%, and size-compensated flows (Vmax
50%/FVC). Further negative impacts were due to parental smoking, especially as it interacts
with active smoking and respiratory disease. Measures of flow (Vmax 50%, Vmax 50%/FVC) were
more sensitive than FEVt to the effects of concurrent disease and smoking and were better
indicators of a long-term effect persisting into early adulthood.
Stern and colleagues (Section 5.4.3.) found a statistically significant 0.7% decrement in
FEVj associated with maternal smoking during the first two years of life, but no effect on FVC
was observed. The small study by Teculescu and others (Section 5.4.3.) also reported a
significant decrease in forced expiratory flows, with effects more marked in boys. The single-
breath nitrogen washout test, a sensitive test of small airways obstruction in adults, did not
detect any effect of passive smoking in this limited sample (46 nonsmokers of ages 10 to 16
whose parents were smokers).
5-30 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
5.6.4. Summary and Discussion of Pulmonary Function
The NRC report has previously noted the difficulties in measuring lung function in small
children and then sorting out the effect attributable to ETS exposure. Height, age, and sex are
factors to control in tests of pulmonary function. Factors related to familial resemblance
(Kauffmann et al., 1989), such as parental smoking and lung function, temporal trends in parent-
child correlations as a child's age increases, and different growth patterns between boys and girls
are further complicating factors. The evidence suggests that any effect of passive smoking on
lung function is likely small. In view of these conditions, it is not surprising that study results
have been somewhat mixed and difficult to assess overall.
If a health effect is associated with ETS, it is generally more apt to be observed if the
differential exposure to ETS between study subjects being compared is large than if it is small.
This is particularly true when factors affecting the variability of study data are poorly controlled
and difficult to assess. This observation may be particularly relevant for studies of pulmonary
function, as illustrated by comparative analyses of two major longitudinal studies. Lebowitz and
Holberg (1988) and Tager et al. (1987) both reanalyzed the longitudinal data from the East
Boston Study and the Tucson Study. The former study found an effect of maternal smoking on
FEVj and the latter did not. Both reviews concluded that the disparate results were not
attributable to the different methods of analysis originally used. The difference in ETS exposure
between the "exposed" and "unexposed" groups appears likely to be much larger in one study
than in the other, attributed (at least in part) to the differences in climate (which affects the
amount of ventilation from outdoor air).
Differences between studies that affect exposure levels and the influence of host-related
variables affecting measured responses that cannot be fully controlled make it difficult to assess
the overall evidence from studies on pulmonary function. In reviewing previous studies, Tager
(1986) notes that some consistency emerges if one focuses on FEVj and FEF^.^. The recent
5-31 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
study of Kauffmann et al. (1989) found that maternal smoking was related to FEVj and
but not to FVC, consistent with Tager's observation. Two other conclusions that appear
frequently in the cumulative studies on lung function are the greater susceptibility of boys and
the significance of age in subjects up to adulthood.
Recent studies provide additional information on the question of whether childhood
exposure to ETS has implications for long-term effects in lung function. As noted by the NRC,
it is not possible to determine whether ETS is directly causing the decreased lung function
observed in children of smoking parents or if an increased infection rate in these children is
responsible for the decrease. Epidemiologic data of Paoletti et al. (1989) support the hypothesis
that childhood history of respiratory infection (prior to age 12) and adolescent-adult history are
related to increased prevalence of a number of detrimental health conditions in adulthood,
including reduced lung function and chronic obstructive lung disease (independent of whether
parental smoking is implicated in the childhood history of respiratory illness). Stern et al.
(1987) conclude that the strong relationship between hospitalization before the age of two years
for a chest illness and subsequent respiratory symptoms and decreased pulmonary function later
in childhood suggests that there are definite carryover effects of early acute respiratory illness.
The longitudinal study of Lebowitz and Holberg (1987) links early respirator)' disorders with
long-term pulmonary effects, specifically in the small airways. The study of Masi et al. (1988)
suggests a permanent affect in young men (but not women). If passive smoking in childhood is
causally associated with respiratory illness (only "association", not "causal association", is
concluded in this report), then these studies support the hypothesis of a long-term effect on lung
function.
Based on the cumulative evidence available, this report concludes that passive smoking in
early childhood is associated with decreased lung function in childhood and with a small
reduction in their rate of pulmonary growth and development.
5-32 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
5.7. RELATED RESULTS
A number of studies have investigated the potential for health hazards to children from
parental smoking that have not been described under the sections on respiratory symptoms,
respiratory illness, or pulmonary function. In most cases there is less evidence on these health
related endpoints, but that does not imply that they are necessarily less important-just less
thoroughly studied at this time. Some of these studies and their topics of inquiry are
summarized in this section. This report concludes that household smoking is associated with
excess incidence of middle ear effusion (Section 5.7.1.), but no further conclusions are drawn
from the study results described. There is increasing evidence that maternal smoking may be
associated with increased prevalence of asthma, particularly before the age of one year.
Potential bias in parent-reported data and confounding by in utero exposure, however, are
difficult to assess. Based on the overall statistical evidence, it appears unlikely that maternal
smoking has a very large effect on asthmatic conditions in children. No conclusions are
warranted, however, as study results have been inconsistent and ambiguities complicate
comparisons.
5.7.1. Middle Ear Effusion
The U.S. SG report includes reviews of five studies that demonstrate an excess of chronic
middle ear disease (including middle ear effusion, a sign of chronic middle ear disease) in
children exposed to parental cigarette smoke. A causal mechanism is unknown, however, and
potential confounding factors may be important. The long-term implications of excess middle ear
disorders need further study. The U.S. SG and NRC reports are similar on this topic; neither
draws a firm conclusion.
Additional study data subsequent to these two reports adds to the weight of evidence. In
particular, a dose-response relationship is reported by Reed and Lutz; Strachan and coworkers
5-33 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
have shown a relationship with salivary cotinine concentrations. Based on the cumulative
evidence, this report concludes that parental smoking is associated with increased incidence of
middle ear effusion. Recent studies are described below.
Reed and Lutz (1988) evaluated the association between middle ear effusion and
household smoke exposure in children seen in an outpatient office. They reported an increasing
percentage of middle ear effusion in children with progressive levels of smoke exposure.
Strachan and colleagues (1989) assayed saliva of 7-year-olds for cotinine concentrations
(apparently the first study to use biochemical data to evaluate exposure to ETS in primary
schoolchildren). They concluded that about one-third of the cases of middle ear effusion were
statistically attributable to exposure to tobacco smoke. By contrast, Kallail et al. (1987) found
that exposure to cigarette smoke apparently was not a risk factor for middle ear problems in a
survey of primary schoolchildren. The discrepancy in conclusions is possibly attributable to
study differences. Kallail and colleagues grouped children according to outcomes on a school's
hearing test. All members of the experimental group were diagnosed by a physician as
manifesting a middle ear problem. By contrast, Reed and Lutz (1988) and Strachan et al.
(1989) both addressed a specific middle ear problem (effusion) as indicated by abnormal
tympanograms.
The association of passive smoking with middle ear effusion is further supported by
Hinton (1989), who conducted a study to ascertain whether there is any relationship between
parental smoking and various factors in children undergoing surgery for otitis media with
effusion. That study included 115 children of ages 1 to 12 years who were admitted for
grommet insertion. The admission rate for grommet insertion was statistically higher for
children with at least one parent who smokes.
5-34 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
5.7.2. Acute upper-respiratory-tract illness
Risk factors for acute upper-respiratory-tract disease in childhood were evaluated in a
population-based sample of the Atlanta metropolitan area by Fleming and colleagues (1987).
Mothers of children less than 5 years of age were questioned about upper-respiratory-tract
infection and ear infection occurring in their children during the preceding two weeks. Maternal
smoking was a risk factor for a child's having upper-respiratory-tract infection (odds ratio = 1.7,
p = 0.01). The small study by Teculescu and colleagues (Section 5.4.3.) found that children of
parents who smoke had more frequent upper airway infections.
A study by Willatt (1986) found that sore throats were predicted by maternal passive
smoking in children of ages 2 to 15 years. A regression model indicated a dose-response
relationship between sore throats and the number of cigarettes the mother smoked. The author
noted that active smoking in the older children could not account for these results since the
relationship between sore throats and smoke exposure was strongest for children under the
median age of 6.9 years. As the author also observes, few studies have looked at the effects of
passive smoking on children's upper respiratory tract.
5.7.3. Asthma
The two central issues related to ETS and asthma in children are whether parental
smoking increases the prevalence of asthma cases and whether passive smoking exacerbates
conditions in asthmatic children. The populations differ with regard to these two questions so
the first issue is discussed in this section and the other one is addressed in the next section.
The U.S. SG report found no consistent relationship between the report of a doctor's
diagnosis of asthma and exposure to ETS. It noted that the variability in results may reflect
differing ages of the children studied, differing exposures, or uncontrolled bias. Recent studies
5-35 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
offer some additional evidence suggesting increased prevalence of asthma in children with
smoking parents, but the cumulative evidence remains inconclusive.
Stern and coworkers (1987) found that asthma is more prevalent before the age of two in
smoking households. SomerviJle and colleagues (1988) reported increased asthma attacks and
morning cough in children 5 to 11 years old, but the results were not statistically significant.
Chen et al. (1988) found that asthma was reported more frequently for children in smoking
families, but the increase was not statistically significant.
A very recent article by Weitzman et al. (1990) reports significant increase in childhood
asthma with maternal smoking. Data from the Child Health Supplement to the 1981 National
Health Interview Survey were analyzed with information about 4331 children aged 0 to 5 years
to study the relationship between maternal smoking. It was found that maternal cigarette
smoking is associated with higher rates of asthma, an increased likelihood of using asthma
medications, and an earlier onset of the disease, independent of a number of potentially
confounding variables. Children whose mothers smoke one-half pack or more per day are twice
as likely to have asthma and are four times as likely to use asthma medications as children of
mothers who do not smoke. The authors caution that all information in the study is based on
parents' reports (which may be a source of bias as discussed previously).
A study relating bronchial responsiveness in parental smoking was conducted by Martinez
and colleagues (1988). Questionnaires were administered to parents of 172 Italian children 9
years old regarding parental smoking habits, the child's and family's history of respiratory illness
and symptoms, the number of persons living in the house, the number of rooms in the house,
and the type of heat. Skin prick tests and a flow-volume spirometric test were also
administered. Male children with smoking parents had a statistically significant increase in
bronchial responsiveness (BR) when compared to those whose parents did not smoke (odds
ratio, OR = 4.3). No significant increase in BR was found in female children of smoking
5-36 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
parents (OR = 0.5). The relationship between BR in children and smoking in parents was
stronger in asthmatics (p = 0.02) and remained significant after controlling for asthma and
atopy. BR was significantly correlated with atopy. This was also true for nonasthmatic children
and for both males and females separately. Male children of smoking parents had increased
reactivity to allergens as assessed by the skin prick test index (p = 0.001). It was hypothesized
that passive smoking, by increasing the frequency of BR and of atopy, may increase the risk of
asthma in childhood, particularly in boys.
The following study by Geller-Bernstein et al. (1987) is included in this section on asthma
because, as the authors discuss, atopic children with post-infancy wheezing often suffer from
asthma throughout childhood. The authors recorded the clinical course and sequential IgE
values in a 4-year prospective study of 80 atopic wheezing children between the ages of 6
months and 5 years. Although there was no correlation between increase of IgE levels and type
of feeding or exposure to cigarette smoke, statistical data confirmed that bottle feeding and
parental smoking lead to persistent wheezing in atopic children.
Neuspiel et al. (Section 5.4.3.) found no evidence of increased prevalence of asthma in
children of mothers who smoke, but did find a significant increase in wheezy bronchitis in those
children up to 10 years of age.
5.7.4. Symptoms in Asthmatics
The U.S. SG and NRC found some evidence that ETS exposure may increase the
frequency or severity of attacks of bronchoconstriction in asthmatic children, but results have
been inconsistent and difficult to compare. The stability and the mechanisms of
bronchoconstriction differ among asthmatics and study populations have not always been fully
characterized. The NRC notes several unresolved issues. For instance, what proportion of a
clearly defined population of asthmatics do react to ETS? If the patients are selected according
5-37 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
to methylcholine or histamine responsiveness, criteria should be given for the extent of
responsiveness since it is a continuum. These issues remain unresolved, although recent studies
contribute additional results to the accumulating evidence on this general topic.
Murray and Morrison (1989) reported in a previous article (1986) that asthmatic children
with mothers who smoke have substantially reduced lung function compared to those whose
mothers do not smoke. Their current study of 415 nonsmoking asthmatic children found that
asthma symptoms (based on an asthma score) were more severe if the mother smoked, with
boys apparently more affected than girls. Compared to boys with nonsmoking mothers, there
was also a significant decrease in FEV,, FEF^.^, and PC20 in boys with mothers who smoke.
Maternal smoking was not significant for lung function tests in girls. When analyzed by age
categories, 1 to 6, 7 to 11, and 12 to 17, an age effect became apparent. In the youngest
category, there was no significant effect of maternal smoking nor any indications of asthma
severity. In the intermediate age category, there was a significant difference in asthma severity
(as measured by an asthma score) but not in FVC, FEV,, FEF^.^, or PC20. In the oldest
category, however, maternal smoking was significant for all but FVC. These results are in
contrast to those for asthmatics with nonsmoking mothers where lung function improved
significantly with age.
Evans and colleagues (1987) evaluated 276 asthmatic children for association of ETS
exposure with frequency of emergency room visits, hospitalizations, and impaired pulmonary
function. Although a strong association was found for emergency room visits, no association of
passive smoking was detected for hospitalizations or abnormalities of pulmonary function (FEVj,
FEF25.75, and PEFR). The analysis of data from the National Health Interview Survey by
Weitzman et al. (Section 5.7.3.) found no relationship between maternal smoking and
hospitalizations of asthmatic children up to five years of age.
5-38 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
5.7.5. Non-Specific Ailments
Ostro (1988) reviewed five years of data from the annual Health Interview Survey
(conducted by the National Center for Health Statistics) and found that the number of bed
disability days for children of age 0 to 6 years is 20% higher in households with a pack-a-day
smoker. A similar result was obtained for adult nonsmokers with a spouse who smokes. (The
author notes that the use of disability days in bed as an indicator of acute morbidity is not a
strict measure of respiratory impairment.)
Charlton and Blair (1989) found that children's absence from school for minor ailments
(e.g., colds, flu, tonsillitis) could be predicted on the basis of their own and their parents'
smoking habits four months earlier. The sample consisted of 2885 English children, 12 to 13
years old. Passive smoke exposure was defined as neither parent smoked, only the father
smoked, only the mother smoked, or both parents smoked. Logistic regression indicated that
whatever the children's smoking habits, the proportion who were absent was higher when both
parents or at least the mother smoked. For children who never smoked the proportions absent
were 17% if neither parent or only the father smoked vs. 21% if both parents or only the
mother smoked. When children smoked "regularly," the proportion absent was 37% if neither
parent smoked vs. 46% if both or only the mother smoked. Sex and social background had little
effect. Although the authors relate absenteeism to ETS exposure, the evidence for a causal
relationship is not apparent.
5-39 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
APPENDIX A
SUMMARY DESCRIPTIONS OF ELEVEN CASE-CONTROL STUDIES
BROW. The case-control study of risk factors for adenocarcinoma by Brownson et al. (1987)
includes 23 never-smoker cases (19 females) among the 102 cases interviewed. All subjects were
white, had microscopically confirmed cancers incident from 1979 to 1982, and were identified
through the Colorado Central Cancer Registry which covers the five county Denver
metropolitan area. In the study as a whole, interviewed cases represented 68.5% of the 149
cases meeting eligibility criteria. Controls were chosen from persons with cancer at sites
unassociated with cigarette smoking and were matched to the cases on age and sex. Of the 169
eligible controls, 131 (77.5%) were interviewed. Sixty-nine percent of the cases and 39% of the
controls required surrogate respondents.
Passive smoke exposure was analyzed both as a dichotomous variable based on the
smoking status of the spouse and as a stratified variable based on the hours per day that the
subject was in the presence of persons smoking. Other variables pertain to previous smoking,
education, income, occupation, and residence history as an indirect measure of exposure to total
suspended particulates.
The relative risk for adenocarcinoma among female never-smokers exposed four or more
hours per day relative to a lower exposure was 1.68 (95% C.I. = 0.39 - 2.97) after adjustment
for age, income, and occupation. Similar nonsignificant risk estimates were shown when
smoking by the spouse was considered as a dichotomous variable. The high proportion of
surrogate source data led the authors to conduct parallel analyses limited to self-reported data.
Results from those analyses were described as highly comparable and indicated possibly higher
risks than those reported for all respondents.
A-l 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Note: The number (19 female cases) of never-smokers in this study is much too small to
make even a large observed odds ratio (1.68) statistically significant. Further, combining ever-
smokers and never-smokers (possibly to increase the sample size) makes the results difficult to
compare with previous findings.
GAO. Gao et al. (1987) report the results of a large (1407 subjects) population-based case-
control study of lung cancer etiology in Shanghai China, where lung cancer rates for women are
among the highest in the world. Potential cases included all female patients with newly
diagnosed primary lung cancer incident between February 1984, and February 1986, who were
35 to 69 years of age at the time of diagnosis and were residents of urban Shanghai. After
exclusion of 93 patients who died, the remaining 672 cases were interviewed. Eighty-one
percent were diagnosed by tissue biopsy or cytology and 19 percent by repeated x-ray.
Adenocarcinoma was the predominant (61%) diagnosis. Controls were frequency-matched
within five-year age strata and randomly selected from the general population of the Shanghai
urban area. Of the total of 735 controls interviewed, only 9.7% were secondary controls, chosen
mainly because the first selected control had moved from the Shanghai urban area or was found
to be outside the eligible age range. The study includes 246 cases and 375 controls who were
nonsmokers (presumably had never smoked cigarettes). Logistic models were used to estimate
relative risks of disease adjusted for other study factors.
Among all subjects no significant increase in risk was observed for overall ETS exposure
during childhood (OR = 1.1, 95% C.I. = 0.7 - 1.7) or adult life (OR = 0.9, 95% C.I. = 0.6 -
1.4). For these calculations, exposure was said to have occurred if the subject had ever lived
with a smoker. However, when exposure was defined in terms of husbands' smoking and the
analysis was limited to nonsmoking women, lung cancer risks tended to increase with the
number of years of exposure, with the highest observed risk (OR = 1.7, 95% C.I. = 1.0 - 2.9)
A-2 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
occurring in the comparison of those with 40+ years of exposure to those with 20- years
exposure, after adjustment for age and education (Table II). The relative risk in this
comparison was higher (OR = 2.9, 95% C.I. = 1.0 - 8.9) for squamous- and oat-cell carcinoma
alone. No test for trend over levels of ETS exposure was reported.
In the discussion of the results, the authors note the upward trend in risk associated with
increasing years of exposure among nonsmoking women married to smokers. They conclude that
ETS may be a contributing causative agent, but that other factors need to be considered as well,
e.g., pre-existing lung disease, hormonal conditions, and especially exposure to cooking oil
vapors.
Note: The study was not undertaken specifically to look at ETS lung association. Despite
the large number of nonsmokers, it was not possible (or the authors chose not) to use women
married to nonsmokers as a comparison group in their Table II. That may have been
necessitated by the high prevalence of cigarette smoking among Chinese males.
GENG. In a brief article describing work similar in design and purpose to Gao et al, Geng et al.
(1987) report the results of their study of lung cancer risk factors among women living in
Tianjin, where the rates of lung cancer mortality are the highest in China. All 157 female cases
were resident in Tianjin for at least ten years and were pair-matched to 157 controls by sex,
race, age (within 2 years), and marital status. Diagnosis was predominantly by histologic or
cytologic review (84.7%), although computerized tomography (10.8%) and clinical or x-ray
(4.5%) methods were also used to identify cases. The authors describe the case group as
representative of Tianjin female lung cancer patients in terms of age and distribution of
residents. They further state that the prevalence of smoking among the controls (40.8%) is
similar to that seen among the Tianjin adult female population. The participation rates for cases
and controls is not given, but other studies from China have reported very high response rates.
A-3 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The study report available in the literature is fairly brief. Neither the method for
assigning ETS exposure nor information about personal smoking status are discussed. Both
multiple conditional regression and stratified analytic techniques were used to calculate reported
risk estimates, but the authors do not stipulate which variables were controlled for in the
analyses.
The authors report that among the odds ratios of passive smoking from husbands,
fathers, mothers, and colleagues, only that from husbands is significant. However, it is not clear
whether this applies to smokers and nonsmokers combined in the same analysis or whether the
analyses of ETS were restricted to nonsmokers only. The authors do explicitly state in Table 5
that the odds ratio for lung cancer in nonsmoking women married to smokers is 2.16 (95% C.I.
= 1.05 - 4.53), but it is not clear why this estimate differs from the odds ratio of 1.86 for
nonsmoking wives with smoking husbands in Table 7. The odds ratios for lung cancer increase
with the number of cigarettes smoked per day by the husband and the duration of exposure to
the husband's smoking (Table 6). No tests for trend are provided, however, and whether these
findings apply to all subjects as a group or only to the nonsmokers is not clear.
One interesting finding in Table 7 of this brief report is the similarity of estimated
effects associated with ETS exposure from a husband only (OR = 1.86, 95% C.I. = 1.04 - 3.5)
and active smoking by the wife only (OR = 2.61, 95% C.I. = 1.4 - 4.6). Further, these
independent risks can be seen to interact on a multiplicative scale among smoking women
married to smoking husbands (OR = 4.9, 95% C.I. = 1.8 - 9.5). The authors did not state
whether these estimates were adjusted for other factors.
HUMB. The study by Humble and colleagues (Humble et al., 1987) includes 28 incident cases
described by interview to be lifelong nonsmokers (8 men, 20 women). Cases were identified
through the population-based New Mexico Tumor Registry while controls (130 men, 162
A-4 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
women) were chosen through randomly generated phone numbers and Health Care Financing
Administration rosters of Medicare participants. Controls were frequency-matched to cases by
ten-year age groups and by sex. Subjects were the nonsmoking subset in a larger study of lung
cancer risk factors in which 88.5% of cases and 83.1% of controls eligible for interview had
participated. Of the 28 lung cancers among nonsmokers, 24 had a histologic diagnosis in the
Tumor Registry record. However, in a separate review of histologic materials for 17 of these
cases, only eight cell types concurred with the Registry.
Subjects or their proxies were interviewed regarding their personal smoking habits,
smoking by their spouses, and their occupational exposures. Surrogate interviews (usually with
the spouse) were necessary for 19 of the 28 cases, but for only 13 of the 292 controls. No effect
of information source was noted when analyses were run separately for self-reported and
surrogate-reported cases using self-reported controls as the comparison group. Small numbers
precluded a separate reporting of the OR for males.
An elevated risk of lung cancer was reported for all subjects combined and for females
separately. Logistic models, which included adjustment for age and ethnicity and sex when
appropriate, calculated ORs of 2.6 (90% C.I. = 1.2 - 5.6) for all subjects and 2.2 (90% C.I. = 0.9
- 5.5) for females. Risk increased with the duration of spousal smoking (chi-squared statistic for
linear trend equals 2.01 for all subjects and 1.23 for females alone) in cross-tabular analyses, but
not in results from multiple logistic models. No trend was seen over the average number of
cigarettes smoked per day by the spouse. Separate analyses for current and former smokers
revealed no increased risk associated with marriage to a smoker.
Cell-line specific analyses were precluded by the small number of cases with histologic
confirmation of their diagnosis, the poor concordance of histologic designations in the Registry
file, and the special review. The high proportion of cases with surrogate respondents may
actually have improved the quality of data regarding exposure to a spouse's cigarette smoking, as
A-5 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
spouses were the principal source of surrogate data. Exclusion of four former smokers (by
information from other sources) did not alter the results. Size of the case series allowed only
crude stratification of duration and amount when testing for trends, and may explain the
marginal significance of findings reported separately for women.
INOU. In a case-control study of smoking and lung cancer in two Japanese cities, Inoue and
Hirayama (1987) identified 37 women who died with lung cancer. Twenty-eight of these women
(75.7%) were nonsmokers (definition not given). Cases were matched for age, year of death
(within 2.5 years), and residential district to 74 controls who had died of cerebrovascular disease.
Sixty-two (83.8%) of the controls were nonsmokers. Husbands' smoking status was available for
29 of the 37 cases and 54 of the 74 controls. Interviews were used to gather data for analysis,
but the authors do not describe the characteristics or degree of relatedness of the surrogate
respondents. Neither do they describe the degree of cooperation among the study subjects.
The Mantel-Haenszel procedure was used to estimate the relative risks of disease
associated with ETS, adjusted for age alone and for age and residential district (due to
differences in socio-economic status of the two areas). The odds ratios, stratified by age and
district, are 2.58 (90% C.I. = 0.44 - 5.7) when husbands smoked less than 19 cigarettes a day,
and 3.09 (90% C.I. = 1.04 - 11.81) when husbands smoked 20 or more cigarettes a day. The
chi-squared test for trend is significant (p < 0.05).
LAMT. The large case-control study by T.H. Lam and colleagues (Lam et al., 1987) assessed
the respective roles of active and passive smoking in lung cancer etiology among women living in
Hong Kong. Only patients with a pathologist's confirmation (98% by histological or cytological
review) were included. Those with rare tumors, e.g., carcinoids, were excluded. Women were
interviewed in the hospital and then age-matched to healthy female controls selected from
A-6 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
within their own neighborhoods. Interviews took place between 1983 and 1986 and
approximately 99% of all eligible subjects responded. Never-smoker status for both subjects and
their husbands was defined as having never smoked as much as one cigarette a day, or its
equivalent in other tobacco products, for at least one year. A woman was considered exposed to
her husband's tobacco smoke if she had lived with her smoking husband in the same household
continuously for at least one year. If the husband was an ever- smoker, information on the type
of tobacco and amount usually smoked per day by the husband and the duration of exposure
was obtained. Never-married women were included as nonexposed to ETS. The authors
describe the results of separate analyses on cigarettes only and on all forms of tobacco as similar
and only report the latter. RR and 95% confidence intervals were calculated for each level of
ETS exposure. The Fisher's Exact Test (two-sided) was used to check whether the RR was
significantly different from unity. Multivariate methods do no appear to have been used.
Among the total of 444 cases and 443 controls were 199 cases and 335 controls who had
never smoked and for whom data on husbands' smoking were available. For never-smokers the
RR for lung cancer of all types from ETS exposure is 1.65 (95% C.I. = 1.16 - 2.35); for
adenocarcinoma the RR is 2.12 (95% C.I. = 1.32 - 3.39). The risks for small and large cell
carcinomas are 3.00 and 3.11, respectively, but these estimates are not statistically significant.
Trends in relative risk for cancer at all sites, and for adenocarcinoma by the amount of tobacco
smoked daily by the husband, are both significant with p < 0.001. The authors discount the
possibility that misclassification bias could have lead to the observed results, given the low
prevalence of smoking (4.1%) among women in Hong Kong and the strength of the findings in
the present study.
LAMW. The dissertation of Lam (Lam, 1985) was the third case-control study of risk factors
for lung cancer among females in Hong Kong. The nonsmoker cases, all with histologic or
A-7 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
cytologic confirmation of adenocarcinoma, were part of a larger case series of 161 interviewed
Chinese female lung cancer patients diagnosed at a large, regional general hospital between
January 1981 and April 1984. Fifteen cases with three other lung cancer histologies, as well as
any patients with metastatic disease, were not included. Nonsmoking controls (n = 144) were
part of a larger series of 185 Chinese, mostly lower income female patients admitted to the
orthopedic wards between 1982 and 1984. Cooperation of potential subjects exceeded 99%.
There was little difference in the ages, occupations, years of schooling, or recent
residences of the 161 cases and 185 controls, so the author deemed it unnecessary to control for
(stratify on) these variables in the analysis of the 60 nonsmoking cases with adenocarcinoma and
144 nonsmoking controls. Exposure to ETS was categorized separately for husbands and other
sources, e.g., cohabitating relatives or coworkers. Subjects were also queried regarding exposure
to smoke from kerosene stoves and incense. The author interviewed all cases and, with a single
research assistant, all controls. Thus, one may assume that interviews were not "blind."
The strongest and most statistically significant associations of ETS were with peripheral
adenocarcinoma, with the highest odds ratio (2.64) occurring when exposure was based solely on
husbands' smoking behavior. Estimates of relative risks of 1.6 and 1.7 were found for centrally
located tumors when ETS was based on the husband's habits and total exposure to passive
smoking, respectively. When data from Table 7.5 of the study are summed over sites, relative
risks of approximately 2.0 are obtained with p < 0.05, regardless of exposure classification
scheme. All odds ratios appear to be unadjusted for any other study factors. No statistically
significant risks from kerosene or incense were found. The author concludes that the small
sample size and use of only a single hospital source for subjects are limitations. Logistic
regression was used in the statistical analysis, along with Bayesian risk-ratio procedure.
A-8 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
SHIM. Shimizu and his colleagues (Shimizu et al., 1988) use a hospital-based case-control study
of lung cancer in women to examine the effect of involuntary exposure to tobacco smoke from a
variety of sources. Among 118 female patients with histologically confirmed lung cancer, 90
reported having never smoked cigarettes. Cases were matched on hospital, age (within 1 year),
and date of admission to patients being seen for conditions generally unrelated to tobacco use.
All subjects were asked to complete a questionnaire about occupational history, kinds of fuels
used for cooking and heating, and smoking habits, including number of cigarettes smoked daily
by parents, siblings, the husband, and the husbands' parents in the home, as well as the amount
of time spent in the same room with the husband, and the duration of marriage. ETS exposure
at work was simply categorized by presence or absence of smokers.
No association was observed between risk of lung cancer and smoking by husbands,
fathers, siblings, or coworkers. However, increased odds ratios were seen for smoking by
subjects' mothers (OR = 4.0, p < 0.05) and by their husbands' fathers (OR = 3.2, p < 0.005).
Dose-response relationships were not apparent for exposure by the mother or the husband's
father, but the authors suggest that subjects may have been unable to recall the exact number of
cigarettes in some cases (especially in childhood).
It is not clear whether variables such as occupational exposure to iron and other metals,
or type of heating fuel, were assessed. Neither is there mention of cooperation rates by cases
and controls. Adjustment of odds ratios for smoking by mother, smoking by husbands' father,
and occupational exposures to iron and other metals, caused modest reductions in the point
estimates, although smoking by husband's father in the home, (adjusted OR = 3.2) is still
significant with p < 0.005. The authors describe this association as plausible since a high
proportion of Japanese wives live with their in-laws after marriage and their father-in-law may
have already retired.
A-9 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
SVEN. The study of lung cancer etiology in women who had never been regular smokers by
Svenson et al. (1988) includes 34 cases with microscopically confirmed non-carcinoid cancer.
Cases were patients referred to one of four clinical departments that diagnose or treat lung
cancer in Stockholm county, Sweden. Only patients who would not benefit from specialist care,
or who were not in physical or mental condition to allow an interview, were excluded from
eligibility. Cases were matched on age using random selection from the population register in
Stockholm County. Only seven subjects refused to be interviewed, resulting in a sample of 210
cases and 209 controls. Cooperation of nonsmoking cases and their matched controls was
presumably high as well.
Four physicians completed all interviews using a structured questionnaire that included
ETS exposure during childhood, as well as domestic and work environment exposure during
adulthood. Other questions concerned the consumption of foods rich in vitamins A and C, and
information about the dwellings where a subject had lived for more than two years. No
surrogate sources of information were used and squamous/small cell carcinomas constituted
57.9% and 20.6% of the case histologies, respectively.
Women who lived with a smoking mother as children (RR = 3.3), or were exposed to
ETS both at home and at work (RR = 2.1), or were exposed both as children and as adults (RR
= 1.9), showed the highest risks. However, all estimates had very wide confidence intervals
owing to the small sample size, and tests of association between ETS exposure and lung cancer
incidence and tests for trend were all nonsignificant.
The authors describe the results for ETS as inconclusive, but note that most estimates of
relative risk are greater than unity. The statistical power to detect an increased risk of 50% from
exposure to ETS was only about 0.1. The author suggests that information bias may have
precluded the identification of statistically significant small increases in risk. Specifically, no
information on the duration or intensity of ETS exposure was obtained in the study, so it was
A-10 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
difficult to assess the relative importance of domestic and workplace exposures. Several
statistical methods were applied to the data, including stratified analyses and multiple logistic
regression.
VARE. The case-control study described by Varela in his 1987 dissertation (Varela, 1987) is
based on 439 histologically confirmed primary lung cancer cases incident in nonsmokers over an
18-month period in upstate New York. Sample size requirements were set large enough that
detection of a relative risk of the size reported by Hirayama and Trichopoulos would be likely.
However, to reach the calculated requirement of 450 matched case-control pairs, it was
necessary to include former smokers (55% of sample) in addition to never-smokers. Cases were
identified through a special rapid reporting system in all participating hospitals and through
periodic review of the New York State Cancer Registry. Controls were matched to cases on
residence, age (within 5 years), sex, smoking history, and whether the interview was with the
subject (67%) or with a surrogate (33%). Standardized interviews were conducted to collect data
describing exposure to a spouse's cigarette smoke in terms of cig./day, total years of smoke
exposure, and total cigarettes smoked during the marriage. Information was also collected on
total exposure from all smokers in the household, from coworkers on the job, and from exposure
in social circumstances. The potentially confounding variables considered in the analysis include
religion, income, marital status, other occupational exposures, and number of cigarettes
smoked/day for former smokers. The study's total of 439 cases represents a cooperation rate of
84% among those selected for interviews.
The author provides a systematic and exhaustive analysis based on linear logistic models
for pairwise matched data. These data were collected as continuous values to allow analysis by
source of exposure, e.g., spouse, other household smokers, coworkers, and social encounters,
using methods for both continuous data and for categorical data. Analysis of household exposure
A-11 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
was further complicated by the use of two alternative assumptions regarding missing data for
exposure at previous residences.
After extensive analyses no index of exposure to spouse's tobacco or smoking by
coworkers was associated with an increased risk for lung cancer. However, person-years of total
exposure from all smoking household members showed a statistically significant linear trend.
When exposure was fitted as a continuous variable, the unadjusted odds ratio associated with
150 person-years of exposure was 1.86 (95% C.I. = 1.22 - 2.83). Adjustment for the potentially
confounding variables listed above reduced the OR for 150 person-years of exposure to 1.56
(95% C.I. = 1.00 - 2.41). Exposure to passive smoke in social situations showed an anomalous
protective effect in both adjusted and unadjusted models (Tables 20 to 22 and Figures 25 to 28).
Note: The study contains extensive statistical analyses of which only a small part have
been described here. When a large number of tests are made, the likelihood that one or more
statistically significant results will occur by chance alone increases. This can cause results to be
interpreted as more significant than may be justified.
The author suggests that his own finding of no effect from exposure to spouses' smoke is
understandable because the smoking habits of a spouse may not accurately describe true
exposure to passive smoke. By contrast, the household exposure variable which was designed to
more fully capture exposure in the home was the only index that was associated with increased
risk of disease in this study. The greater association of household exposures with epidermoid
and small cell histologies (Tables 12, 13, 15, 16) is not inconsistent with the apparent specificity
of effect observed in PERS and GARF. One difficulty with comparing the Varela study with
other case-control studies is the inclusion of either males with females or ex-smokers with never-
smokers, in the reported results. Although the analysis is very comprehensive, no reports for the
risk of female never-smokers alone were found. The author suggests that differences in past
smoking habits of cases and controls may have a confounding effect. Although identical
A-12 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
proportions of cases and controls were former smokers, cases had smoked a larger number of
cig./day (28.9 vs. 23.8, p = 0.0002). Former smokers were not included, however, unless they
had stopped smoking at least ten years prior to the interview. The author questions the validity
of an apparent significant protective association from ETS in social circumstances, suggesting
the possibilities of biased reporting and questionnaire artifacts as alternative explanations for
this finding.
WU. Wu and her coauthors (Wu et al., 1985) report the effects of ETS exposure as part of a
larger study of determinants of lung cancer among white women living in Los Angeles County.
Eligible cases included only patients with microscopically diagnosed primary adenocarcinoma
(ADC) or small cell carcinoma (SCC) of the lung, incident between April 1, 1981, and August
31, 1982. Subjects also had to be English-speaking residents and less than 76 years old at the
time of diagnosis. One neighborhood control was individually matched to each interviewed case
using date of birth (within five years).
From a total of 490 eligible cases (smokers and nonsmokers), 190 were dead or too ill to
participate, eight could not be located and 44 refused to be interviewed, leaving 220 (44.9%) as
the interviewed case group. After replacement of 85 potential controls who refused to
participate, 220 controls were also interviewed. Surrogate respondents were not used because
they were thought to be an unreliable source of information for ETS exposures and dietary
practices in childhood, this article reports 29 cases (ADC) with 62 controls, but does not
include the percentages exposed to spousal smoking. Also, it is noted that 15 pairs of the ADC
were deleted from the analysis because either the case or control was never married.
Cases and controls were interviewed by telephone regarding personal smoking habits,
exposure to ETS, history of lung diseases, dietary intake of vitamin A, types of heating and
cooking fuels used, and reproductive history. Information obtained about childhood exposure to
A-13 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
ETS included the amount and years of smoking by fathers, mothers, and other household
members. Questions on exposure in adulthood pertained to smoking habits of spouses and other
household members.
Study data were adjusted for potential confounding variables by application of logistic
regression. Estimates for the relative risk of ADC are provided separately for nonsmokers,
ex-smokers, and current smokers, but a small number of occurrences precluded the
corresponding calculations for SCC. For ADC and SCC among smokers and nonsmokers
combined, no significantly increased risks were observed due to smoking by the subject's mother,
father, spouse, or coworkers after adjustment for personal smoking habits. For the 29
nonsmoking ADC cases exposed to passive smoke, no significant elevated risk was associated
with ETS exposure from a mother who smoked, a father who smoked, a spouse who smoked, or
from the workplace. The observed relative risk for ADC increases with the number of years of
adult ETS exposure from spouse(s) and coworkers, but a test for trend is not statistically
significant. The authors attribute the ambiguous nature of their results to the lesser etiologic
role of ETS for ADC compared to SCC. Further, 12 (41%) of the 29 ADC cases are
bronchoalveolar cell carcinomas, which Correa et al. (1983) found to have a relatively weaker
association with passive smoking.
A-14 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
APPENDIX B
MATHEMATICAL FORMULAS AND RELATIONSHIPS
ADJUSTING RELATIVE RISK FOR SMOKER MISCLASSIFICATION AND
BACKGROUND ETS EXPOSURE
The formula relating observed relative risk (RRO) and the value after adjustment for
misclassification (RRM) is shown in Equation Bl below, with terms described in Tables B-l and
B-2. The calculational procedure is similar to that of Wald et al. (1986), except that separate
terms for former smokers and current smokers are retained (instead of being combined into a
single term for ever-smokers) and distinction is made between "correct" values and "reported"
values, e.g., for the number of never-smokers. It may be noted that an assumption of additive
risks due to current or former smoking, and exposure to spousal smoke, is implicit.
RRO = UY/VX (Bl)
where U = [(Cc-Cr)/NJP(E/C)RR(E/C)
+ [NC/NJP(E/N)RR(E/N)
+ [(Fc-Fr)/NJP(E/F)RR(E/F),
V is The same as U with the terms for RR omitted,
X = [(Cc-Cr)/NJ [1-P(E/C)] RR(E/C)
+ [Nc/Nr] [1-P(E/N)] RR(E/N)
+ [(Fe-Fr)/NJ [1-P(E/F)] RR(E/F),
Y is The same as X with the terms for RR omitted,
RR(E/N) = RRM, the relative risk after adjustment for misclassification, and
RR(E/C) = RR(U/C) + (RRM - 1),
RR (E/F) = RR (E/F) + (RRM - 1).
B-l 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE B-l. DEFINITION1 OF TERMS IN EQUATION Bl RELATING OBSERVED
RR (RRO) AND ITS VALUE ADJUSTED FOR MISCLASSIFICATION (RRM)
Term
Description
Nr
N
P(E/N)
P(E/C)
P(E/F)
RR(E/N)
RR(E/C)
RR(E/F)
RR(U/C)
RR(U/F)
Reported number of NS (never-smokers)
Correct number of NS (never-smokers)
Reported number of CS (current smokers)
Correct number of CS (current smokers)
Reported number of FS (former smokers)
Correct number of FS (former smokers)
Proportion of NS exposed2
Proportion of CS exposed2
Proportion of FS exposed2
Risk of lung cancer death (LCD) for NS
exposed, relative to NS unexposed3
Same as RR(E/N) except for CS exposed3
Same as RR(E/N) except for FS exposed3
Same as RR(E/N) except for CS unexposed3
Same as RR(E/N) except for FS unexposed3
1 Table applies to marrieds.
2 "Exposed" means married to a smoker.
3 RR(E/N) equals RRM in the text notation.
RR(U/C) is the relative risk of smoking, a parameter value.
RR(U/F) is the relative risk of former smoking, a parameter value.
RR(E/C) = RR(U/C) + (RR(E/N) - 1). Assumes relative risk of exposed smoker
is the sum due to smoking and spousal exposure to ETS.
RR(E/F) = RR(U/F) + (RR(E/N) - 1). Same assumption as for RR(E/C) except
applied to former smokers.
B-2
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE B-2. PARAMETERS AND THEIR ALTERNATIVE SPECIFICATIONS
REQUIRED FOR EQUATION (Bl)
Parameter
Identifier
VI
V2
V3
V4
V5
V6
V7
V8
V9
V10
Vll
V12
V13
V14
V15
V16
V17
V18
V19
Appears in
(Bl)?
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
N
N
N
N
N
Parameter1
(Q-Cr)/Nr
(Fc-Fr)/Nr
Nc/Nr
P(E/N)
P(E/C)
P(E/F)
RR(U/C)
RR(U/F)
RR(E/N)
RR(E/C)
RR(E/F)
Cr/T
Q/T
(cc-cr)/cc
Nr/T
NC/T
Fr/T
FC/T
(Fc-Fr)/Fc
Description
Proportion of reported never-smokers
(NS) who are current smokers (CS)
Proportion of reported NS who are
former smokers (FS)
Proportion of reported NS who are NS
Proportion of NS exposed (married to
a smoker)
Proportion of CS exposed
Proportion of FS exposed
Risk of misclassified CS, not exposed,
relative to NS, not exposed
Risk of misclassified FS, not exposed,
relative to NS, not exposed
Risk of NS, exposed, ( = RRM) relative
to NS, not exposed
Risk of misclassified CS, exposed, relative
to NS, not exposed
Risk of misclassified FS, exposed, relative
to NS, not exposed
Proportion of subjects reported to be CS
Proportion of subjects correctly classified
CS
Proportion of CS reported to be NS
Proportion of subjects reported to be NS
Proportion of subjects correctly classified
asNS
Proportion of subjects reported to be
former smokers
Proportion of subjects correctly classified
FS
Proportion of FS reported to be NS
1 "T" is the total number of subjects.
B-3
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
RRO is just a ratio of weighted relative risks. The term U/V applies to exposed subjects
and X/Y applies to unexposed subjects. To make the relationship more transparent, consider
the special case where the proportion of all subjects (regardless of their own smoking habits)
married to a smoker is one-half and there is no misclassification of former smokers. Then RRO
can be written as RRO = A/B, where A = (no. of misclassified CS exposed) times (their
relative risk) + (no. of correctly classified NS exposed)times(their relative risk) and B is the
same as A except with "exposed" replaced by "unexposed." All of the relative risk terms are
relative to the same value (the lung cancer risk in unexposed never-smokers) so the ratio A/B is
the risk observed in exposed subjects relative to the risk observed in unexposed subjects.
The terms in expression Equation Bl, stated as proportions of the number of reported
never-smokers (Nr), may need to be converted from alternative parameter specifications. For
example, the misclassification rate of current smokers, (Cc - Cr)/Cc), may be specified instead of
the proportion of reported misclassified current smokers, (Cc - Cr)/Nr). In the literature,
"misclassification" is usually referred to as a percentage or proportion of a reference group, but
authors do not all have the same reference group in mind. Conversion of parameters make
some values more interpretable, as well. Formulas for conversion of parameters between
reported and correct classifications are given in Table B-3.
The procedure used to account for ETS exposure from sources other than spousal
smoking is described in the NRC report (1986). It is assumed that lifetime lung cancer risk
from exposure to ETS is linear in the range of environmental exposures to ETS. RRM is the
risk of never-smokers (NS) exposed (e.g., married to smoker) to ETS relative to unexposed NS
(e.g., married to a never-smoker). Both the exposed and unexposed NS experience ETS from
sources other than what differentiates their classification in epidemiologic studies (typically,
"exposed" means married to a smoker).
B-4 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The terms RRB and RRM denote the risk of exposed NS relative to NS with zero
exposure to ETS, and relative to an unexposed NS, respectively, where an "unexposed" NS is not
married to a smoker but experiences ETS from other sources collectively referred to as
background sources.
TABLE B-3. CONVERSION OF PARAMETERS BETWEEN REPORTED
AND CORRECT CLASSIFICATIONS
From Correct Values to Reported Values
Let W = (V13)(V14) + V16 + (V18)(V19)
VI = (V13)(V14)/W
V2 = (V18 + V19)/W
V3 = V16/W
From Reported Values to Correct Values
V14 =
(V1)(V15) + V12
V16 = V15 - (V1)(V15) - (V2)(V15)
V19 = (V2)(V15)
(V2)(V15) + V17
B-5
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
If R(E), R(U), and R(TU) denote the absolute risk (lifetime probability of lung cancer) for
exposed, unexposed, and truly unexposed (no background exposure to ETS) NS, respectively,
then RRM = R(E)/R(U), RRB = R(E)/R(TU), and R(U)/R(TU) = RRB/RRM.
Let Z be the ratio of (1) the excess risk of exposed NS relative to truly unexposed NS, and
(2) the excess risk of unexposed NS relative to a truly unexposed NS. Then, Z = (l)/(2) =
(RRB-1)/(RRB/RRM - 1). In this report, Z = 3 is assumed (see Section 4.4.4). RRM is the
observed relative risk (RRO) after adjustment for smoker misclassification. For values of Z and
RRM,
RRB = (1 - Z)/(l - Z/RRM) (B2)
POPULATION-ATTRIBUTABLE RISK (PAR)
Let RRB and RRM be as defined previously. The population-attributable risk is the ratio
of the excess risk due to ETS exposure to the total risk from all sources.
PAR = PfE/NVRRM - 1) + fRRB-RRMI (B3)
P(E/N)(RRM - 1) + (RRB-RRM) + 1
where P(E/N) is the proportion of NS exposed to ETS. Note that all NS are at risk from
background exposure (the term RRB - RRM) and the exposed persons have an additional risk
from ETS exposure (the term RRM - 1).
LUNG CANCER DEATHS IN FORMER SMOKERS ATTRIBUTABLE TO ETS
It is assumed that the RR of lung cancer from exposure to ETS is the same for former
smokers (FS) and never-smokers (NS). The number of lung cancer deaths per year in FS due
to ETS exposure is approximated from the ratio of the number of FS to the number of NS in
the U.S. population times the estimated number of lung cancer deaths for NS (1750 women, 810
men). The number of NS (FS) in the 1985 U.S. population is 55.4 million (17.1 million) women
B-6 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
and 32.8 million (29.1 million) men (Table 2, U.S. SG [1989]). The estimated number of lung
cancer deaths in FS is 540 women and 720 men, for a total of 1260.
B-7 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
APPENDIX C
DOSIMETRY OF ENVIRONMENTAL TOBACCO SMOKE
C.I. INTRODUCTION
The biological relationship between exposure to tobacco smoke and lung cancer risk has
been the subject of much research. The constituency of tobacco smoke is a complex mixture of
chemicals, a number of which have been classified as carcinogens, with varying
weights-of-evidence. Research continues toward identifying the agents of tobacco smoke, and
their combinations, that account for the carcinogenic risk to the lung and other organs. In
addition to knowledge of the chemical agents of interest, a part of the biological puzzle
concerns the intake, uptake and organ deposition of the chemicals. Once organ dose is
determined, the problem concerns the process by which dose poses a cancer risk. In this last
step, pharmacological research, dose-response data from animal or epidemiological studies, and
quantitative models all contribute toward estimating the magnitude of increased cancer risk
associated with environmental exposure levels. The following discussion addresses the second of
the three steps above, the determination of target organ dose (or surrogate) of chemicals
present in tobacco smoke, in particular the dose to the lung. A general mathematical
framework is given that applies to both active and passive smoking. It will be helpful first to
review briefly the current knowledge base regarding carcinogens in tobacco smoke.
The constituents of tobacco that have been identified as carcinogens, largely in animal
studies, are discussed in several sources, e.g., NRC, 1986; U.S. SG, 1986; IARC, 1987; Hoffmann
and Hecht, 1989. The relative concentrations in sidestream (SS) and mainstream (MS) smoke
vary over a range of severalfold, for both the particulate and vapor phases. Hoffmann and
Hecht (1989) classify the tumorigenic agents in tobacco and tobacco smoke as polycyclic
aromatic hydrocarbons (PAH), aza-arenes, N-nitrosamines, aromatic amines, aldehydes,
C-l 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
inorganic compounds, and miscellaneous organic compounds. Particularly relevant to active
smoking are the PAH, N-nitrosamines (notably NNK or NNK-4-(methyllnitrosamino)-l-(3-
pyridyl)-l-butanone), and the aldehydes. Based on animal bioassays, the levels of exposure to
active smokers of PAH and NNK are sufficient to be potential causative agents of respiratory
tract cancer. More specifically, likely causative agents for cancers of the lung or larynx from
active smoking include: PAH, with enhancing agents catechol (a cocarcinogen) or a weakly
acidic tumor promoter; NNK, with enhancing agents acrolein or crotonaldehyde (some
uncertainty in the latter); acetaldehyde; formaldehyde; and Polonium-210, a minor factor (Table
4, Hoffmann and Hecht, 1989).
Of course it is unknown exactly which constituents of tobacco smoke, active separately or
in combination, account for the lung cancer risk in active smoking and may pose a risk to
passive smokers. It is unlikely that causative agents are exclusive to either the particulate or
vapor phase, so both phases need to be considered. The level of detail that can be included in
modeling lung exposure requires knowledge of parameter values and information on biological
mechanisms to describe them by equations. The most refined level, biologically-based modeling,
potentially provides a sensitive means by which to compare and contrast features of active and
passive smoking. This is particularly relevant because active smoking has been the subject of
much research in the past. Typical exposure levels to environmental tobacco smoke (ETS) can
be evaluated under suitable models, the sensitivity of selected parameters tested, and parameters
identified that may help to characterize potentially hypersusceptible subpopulations. Aside from
the differences in the composition of SS and MS smoke, active and passive smoking ostensibly
differ in fundamental ways that must be considered. Active smokers are exposed to high
concentrations of inhaled smoke for short durations while passive smokers are exposed to lower
concentration at their typical volumetric breathing rate over a more extended time period.
Additionally, the active smoker is generally a passive smoker as well.
C-2 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
There are some common features of intake, uptake, and lung deposition for agents in the
particulate phase of tobacco smoke that depend on particle size and density. This makes it
useful to consider the particulate and vapor phases separately in a mathematical model. Unlike
the particulate phase, constituents of the vapor phase may not have any parameter values in
common and thus have to be treated individually. When a chemical occurs in both phases, the
biokinetics are determined separately for the two phases, at least prior to entry into cells and
tissues.
For the general framework described in the next section to be useful for cancer risk
assessment, it is necessary to assume that the contribution of ETS to cancer risk in the lung
depends primarily on dose to the lung (although host factors will modify this risk). The model
identifies the parameters needed for dose determination and their interrelationship. Since
values for parameters, or data from which to estimate them, are not always available in practice,
several measures related to exposure are provided to accomodate the level of detail in the
information available. These include, in increasing order of refinement and information
required: exposure concentration (in room air), cumulative exposure, lung intake, lung uptake,
lung burden, lung dose, and dose distribution in systemic organs.
Although many aspects of the biokinetics of passive smoking and active smoking are not
fully understood, much is known about critical and separate features of passive and active
smoking. Quantitative modeling reveals the structure and interrelationship of the basic
biokinetic features, serves to identify areas of research needed, and shows where assumptions
are required to bridge current gaps in knowledge regarding mechanisms or chemical properties.
In this sense a quantitative model that integrates the current knowledge base over several
disciplines is a useful guide to the current state-of-knowledge and future research needs. It also
contributes to evaluation of dose surrogates for ETS and the potential comparative basis for
lung cancer risk from active and passive smoking.
C-3 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
C.2. GENERAL ISSUES
C.2.1. The Lung
Suppose we are interested in a mixture of chemicals such as ETS, where these chemicals
may exist either in the vapor or particulate phase. In general, the vapor phase chemicals are
inhaled and absorbed in the lung with characteristics specific to that chemical, i.e., the diffusion
coefficient and solubility coefficient for that chemical in lung tissue. The chemicals existing in
the particulate phase, however, are inhaled and deposited with characteristics specific to the
particle size with which the chemicals are associated. For this reason, it is necessary to
understand the aerodynamic properties of inhaled particles. To differentiate between the vapor
and particulate phase of a chemical, the subscript v (for vapor) will be employed, while the
subscript d (for the particle's aerodynamic diameter) will be employed for the particulate phase
chemicals.
The simplest, but most approximate, measure of dose from a chemical is its concentration
in air. This concentration typically is referred to as the exposure intensity and denoted by C.
An index i will refer to the ith chemical of a mixture contained in ETS. The subscripts v and d
will denote the vapor phase and particle phase, respectively, with d taking a particular value for
particle diameter. For example, CNOS is the concentration of the N"1 chemical, e.g., nicotine
(g/m3), attached to aerosol particles of diameter 0.5/j. Then CNv would denote the
concentration of nicotine (also in g/m3) in the vapor phase. The total exposure intensity for
nicotine would be the sum of CN05 and CNv, although this should not be interpreted as a
reduction of the two phases to a common measure of dose. The subscripts for chemical and
vapor phase will be omitted for ease of reading, except as needed. Since the incidence of effect
per unit concentration can be quite different for these two components, total exposure intensity
may act as a poor measure of risk.
C-4 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The cumulative exposure, 0, over a time interval [0, T] is
C(t)dt, (1)
where C(t) is the concentration of the chemical in air at time t. At constant concentration C,
The total cumulative exposure over [0,T] is obtained by summing all contributions from the
various phases in units of gram-days per liter of air (g-d/L).
The total amount of inhaled chemical will be referred to as the lung intake, I. Let V be
the volumetric breathing rate (L/min), which equals the product of the tidal volume (L) and the
breathing frequency (min"1). The value of V may change with age (Crawford-Brown, 1987), in
which case the dependence of V on an individual's age can be made explicit in the notation.
The lung intake for the ith chemical during an interval of length T is
7(7) = V x tfi (7). (3)
The intake is in grams of the chemical considered.
With the exception of radionuclides, an inhaled chemical must deposit onto the walls of
the lung to yield damage. The quantity of a chemical deposited in the lung over a time period
[0,T] is the lung uptake, U(T). Typically, the relationship between intake and uptake will vary
dramatically between the phases of a chemical due to differences in deposition and absorption
processes. The uptake to the lung equals the product of the intake and the fraction "f" of the
chemical deposited onto the surface of the lung. The total uptake to the lung then is
U(T) = f x
C-5 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
For the vapor phase, the value of "f' depends on the specific chemical and often is
unknown (although it may be infered from solubility coefficients if these are developed in future
studies). For the particulate phase, the dominant environmental factor affecting f is the
aerodynamic diameter, d, of the particle. In general, fd will also be a function of the age of the
individual, the state of activity, i.e., whether resting or working, the hygroscopicity of the
particles, i.e., their ability to grow by gaining water in the lung, and the state of health of the
individual. The uptake U is in grams of the chemical considered.
Lung uptake must be refined further in order to account for the structure (anatomy) of
the lung. As described elsewhere (Weibel, 1963), the lung may be depicted as a series of
bifurcating passageways that branch into smaller passageways as one moves from the proximal
(near the mouth) to the distal (deep lung) locations in the lung. A simplified version of this
branching scheme has been adopted by the International Commission on Radiological Protection
(ICRP) in its report on lung modeling (ICRP, 1966). The ICRP model divides the lung into
three distinct subsections or regions, each with a specific value of f. These regions are the
naseopharyngeal (NP) region, consisting of the nose and pharynx; the tracheobronchial (TB)
region, which extends from the trachea down to the terminal bronchioles; and the pulmonary (P)
region of the lung, where gas exchange occurs between the alveolar sacs and the bloodstream. A
schematic of the lung is shown in Figure C-l.
The value of f differs between the three regions of the lung due to differences in airflow
and the size of passageways. To reflect this situation, let fNP, fTP, and fp be the deposition
fractions in the NP, TB, and P regions, respectively. Their values will depend upon the phase of
the inhaled chemical. The uptake in each of the three regions over [0,T] is obtained from
Equation 4
Up(T)=fp*I(T) (5)
C-6 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TRACHEA
BRONCHIOLES
BRONCHUS
FIGURE C-l. THE GENERAL ANATOMY OF THE LUNG FROM THE TRACHEA
DOWN TO THE DISTAL BRONCHIOLES
C-7
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
UTB(T)=fTB*I(T] (6)
UNP(T)=fNP*I(T). (7)
Each term may be subscripted further to indicate the chemical and its phase. The total lung
uptake equals the sum of Equations 5 through 7, although this sum should not be construed as
an appropriate measure of risk to specific cells of the lung.
After deposition in the lung tissue, a chemical must interact with the cells in order to
produce an effect. In general, the probability of this interaction increases as the residence time
of the chemical in the tissue increases. Because of this factor, it is necessary to specify a
retention function R(t), describing the fraction of a chemical remaining in that tissue at a time,
t, after uptake. For the pulmonary region, R(t) results from the translocation of the chemical
across the alveolar membrane and into the bloodstream (with some contribution by engulfment
into macrophages). For the TB and NP regions, R(t) is controlled by the movement of the
mucociliary blanket towards the esophagus and, ultimately, into the gastrointestinal (GI) tract.
For the TB region, the retention can be influenced by the particular pattern of deposition
within the separate generations of that region. A further complication arises due to the
possibility of enhanced deposition at the bifurcations of airways (Martonen and Hoffmann,
1986), where movement of the mucociliary blanket will be slower. We will assume that a
chemical deposited in the TB region may be characterized by a single retention function,
however, since more refined characterizations are not feasible at present.
The amount of a chemical present in a region of the lung at time, t, is the organ burden,
B(t). For the case of an acute, i.e., instantaneous, uptake denoted by U0, the burden is
B(t) = U0 x R(t), (8)
C-8 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
where the unit of B(t) is grams. The quantities B, U, and R may be subscripted by chemical
and phase, and may be divided into the three separate regions of the lung.
For protracted exposures over a time interval of length T > 0, the calculation of B(T)
requires a convolution integral (Checkoway et al., 1989). The burden of a chemical in the lung
regions is described by
T VP(t)Rp(T-t)dt, (9)
B™(T)= C 0TB(t)RTB(T-t)dt, (10)
and
BNP(T) = VNP(f)RNp(T-f)dt. (11)
In these equations 0 is the rate of uptake into the separate lung regions. When concentration
C is constant over time, simplifications follow as demonstrated earlier. The unit of burden in
the lung region of interest is grams. Further subscripting to account for chemical and phase
continues to be omitted.
The rate at which damage is produced in a tissue at time t is assumed to be proportional
to the dose-rate, D(t), in the tissue (Checkoway et al., 1989), and the dose-rate is assumed to be
proportional to the burden, B(t). Under these assumptions,
= K x B(t),
where K is a proportionality constant that depends upon the particular chemical and tissue. In
essentially all cases, with the exception of radionuclides, it is not possible to specify a value of K
for chemicals because the important molecular damage is neither specified nor measured.
C-9 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Unfortunately, not knowing the value of K for different chemicals precludes combining burdens
from across chemicals to yield a single estimate of the dose-rate from ETS.
The total dose to an organ (or lung region) is the integral of the dose-rate over the time
interval of interest
D(T) = \T£>(t)dt. (13)
Substituting Equations 9, 10, or 11 into Equation 12, and substituting this into Equation 13,
yields
D(T) = K x Vi { CMfRft - r)drdt, (14)
Jo Jo
which for cases of constant concentration in air reduces to
D(T) = K x C x V x /J T^R(t - r)drdt. (15)
Again, all terms may be subscripted by chemical, phase, and lung region. The units of dose are
gram-days in the lung region of interest. When K is unknown (as is true for ETS), it is ignored
and the dose is replaced by the integral of the organ burden, IB.
Retention functions can take on a variety of forms that depend upon the physical
processes involved in removal from an organ. These functions, however, usually are
approximated by an exponential function or a sum of such functions
R(t) = e-*,
where \ is the removal rate constant (unit of time'1) from the organ or region of interest.
Substitution of this retention function into either of Equations 9 through 11 for constant
concentration C gives
C-10 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
D(T) - C x V x /f r e-*# = CxFx/xQ-e-^ (17)
Jo A
ignoring the subscripts for lung region. The dose may be obtained by substituting Equation 17
into Equation 12, and then inserting Equation 12 into Equation 13:
D(T) = {
= A:xCxKx/x(7-(l-
As before, the integral organ burden over T is obtained by ignoring K in Equation 18.
An additional complication arises in distinguishing between an organ burden (or dose) and
a biologically active organ burden. For some chemicals, e.g., nicotine, biotransformation may
occur in the body (see, e.g., Jacob et al., 1988; Hoffmann and Hecht, 1989; Hoffmann et al.,
1987; Hoffmann and Wynder, 1986). The chemical form of a substance may be altered,
producing a new molecule (such as cotinine) of either greater or lesser potential for harm. In
that case, the burden (or dose) of interest would be the one from the active form of the
chemical. This form may or may not be the form present in the environment. The biologically
active burden BB, will be equal to the burden of inhaled material, B, times a scaling factor, kA,
for activation of that material. When the rate constant for activation is small compared to the
rate constant for elimination of the active form, it may be necessary to perform specific
calculations of BB. In general, kA will be the fraction of the inhaled chemical biotransformed
into the active form. Similarly, DB will equal the biologically active dose and is obtained by
multiplying D by kA (or by calculating DB directly). For most chemicals (particularly those in
ETS), kA is unknown and BB or DB must be approximated by B or D as described earlier. An
exception is the conversion of nicotine to cotinine.
C-ll 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
C.2.2. Translocation to Systemic Organs
As shown in Figure C-2, the lung is coupled to the other organs of the body through the
bloodstream and the GI tract. Material deposited in the P region of the lung tends to be
translocated into the bloodstream, where it is carried to the systemic organs or excreted,
primarily in the urine. This process may be viewed as essentially catenary (Crawford-Brown,
1984), in which a chemical moves from the lung, into the blood, into an organ, and then into the
urine. Flow is assumed to be unidirectional, avoiding the complications introduced by
recirculation and exchange between organs.
Material deposited in the NP region is removed primarily to the G.I. tract, with little
absorption of ETS chemicals directly into the bloodstream (Wald et al., 1981). For deposition in
the TB region, there is evidence of high absorption of nicotine into the blood, at least in dogs
(Herrmann et al., 1989). From the G.I. tract, the chemicals are absorbed to a limited degree
into the bloodstream, where they are expected to behave in a manner similar to the material
entering the blood from the P region. Slight differences can occur due to the proximity of the
liver to the G.I. tract, but there is not sufficient information available to consider that further.
Chemicals unabsorbed by the GI tract will be excreted in the feces.
Let fpb equal the fraction of a chemical leaving the P region and entering the blood. This
fraction is determined with respect to the uptake and not the intake. Similarly, let fTB b and fNPb
be the fraction leaving the TB and NP regions, respectively, and entering the blood. The latter
fraction will be quite small and is ignored here. The fractions f^ GI and f^^ represent the
fraction of regional uptake entering the G.I. tract. The latter fraction is assumed to equal unity
here. The term fGI b equals the fraction of a chemical in the G.I. tract which crosses into the
bloodstream. The total uptake of a chemical entering the bloodstream is
(19)
C-12 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Ci
'NP
(TB,
NP
rNP,b
TB
t
P
^TB.b
G..
T
R
?
, 1
SYSTEMIC
ORGANS
t
EXCRETA
FIGURE C-2. A COMPARTMENTAL MODEL OF THE HUMAN BODY, DISPLAYING
ORGANS, TISSUES, FLUIDS, AND THEIR INTERCONNECTIONS
C-13
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The expressions for Up, Ura and UNP may be found in Equations 5 through 7.
Material in the blood is cleared into the systemic organs. The fraction of material moving
from the blood into a specific organ, indexed by j, will be denoted by fbj. Its value is assumed to
be independent of the route into the blood, although this assumption is untested at present.
The uptake into organ j is described by
Ul = U" x fbf (20)
where Ub is given by Equation 19.
The retention function for a substance in organ j is given by Rj(t). The dose to an organ
following an acute uptake to the organ then is
D> =#x W x (TR.(t]dt. (21)
This dose may be converted to a biologically active dose by multiplying Equation 21 by kA
(unknown at present). Since K also is unknown, D must be replaced by the integral organ
burden.
For protracted exposure to a chemical at a constant concentration, C in air, uptake to
organ j in the time interval [0,T] is
D>(T) = * x C x F x [f/?6 + fT/TBib + fN/NP,G/GI,]fbj x rf'/?/; - r)drdt. (22)
Implicit in the equation is the assumption that material clears rapidly from the lung and
into organ j, at least with respect to the interval of exposure. If it is assumed that the NP region
does not contribute significantly to systemic doses, due to low values of fGI b Equation 22 reduces
to:
fT/TBb] x fh. x J^R.(t - T)drdt. (23)
C-14 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
If it is assumed further that Rj(t) is a single exponential function with rate constant, X, then:
D>(T) -KxCxVx{f/ff+ Umb] x fbj x (T -(l-e-^/\)/\. (24)
Unfortunately, organ specific values for fbj are not known for the chemicals in ETS.
There are, however, limited data concerning the distribution of nicotine, a component of ETS,
within various tissues. The steady-state distributions of nicotine in those organs will be
approximately proportional to the dose-rate and, hence, the dose. The measurements, taken
from a report by the U.S. Surgeon General (U.S. DHHS, 1988) are displayed in Table C-l. It
will be noted that nicotine accumulates primarily in the kidney, followed by the liver, heart,
brain, muscle and adipose tissue. If the dose to the blood is calculated for nicotine, therefore,
the dose to other organs or tissues may be obtained by multiplying by the ratios in Table C-l. It
is unlikely, however, that the same ratios will apply to other chemicals in ETS.
C.2.3. Summary
A measure of exposure to a given chemical in ETS could take several forms:
1) Exposure Intensity, C, in units of g/m3.
2) Cumulative Exposure, 0, in units of g-days/m3 (see Equations 1 and 2).
3) Lung Intake, I, in units of grams of chemical (see Equation 3).
4) Total Lung Uptake, U, or Uptake to the P region, Up, TB region, UTB, or NP
region, UNP, in units of grams of chemical (see Equations 4-7).
5) Total Lung Burden, B, or Burden in the P region, Bp, TB region, BTB, or NP
region, BNP, in units of grams of chemical (see Equations 8-11).
6) Integral Organ Burden, IB, in units of gram-days of chemical.
7) Total Lung Dose, D, or Dose to the P region, Dp, TB region, DTB, or NP region,
DNP, in units of gram-days of chemical.
C-15 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-l. STEADY-STATE RATIO OF CONCENTRATIONS OF
NICOTINE IN BODY TISSUES OR ORGANS
Tissue
Blood
Brain
Heart
Muscle
Adipose
Kidneys
Liver
Lung
Gastrointestinal
Ratio*
1.0
3.0
3.7
2.0
0.5
21.6
3.7
2.0
3.5
* Relative to blood.
Source: U.S. Surgeon General (1986).
C-16
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
obtained from Intake or Uptake (above) or:
8) Blood Uptake, Ub, in units of grams of chemical (see Equation 19).
9) Organ Uptake, Uj, in units of grams (see Equation 20).
10) Integral Organ Burden, IB*, in units of gram-days of chemical (see Equations 21-24
divided by K).
11) Organ Dose, DJ, in unit of gram-days of the chemical in the organ (see Equations
21-24). This generally will not be possible to calculate.
12) Biologically Active Organ Dose, D^, in units of gram-days of the biologically active
form. This generally will not be possible to calculate.
The measure of exposure to a chemical depends upon the level of available information. In
cases where parameter values are unknown, it will not be possible to calculate values that
depend upon those parameters. In such cases, the measure further "upstream" in the chain of
calculations must be used. For example, the retention functions for many of the chemicals in
ETS are not available at present. For these chemicals, uptake is the most highly developed
measure of dose possible. For vapor phase chemicals, the deposition fractions or equilibrium
concentration ratios (tissue:air) have not been measured to date, leaving intake as the best
available measure of exposure.
C.3. ASSUMED EXPOSURE CONDITIONS AND INTAKES
Exposure to ETS may vary widely due to differences in cigarette type, rate of smoking,
ventilation conditions, room volume, etc. No attempt is made here to develop calculations
under the immense range of conditions likely to be found in society. Instead, calculations are
presented for a simplified case that is typical of exposure conditions. The exposure duration, T,
for both active and passive smokers is taken to be one day, with exposure to ETS at a constant
concentration, C (see the equations in Section 2 of this report). Of interest is the dose
C-17 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
delivered to tissues as a result of this single day of exposure. Predictions at any other
concentration, C*, and any other exposure length, T*, may be obtained by multiplying the
reported value by:
ClxTI
C T
For this example it is assumed that an active smoker smokes one pack (20 cigarettes) in a
structure of volume 150 m3 over a one day period with a passive smoker present. The NRC
(1986) reports an average of 26 mg of respirable suspended particulate (RSP) matter per
cigarette in the sidestream smoke (SS), giving an emission rate of 22 mg/hour. With an air
exchange rate of 1 per hour, an approximate U.S. average, in a room volume of 150 m3, the
concentration of RSP in the air will be approximately 200 /jg/m3 (see Figures 5-4 and 5-6 in
NRC, 1986). Assuming a tidal volume of 750 mL and a breathing frequency of 15 per minute
(Crawford-Brown, 1987), the total daily intake of RSP for the passive smoker will be
approximately 3 mg.
Rickert et al. (1984) measured the RSP in mainstream smoke (MS) and found a range of
0.7 to 17 mg per cigarette, for cigarettes primarily low in tar. By contrast, the NRC (1986)
report 15 to 40 mg for non-filter cigarettes. A moderate value of 12 mg will be assumed here.
The daily intake of RSP for an active smoker of 20 cig./day will be 240 mg. The values of 240
mg in the active smoker and 3 mg in the passive smoker correspond to values assumed by Wells
(1988).
The distributions of chemicals by mass in the MS from one nonfilter cigarette are
displayed in Table C-2 by vapor and particulate phase (NRC, 1986). In this table, a
representative value from the NRC report (1986) is reported. Also shown in the table are the
ratios of the amount of each chemical leaving the cigarette in diluted SS and in MS. The total
C-18
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-2. APPROXIMATE COMPOSITION OF MAINSTREAM (MS) AND
DILUTED SIDESTREAM SMOKE (SS) FROM ONE NON-FILTER CIGARETTE*
Constituent
Vapor Phase in MS
Carbon monoxide
Carbon dioxide
Carbonyl sulfide
Benzene
Toluene
Formaldehyde
Acrolein
Acetone
Pyridine
3-Methylpyridine
3-Vinylpyridine
Hydrogen cyanide
Hydrazine
Ammonia
Methylamine
Dimethylamine
Nitrogen oxides
N-nitrosodimethylamine
N-nitrosodiethylamine
N-nitrosopyrrolidine
Formic acid
Acetic acid
Methyl chloride
Particulate Phase in MS
RSP
Nicotine
Anatabine
Phenol
Catechol
Hydroquinone
Aniline
2-Toluidine
2-Naphthylamine
4-Aminobiphenyl
Benz(a)anthracene
Average MS
15 mg
30 mg
30 Mg
30 Mg
150 Mg
85 Mg
80 Mg
175 Mg
30 Mg
25 Mg
20 Mg
450 Mg
30 Mg
90 Mg
20 Mg
9Mg
400 Mg
30 ng
20 ng
20 ng
350 Mg
500 Mg
350 Mg
24 mg
14 mg
10 Mg
100 Mg
200 Mg
200 Mg
360 ng
160 ng
2ng
5 ng
50 ng
Average SS/MS
3
10
0.1
8
7
20
12
3
13
13
10
30
3
110
5
4
7
60
30
18
1.5
3
2.5
1.5
3
0.3
2.5
0.7
0.8
30
19
30
31
3
(continued on following page)
C-19
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-2. (continued)
Constituent
Benzo(a)pyrene
Cholesterol
A-Butyrolactone
Quinoline
Harman
N-nitrosonornicotine
NNK
N-nitrosodiethanolamine
Cadmium
Nickel
Zinc
Polonium-210
Benzoic acid
Lactic acid
Glycolic acid
Succinic acid
Average MS
30 ng
20 pg
15 pg
1 pg
2 pg
1500 ng
500 ng
50 ng
100 ng
50 ng
60 ng
0.1 pQ
20 pg
100 pg
100 pg
120 pg
Average SS/MS
3
0.9
4
10
1
2
3
1
7
20
7
3
0.8
0.6
0.8
0.5
* Adapted from Table 2-2 of NRC (1986).
RSP inhaled in one day by the passive smoker in our example is proportional to its
concentration in air of 200 pg/m3. As noted above, using that value and Table C-2, the
concentration of other chemicals inhaled from SS can be obtained from the relation
C, = 200 x
M..
M
RSP
R,
R
RSP
(25)
where Q is the concentration of chemical i in the room air (pg/m3), Mj is the average mass of
chemical i in MS and MRSP is the mass of RSP in MS (both taken from Table C-2), R( is the
ratio SS/MS for chemical i from Table C-2 and RRSP is the ratio SS/MS for RSP. Computed
values of C, and the intakes for the passive smoker and the active smoker are shown in Table
C-3 for known carcinogens in tobacco smoke.
C-20 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The calculations in Table C-3 presume fresh diluted SS, which ages over time. This aging
may change the physicochemical properties of the ETS due to plating, ventilation, metabolism,
etc. (NRC, 1986). Little is known, however, about the effect of aging. For reference, the
average concentration of several airborne components of ETS measured under diverse
environmental conditions are shown in Table C-4 (adapted from Repace, 1987).
Since the results in Table C-4 were obtained under such a wide range of conditions,
absolute concentrations of chemicals in aged air are difficult to specify for the environmental
conditions assumed in this report. Still, a limited comparison of relative values can be made by
focusing on nicotine, benzene, N-nitrosodimethylamine and N-nitrosodiethylamine, since these
measuresurements were made under roughly similar conditions in a room with a large number
of smokers (Badre et al., 1978; Brunnemann et al, 1978; Stehlik et al., 1982). From these
measurements, the relative concentrations of nicotine: benzene: N-nitrosodimethylamine:
N-nitrosodimethylamine are 1:0.2:0.0002:0.0001. These values may be compared against the
predictions using fresh diluted SS in Table C-3, which suggest values of 1:0.06:0.00004:0.0001.
The relative concentrations of these four chemicals, therefore, do not appear to have been
significantly affected by aging, with the possible exception of benzene. The reason for the
increase in benzene after aging is unknown. The benzene estimate is based on a single small
sample, which may be a factor. Subsequent calculations will use the concentrations and intake
values in Table C-3 based on fresh diluted SS. There is need for more research on the effects
of aging for ETS.
C-21 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-3. SUMMARY OF CONCENTRATIONS AND DAILY INTAKES FOR
CONSTITUENTS OF CIGARETTE SMOKE, ASSUMING FRESH SS
Constituent*
Benzene
Hydrazine
N-nitrosodimethylamine
N-nitrosodiethylamine
N-nitrosopyrrolidine
RSPa
Nicotine*5
2-Naphthylaminea
4-Aminobiphenyla
Benz(a)anthracenea
Benzo(a)pyrenea
A-Butyrolactonea
N-nitrosonornicotinea
N-nitrosodiethanolaminea
Nickel8
Polonium-210a
cr
1.3 pg/m3
0.5 ng/m3
30.0 ng/m3
3.0 ng/m3
2.0 ng/m3
200.0 AJg/m3
23.0 /jg/m3
0.3 ng/m3
0.9 ng/m3
0.8 ng/m3
0.5 ng/m3
0.3 pg/m3
17.0 ng/m3
0.3 ng/m3
6.0 ng/m3
2.0 nCj/m3
U
21 ^g
8ng
160 ng
48 ng
32 ng
3 mg
370 pg
5 ng
14 ng
13 ng
8ng
5 ng
270 ng
5ng
96 ng
32 nQ
Ttt
A?
300 pg
300 ng
300 ng
200 ng
200 ng
240 mg
14 mg
20 ng
50 ng
500 ng
300 ng
150 Pg
15 pg
500 ng
500 ng
ipc,
* Only constituents listed as human carcinogens, suspected human carcinogens or animal
carcinogens (NRC, 1986) are listed, with the exception of nicotine (a precursor to
carcinogens).
** For passive exposures only.
t Intake for passive exposure.
^ Intake for active exposure.
a Chemicals located in the particulate phase for both active and passive smokers.
b Nicotine is assumed to be entirely in the particulate phase for active smokers and entirely in
the vapor phase for passive smokers.
C-22 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-4. MEASUREMENTS OF ETS CONSTITUENTS
IN ENVIRONMENTAL SETTINGS*
Constituent Setting
Acrolein Varied
Varied
Benzene Varied
Toluene Varied
Benzo(a)pyrene Arena
Restaurant
Coffeehouses
Restaurant
Public places
Carbon monoxide Varied
Varied
Rooms varied
Taverns
Planes
Arenas
Restaurant
Restaurant
Varied
Average
concentration
0.1 mg/m3
8.0 ppb
0.1 mg/m3
1.0 mg/m3
10.0 ng/m3
6.0 ng/m3
5.0 ng/m3
10.0 ng/m3
100.0 ng/m3
20.0 ppm
8.0 ppm
5.0 ppm
12.0 ppm
3.0 ppm
15.0 ppm
3.0 ppm
5.0 ppm
10.0 ppm
Reference
Badre et al., 1978
Fischer et al., 1978, and
Weber et al., 1979
Badr<§ et al., 1978
Badre et al., 1978
Elliot and Rowe, 1975
Galuskinova, 1964
Just et al., 1972
Husgafvel-Pursiainen
et al., 1986
Perry, 1973
Badre et al., 1978
Chappell and Parker, 1977
Coburn et al., 1965
Cuddleback et al., 1976
USDOT, 1971
Elliott and Rowe, 1975
Weber et al., 1979
Fischer et al., 1978
Godin et al., 1972
(continued on following page)
C-23
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-4. (continued)
Constituent Setting
Office
Car
Train
Public places
Rooms
Varied
Conference
rooms
Offices
Nicotine Varied
Submarines
Train
Varied
Restaurants
N-nitrosodi- Varied
methylamine
Restaurants
Average
concentration
5.0 ppm
40.0 ppm (peak)
20.0 ppm
8.0 ppm
15.0 ppm
10.0 ppm
8.0 ppm (peak)
3.0 ppm
100.0 pg/m3
30.0 A/g/m3
2.0 ^g/m3
6.0 /.g/m3
15.0 ug/m3
100.0 ng/m3
25.0 ng/m3
Reference
Harke, 1974
Harke and Peters, 1974
Harmsen & Effenberger,
1957
Perry, 1973
Portheine, 1971
Stebben et al., 1977
Slavin and Hertz, 1975
Szadkowski et al., 1976
Badre et al., 1978
Cano et al., 1970
Harmsen & Effenberger,
1957
Hinds & First, 1975
Muramastu et al., 1984
Brunnemann & Hoffmann,
1978, and Brunnemann
et al., 1978
Stehlik et al., 1982
(continued on following page)
C-24
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-4. (continued)
Constituent
Setting
Average
concentration
Reference
RSP
Acetone
Sulfates
Varied 200.0 ^g/m3
Varied 300.0 /ug/rn3
Coffeehouses 1000.0 pg/m3
Hospital 30.0 /jg/m3
Residences 60.0
Offices 130.0
Offices 50.0
Restaurants 1000.0
Houses
Tavern
Residences
Arenas
Varied
Residences
150.0 /jg/m3
600.0
30.0
400.0
1.0 mg/m3
5.0
Repace and Lowrey, 1980
Repace and Lowrey, 1982
Just et al., 1972
Neal et al., 1978
Spengler et al., 1981
Weber and Fischer, 1980
Nelson et al., 1982
Husgafvel-Pursiainen
et al., 1986
Brunekreef and Beleij, 1982
Cuddleback et al, 1976
Dockery and Spengler, 1981
Elliott and Rowe, 1975
Badr6 et al., 1978
Dockery and Spengler, 1981
"Adapted from Repace (1987).
C-25
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Application of the dosimetry model of ETS beyond the initial step of predicting inhalation
becomes hampered by the large amount of detailed biological information required for the
carcinogens in tobacco smoke. This Limitation is more applicable to the vapor phase, however,
than to the particulate phase. In the latter case, many of the dosimetric characteristics are
largely dependent on the distribution of the size and density of particulates rather than
chemical-specific properties. To continue illustration of our example as possible, and also to
identify where information is available and where it is needed, we will consider the particulate
phase further but not the vapor phase. The information available for calculation by lung regions
is disparate, so assumptions will be made explicit as required to complete calculations for lung
dose from the particulate phase for our example. We have stopped short of introducing
assumptions that do not seem "reasonable," however, simply for the sake of illustration. The
calculated values may be viewed as approximations, vis-a-vis the assumptions used. In any
event, intake of vapor phase components is included in Table C-3.
Fortunately, one of the major constituents of interest in tobacco smoke, nicotine, has been
sufficiently studied that much of the information required for prediction of lung and systemic
organ dose of nicotine and the metabolite cotinine can be calculated for our example, including
both the vapor and particulate phases for active and passive smoking. Nicotine dosimetry is
particularly relevant because it is the addictive factor in active smoking and is a pre-cursor of
tobacco specific nitrosamines, at least one of which (NNK) is a potent carcinogen (Hoffmann
and Hect, 1989). Also, nicotine forms the tobacco-specific metabolite cotinine, widely
considered to be the preferred biomarker for ETS exposure. Calculations for nicotine/cotinine
are in Section C.6.
C-26 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
C.4. UPTAKE OF PARTICULATE PHASE CHEMICALS.
Due to lack of data on uptake of vapor phase components by lung tissue, only the uptake
of particulate phase chemicals can be considered here. For chemicals in the vapor phase, the
proportionality constant to convert from intake to uptake may differ between chemicals. The
lack of chemical-specific data on uptake from the vapor phase is a major limitation for
comparison of carcinogenicity of tobacco smoke to active and passive smokers. There is a
pressing need for research on concentration ratios (airtissue) for vapor phase components of
ETS.
In calculating uptake of particulate phase chemicals, it is necessary to specify regional
deposition fractions (Equations 5 through 7). A primary environmental determinant of these
fractions is particle diameter. The mean diameter for MS has been reported to range from 0.1/j
to 1/j (Carter and Hasegawa, 1975; Killer et al., 1982), and from 0.01/^ to 0.8/j for SS. For the
calculations reported here, a Mass Median Aerodynamic Diameter or MMAD of 0.7/j is used
for MS (Stober, 1984) and a MMAD of 0.4p is assumed for fresh diluted SS (Wells, 1988). The
particle diameters are assumed to be distributed lognormally with a geometric standard
deviation (GSD) of 1.5 (Stober, 1984).
Aged air, however, may contain a different distribution of aerosol sizes. Several authors
(Keith and Derrick, 1960; Wynder and Hoffman, 1967; Ingebrethsen and Sears, 1985) have
demonstrated that the MMAD for cigarette smoke decreases by a factor of 2 to 3 due to aging.
This appears to be due to the loss of large particles from the suspended aerosol, as may be seen
in the measured and predicted distributions published by Nazaroff and Cass (1989). An
additional factor may be the "boiling off of chemicals from the RSP. The present report,
therefore, assumes that the MMAD for aged ETS is on the order of 0.15p, and that for direct
smoking is 0.1/u.
C-27 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Total deposition of smoke in the lung has been reviewed by Stober (1984), based primarily
on studies by Killer et al. (1982), Mitchell (1962), and Polydorova (1961). These results suggest
that as much as 80% of the MS particulates are deposited in the lung (i.e., fNP + f^ 4- fp = 0.8),
while 10% to 20% of the ETS particulates are deposited. The value for ETS is consistent with
the predictions of total lung deposition from an age-dependent model by Crawford (1982, 1983),
which yields values of 1%, 4%, and 10% for f^,, fpg, and fp, respectively, for a MMAD of 0.15p.
The very high value of total deposition in active smokers appears to arise from several
factors. The first is hygroscopic growth, which may be expected to double the size of particulate
MS from 0.7/j to 1.4^ (Ishizu et al., 1980). The second factor is breath-holding, in which
cigarette smoke is held in the lungs for several seconds prior to exhalation. If the model of
Crawford (1982, 1983) is used with a breath-holding period of 3 seconds, particle diameters of
lAfj are predicted to yield values of 1%, 15%, and 60% for f^, f^, and fp, respectively. Since
these sum to approximately the 80 percent reported in experiments, these values will be
assumed here. Hygroscopic growth of ETS particles will not be assumed, since the inhaled and
exhaled particles appear to be of the same diameter (Killer et al., 1982).
As described previously, the total intake of RSP is assumed to be 240 mg in an active
smoker and 3 mg in a passive smoker in our example. Using these values in conjunction with
the estimates of f^, f^, and fp from Equations 5 through 7, the calculated daily uptakes in mg
by lung region are 12 mg, 36 mg and 144 mg for the NP, TB and P regions of the active smoker;
0.03 mg, 0.12 mg and 0.3 mg for the NP, TB and P regions of the passive smoker.
C.5. INTEGRAL ORGAN BURDENS FOR THE LUNG
C.5.1. Integral Organ Burden from RSP
Due to the very low assumed deposition fractions in the NP region, the focus of this
discussion will be on the TB and P regions.
C-28 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Translocation of the particles is on the mucus layer, which is driven forward towards the
esophagus by the cilia. The velocity of this mucus blanket decreases dramatically in the deeper
sections of the TB region. As a result, the length of time a particle resides in the lung depends
critically on the site of deposition. Data on the removal of radiolabeled particles, however, show
that removal from the TB region generally may be characterized by two phases. The first is a
rapidly cleared phase, dominated by particles deposited on the mucus of the upper passageways.
The second is dominated by particles deposited on the slowly moving mucus of distal
passageways. Both phases are controlled primarily by the movement of the mucus and the site
of deposition of the particles rather than on the chemical nature of the particles. It is possible,
therefore, to use the results of the radiolabeled aerosol studies to estimate the retention of
particulate ETS in the TB region.
Crawford and Eckerman (1983) have used the deposition model of Crawford (1982) and a
model of mucus movement, in conjunction with measurements of retention of radiolabeled
aerosol particles in healthy (non-smoking) human lungs, to develop predictive equations of
retention. These retention functions contain two exponentials, corresponding to the two removal
phases described above. The parameters in these equations depend upon the aerosol diameter
and breathing characteristics. The general form of the equation is:
where t is the time since uptake into the TB region (in minutes). The parameter b is a function
of median aerosol diameter, age, the GSD of the particle distribution and breathing
characteristics. This parameter value equals the fraction of deposited particles found in the
slowly removed component. The parameters C, and C2 are the removal half-times for the rapid
and slow components, respectively.
C-29 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
: As described ear ilex, the MMAD far ETS particles is 0.15^, with a OSD of 1.5. Applying
the results of Crawford and Eckerraan (1983) to ETS, yields values for b, Ch and C, of 0.98,
450. and'710.minutes, respectively, For a MMAD of. 1.4y and a GSD of 1,5, the values of b, C,,'
and ;Q are 0.82, 280, and 700 minutes, respectively. The flow of mucus in the TB region of
active smokers, however, is-reduced by a factor of 2 (Albert et a!., 1975; Wanner ei al, 1973). -(f
the parameter-values for !.4,u in normal (non-smoking) -lungs are changed to reflect the'
condition of slowed mucus, the half-times in the retention function would be doubled.' For
active smokers, therefore, Cj and C2 would he 560 minutes'.and 1400 minutes, respectively. ..The
value of b should be unchanged. •
The daily integral organ harden to the TB region from RSP may be obtained from
Equation 1.5 by setting K equal to unity, f equal to frB. C x V equal to the daily intake of RSP,
and T equal to 5440 minutes-(24 hoars). Using these values, the daily RSP integral organ
burden to the TB region is 64,873 and 122 mg-minutes for !he active and passive smokers in our
example, respectively,
, , Solubtiizatiori and engulfrnent by macrophages generally dominate removal from the P
region, of the lung. Unfortunately, few data are available on the removal of RSP from the deep
lung, it is known, however that the constituent chemical nicotine deposited in active-smokers is
highly soluble in lung fluid (Janoff et al., 1987). Black and Pritchard (1984') have found an
alveolar retention half-time of .17 hours for RSP in active smokers, which will be used in our
example, Similar measurements in passive srnokws are: not available.. At this time, we use the
same half-time for both passive and active smokers. Additional research is needed to accurately
quantify removal of RSP in passive smokers. As a first approximation, a half-time of 17 hours
for RSP removal will he assumed for bolh active and passive smokers (Wells, 1988). Tiie
rete.rstic-u function for the P region is
R(t) = e-«.«3./i^
C-30 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
where t is in minutes since deposition. From Equation 15 for fp and the daily intakes of RSP
described earlier, the daily integral organ burden to the P region is 1.9 x 105 mg-minutes in
active smokers and 390 mg-minutes in passive smokers.
C.5.2 Lung Integral Organ Burden from Particulate Chemicals
Chemicals should solubilize from the RSP at different rates, thereby affecting dose rate to
the lung. Data to differentiate between chemical dose rates, however, are not available. The
retention half-times used earlier will, therefore, be used here for other chemicals contained in
particles. The ratio of integral organ burdens from chemical components, relative to RSP
values, may be obtained from the ratio of intakes of those chemicals in the particulate phase
shown in Table C-3. Daily integral organ burden to the lung by chemicals in the particulate
phase have been calculated for the active and passive smoker of our example and are displayed
in Table C-5.
C.6. CALCULATIONS FOR NICOTINE AND COTININE
The intake of nicotine by the active and passive smoker in the example described come
from the particulate and vapor phases, respectively (Eudy, 1986). Leaderer (1988) gives the
percentage of nicotine in the vapor phase of ETS as 95 + , while Pritchard (1990) gives it as 70%.
C-31 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-5. DAILY INTEGRAL ORGAN BURDENS* FOR PARTICULATE
PHASE CHEMICALS AS CALCULATED IN THIS REPORT
All Doses Are in mg-Minutes
Constituent
RSP
Nicotine
2-Naphthylamine
4-Aminobiphenyl
Benz(a)anthracene
Benzo(a)pyrene
X-Butyrolactone
N-nitrosonornicotine
N-nitrosodiethanolamine
Nickel
IB™
122
0
2x
5.6
5.2
3.2
2 x
1.1
2x
3.8
10"4
x lO"4
x 1Q-4
x IO-4
10-*
x ID'2
10-
x 10'3
mP TTlTB TT>P
p •"'A '-"A
390
0
6
6.5 x IO4
3.9 x IO3
2 x
1.7 x
1
9
5 x
3x
5.8 x
3
5
1
1 x
8x
1 x
io-3
io-3
ID'3
io-4
10-*
io-2
10-*
io-2
5
1
0
0
0
0
.4 x ID'3
.4 x ID'2
.14
.08
.04
.004
0.14
0
.14
1.9 x 10s
1.1 x IO4
0.016
0.04
0.4
0.24
0
0.
12
012
0.4
0.
4
Superscript on IB indicates lung region (TB for tracheobronchial and P for pulmonary), and
the subscript indicates passive (P) or active (A) smoker.
C-32 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Our calculations, which presume that nicotine is entirely in the particulate phase for active
smokers and entirely in the vapor phase for passive smokers, have been given in Table C-3.
Nicotine concentrations have been measured in body fluids of active and passive smokers
following known experimental exposure, providing information that can be applied to dosimetry
calculations. Jarvis et al. (1984) have reported the results of nicotine measurements, which are
summarized in Table 8-3 of the NRC report (1986). The nicotine/cotinine in the body fluids of
passive smokers tends to be about 1% of the levels in active smokers.
As measured by Jarvis et al. (1988), Sepkovic et al. (1986), Kyerematen et al. (1982), and
Benowitz et al. (1983), the clearance half-time for cotinine from the systemic body organs is on
the order of 15 hours in active smokers. Sepkovic et al. (1986) suggest that this value for
passive smokers is close to 45 hours. Their published data, however, indicate a half-time closer
to 25 hours, a conclusion agreed upon by Jarvis et al. (1988). The half-time of 25 hours is more
consistent with the measurements of urine excretion (Jarvis et al., 1988), where the half-time in
passive smokers was 33 hours and in active smokers was 22 hours. This suggests that the
passive-to-active ratio of half-times for removal of cotinine is about 1.5. For this ratio of
excretion half-times, the data on body fluids (Jarvis et al., 1984) suggest that passive smokers
take nicotine into their blood at a rate of 0.01/1.5 times the rate in active smokers, i.e., about
0.7% of the rate in active smokers. If it is assumed further that nicotine requires the same
length of time to traverse the alveolar cells in passive and active smokers, a topic on which no
data are available, then the ratio (active :passive) of integral organ burden for blood will equal
approximately the ratio of rates into the bloodstream. This conclusion requires the assumption
that the GI tract does not contribute significantly to the nicotine in the bloodstream. Data on
nicotine absorption are not available at present. Since the nicotine dose to the P region is
11,000 mg-minutes for active smokers (see Table C-5), the integral organ burden to the lungs of
passive smokers from vapor phase nicotine to this region will be approximately 80 mg-minutes
C-33 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
(or 11,400 x 0.007). For the TB region, the integral organ burden to the lungs will be 3900 x
0.007 or 27 mg-minutes. Neither the TB nor P region integral organ burdens are affected if the
fraction of nicotine in the vapor phase is set equal to 70% instead of 100%. This invariance is
due to the reliance on measured blood uptake for the calculation of integral organ burdens.
There is little available information on the uptake of chemicals to the bloodstream from
which to calculate systemic organ doses. Again, nicotine/cotinine is an exception since
measurements of body fluid concentrations and clearance half-times are available. As described
above, nicotine is highly soluble in lung tissue, which implies that fpb and fTBb in Equation 19 are
approximately 1. Thus, the uptake of nicotine to the bloodstream then equals the uptake into
the P and TB regions of the lung. Once in the bloodstream, nicotine is converted to cotinine, as
described by Jacob et al. (1988). This metabolic model is shown in Figure C-3, from which it
may be seen that 70% of the nicotine is converted to cotinine, 9% goes directly to the urine, and
4% is metabolized to nicotine n-oxide. The remaining 17% is unaccounted for at present.
Nicotine is removed from the blood with a half-time of 2 hours in smokers (Jacob et al.,
1988; Benowitz et al., 1982). As described above, the ratio of removal half-times for cotinine in
passive and active smokers is 1.5. If this same ratio applies to the conversion of nicotine,
passive smokers would display a removal half-time of 3 hours. If this ratio does not apply, both
groups would possess a half-time of 2 hours. The retention function for nicotine then is either
R(t) = e-°-693t/2
or
R(t) = e-°-693t/3,
depending on whether the removal half-time is 2 or 3 hours, respectively; t is in hours. Applying
Equation 23, the daily integral organ burden from nicotine to the systemic organs of active
C-34 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
17%
4%
NICOTINE-
N-OXIDE
UNKNOWN
60%
OTHER
METABOLITES
FIGURE C-3. A METABOLIC MODEL FOR THE CONVERSION OF NICOTINE
AND THE EXCRETION
C-35
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
smokers is 2000 mg-minutes. The daily dose from nicotine to the systemic organs of passive
smokers, using 0.7% of the value in active smokers, as discussed above, is approximately 16
mg-minutes if the half-time is 2 hours, and 27 mg-minutes if the half-time is 3 hours. Since the
data of Kyerematen et al. (1982) and Lee et al. (1987) suggest that the rate of metabolism of
nicotine is higher in active smokers, the latter value of 27 mg-minutes appears to be the best
estimate.
The calculation of systemic organ doses from cotinine is more complicated than Equation
15, since cotinine is a metabolic product. The burden of nicotine in the blood at any time, t,
after an uptake Un is
Bn(t) = U.e'"^,
where T is the removal half-time for nicotine (either 2 or 3 hours). The differential equation
describing the rate of change of the burden of cotinine, Bc(t), in blood then is:
= 0.693Bn(t)/Tc - Q.693Bc(t)/Te, (26)
where Tc is the conversion half-time from nicotine to cotinine and Te is the elimination half-time
for cotinine from the body (15 hours in smokers and 25 hours in passive smokers). The burden
of cotinine is obtained by solving Equation 26 to yield
A U f>~A' - f^
B(i) - (27)
where
\c = 0.693/7,
\e = 0.693/7c
A = 0.693/7
C-36 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
and Un is 10.6 mg in the active smoker and 0.08 mg in the passive smoker. The dose from
cotinine then is obtained from Equations 12 and 13. The value of Tc is equal to T/0.7, where
0.7 is the fraction of nicotine converted to cotinine. The daily integral organ burden to the
systemic organs of the active smoker then is 7660 mg-minutes. The daily integral organ burden
to the passive smoker is approximately 145 mg-minutes if T is three hours and 140 mg-minutes
if T is two hours. The measures calculated for nicotine and cotinine, and their ratios in MS to
ETS, are included in Table C-6 and C-7.
C-37 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-6. SUMMARY OF DOSE MEASURES CALCULATED IN THIS REPORT
(FOR OTHER PARTICULATE PHASE DOSES, SEE TABLE C-5)
Constituent
Measure
Value
1. Total RSP
Intake
Uptake
a. NP region
P* = 3 mg
A* = 240 mg
P = 0.03 mg
A = 12 mg
2. Nicotine
(particulate)
b. TB region
c. P region
Integral Organ Burden
a. NP region
b. TB region
c. P region
Intake
Uptake
a. NP region
b. TB region
c. P region
P = 0.12 mg
A = 36 mg
P = 0.3 mg
A = 144 mg
P = na**
P = 122 mg-min.
A = 64,873 mg-min.
P = 390 mg-min.
A = 1.9 x 105 mg-min.
P = 0 mg
A = 14 mg
P = 0 mg
A = 0.72 mg
P = 0 mg
A = 2.2 mg
P = 0 mg
A = 8.6 mg
(continued on following page)
C-38
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-6. (continued)
Constituent
Measure
Value
3. Cotinine
(paniculate)
4. Nicotine (vapor)
Integral Organ Burden
a. NP region
b. TB region
c. P region
d. Systemic organs
Integral Organ Burden to
systemic organs
Intake
Uptake
a. NP region
b. TB region
c. P region
Integral Organ Burden
a. NP region
b. TB region
c. P region
P = Omg
A = na
P = Omg
A = 3800 mg-min.
P = 0 mg
A = 11,000 mg-min.
P = 0
A = 2000 mg-min.
P = 0
A = 7600 mg-min.
P = 0.37 mg
A = 0 mg
P = na
A = 0 mg
P = na
A = 0 mg
P = 0.06 mg
A = 0 mg
P = na
A = 0 mg
P = 27 mg-min.
A = 0 mg-min.
P = 80 mg-min.
A = 0 mg-min.
(continued on following page)
C-39
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-6. (continued)
Constituent
Measure
Value
5. Cotinine (vapor)
6. Nicotine (total)
d. Systemic organs
i. Passive smoking
Conversion Tw = 2 hrs.
ii. Passive smoking
Conversion Tw = 3 hrs.
Integral Organ Burden to
systemic organs
i. Passive smoking
Conversion Tw = 2 hrs.
ii. Passive smoking
Conversion Tw = 3 hrs.
Intake
Uptake
a. NP region
b. TB region
c. P region
Integral Organ Burden
a. NP region
b. TB region
c. P region
P = 16 mg-min.
A = 0 mg-min.
P = 27 mg-min.
A = 0 mg-min.
P = 140 mg-min.
A = 0 mg-min.
P = 145 mg-min.
A = 0 mg-min.
P = 0.37 mg
A = 14 mg
P = na
A = 0.72 mg
P = na
A = 2.2 mg
P = 0.06 mg
A = 8.6 mg
P = na
A = na
P = 27 mg-min.
A = 3800 mg-min.
P = 80 mg-min.
A = 11,000 mg-min.
(continued on following page)
C-40
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-6. (continued)
Constituent
Measure
Value
7. Cotinine (total)
8. Benzene
9. Hydrazine
10. N-nitroso-
dimethylamine
d. Systemic organs
i. Passive smoking
Conversion T^ = 2 hrs.
ii. Passive smoking
Conversion T^ = 3 hrs.
Integral Organ Burden to
systemic organs
i. Passive smoking
Conversion T^ = 2 hrs.
ii. Passive smoking
Conversion ^ = 3 hrs.
Intake
i. Using fresh SS
ii. Using aged SS
Intake
i. Using fresh SS
ii. Using aged SS
Intake
i. Using fresh SS
ii. Using aged SS
P = 16 mg-min.
A = 2000 mg-min.
P = 27 mg-min.
A = 2000 mg-min.
P = 140 mg-min.
A = 7660 mg-min.
P = 145 mg-min.
A = 7660 mg-min.
P = 21 fig
A = 300 p
P = 65 ^
A = 300
P = 8ng
A = 300 ng
P = 8 ng***
A = 300 ng
P = 160 ng
A = 300 ng
P = 80 ng**
A = 300 ng
(continued on following page)
C-41
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-6. (continued)
Constituent
Measure
Value
11. N-nitroso-
diethylamine
12. N-nitroso-
pyrrolidine
Intake
i. Using fresh SS
ii. Using aged SS
Intake
i. Using fresh SS
ii. Using aged SS
P = 48 ng
A = 200 ng
P = 48 ng***
A = 200 ng
P = 32 ng
A = 200 ng
P = 32 ng***
A = 200 ng
* P = Passive, A = Active.
** na = not available for calculation due to insufficient information.
*** Obtained by multiplying the passive smoking intake for fresh SS by R, where R is the ratio
of the concentration of the chemical relative to nicotine in aged SS to the concentration of
the chemical relative to nicotine in fresh SS.
C-42
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-7. SUMMARY OF RATIO OF MEASURES (ETS/MS)
CALCULATED IN THIS REPORT
Measure of Dose
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Intake of Total RSP
Uptake of Total RSP
a. NP region
b. TB region
c. P region
Integral Organ Burden of Total RSP
a. NP region
b. TB region
c. P region
Intake of Particulate Nicotine
Uptake of Particulate Nicotine
a. NP region
b. TB region
c. P region
Integral Organ Burden of Particulate Nicotine
a. NP region
b. TB region
c. P region
d. Systemic organs
Integral Organ Burden of Cotinine
(from particulate nicotine)
a. Systemic organs
Intake of Vapor Nicotine
Uptake of Vapor Nicotine
Integral Organ Burden of Vapor Nicotine
a. NP region
b. TB region
c. P region
d. Systemic organs
Ratio
0.013
0.0025
0.003
0.002
na*
0.002
0.002
0
0
0
0
0
0
0
0
0
Very large*
Very large
Very large
Very large
Very large
Very large
(continued on following page)
C-43
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE C-7. (continued)
Measure of Dose
Ratio
11. Integral Organ Burden of Cotinine
(from vapor nicotine)
a. Systemic organs
12. Total Intake of Nicotine
13. Total Integral Organ Burden of Nicotine
a. NP region
b. TB region
c. P region
d. Systemic organs
i. Passive smoking
Conversion Tw = 2 hours
ii. Passive smoking
Conversion Tw = 3 hours
14. Total Integral Organ Burden of Cotinine
(Systemic organs)
i. Passive smoking
Conversion Tw = 2 hours
ii. Passive smoking
Conversion Tw = 3 hours
15. Intake of Benzene
a. Using fresh SS
b. Using aged SS
16. Intake of Hydrazine
a. Using fresh SS
b. Using aged SS
17. Intake of N-nitrosodimethylamine
a. Using fresh SS
b. Using aged SS
18. Intake of N-nitrosodiethylamine
a. Using fresh SS
b. Using aged SS
19. Intake of N-nitrosopyrrolidine
a. Using fresh SS
b. Using aged SS
Very large
0.03
na
0.01
0.01
0.01
0.015
0.02
0.02
0.07
0.2
0.03
0.03
0.5
0.3
0.2
0.2
0.15
0.15
*na = not applicable due to lack of data.
large occurs when the value for active smokers is zero.
C-44
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
APPENDIX D
ALTERNATIVE APPROACHES FOR ESTIMATING THE YEARLY NUMBER OF
LUNG CANCER DEATHS IN NON-SMOKERS DUE TO ETS
BASED ON DOSE-RESPONSE MODELING
D.I. INTRODUCTION
In Chapter 4 the annual number of lung cancer deaths attributable to ETS was estimated
from epidemiological case-control and cohort studies. This appendix investigates alternative
methods based on dose-response modeling techniques.
In order to use dose-response modeling approaches to directly estimate the number of
lung cancer deaths in nonsmokers attributable to ETS, three elements are required:
1. the distribution of the time-weighted exposure of ETS in the nonsmoking population,
2. the age distribution of the nonsmoking population, and
3. a mathematical dose-response model describing the relationship between the age-
specific lung cancer rate and the independent variables age, sex, race, and ETS
exposure.
The U.S. EPA has already collected sufficient information so that elements 1 and 2 can be
approximated with reasonable accuracy in a straightforward manner. A discussion of potential
methods for the derivation of the dose-response model, element (3), is the subject of this
appendix.
Three independent approaches are identified for estimating the dose-response relationship
between age-specific lung cancer death rates and ETS. Each of these methods has its
advantages and disadvantages in estimating ETS cancer risk. Presently, none of them is
developed in full detail. The purpose of presenting these preliminary approaches is to invite
comment on their relative merit, solicit advice on other potential approaches that might be
investigated, and to help prioritize further research efforts in this area. Much of the material
D-l 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
considered here is based on ongoing research that is not fully documented at this time, and it is
presented as an illustration of the type of approach that could be taken.
The three proposed general approaches for deriving ETS dose-response models are:
1. Establish a dose-equivalent relationship between ETS and a positive control such as
inhaled benzo[a]pyrene (B[a]P) which has an animal-based inhalation cancer dose-
response model associated with it. Heavy use would be made of animal carcinogen test
results in this approach. This approach will be subsequently referred to as the Relative
Potency Approach (RPA).
2. Establish an equivalency relationship between the number of cigarettes smoked per day
and ETS exposure levels in mg/m3 inhaled air. This relationship would then be used to
estimate risk based on a direct state-of-the-art cigarette smoking dose-response model
obtained from multiple sources of epidemiological data. This will be referred to as the
Cigarette-equivalent Approach (CEA).
3. Use ETS epidemiological studies where a dose-dependent increase in the risk of
nonsmoking women is associated with ETS. This will be referred to as the Direct
Approach (DA).
Details concerning these approaches, examples of information that may be used in their conduct,
and an evaluation of their strengths and weaknesses are presented in the following sections.
D.2. RELATIVE POTENCY APPROACH
D.2.1. Overview
The products of incomplete combustion from hydrocarbons (e.g., tobacco products)
contain very complex mixtures of agents including thousands of polycyclic aromatic hydrocarbons
(PAHs), many of which are known or suspected to be carcinogenic. The direct evidence for the
carcinogenicity of hydrocarbon combustion products comes mainly from three types of
information:
1. animal carcinogenicity tests of pure PAHs such as benzofajpyrene (B[a]P), etc., that
are known to be formed as part of the combustion products of hydrocarbons,
2. animal carcinogenicity tests of condensates and various fractions of the condensates
from hydrocarbon combustion products (e.g., coal flue gas, gasoline engine exhaust,
diesel engine exhaust, coke oven emissions, etc.), and
D-2 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
3. human epidemiological studies (e.g., cigarette smoking, roofing tar fumes, coke oven
emissions, etc.).
The EPA has historically used this type of information to help establish air quality criteria
for emissions of complex mixtures of PAHs. In one situation a criterion for coke oven
emissions was set based directly on epidemiological evidence of a dose-dependent increase in
lung cancer (U.S. EPA, 1984). This evidence was gained from a long-term follow-up of black
male workers who were working in close proximity to the coking operations in steel mills
(Redmond et al., 1972). However, in most situations, direct evidence of the combined
carcinogenic potency of the complex products of an emission source is not available. What often
is available is information concerning the relative potency of complex PAH mixtures compared
to a standard (such as B[a]P) obtained in experimental animal test systems (e.g., skin painting,
lung implant, etc.). Data of this nature are not directly extrapolatable to humans due to our
inability to establish equivalent exposure units for the experimental animal and anticipated
human exposure routes. As a result, Albert et al. (1982) devised indirect methods for using
relative potency information to estimate the risk due to inhalation of complex PAH mixtures.
The general approach is to establish the relative potency of the complex PAH mixture compared
to a standard agent that has a known inhalation dose-response model associated with it. Given
the relative potency value, the exposure to the PAH mixture is converted into standard agent
equivalent exposure units by taking the product of the PAH mixture exposure level and the
relative potency. These standard equivalent exposure units are then substituted into the
standard inhalation dose-response model to obtain cancer risk estimates that could be attributed
to the complex PAH mixture. This general approach has been the guiding principal behind
much of the PAH risk assessment research conducted by the EPA in recent years.
One view of ETS is that it is simply another complex mixture of agents containing multiple
carcinogenic PAHs. Although ETS contains many carcinogens other than PAHs, recent research
D-3 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
(Grimmer et al, 1988) strongly suggests that the majority of ETS lung carcinogenic potency is
due to the greater than 3-ring PAHs. This suggests that the same approaches for estimating
lung cancer risk for complex PAH mixtures could also be employed to good advantage for
estimating ETS lung cancer risk. Some of the information that could be useful in obtaining a
cancer dose-response model for ETS using the general approach for PAHs is displayed in Table
D-l.
D.2.2. Estimating ETS Relative Potency
The first step in obtaining an ETS dose-response model is to establish the relative
carcinogenic potency of ETS compared to an appropriate standard (e.g., B[a]P, coke oven
emissions, diesel engine exhaust). All of the available experimental information should be
reviewed and evaluated for its quality and relevance in obtaining ETS relative potency estimates.
One experiment that is a likely candidate for use in obtaining ETS versus B[a]P relative potency
estimates is the lifetime rat lung implant study conducted by Grimmer et al. (1988). Due to its
potential importance the protocol of that experiment is explained briefly. Three-month-old
inbred Osborne-Mendel female rats were used. Various amounts of B[a]P or ETS fractions
were dissolved in residue-free acetone, warmed to 50 degrees C, and a 1:1 mixture of beeswax
and Trioctanoin was added. The acetone was removed by rotary evaporation under reduced
pressure. This material was then warmed to 60 degrees C and introduced by injection into the
left lobe of the lung of Nembutol-anesthetized animals following thoracotomy. Following its
injection, the implant hardened into a pellet from which the test material diffused into the
surrounding lung tissue. Following the test material injection, the thoracotomy aperture was
sutured and the skin incision clipped. No further post-operative treatment was needed;
operative and post-operative mortality was less than 5%. After surgery, rats were observed until
D-4 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE D-l. SELECTED SOURCES OF INFORMATION POTENTIALLY
USEFUL FOR DERIVING A DOSE-RESPONSE
RELATIONSHIP FOR ETS
Agent
Benzo[a]Pyrene
(B[a]P)
3-Methylcholanthrene
(MCA)
Artificial Complex PAH
Mixture
Coal Flue Gas Condensate
Gas Engine Condensate
Diesel Engine Exhaust
Sidestream Cigarette
Smoke (ETS)
Mainstream Cigarette
Smoke
Coke Oven Emissions
Aluminum Smeller
Emissions
Roofing Tar Fumes
Indoor Coal
Wood Combustion
Route of exposure in animal experiments
Skin painting
X
X
X
X
X
X
X
X
X
X
X
Lung implants
X
X
X
X
X
X
X
Inhalation
X
X
X
X
no tumors
induced
X
Human risk models
based on
epidemiological data
where exposure is
inhalation
Under development
Possible to develop
State-of-the-art model
required
Model available
U.S. EPA (1984)
Upper bound under
development
Out of date
Under development
Under development
D-5
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
their natural deaths which was as long as 32 months post exposure. Moribund animals were
killed and when all animals were dead, complete autopsies were performed. The carcinogenic
data obtained in the experiment are displayed in Table D-2. The historical control data for
Osborne-Mendel rats are given in Table D-3, which are useful for obtaining stable non-zero
estimates of the population background cancer rates.
The advantages of the Grimmer et al. (1988) study are:
• the cancer response is at the anticipated target site for ETS (i.e., the lung),
• the animals were observed for their full lifespan,
• the exposure was a continuous, lifelong leaching of the test material out of the
beeswax/Trioctanoin pellet,
• multiple dose levels of the B[a]P positive control were employed,
• the average survival time for the experimental groups are given, which allows
appropriate age adjustments to be made,
• the experiment is one of a series of six on complex PAH mixtures conducted by the
same investigators that allows various hypotheses to be evaluated (e.g., dose additivity,
irritation effects, etc.), and
• the experiments, quality control, and the investigators' reputations are of the highest
order.
The disadvantages of the experiment are:
• exposure levels were most likely exponentially decreasing over time,
• the entire ETS condensate was not evaluated as one total exposure, and as a result,
dose additivity of the ETS fractions must be assumed to obtain a relative potency
estimate for the entire sample,
• the ETS condensate was not as aged as much as the ETS to which humans are
expected to be exposed,
• multiple exposures were not given for the ETS fractions,
D-6 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE D-2. DATA THAT CAN BE USED TO OBTAIN RELATIVE
POTENCY ESTIMATES FOR ETS CONSTITUENTS
Material
PAH-free material and
PAH 2,3 rings
PAH 4 and more rings
Semivolatives (gaseous
phase)
Benzo[a]Pyrene
Controls
Dose
mg x
16.00
1.06
11.80
0.03
0.10
0.30
Historical'
Vehicle
Untreated
Median survival
t
102
105
104
93
98
75
104
102
105
Animals with
epidermoid
carcinomas/ Total
animals
1/35
5/35
0/35
3/35
11/35
27/35
1/1945
0/35
0/35
Source: Grimmer et al. (1988) and 'Goodman et al. (1980)
TABLE D-3. HISTORICAL LUNG TUMOR CONTROL DATA FOR
OSBORNE-MENDEL RATS
Lung tumors
Epidermoid
Carcinomas
Alveolar /Bronchiolar
Adenoma
Alveolar Bronchiolar
Carcinoma
Male
1/975
4/975
3/975
Female
0/970
2/970
3/970
Combined
1/1945'
6/1945
6/1945
Source: Goodman et al. 1980
'Value used in ETS fraction relative potency estimation.
D-7
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
• the ETS mixture contains only the SS component and not the exhaled MS components
of ETS, and
• the tumors were epidermoid carcinomas at the implant site rather than the alveolar-
bronchiolar carcinomas usually associated with cigarette smoke.
The relative potency estimates of the ETS fractions and the theoretical estimates of the
total emission based on the assumption of dose additivity are displayed in Table D-4. The dose
additivity assumption has been shown to be consistent with information obtained in the same
animal model system employing diesel or gas engine exhaust condensate. It is estimated that
more than 70% of the total carcinogenic potency of the ETS is due to the 4 or more ringed
PAHs (i.e. [.03673 x .05833J/.00302 = .7098). The relative potency estimates incorporate a
special case of a two-stage mathematical model where the first stage preneoplastic clone has no
selective advantage over normal tissue in the rat lung. The U.S. EPA is developing this model
to estimate the relative potencies of other complex PAH mixtures whose carcinogenicity has
been evaluated by the lung implant experimental system (Thorslund 1990).
TABLE D-4. RELATIVE POTENCY ESTIMATES OF ETS CONSTITUENTS'
Constituents
PAH-free
PAH 2,3 rings
PAH 4 and
more rings
Semivolatiles
(gaseous
phase)
Weighted
Total
j
1
2
3
Dose mg.
(Xj)
16.00
1.06
11.80
28.86
Fraction
sample
,)
0.00158
0.05833
0.00000
0.00302
= Pr
% of total
carcinogenicity
attributable to
constituent
29.02
70.98
0.00
100.00
*Source of data: Grimmer et al. (1988)
D-8
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Refinements in the estimate and a more complete documentation of the techniques
employed in the analysis are required. Also, results from other experimental systems should be
used to estimate relative potencies if possible. The discussion given here should be regarded
only as an illustration of the type of analysis that can be conducted with some of the available
information.
D.2.3. Inhalation Dose-Response Models for PAHs
Once the relative potency of ETS to the standard agent (e.g., B[a]P) has been estimated,
the result is used to estimate the standard equivalent exposure units which are substituted into
the standard inhalation dose-response model to obtain cancer risk estimates. As indicated in
Table D-l, a number of alternatives exist upon which a standard inhalation dose-response model
could be based. We shall evaluate three potential choices in this section.
D.2.3.1. Hamster Inhalation B[a]P Dose-Response-The only animal pure PAH inhalation
experiment presently available that contains sufficient information to establish a dose-response
relationship was conducted by Thyssen et al. (1981). In that study Syrian golden hamsters were
exposed over their entire lifespan to pure B[a]P via an Nad aerosol. The tumors most closely
associated with B[a]P exposure were malignant and found in the larynx and pharynx.
Summarized results of the study are displayed in Table D-5. Thorslund (1990) demonstrated
that a two-stage model with exponential growth of preneoplastic targets can adequately describe
the experiment. The advantages of using the Thyssen et al. (1981) study are:
• the exposure was well monitored over the entire length of the experiment,
• the average lifetime exposure and age at death was available for each animal in the
experiment,
D-9 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE D-5. EXAMPLE OF ANIMAL INHALATION DOSE-RESPONSE
MODEL SYRIAN GOLDEN HAMSTERS EXPOSED TO B[a]P VIA NaCl AEROSOL
(Thyssen et al. 1981)
Lifetime average
exposure
(mg/m3 B[a]P)
X
Historical Controls
Matched Controls
Total Controls
Low Exposure
Chamber (2 mg/m3)
0.250 mg/m3
Middle Exposure
Chamber (10 mg/m3)
1.016 mg/m3
High Exposure
Chamber (50 mg/m3)
4.292 mg/m3
Average survival
(weeks)
t
80.0
105.0
100.5
102.5
.70.7
Animals
examined
226
22
248
24
23
23
# of hamsters with one or more malignant
laryngeal or pharyngeal tumors
Observed
1
0
1
0
11
17
Predicted (sum of
individual data)
0.642
0.179
0.821
1.23
8.55
17.98
Model*
P(x,t)--\-exp-H(x,t)
H(x,t) - (1 +Sx)2{[exp(GO-1 -Gf\]
G
A-3.865x10-7, S-6.843, G-0.0263
'Thorslund (1990)
D-10
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
• the animals were followed for their entire natural lifespan, and
• careful histopathological examinations were conducted on each animal.
The disadvantages of this study are:
• the hamsters are not humans,
• the tumors did not develop at the site anticipated in humans (i.e., lung),
• the hamsters are resistant to lung tumor formation, and
• the bioavailability of the B[a]P/NaCl aerosol may be different from the bioavailability
of the PAH-matrix to which humans are exposed.
When confronted with inhalation exposures of complex PAH mixtures, the approach used
by the EPA program offices has usually been to assume that the entire PAH mixture is as
potent as B[a]P and to substitute the total exposure units into an earlier version of the B[a]P
dose-response model derived from the Thyssen et al. (1981) data. This approach is recognized
as having numerous uncertainties and as being conservative.
D.2.3.2. Rat Inhalation Diesel Engine Exhaust Dose-Response--The diesel engine exhaust rat
inhalation study of Mauderly et al. (1987) offers another possibility for establishing an inhalation
dose-response model. The advantages of this study are:
• the tumors appeared in the lung,
• the PAH-matrix is reasonably similar to the type one might expect with human
exposures, and
• the lung burden exposure measurements are available.
The disadvantages of this study are:
• the rats are not humans,
• the lung tumors were for the most part not malignant, and
• the relative potency estimates compared to B[a]P for the exact diesel engine emissions
used in the experiment are not available.
D-ll 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The authors of the diesel engine exhaust paper obtained a dose-response model from their
experiment results. However, their derived model would not be consistent with the usual
regulatory assumption of low dose linearity. The same type of model employed for B[a]P could
also be used for diesel engine exhaust. As a result, it would be desirable to acquire the
individual animal data and fit the two-stage model to it to maintain a consistent approach
throughout.
To use this study an estimate of the relative potency of ETS compared to diesel engine
exhaust is required. Several options for obtaining such estimates exist. Perhaps the most direct
approach would be to pool the data obtained in the Grimmer et al. (1987, 1988) papers on ETS
and diesel engine exhaust which both employed the lung implant experimental system.
D.2.3.3. Human Inhalation Coke Oven Dose-Response-As noted previously, a dose-response
model for coke oven emissions has been used by the U.S. EPA (1984). This model is based on
a simple linear absolute risk model where the age-specific lung cancer risk is proportional to a
lag-time adjusted cumulative exposure. The advantages of using this model are:
• it is based on human occupational epidemiological data,
• the coke oven exposure in inhaled air is comparable to how humans are exposed to
ETS, and
• human coke oven inhalation data have been used by EPA to support regulatory
decisions.
The disadvantages are:
• the model should be updated with regard to presently available mortality and exposure
information, which would require considerable effort and resources,
• the cigarette smoking rates of the cohort members are unknown and thus are not
adjusted for and could be an important confounding variable, and
D-12 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
• two different bioassay systems are required to obtain coke oven equivalent exposure
units for ETS (i.e., ETS compared to B[a]P lung implant and coke oven emissions
compared to B[a]P skin painting).
Re-evaluating the coke oven data using the same two-stage model approach employed for
the other PAH data sets and the updated mortality experience appear to be a more scientifically
sound path to follow, but would require substantial resources.
D.3. CIGARETTE-EQUIVALENT APPROACH
D.3.1. Overview
The most obvious approach for obtaining an inhalation dose-response model for ETS is to
use direct cigarette smoking. The cigarette-equivalent approach (CEA) may be viewed as a
special case of the RPA with an added complication. An adjustment is required to equate the
lung deposition of carcinogens achieved by forced deep puffing on cigarettes with that resulting
from normal inhalation of ETS in the surrounding air. In chapter 4 several approaches were
discussed for making such adjustments that were felt to be inadequate. In this section an
alternative general approach based on specific biological markers is suggested.
The three necessary elements required to develop a credible ETS dose-response model
are:
• a state-of-the-art human mainstream smoke (MS) dose-response model,
• a relative potency estimate of ETS compared to MS, and
• a deposition rate equivalency for an appropriate biological marker (e.g., B[a]P-DNA
adduct) between ETS and MS.
Each of these elements are discussed in the following sections.
D-13 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
D.3.2. State-of-the-Art Mainstream Cigarette Dose-Response Model
The ideal study for establishing a dose-response model for mainstream smoking would be
based on a large cohort where many of the members had died of lung cancer, and the following
information on each member of the cohort would be obtainable.
1. detailed smoking history
a. age at start of smoking
b. way of smoking
i. inhalation patterns
ii. average puffs per cigarette
iii. length to which cigarette is smoked
c. smoking intensity (i.e., number of cigarettes smoked per day)
d. changing pattern of cigarette use over time
2. age at the start and end of the observation period and vital status at the end of the
observation period.
3. most detailed pathology information available
4. demographic information
a. race
b. sex
c. population density of domicile
d. job status
5. workplace exposure to other known lung carcinogens
While the information contained in any actual conducted study will not even come close to
conforming to the ideal, the list still is a convenient yard stick to measure potential studies for
possible inclusion in our dose-response development. Different limited studies may be useful in
contributing information concerning as little as one parameter in the eventual dose-response
model.
D-14 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
D.3.2.1. Sources of Epidemiological Information Useful for Constructing a Dose-Response
Model--A number of epidemiological studies may be viewed as primary sources of information
concerning the dose-response relationship between mainstream cigarette smoke and lung cancer.
These studies are briefly described below.
D.3.2.1.1. U.S. American war veterans (AWV). A cohort of about 300,000 American male war
veterans was assembled in 1954 by Dorn and their mortality experience subsequently reported
on by Kahn (1966), Rogot (1974), Whittemore (1988), and Freedman and Navidi (1989). This
study includes information of age at start of smoking and number of cigarettes smoked per day.
It also included ex-smokers. The study has the distinct advantage of having a total of 1266 lung
cancer deaths available for analysis, and some details on individual cohort members are
potentially obtainable from National Cancer Institute (NCI) data tapes. The study's
disadvantages are that exposure information was only obtained at the beginning of the
observation period so that changes in smoking patterns, except for stopping, cannot be taken
into account. Also, there are no women in the cohort.
D.3.2.1.2. American Cancer Society (ACS) volunteers. The ACS enlisted the help of a large
number of volunteer workers to help define and follow a cohort of about 440,000 male and
570,000 female predominantly middle to upper class white Americans. This study was reported
by Hammond (1966) with additional information available from ACS personnel. The study is
particularly useful in that it contains extensive information on women and nonsmokers not
available from other sources. Additional advantages of the study are the large number of lung
cancer deaths in the cohort, 1542 for male smokers and 164 for the females, and information on
age when smoking started and length of follow-up is available for each cohort member. The
D-15 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
main disadvantages of the study are: (1) it took place over a relatively narrow time frame of
8.5 years which increases the potential for a calendar year bias, and (2) the lower social
economic classes are under-represented which may introduce a bias due to the difference in type
and way cigarettes are smoked.
D.3.2.1.3. British male physicians. Doll and Peto (1978) published the data shown in Table D-6
based on the information obtained by following the survival of a cohort of approximately 34,000
British male physicians. The smoking histories of each individual in the cohort were obtained by
questionnaires at three different points in time. Table D-6 is a subset of the total cohort
consisting of subjects who smoked at a nearly constant rate over their smoking lifetime. Due to
the quality of the smoking information and pathology confirmation of most of the cases, this
study is generally acknowledged to be the most informative available for establishing dose-
response relationships. The disadvantages are that the number of observed lung cancer deaths
are relatively small (i.e., 215), no women are included in the sample, and information on ex-
smokers was never published in a form suitable for analysis. Also, a sample of physicians has a
high potential for a sociological bias to be built into it. Ten years of additional observation is
available on the cohort that has not yet been published and could be of considerable importance
in the establishment of a dose-response model.
D-16 05/17/90
-------
TABLE D-6. NUMBER OF LUNG CANCER DEATHS AND PERSON-YEARS OF OBSERVATION FOR
BRITISH MALE PHYSICIANS
Mid-point
age interval
(years)
42.5
47.5
52.5
57.5
62.5
67.5
72.5
77.5
observed lung cancer deaths
person-years
observed lung cancer deaths
person-years
observed lung cancer deaths
person-years
observed lung cancer deaths
person-years
observed lung cancer deaths
person-years
observed lung cancer deaths
person-years
observed lung cancer deaths
person-years
observed lung cancer deaths
person-years
Average exposure (cigarettes per day)
0.0
0.0
17,846.5
0.0
15,832.5
0.0
12,226.0
2.0
8,905.5
0.0
6,248.0
0.0
4,351.0
1.0
2,723.5
2.0
1,772.0
2.7
0.0
1,216.0
0.0
1,000.5
0.0
853.5
1.0
625.0
1.0
509.5
0.0
392.5
1.0
242.0
0.0
208.5
6.6
0.0
2,041.5
0.0
1,745.0
0.0
1,562.5
0.0
1,355.0
1.0
1,068.0
1.0
843.5
2.0
696.5
0.0
517.5
11.3
1.0
3,795.5
1.0
3,205.0
2.0
2,727.0
1.0
2,288.0
1.0
1,714.0
2.0
1,214.0
4.0
862.0
4.0
547.0
16.0
0.0
4,824.0
1.0
3,995.0
4.0
3,278.5
0.0
2,466.5
2.0
1,829.5
2.0
1,237.0
4.0
683.5
5.0
370.5
20.4
1.0
7,046.0
1.0
6,460.5
6.0
5,583.0
8.0
4,357.5
13.0
2,863.5
12.0
1,930.0
10.0
1,055.0
7.0
512.0
25.4
0.0
2,523.0
2.0
2,565.5
3.0
2,620.0
5.0
2,108.5
4.0
1,508.5
5.0
974.5
7.0
527.0
4.0
209.5
30.2
1.0
1,715.5
2.0
2,123.0
3.0
2,226.5
6.0
1,923.0
11.0
1,362.0
9.0
763.5
2.0
317.5
2.0
130.0
38.0
0.0
892.5
0.0
1,150.0
3.0
1,281.0
4.0
1,063.0
7.0
826.0
9.0
515.0
5.0
233.0
2.0
88.5
Source: Doll and Peto (1978)
~J
o
-------
DRAFT-DO NOT QUOTE OR CITE
D.3.2.1.4. Other data sources. Other sources of information that could prove useful in
obtaining information on a dose-response model are Best (1966), Canadian smokers; Bross et al.
(1968), individuals who switched to filter cigarettes; Cederlof et al. (1975), a national probability
sample of Swedish subjects; Graham and Levin (1971), individuals who stopped smoking;
Hirayama (1977), Japanese smokers; Stevens and Moolgavkar (1984), British males; Lubin et al.
(1984), individuals who changed smoking habits; Wald et al. (1988), U.K. smoking statistics; the
IARC monograph on the evaluation of the carcinogenic risk of tobacco smoking to humans
IARC (1986), general information; and the U.S. Public Health Service, Smoking and Health
Report series for various types of smoking related information.
D.3.2.2. Modeling Approach for Cigarette Smoking Data-Various investigators, such as Doll
and Peto (1978), Thorslund and Charnley (1987), Brown and Chu (1987), Gaffney and Altshuler
(1988), Darby and Pike (1988), Freedman and Navidi (1989), and Moolgavkar et al. (1989), were
successful in fitting various forms of multi-stage type models to the British physicians data.
Modeling attempts using the AWV and ACS data have been less successful. Freedman and
Navidi (1989) could not obtain adequate fits using standard multi-stage models to the AWV and
ACS data sets when information on ex-smokers was included. The reasons for this inability
could be either deficiencies in the multi-stage model (hypothesis put forth by the authors) or
some unknown bias in the data that distorts the true dose-response relationship. To clarify the
situation other modeling approaches should be attempted.
Probably the most successful approach for mathematically modeling cigarette smoking data
was put forth by Moolgavkar et al. (1989). This is the only attempt to date to incorporate a
promotional component of cigarette smoke into a dose-response model. Using Moolgavkar's
basic model and the additional simplifying robust assumptions that:
D-18 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
• the number of stem (target) cells are constant over time,
• the ratio S of unit exposure induced to background cell transition rates are equal for
the two cellular transitions (i.e., normal stem to preneoplastic and preneoplastic to
neoplastic), and
• the growth rate of preneoplastic cells is a function, G(x), of the number of cigarettes
smoked per day,
the age-specific lung cancer rate of an individual at age t who has smoked x cigarettes per day
since age t0 can be expressed as:
where A is the product of the background transition rates, G(x)=G(0)[l + (R-l)M(x)] with
R=G(oo)/G(0) being the maximum relative growth rate increase that can be induced by
cigarettes and M(x) is a still to be specified function that defines the fraction of the maximum
growth rate increase that is induced with x cigarettes per day. In the model employed by
Moolgavkar, the simplifying assumption G(x) = G(0)+Ax was made. While this assumed
relationship may be appropriate at low doses, it very likely results in a distortion of the effect
for heavy smokers.
It is proposed that the Moolgavkar (i.e., two-stage) model parameter estimates be
obtained by simultaneously using multiple epidemiological-smoking-lung cancer data sets and the
following modifications and extensions of the above basic model:
• Moolgavkar assumed that the time from the development of a neoplastic cell until
death due to a lung cancer was a constant 3.5 years for each of the lung cancer deaths.
As an alternative this length of time will be estimated by maximum likelihood methods
assuming:
1. it is a constant unknown value for all lung cancer deaths, and
2. it is a random variable with an integer gamma probability distribution.
• Alternative specific forms for G(x) will be specified based on various assumptions of
how binding of smoking product agents with preneoplastic cells induce promotion.
D-19 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
• An adjustment will be made for the difference between British and American cigarettes
and British and American smoking habits.
• An investigation will be made of the hypothesis that G(0) may be different pre- and
post-exposure to accommodate the observation of rapidly falling age-specific rates post
cessation of smoking.
The largest information data base possible will be used in fitting the different variations of the
model. An illustration of how one of the parameters in the model, G(0), could be estimated is
given below.
Hammond (1966) pooled the ACS lung cancer mortality data for men and women
nonsmokers and obtained age-specific death rates for five-year age intervals. This information
is displayed in Table D-7. The justification given by Hammond (1966) for pooling the data was
the inability to reject the hypothesis of equal rates for the sexes on the basis of a statistical test.
Under the assumption of no cigarette smoking, x=0, so the previously described age-specific
rates for the two-stage model has the reduced form:
{exp[G(0)<]-1}
Assuming that the number of lung cancer cases out of the number of person years of
observation was an independent binomial random variable for each age class, maximum
likelihood estimates were obtained for the unknown parameters A and G(0) in the above model.
The adequate fit of the model is displayed in Table D-7 and Figure D-l.
It is reasonable to assume that the parameter G(0) is human population independent and,
perhaps, even species independent taken on a lifetime equivalent time scale. However, the
value A would most likely be dependent on the environmental conditions an individual is living
under. Therefore, different values for U.S. and British populations should be estimated.
D-20 05/17/90
-------
TABLE D-7. LUNG CANCER DEATH RATES PER 100,000 PERSON-YEARS AND
OBSERVED AND PREDICTED NUMBER OF LUNG CANCER DEATHS AMONG
MEN AND WOMEN WHO NEVER SMOKED REGULARLY
Age group
L to L+5
40-44
45-49
50-54
55-59
60-64
65-69
70-74
75-79
80-84
Total
Combined men and women
Number of lung cancer deaths
n
Observed
4
16
16
30
32
26
18
21
14
177
Predicted
5.40
14.01
20.05
24.52
27.53
29.43
25.84
18.82
11.41
Death rate
dr
2.3
5.0
4.9
10.5
13.9
14.7
16.1
35.8
54.6
Population size
N
(person-years)
173,913
320,000
326,531
285,714
230,216
176,871
111,801
58,659
25,641
o
o
1
o
o
jo
O
Source: Hammond (1966)/ACS Study
* yv
nxW6
dr
calculated from dataXS = 7.036 p = 0.425
^J
o
-------
DRAFT-DO NOT QUOTE OR CITE
FIGURE D-l.
GOODNESS-OF-FIT OF TWO-STAGE MODEL TO NON-SMOKERS
AGE-DEPENDENT LUNG CANCER DATA
rt
"I
•C o
C
a
i
Lung Cancer Mortality among Male and Female Non-Smokers,
Obverwd Ratov (point*};
Expected fabM (Una)
A = 1,35 * 1Q-7
0(0} - O.DB6J
h(t) = A ' {exp[G(0)tl - 13 / G(0)
45 50 55 6D 65 70
Age in yea re (t)
75
SO
Source of data: Hammond et al. (1966).
D-22
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
The obvious advantage of the proposed CEA is that it is based on the most extensive body
of information concerning the dose-dependent effects of an environmental agent on a human
cancer response that exists. The main disadvantages are the complexity of the analysis and the
possibility of not establishing a credible ETS equivalency relationship. The latter factor is
discussed in the next sections.
D.3.3. Estimation of the Relative Potency of ETS Compared to MS
Previous approaches for establishing ETS/cigarette equivalency (e.g., Darby and Pike,
1988) have made the implicit assumption that the ratio of the potency of emissions to some
surrogate measure of internal exposure (e.g., nicotine, cotinine, etc.) is the same for ETS and
MS. The large variability in relative potency estimates of complex-PAH mixtures that are
displayed in Table D-8 suggests that the implicit assumption of equal potency is suspect.
Several methods can be used to estimate the ETS compared to MS relative potency. The
inhalation studies in Syrian golden hamsters where laryngeal carcinomas were elicited from MS
(Dontenwill et al., 1973; 1977) and from B[a]P (Thyssen, 1981) can be used to obtain a MS-to-
B[a]P relative potency estimate. Dividing this obtained potency value into the ETS-to-B[a]P, the
relative potency obtained from the lung implant studies discussed in Section D. 1.3.1 would give a
relatively potent estimate of ETS to MS. Stanton et al. (1972) conducted a lung implant study
using cigarette smoke condensate (CSC). Unfortunately for our present purposes, 3-
methylcholanthrene (MCA) was used as the positive control in the experiment so direct
comparison with ETS is not possible. However, Grimmer and his colleagues for the most part
closely adopted Stanton's experimental protocol for conducting lung implant studies. Thus, a
direct pooling of the data in the Stanton and Grimmer experiments could logically be used to
obtain a potency estimate. As an alternative, the two-step approach of estimating the potency of
D-23 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE D-8. RELATIVE POTENCY ESTIMATES OF COMPLEX MIXTURES
OF INCOMPLETE COMBUSTION PRODUCTS OF HYDROCARBONS
COMPARED TO B[a]P
Complex PAH exposure
Coal Flue Gas Condensate
Gas Engine Condensate
Diesel Engine Exhaust
Sidestream Cigarette Smoke
Coke Oven Emissions
Direct bioassay estimate + of relative
potency
0.05444
0.02190
0.00230
0.00302
0.03180'
+Lung implant studies
'Skin painting
CSC compared to MCA from the Stanton experiment and then establishing the relative potency
of MCA compared to B[a]P in another assay system (e.g., subcutaneous injection, skin painting,
etc.) could be employed. A final alternative might be to compare the weighted relative potency
estimates of the known constituents in the MS and ETS samples that have stable established
estimates of their carcinogenic potency compared to B[a]P. One potential list of stable relative
potency estimates developed by Thorslund (1990) is shown in Table D-9.
The last piece of information required to obtain an ETS risk model based upon the CEA
is a deposition ratio estimate between MS via active smoking and ETS under normal inhalation
conditions. One promising approach of using B[a]P-DNA-adducts and other endpoints as
biomarkers is discussed in the next section.
D.3.4. Deposition Differences of Chemicals from Cigarette Smoke in Smokers and Nonsmokers
To obtain an equivalency relationship between MS and ETS, both potency and deliverable
dose conversion factors are needed in order to use the MS-lung cancer data as a surrogate for
lung cancer induced by ETS.
D-24
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE D-9. RELATIVE POTENCY ESTIMATES OF AGENTS
COMPARED TO B[a]P
Agent
Anthracene
Fluoranthene
Pyrene
Benzo[b]fluoranthene
Benzo[k]fluoranthene
Benzo[j]fluoranthene
Benzo[e]pyrene
Benzo[a]pyrene
Indeno Pyrene
Benzo[ghi]perylene
Anthanthrene
Relative potency
0.00000
0.00000
0.00000
0.12277
0.05322
0.05232
0.00704
1.00000
0.27800
0.02124
0.31598
Source of estimate
IARC adequately studied; no
indication of carcinogenic effect
category
Deutsch-Wenzel et al. (1983)
(Grimmer's group)
lung implant data
Thorslund (1990) estimates
Under the assumption that the PAHs possess most of the carcinogenic potency in MS and
ETS, the deliverable target dose can be estimated by directly measuring the number of DNA
adducts formed in people smoking different numbers of cigarettes per day and in people who
are nonsmokers in the presence of smokers with different frequencies of smoking.
Specific adducts, such as the DNA 7,8-diol-9,10-epoxide of B[a]P which is present in both
MS and ETS, can be detected using sensitive immunoassays or postlabelling DNA techniques
(Shamsuddin et al., 1985; Randerath et al., 1986). Differences in adduct formation between
smokers and nonsmokers varied depending on the experiment but was as high as 400-fold when
DNA from oral mucosa was analyzed using the postlabelling technique. Hemoglobin adducts as
markers of genotoxicity have been analyzed in smokers and nonsmokers where smokers had
about a 7-fold greater number of adducts than nonsmokers.
D-25
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Indirect measures of dose between smokers and nonsmokers may also be available in
which gene mutations can be measured in peripheral leukocytes at the Hypoxanthine
phosphoribosyl transferase locus as well as other loci. In fact, such a test could conceivably be
used directly to obtain a cigarette equivalence estimate without making potency difference
adjustments. Other genetic damage tests, such as chromosomal aberrations and sister chromatid
exchanges, may also be useful in determining deliverable target dose information for smokers
and nonsmokers exposed to ETS.
To obtain an equivalency relationship of deliverable dose between smokers and
nonsmokers, a thorough review of the literature for articles that show dose-response
relationships between MS/ETS and DNA adducts, protein adducts, and gene mutations should
be conducted and the most appropriate endpoints selected for use in the equivalency estimate.
The main advantage of the approach is the high suspected correlation of the endpoint with the
cancer response. The main disadvantage is the discounting of potential agents that act
exclusively as promoters.
D.4. DIRECT APPROACH
The most straightforward approach for estimating ETS lung cancer risk is to estimate ETS
exposure in a suitable cohort and follow the resulting mortality pattern over time. As of yet, no
directly measured ETS exposure data exist on a cohort. The ideal in this regard would be
personal monitoring data obtained from nonsmokers for an agent such as cotinine which is
closely and uniquely associated with cigarette smoke. In this application, the use of cotinine is
appropriate as long as it is linearly related to total ETS air levels. In lieu of such information,
investigators have attempted to obtain surrogate measures of ETS. One such measure is the
number of cigarettes smoked per day by the spouses of nonsmoking individuals. The quality of
such a surrogate measurement depends upon: (I) the extent that nonsmokers are exposed to
D-26 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
TABLE D-9. RELATIVE POTENCY ESTIMATES OF AGENTS
COMPARED TO B[a]P
Agent
Anthracene
Fluoranthene
Pyrene
Benzo[b]fluoranthene
Benzo[k]fluoranthene
Benzo[j]fluoranthene
Benzo[e]pyrene
Benzo[a]pyrene
Indeno Pyrene
Benzo[ghi]perylene
Anthanthrene
Relative potency
0.00000
0.00000
0.00000
0.12277
0.05322
0.05232
0.00704
1.00000
0.27800
0.02124
0.31598
Source of estimate
IARC adequately studied; no
indication of carcinogenic effect
category
Deutsch-Wenzel et al. (1983)
(Grimmer's group)
lung implant data
Thorslund (1990) estimates
Under the assumption that the PAHs possess most of the carcinogenic potency in MS and
ETS, the deliverable target dose can be estimated by directly measuring the number of DNA
adducts formed in people smoking different numbers of cigarettes per day and in people who
are nonsmokers in the presence of smokers with different frequencies of smoking.
Specific adducts, such as the DNA 7,8-diol-9,10-epoxide of B[a]P which is present in both
MS and ETS, can be detected using sensitive immunoassays or postlabelling DNA techniques
(Shamsuddin et al., 1985; Randerath et al., 1986). Differences in adduct formation between
smokers and nonsmokers varied depending on the experiment but was as high as 400-fold when
DNA from oral mucosa was analyzed using the postlabelling technique. Hemoglobin adducts as
markers of genotoxicity have been analyzed in smokers and nonsmokers where smokers had
about a 7-fold greater number of adducts than nonsmokers.
D-25
05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Indirect measures of dose between smokers and nonsmokers may also be available in
which gene mutations can be measured in peripheral leukocytes at the Hypoxanthine
phosphoribosyl transferase locus as well as other loci. In fact, such a test could conceivably be
used directly to obtain a cigarette equivalence estimate without making potency difference
adjustments. Other genetic damage tests, such as chromosomal aberrations and sister chromatid
exchanges, may also be useful in determining deliverable target dose information for smokers
and nonsmokers exposed to ETS.
To obtain an equivalency relationship of deliverable dose between smokers and
nonsmokers, a thorough review of the literature for articles that show dose-response
relationships between MS/ETS and DNA adducts, protein adducts, and gene mutations should
be conducted and the most appropriate endpoints selected for use in the equivalency estimate.
The main advantage of the approach is the high suspected correlation of the endpoint with the
cancer response. The main disadvantage is the discounting of potential agents that act
exclusively as promoters.
D.4. DIRECT APPROACH
The most straightforward approach for estimating ETS lung cancer risk is to estimate ETS
exposure in a suitable cohort and follow the resulting mortality pattern over time. As of yet, no
directly measured ETS exposure data exist on a cohort. The ideal in this regard would be
personal monitoring data obtained from nonsmokers for an agent such as cotinine which is
closely and uniquely associated with cigarette smoke. In this application, the use of cotinine is
appropriate as long as it is linearly related to total ETS air levels. In lieu of such information,
investigators have attempted to obtain surrogate measures of ETS. One such measure is the
number of cigarettes smoked per day by the spouses of nonsmoking individuals. The quality of
such a surrogate measurement depends upon: (1) the extent that nonsmokers are exposed to
D-26 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
smokers other than their spouses, (2) the consistency within the cohort of the husbands' and
wives' spatial and time closeness, and (3) the consistency within the cohort of the fraction of the
total cigarettes that are smoked by the spouse in the home. Due to sociological factors
regarding a woman's place in Japan, the homogeneity of the Japanese society, and the small,
close living arrangements of Japanese couples, probably the best surrogate measure of ETS
exposure available is the number of cigarettes smoked per day by the husbands of Japanese
women. The person-years of observation and the number of lung cancer deaths for Japanese
women classified in regard to their husband's age and smoking habits obtained in the
prospective study conducted by Hirayama (1984) is displayed in Table D-10. Under the
assumption that all the excess lung cancer risk in Japanese women was due to husband-produced
ETS exposure in the home, crude risk models can be generated from the information supplied
in Table D-10. Better estimates could be obtained if information such as the length of marriage,
wife's age, age husband started smoking, and smoking habits of wife's parents were available for
individual cohort members. A fair amount of such information has been generated by Hirayama
(1984) but presently is not reported in the open literature. Gaining access to the data could
prove valuable.
D-27 05/17/90
-------
O
oo
TABLE D-10. LUNG CANCER MORTALITY IN JAPANESE WOMEN BY HUSBAND'S AGE GROUP
AND SMOKING HABITS (PATIENT HERSELF A NON-SMOKER)*
Husband's
age group
40-49
50-59
60-69
70-79
Total
Husband's smoking habit
Non-smoker
4
10
18
5
37
6,229
7,791
7,120
755
21,895
Ex-smoker
1
3
11
2
17
1,255
1,922
2,687
348
6,212
1-14/day
8
20
28
2
58
8,621
9,668
7,243
612
26,144
15-19/day
6
8
9
1
24
5,158
4,052
2,513
105
11,828
20+ /day
16
24
23
1
64
10,764
9,820
4,651
226
25,461
Total
35
65
89
11
200
32,027
33,253
24,214
2,046
91,540
"Number of lung cancer deaths out of number of wives in the same cross classification cell.
Source: Hirayama (1984).
O
O
*
s
O
O
a
n
-~j
O
-------
DRAFT-DO NOT QUOTE OR CITE
REFERENCES AND RELATED BIBLIOGRAPHY
Agresti, A. (1990) Categorical data analysis. John Wiley & Sons, Inc.
Akiba, S.; Kato, H.; Blot, W.J. (1986) Passive smoking and lung cancer among Japanese women.
Cancer Research 46:4804-4807.
Albert, R.E.; Lippman, M.; Briscoe, W. (1969) The Characteristics of bronchial clearance in
humans and the effect of cigarette smoking. Arch. Environ. Health 18:738-755.
Albert, R.E.; Peterson, H.T.; Bohning, D.E.; Lippmann, M. (1975) Short-term effects of
cigarette smoking on bronchial clearance in humans. Arch. Environ. Health 30:361-367.
Albert, R.E., Lewtas, J., Nesnow, S., Thorslund, T., Anderson, E. (1982) Comparative potency
method for cancer risk assessment: Application to diesel particulate emissions. J. Risk
Anal 101-107.
Altschuler, B.; Nelson, N.; Kuschner, M. (1964) Estimation of lung tissue dose from the
inhalation of radon and daughters," Health Phys. 10:1137-1161.
American Academy of Pediatrics, Committee on Environmental Hazards (1986) Involuntary
smoking--A hazard to children. Pediatrics 77(5):755.
Angioni, A.M.; Fanciulli, G.; Corchia, C. (1989) Frequency of and risk factors for allergy in
primary school children: Results of a population survey. Pediatric and Perinatal
Epidemiology 3:248-255.
Air Pollution Control Association (APCA) (1986) Indoor Radon, APCA, Pittsburgh, PA.
Aronow, W.S. (1978) Effect of passive smoking on angina pectoris. New England Journal of
Medicine 299(l):21-24.
Arundel, A.; Sterling, T.; Weinkam, J. (1987) Never smoker lung cancer risks from exposure to
particulate tobacco smoke. Environ. Int. 13:409-426.
Badre, R.; Guillerm, R.; Abran, N.; Bourdin, M.; Duman, C. (1978) Atmospheric pollution by
cigarette smoking. Ann Pharm Fr. 36:443-452.
Baiter, N.J.; Schwartz, S.L.; Kilpatrick, S.J.; Witorsch, P. (1986) Causal relationship between
environmental tobacco smoke and lung cancer in non-smokers: a critical review of the
literature. Proceedings of the Air Pollution Control Association, 79th annual meeting.
Banks, M.H.; Bewley, B.R.; Bland, M.; Pollard, V. (1978) Effect of children's and parents'
smoking on respiratory symptoms. Archives of Disease in Childhood 53(2): 100-105.
Barker, D.; Osmond, C. (1986) Childhood respiratory infection and adult chronic bronchitis in
England and Wales (UK). British Medical Journal 293(6557): 1271-1275.
R-l 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Bassi, J.A.; Rosso, P.; Moessinger, A.C.; Blanc, W.A.; James, L.S. (1984) Fetal growth
retardation due to maternal tobacco smoke exposure in the rat. Pediatric Research
18(2): 127-130.
Battista, S.P.; Kensler, C.J., (1970) "Mucus Production and Ciliary Transport Activity," Arch.
Environ. Health 20:326-338.
Benowitz, M.L.; Jacob, P.; Jones, R.T.; Rosenberg, J. (1982) Inter-individual variability in the
metabolism and cardiovascular effects of nicotine in man. J. Public Health 221:368-372.
Benowitz, N.L.; Jacob, P. (1984) Daily intake of nicotine during cigarette smoking. Clin.
Pharmacol. Ther. 35:499-504.
Bergman, H.; Axelson, A. (1983) Passive smoking and indoor radon daughter concentration,
Lancet 8362:1308-1309.
Bergman, H.; Edling, C; Axelson, O. (1986) Indoor Radon Concentrations and Passive
Smoking, Environ. Int. 12:17-19.
Berkey, C.S.; Dockery, D.W.; Ferris, B.G.,Jr.; Speizer, F.E.; Ware, J.H. (1986) Indoor air
pollution and pulmonary function growth in preadolescent children. American Journal of
Epidemiology 123(2):250-260.
Berkey, C.S.; Ware, J.H.; Speizer, F.E.; Ferris, B.G.Jr. (1984) Passive smoking and height growth
of preadolescent children. International Journal of Epidemiology 13(4):454-458.
Best, E.W.R. (1966) A Canadian study of smoking and health. Department of National Health
and Welfare, Ottawa.
Bewley, B.R., Bland, J.M., Harris, R. (1974) Factors associated with the starting of cigarette
smoking by primary children. British Journal of Preventive Social Medicine 28:37-44.
Black, N. (1978) The aetiology of glue ear: A case-control study. International Journal of
Pediatric Otorhinolaryngology 9(2): 121-133.
Black A.; Pritchard, J. N. (1984) A comparison of the regional deposition and short-term
clearance of tar particulate material from cigarette smoke, with that of 2.5 %m
polystyrene spheres. Aerosol Sci. 15:224-227.
Bland, M.; Bewley, B.R.; Pollard, V.; Banks, M.H. (1978) Effect of children's and parents'
smoking on respiratory symptoms. Archives of Disease in Childhood 53(2): 100-105.
Blot, W.J.; Fraumeni, J.F. (1986) Passive smoking and lung cancer. Journal of the National
Cancer Institute 77(5):993-999.
Blot, W.J.; McLaughlin, J.K. (1985) Practical issues in the design and conduct of case-control
studies: use of next-of-kin interviews. Blot, W.J., Hirayama, T., Huel, O.G., eds.
Statistical Issues in Cancer Epidemiology. Hiroshima: Sanei, pp. 46-62.
R-2 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Breslow, N.E.; Day, N.E. (1980) Statistical methods in cancer research. Lyon: IARC Scientific
Publications No. 32.
Bross, I.D.J.; Gibson, R. (1968) Risks of lung cancer in smokers who switch to filter cigarettes.
Am J Publ Hlth 58:1369-1403.
Brown, C.C.; Chu, K. (1987) Use of multistage models to infer stage affected by carcinogenic
exposure: Example of lung cancer and cigarette smoking. J Chronic Dis 40(Suppl):171S-
179S.
Brown, C.C.; Chu, K. (1987) Use of multistage models to infer stage affected by carcinogenic
exposure: Example of lung cancer and cigarette smoking. J. Chronic Dis.
40(2):171A-179A.
Brownson, R.C.; Reif, J.S.; Keefe, T.J.; Ferguson, S.W.; Pritzl, J.A. (1987) Risk factors for
adenocarcinoma of the lung. Am J Epidemiol 125:25-34.
Brunekreef, B.; Boleij, J.S. (1982) Long-term average suspended particulate concentrations in
smokers' homes. Int Arch Occup Environ Health 50:299-302.
Brunekreef, B.; Fischer, P.; Remijn, B.; Van der Lende, R.; Schouten, J.; Quanjer, Ph. (1985)
Indoor air pollution and its effects on pulmonary function of adult non-smoking
women:III. Passive smoking and pulmonary function. International Journal of
Epidemiology 14(2):227-230.
Brunnemann, K.D.; Adams, J.D., Ho, D.P.; Hoffmann D. (1978) The influence of tobacco
smoke on indoor atmospheres. II. Volatile and tobacco-specific nitrosamines in main and
sidestream smoke and their contribution to indoor pollution. Proceedings of the 4th Joint
Conference on Sensing of Environmental Pollutants. Washington, DC. American
Chemical Society, pp. 876-880.
Brunnemann, K.D.; Hoffmann D. (1978) "Chemical studies on tobacco smoke LIX. Analysis
of volatile nitrosamines in tobacco smoke and polluted indoor environments." Walker,
E.A., Castegnaro, M., Griciute, L. and Lyle, R.E., eds. Environmental aspects of N-nitroso
compounds Lyon: IARC Scientific Publications No. 19, pp. 343-356.
Buffler, P.A.; Pickle, L.W.; Mason, T.J.; and Contant, C. (1984) The causes of lung cancer in
Texas. Mizell, M. and Correa, P., eds. Lung Cancer: Causes and Prevention. New York:
Verlag Chemie International, pp. 83-99.
Burchfiel, CM.; Higgins, M.W.; Keller, M.W.; Howatt, W.F.; Butler, W.J.; Higgins, I.T.T. (1986)
Passive smoking in childhood - Respiratory conditions and pulmonary function in
Tecumseh, Michigan. Am. Rev. Respir. Dis. 133:966-973.
Cameron, P.; Kostin, J.S.; Oselett, B.; Stocker, R.; Tighe.G.; Winton, J.; Wolfe, J.H.; Zaks, J.M.
(1969) The health of smokers' and nonsmokers' children. Journal of Allergy 43(6):336-
341.
R-3 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Camner, P.; Philipson, K.; Arvidsson, T. (1971) Cigarette smoking in man, short-term effect on
mucociliary transport. Arch. Environ. Health 23:421-426.
Canadian Pediatric Society (1986) Secondhand cigarette smoke worsens symptoms in children
with asthma. CMAJ 135(4):321-323.
Cano, J. P.; Catalin, J.; Bnadre, R.; Dumas, C; Viala, A.; Guillerm, R. (1970) Analysis of
nicotine by gas chromatography: II applications. Ann Pharm Fr 28:633-640.
Carter, S.L.; Hasegawa. (1975) Fixation of tobacco smoke aerosols for size distribution studies.
J Colloid Interface Science 53:134-141.
Cedarlof, R., Friberg, L., Hurbec, Z., Lorich, U. (1975) The relationship of smoking and some
social covariables in mortality and cancer morbidity: A ten year follow-up on a probability
sample of 55,000 Swedish subjects age 18-69, Parts 1 and 2. Stockholm: Department of
Environmental Hygiene, The Karolinska Institute.
Chan, W.C.; Colbourne, M.J.; Fung, S.C.; Ho, H.C. (1979) Bronchial cancer in Hong Kong,
1976-1977. Br. J. Cancer 39:182.
Chan, W.C.; Fung, S.C. (1982) Lung cancer in non-smokers in Hong Kong. Cancer Campaign,
Vol.6, Cancer Epidemiology, (Grundmann, E., ed.) Stuttgart: Gustav Fischer Verlag, pp.
199-202.
Chan, K.N.; Elliman, A.; Bruan, E.; Silverman, M. (1989) Respiratory sypmtoms in children of
low birth weight. Arch. Dis. Child. 64(9): 1294-1304.
Chappell, S.B.; Pasker, R.J. (1977) Smoking and carbon monoxide levels in enclosed public
places in New Brunswick. Can J of Public Health 68:159-161.
Charlton, A. (1984) Children's coughs related to parental smoking. British Medical Journal
288:1647-1649.
Charlton, A.; Blair, V. (1989) Absence from school related to children's and parental smoking
habits. British Medical Journal 298:90-92.
Checkoway, H.; Pearce, N.; Crawford-Brown, D. (1989) Research methods in epidemiology.
New York, NY: Oxford University Press.
Chen, Y.; Li, W.-X. (1986a) The effect of passive smoking on children's pulmonary function in
Shanghai. American Journal of Public Health 76(5):515-518.
Chen, Y.; Li, W.-X.; Yu, S. (1986b) Influence of passive smoking on admissions for respiratory
illness in early childhood. British Medical Journal 293:303-306.
Chen, Y.; Li, W.-X.; Yu, S.; Qian, W. (1988) Chang-Ning epidemiological study of children's
health: I: Passive smoking and children's respiratory diseases. International Journal of
Epidemiology 17(2):348-355.
R-4 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Chen, Y. (1989) Synergistic effect of passive smoking and artificial feeding on hospitalization for
respiratory illness in early childhood. Chest 95(5): 1004-1007.
Coburn, R.F.; Forster, R.E.; Kane, P.B. (1965) Considerations of the physiological variables
that determine the blood carboxyhemoglobin concentration in man. J Clin Invest 44:1899-
1910.
Colley, J.R.T. (1971) Respiratory disease in childhood. British Medical Bulletin 27(1):9-14.
Colley, J.R.T. (1974) Respiratory symptoms in children and parental smoking and phlegm
production. British Medical Journal 2(5912):201-204.
Comroe, J.H.; Forster II, R.E.; Dubois, A.B.; Briscoe, W.A.; Carlsen, E. (1962) The lung:
Clinical physiology and pulmonary function tests. Chicago: Year Book Medical Publ.,
Inc., pp. 323-364.
Comstock, G.W.; Helsing, K.J.; Meyer, M.B.; Tockman, M.S. (1981) Respiratory effects of
household exposures to tobacco smoke and gas cooking. American Review of Respiratory
Disease 124(2): 143-148.
Copeland, K.T.; Checkoway, H.; McMichael, A.J.; Holbrook, R.H. (1977) Bias due to
misclassification in the estimation of relative risk. Am J of Epidemiol. 105(5):488-95.
Cornfield, J. (1951) A method of estimating comparative rates from clinical data: Applications
to cancer of the lung, breast, and cervix. J. Nat'l Cancer Inst. 11:1269-1275.
Correa, P.; Fontham, E.; Pickle, L.; Lin, Y.; Haenszel, W. (1983) Passive smoking and lung
cancer. Lancet 2:595-597.
Correa, P. (1986) The passive smoking-cancer controversy. In: Mechanisms in Tobacco
Carcinogenesis. Banbury Report 23:343- 360.
Cothern, C. (1987) "Properties," in Environmental Radon, edited by C. Couthern and J. Smith,
Plenum Press, New York, p. 1-29.
Cothern, C.; Smith, J., editors,(1987) Environmental Radon, Plenum Press, New York.
Coultas, D.B.; Howard, C.A.; Peake, G.T.; Skipper, B.J.; Samet, J.M. (1987) Salivary cotinine
levels and involuntary tobacco smoke exposure in children and adults in New Mecico. Am
Rev Respir Dis 136:305-309.
Covey, L.S., Mushinski, M.H., Wynder E.L. (1983) Smoking habits in a hospitalized population:
1970-1980. American Journal of Public Health 73:1293-1297.
Crawford, D.J. (1982) Identifying critical human subpopulations by age groups: radioactivity
and the lung. Phys. Med. Biol. 27:539-552.
Crawford-Brown, D.J. (1984) A unified approach to age-dependent metabolic modeling.
Health Physics 46:809.
R-5 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Crawford-Brown, D.J. (1987) Dosimetry. Cothern C.R. and Smith, I.E., Jr., eds.
Environmental Radon. New York, NY: Plenum Press, pp. 173-213.
Crawford, D.J. (1982), Identifying critical human subpopulations by age groups: Radioactivity
and the lung, Phys. Med. Biol. 27, 539-552.
Crawford, D.J.; Eckerman, K. (1983) Modifications of the ICRD task group lung model to
reflect age dependence. Rad Prot Dosim 2:209-220.
Cuddeback, J.E.; Donovan, J.R.; Burg, W.R. (1976) Occupational aspects of passive smoking.
Am Ind Hyg Assoc 37:263-267.
Cummings, K.M.; Markello, S.J,; Mahoney, M.; Bhargava, A.K.; McElroy, P.D.; Marshall, J.R.
(1986) Measurement of current exposure to environmental tobacco smoke. Am J
Epidemiol 130(1):122-132.
Cummings, K.M.; Markello, S.J.; Mahoney, M.C.; Marshall, J.R. (1989a) Measurement of
lifetime exposure to passive smoke. Am J Epidemiol 30:122.
Dahms, T.E.; Bolin, J.F.; Slavin, R.G. (1981) Passive smoking: Effects on bronchial asthma.
Chest 80(5):530-534.
Dalager, N.A.; Pickle, L.W.; Mason, T.J.; Correa, P.; Fontham, E.; Stemhagen, A.; Buffler, P.A.;
Ziegler, R.G.; Fraumeni, J.F. (1986) The relation of passive smoking to lung cancer.
Cancer Research 46:4808-4811.
Darby, S.C.; Pike, M.C. (1988) Lung cancer and passive smoking: Predicted effects from a
mathematical model for cigarette smoking and lung cancer. Br J Cancer 58:825-831.
Deutsch-Wenzel, R.P.; Brune, H.; Grimmer, G.; Dettbarn, G.; Misfeld, J. (1983) Experimental
studies in rat lungs on the carcinogenicity and dose-response relationships of eight
frequently occurring environmental polycyclic aromatic hydrocarbons. J Natl Cancer Inst
71:539-544.
Dockery, D.W.; Spengler, J.D. (1981) Indoor-outdoor relationships of respirable sulfates and
particles. Atmos Environ 15:335-344.
Dockery, D.W.; Spengler, J.D.. (1981) Personal exposure to respirable particles and sulfates. J.
Air Pollut Control Assoc 31:153-159.
Dodge, R. (1982) The effects of indoor pollution on Arizona children. Archives of
Environmental Health 37(3): 151-155
Doll, R.; Peto, R. (1976) Mortality in relation to smoking: 20 years' observations on male
British doctors. Br. Med. J. ii, 1525-1536.
Doll, R.; Peto, R. (1978) Cigarette smoking and bronchial carcinoma: dose and time
relationships among regular smokers and lifelong non-smokers. J. Epidem. Commun.
Health 32: 303-313.
R-6 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Doll, R.; Peto, J. (1986) Passive smoking. Br J Cancer 54:1019-1021.
Dontenwill, W.; Chevalier H.-J.; Harke H.-P.; Lafrenz, U.; Reckzeh, G.; Schneider, B. (1973)
Investigations on the effects of chronic cigarette smoke inhalation in Syrian
goldenhamsters. J. Natl Cancer Inst 51:1781-1832.
Dontenwill, W.; Chevalier, H.-J.; Harke, H.-P.; Klimisch, H.-J.; Kuhnigle, C.; Reckzeh, G.;
Schneider, B. (1977) Studies on the effect of chronic cigarette smoke inhalation in Syrian
golden hamsters and the importance of vitamin A on morphological alterations after
smoke exposure. Z. Krebsforsch 89:145-151.
Ebert, R.V.; Terracio, M.J., (1975), The Bronchiolar Epithelium in Cigarette Smokers,
American Rev. Resp. Dis. 111:4-11.
Ekwo, E.E.; Huntley, W.H.; Lachenbruch, P.A.; Weinberger, M.M. (1983) Relationship of
parental smoking and gas cooking to respiratory disease in children. Chest 84(6):662-668.
Elliot, L.P.; Rowe, D.R. (1975) Air quality during public gatherings. J. Air Pollut Control
Assoc 25:635-636.
Eriksen, M.P.; LeMaistre, C.A.; Newell, G.R. (1988) Health hazards of passive smoking. Ann
Review of Public Health 9:47-70.
Eudy, L.W.; Thomas, F.A.; Heavner, D.L.; Green, C.R.; Ingebrethsen, B.J. (June 22-27, 1986)
Studies on the vapor-particulate phase distribution of environmental nicotine by selective
trapping and detection methods. Proc of the Air Pollution Control Association 79th
annual meeting, pp. 2-14.
Evans, R.E. (1968) Engineers guide to the elementary behavior of radon daughters. Health
Phys. 38:1173-1197.
Evans, D.; Levison, M.J.; Feldman, C.H.; Clark, N.M.; Wasilewski, D.; Levin, B.; Mellins, R.B.
(1987) The impact of passive smoking on emergency room visits of urban children with
asthma. American Review of Respiratory Disease 135:567-572.
Federal Register (1986) Guidelines for carcinogen risk assessment. Vol. 51, No. 185, p. 33994-
34003, Sept. 24, 1986. Also available as U.S. Environmental Protection Agency (EPA).
(1987) The risk assessment guidelines of 1986. Office of Health and Environmental
Assessment, Washington, DC. EPA/600/8-87/045.
Fergusson, D.M.; Honvood, L.J.; Shannon, F.T., Taylor, B. (1981) Parental smoking and lower
respiratory illness in the first three years of life. Epidemiology and Community Health
35(3):180-184.
Fergusson, D.M.; Hons, B.A.; Horwood, L.J. (1985) Parental smoking and respiratory illness
during early childhood: A six-year longitudinal study. Pediatric pulmonology 1:99-106.
Ferris, E.G.; Ware, J.H.; Berkey, C.S.; Dockery, D.W.; Spiro, A. Ill; Speizer, F.E. (1985) Effects
of passive smoking on health of children. Environmental Health Perspectives 62:289-295.
R-7 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Fischer, T.; Weber, A.; Grandjean, E. (1978) Air pollution due to tobacco smoke in
restaurants. Int Arch Occup Environ Health 41:267-280.
Fleming, D.W.; Cochi, S.L.; Hightower, A.W.; Broome, C.V. (1987) Childhood upper
respiratory tract infections: To what degree is incidence affected by day-care attendance?
Pediatrics 79:55-60.
Fong, P. (1982) The hazard of cigarette smoke to nonsmokers. J. Biol. Phys. 10:65-.
Freedman, D.A.; Navidi, W. (1987a) On the multistage model for carcinogenesis. Tech.
Report. No. 47, University of California.
Freedman, D.A.; Navidi, W. (1987b) On the risk of lung cancer for ex-smokers. Tech. Report.
No. 135, University of California.
Freedman, D.A.; Navidi, W.C. (1989) Multistage models for carcinogens. Environ Health
Perspect 81:169-188.
Fried, P.A.; Watkinson, B.; Dillon, R.F.; Dulberg, C.S. (1987) Neonatal neurological status in a
low-risk population after prenatal exposure to cigarettes, maijuana, and alcohol. In: J.
Dev. Behav. Pediatr. 8(6):318-326.
Friedman, G.D.; Petitti, D.B.; Bawol, R.D. (1983) Prevalence and correlates of passive smoking.
Am J Public Health 73:401-.
Galvskinova, V. (1964) 3-4 benzopyrene determination in the smoky atmosphere of social
meeting rooms and restaurants: a contribution to the problem of the noxiousness of so-
called passive smoking. Neoplasma 11:465-468.
Gao, Y.; Blot, W.J.; Zheng, W.; Ershow, A.G.; Hsu, C.W.; Levin, L.I.; Zhang, R.; Fraumeni, J.F.
(1987) Lung cancer among Chinese women. Int'l J Cancer 40:604-609.
Gaffney, M.; Altshuler, B. (1988) Examination of the role of cigarette smoke in lung
carcinogenesis using multistage models. J. Nat'l Cancer Inst. 80:925-31
Garfinkel, L. (1981) Time trends in lung cancer mortality among nonsmokers and a note on
passive smoking. J. Nat'l Cancer Inst. 6:1061-1066.
Garfinkel, L. (1984) Passive smoking and cancer: American experience. Prev Med 13:691-697.
Garfinkel, L. (1985) Selection, follow-up, and analysis in the American Cancer Society
Prospective Studies. Presented at a Workshop on the Selection, Follow-up, and Analysis
in Prospective Studies held at the Waldorf-Astoria Hotel, New York, N.Y., October 3-5,
1983.
Garfinkel, L.; Auerbach, O.; Joubert, L. (1985) Involuntary smoking and lung cancer: a
case-control study. J Nat'l Cancer Inst. 75:463-469.
R-8 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Gastineau, R.M.; Walsh, P.J.; Underwood, N. (1972) Thickness of bronchial epithelium with
relation to exposure to radon. Health Phys. 23:857-860.
Geller-Bernstein; Kenett, R.; Weisglas, L; Tsur, S.; Lahav, M.; Levin, S. (1987) Atopic babies
with wheezy bronchitis. Allergy 42:85-91.
Geng, G.; Liang, Z.H.; Zhang, G.L. (1988) On the relationship between smoking and female
lung cancer. Smoking and Health, Elsevier Science Publishers, pp. 483-86.
George, A.; Knutson, E.; Tu, K.W. (1983) Radon Daughter Plateout. Health Phys. 45:439-444.
Gies, R.A.; Cross, F.T.; Dagle, G.E. (1972) A histological study of the influence of cigarette
smoking in suppressing Rn daughter carcinogenesis in dogs. Health Phys. 23:857-860.
Gillis, C.R.; Hole, D.J.; Hawthorne, V.M.; and Boyle, P. (1984) The effect of environmental
tobacco smoke in two urban communities in the west of Scotland. Eur J Respir Dis
133(Suppl.): 121-126.
Godin, G.; Wright, G.; Shephard, R.J. (1972) Urban exposure to carbon monoxide. Arch
Environ Health 25:305-313.
Goldstein, G.M.; Collier, A.; Etzel, R.; Lewtas, J.; Haley, N. (1987) Elimination of urinary
cotinine in children exposed to known levels of side-stream cigarette smoke. Seifert, H.;
Esdorn, M.; Fischer, H.; Ruden, J.; Wegner, eds. In: Indoor Air '87, Vol 2, B.
Oraniendruck GmbH, Berlin, pp. 61-67.
Goodman, R. M.; Yergin, B. M.; Landa, J. F.; Golinvaus, M. H.; Sackner, M. A. (1978)
Relationship of smoking history and pulmonary function tests to tracheal mucus velocity in
non-smokers, young smokers, ex-smokers and patients with chronic bronchitis. Am Rev
Resp Dis 117:205-214.
Goodman, D.G.; Ward, J.M.; Squire, R.A.; Paton, M.B.; Reichardt, W.D.; Chu, K.C.; Linhart,
M.S. (1980) Neoplastic and nonneoplastic lesions in aging Osborne-Mendel rats. Toxicol
Appl Pharmacol 55:433-447.
Graham, S.; Levin, M.L. (1971) Smoking withdrawal in the reduction of risk of lung cancer.
Cancer 27:865-871.
Greenberg, R.A.; Bauman, K.E.; Glover, L.H.; Strecher, V.J.; Kleinbaum, D.G.; Haley, N.J.;
Steadman, H.C.; Fowler, M.G.; Loda, F.A. (1989) Ecology of passive smoking by young
infants. J Pediatr 114:774-780.
Grimmer, G.; Brune, H.; Dettbarn, G.; Naujack, K.W.; Mohr, U.; Wenzel-Hartung, R. (1988)
Contribution of polycyclic aromatic compounds to the carcinogenicity of sidestream smoke
of cigarettes evaluated by implantation into the lungs of rats. Cancer Letters 43:173-177.
R-9 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Grimmer, G.; Brune, H.; Deutsch-Wenzel, R.; Dettbarn, G.; Jacob, J.; Naujack, K.W.; Mohr, U.;
Ernst, H. (1987) Contribution of polycyclic aromatic hydrocarbons and nitro-derivatives
to the carcinogenic impact of diesel engine exhaust condensate evaluated by implantation
into the lungs of rats. Cancer Letters 37:173-180.
Hammond, E.G. (1966) Smoking in relation to the death rates of one million men and women.
In: Haenszel, W., ed. Epidemiological Approaches to the Study of Cancer and Other
Chronic Diseases. National Cancer Institute Monograph No. 19, Washington, D.C., pp.
127-204.
Hammond, B.C.; Garfinkel, L. (1975) Aspirin and coronary heart disease: findings of a
prospective study. Br Med J 2:269- 271.
Hammond, E.G.; Garfinkel, L.; Seidman, H.; Lew, E.A. (1976) 'Tar' and nicotine content of
cigarette smoke in relation to death rates. Environmental Research 12:263-274.
Harke, H. P. (1974) The problem of passive smoking. In: The Influence of smoking on the CO
concentration of office rooms. Int Arch Arbeitsmed 33:199-206.
Harlap, S.; Davies, A.M. (1974) Infant admissions to hospital and material smoking. Lancet
l(7857):529-523.
Harmsen, H.; Effenberger, E. (1957) Tobacco smoke in public transportation, dwellings and
work rooms. Arch Hug 141:383-400.
Harley, N.H.; Pasternack, B.S. (1981) A model for predicting lung cancer risks induced by
environmental levels of radon daughters. Health Phys. 40:307-316.
Harley, N.H.; Pasternack, B.S. (1982) Environmental radon daughter alpha dose factors in five-
lobed human lung. Health Phys. 42:788-799.
Hasselbad, V.; Anderson, H.S.; Humble, C.G.; Graham, M.G. (1981) Indoor environmental
determinants of lung function in children. American Review of Respiratory Disease
123(5):479-485.
Henderson, F.W.; Reid, H.F.; Morris, R.; Wang, O.; Hu, P.C.; Helms, R.W.; Forehand, L; et al.
(1989) Home air nicotine levels and urinary cotinine excretion in preschool children.
American Review of Respiratory Disease 140:197-201.
Henderson, F.W.; Morris, R.; Reid, H.F.; Hu, P.C. (1987) Serum and urine cotinine as
quantitative measures of passive tobacco smoke exposure in young children. University of
North Carolina, Chapel Hill, N.C.
Herrmann, D.R.; Olsen, K.M.; Hiller, F.C. (1989) Nicotine absorption following pulmonary
installation. Presented at the American Thoracic Society Meeting. Cincinnati, Ohio. May
1989.
R-10 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Killer, F.C.; McCusker, K.T.; Mazumder, M.D.; Wilson, J.D.; Bone, R.C. (1982) Deposition of
sidestream cigarette smoke in the human respiratory tract. Am Rev Resp Dis 125:406-
408.
Hinds, W.C.; First, M.W. (1975) Concentrations of nicotine and tobacco smoke in public
places. New Engl J Med 292:844-845.
Hinton, A.E. (1989) Surgery for otitis media with effusion in children and its relationship to
parental smoking. J Laryngol Otol 103(6):559-61.
Hirayama, T. (1977) Smoking and cancer: A prospective study on cancer epidemiology based on
a census population in Japan. In: Smoking and Health, vol. II. Washington DC, U.S.
Public Health Service, DHEW Publ. No. (NIH)77-1413, pp. 65-72.
Hirayama, T. (198la) Non-smoking wives of heavy smokers have a higher risk of lung cancer:
A study from Japan. Br Med J 282: 183-185.
Hirayama, T. (1981b) (Letter) Br Med J 283:916-917.
Hirayama, T. (1983a) Passive smoking and lung cancer: consistency of association. Lancet
2:1425-1426.
Hirayama, T. (1983b) Passive smoking and lung cancer. Presented at the 5th World
Conference on Smoking and Health, Winnipeg, Canada.
Hirayama, T. (1984) Cancer mortality in nonsmoking women with smoking husbands based on
a large-scale cohort study in Japan. Prev Med 13:680-690.
Hoffmann, D.; Wynder, E.L. (1986) Chemical constituents and bioactivity of tobacco smoke.
Zavidze, D.G. and Peto, R., eds. Tobacco: A Major International Health Hazard. Lyon,
France: IARC Scientific Pub. No. 74, pp. 145-165.
Hoffmann, D.; Wynder E.L.; Hecht, S.S.; Brunnemann, K.D.; LaVoie, E.J.; Haley, N.J. (1987)
Chemical carcinogens in tobacco. Bannasch, P., ed. In: Cancer Risks: Strategies for
Elimination. Berlin: Springer-Verlag, pp. 101-113.
Hoffman, D.; Brunnemann, K.D.; Haley, N.J. (1989) Absorption of smoke constituents by
nonsmokers. In: Reducing Workplace Exposures to Environmental Tobacco Smoke.
EPA-HHS Manual (in press).
Hoffmann, D.; Hecht, S.S. (1989) Advances in tobacco carcinogenesis. In: Springer Handbook
of Experimental Pharmacology: Chemical Carcinogenesis and Mutagenesis (in press).
Hole, DJ; Gillis, C.R.; Chopra, C; Hawthorne, V.M. (1989) Passive smoking and
cardiorespiratory health in a general population in the west of Scotland. British Med.
Journal 299:423-427.
Hollander, M.; Wolfe, D.A. (1973) Nonparametric statistical methods. New York: John Wiley
and Sons.
R-ll 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Houthuijs, D.; Remijn, B.; Brunekreef, B.; de Koning, R. (1987) Exposure to nitrogen dioxide
and tobacco smoke and respiratory health of children. In: Indoor Air '87: Proceedings of
The 4th International Conference on Indoor Air Quality and Climate 1:463-467.
Hulka, B.S. (1988) The health consequences of environmental tobacco smoke. Environ
Technol Lett 9(6):531-538.
Humble, C.G.; Samet, J.M.; Pathak, D.R. (1987) Marriage to a smoker and lung cancer risk.
Am J Public Health 77:598-602.
Husgafvel-Pursiainen, K.; Sorsa, M; Moller, M; Benestad, C. (1986) Genotoxicity and
polynuclear aromatic hydrocarbon analysis of environmental tobacco smoke samples from
restaurants. Mutagenesis 1:287-291.
Ingebrethson, B.J.; Sears, S.B. (October 2-5, 1985) Particle size distribution measurements of
sidestream cigarette smoke. Presented at the 39th Tobacco Chemists' Research
Conference, Montreal, Canada.
Inoue, R.; Hirayama, T. (1988) Passive smoking and lung cancer in women. In: Smoking and
Health. Elsevier Science Publishers, pp. 283-285.
International Agency for Research on Cancer (IARC). (1987) Environmental carcinogens-
methods of analysis and exposure measurement Volume 9: Passive Smoking (O'Neill,
I.K.; Brunnemann, K.D.; Dodet, B.; and Hoffmann, D. eds.) Lyon: IARC Scientific
Publications No. 81.
International Commission on Radiological Protection, ICRP Task Group on Lung Dynamics.
(1966) Deposition and retention models for internal dosimetry of the human respiratory
tract. Health Phys. 12:173-207.
Ishizu, Y.; Ohta, K.; Okada, T. (1980) The effect of moisture on the growth of cigarette smoke
particles. Beitr Tabakforsch 10:16.
Jacob P.; Benowitz, N.; Alexander, T. (1988) Recent studies of nicotine metabolism in humans.
Pharmacology Biochemistry and Behavior 30:249-253.
Janoff, A.; Pryor, W.A.; Bengali, B.H. (1987) Effects of tobacco smoke components on cellular
and biochemical processes in the lung. Am Rev Resp Dis 36:1058-1064.
Jarvis, M.D.; Turnstall-Pedoe, H.; Feyerabend, C; Vesey, C.; Saloojee, Y. (1984) Biochemical
markers of smoke absorption and self-reported exposure to passive smoking. J Epid
Comm Health 38:335-339.
Jarvis, M.J.; Russell, M.A.; Feyerband, C. et al. (1985) Passive exposure to tobacco smoke:
Saliva cotinine concentrations in a representative population sample of nonsmoking school
children. Br. J. Med. 291:927-929.
R-12 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Jarvis, M.J.; Russell, M.A.H.; Feyerabend, C; Eiser, J.R.; Morgan, M.; Gammage, P.; Gray,
E.M. (1985) Passive exposure to tobacco smoke: Saliva cotinine concentrations in a
representative population sample of nonsmoking schoolchildren. Br. Med. J. 291:929.
Jarvis, M.J.; McNeill, A.D.; Russell, M.A.H.; West, A.; Bryant, A.; Geyerabend, C. (1987)
Passive smoking in adolescents: One-year stability of exposure in the home. Lancet
1:1324-1325.
Jarvis, M.; Russell, M.A.; Benowitz, N.L.; Feyerabend, C. (1988) Elimination of cotinine from
body fluids: implications for noninvasive measurement of tobacco smoke exposure. Am J
Public Health 78:696-698.
Jarvis, MJ. (1989) Application of biochemical intake markers to passive smoking measurement
and risk estimation. Mutation Research 222:101-110.
Just J.; Borkowska, M.; Maziarka, S. (1972) Air pollution by tobacco smoke in Warsaw coffee
houses. Rocz Pantstw Zak Hig 23:129-135.
Kabat, G.C.; Wynder, E.L. (1984) Lung cancer in nonsmokers. Cancer 53:1214-1221.
Kahn, H.A. (1966) The Dorn study of smoking and mortality among U.S. veterans: Report on
eight and one-half years of observation, In: Haenszel, W., ed. Epidemiological
Approaches to the Study of Cancer and Other Chronic Diseases. National Cancer
Institute Monograph No. 19, Washington, D.C., pg. 1.
Kallail, K.J.; Rainbolt, H.R.; Bruntzel, M.D. (1987) Passive smoking and middle ear problems in
Kansas public school children. J. Commun. Disord. 20:187-196.
Kasuga, H.; Hasebe, A.; Matsuki, H.; Osaka, F. (1979) Respiratory symptoms in school children
and the role of passive smoking. Tokai Journal of Experimental and Clinical Medicine
Kauffmann, F.; Oriol, P.; Tessier, J.-F. (1983) Adult passive smoking in the home environment:
A risk factor for chronic airflow limitation. American Journal of Epidemiology
117(3):269-280.
Kauffmann, F.; Tager, I.E.; Munoz, A.; Speizer, F.E. (1989) Familial factors related to lung
function in children aged 6-10 years: Results from the PAARC epidemiologic study. Am
J Epidemiol 129(6): 1289-1299.
Keith, C.H.; Derrick, J.D. (1960) Measurement of the particle size distribution and
concentration of cigarette smoke by the confuge. J Colloid Sci 15:340-356.
Kentner, M.; Triebig, G.; Weltle, D. (1984) The influence of passive smoking on pulmonary
function: A study of 1351 office workers. Preventive Medicine 13(6):656-669.
Kilpatrick, S.J. (1987) Misclassification of environmental tobacco smoke exposure: its potential
influence on studies of environmental tobacco smoke and lung cancer. Toxicology Letters
35:163-168.
R-13 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Kleinman, J.C.; Pierre, M.B.; Madans, J.H.; Land, G.H.; Schramm, W.F. (1988) The effects of
maternal smoking on fetal and infant mortality. Am. J. Epidemiol. 127(2) :274-282.
Knoth, A.; Bohn, H.; Schmidt, F. (1983) Passive smoking as a causal factor of bronchial
carcinoma in female nonsmokers. Medizinisch Klin 78:66-69.
Koo, L.C.; Ho, J.H.; Saw, D.; Ho, C.Y. (1987) Measurements of passive smoking and estimates
of lung cancer risk among non-smoking Chinese females. Int J Cancer 39:162-169.
Koo, L.C.; Ho, J.H-C; Matsuki, H.; Shimizu, H.; Mori, T.; Tominaga, S. (1988) A comparison
of the prevalence of respiratory illnesses among nonsmoking mothers and their children in
Japan and Hong Kong. Am Rev Respir Dis 138:290-295.
Kraemer, M.J.; Bierman, C.W.; Furukawa, C.T.; Pierson, W.E.; Richardson, M.A.; Shapiro,
G.G.; Weiss, N.S. (1983) Risk factors for persistent middle-ear effusions: Otitis media,
catarrh, cigarette smoke exposure, and atopy. Journal of the American Medical
Association 249(8): 1022-1025.
Kraus, J.F.; Greenland, S.; Bulterys, M. (1989) Risk factors for sudden infant death syndrome in
the US Collaborative Perinatal Project. Int. J. Epidemiol. 18(1): 113-120.
Kuller, L.H.; Garfinkel, L.; Correa, P.; Haley, N.; Hoffmann, D.; Preston-Martin, S.; Sandier, D.
(1986) Contribution of passive smoking to respiratory cancer. Environ Health
Perspectives 70:57-69.
Kyermater, G.A.; Damiano, M.D.; Dovorchik, B.H.; Vesell, D.S. (1982) Smoking-induced
changes in nicotine disposition: application of a new HPLC assay for nicotine and its
metabolites. Clin Pharm Ther 32:769-780.
Lam, T. H.; Kung, I.T.M.; Wong, CM.; Lam, W.K.; Kleevens, J.W.L.; Saw, D.; Hsu, C;
Seneviratne, S.; Lam, S.Y.; Lo, K.K.; Chan, W.C. (1987) Smoking, passive smoking and
histological types in lung cancer in Hong Kong Chinese women. Br J Cancer 6:673-378.
Lam, W.K. (1985) A clinical and epidemiological study of carcinoma of lung in Hong Kong.
Thesis submitted to the University of Hong Kong for the degree of Doctor of Medicine
Layard, M.W.; Viren, J.R. (1989) Assessing the validy of a Japanese cohort study. Elsevier
Science Publishers B.V. (Biomedical Division).
Leaderer, B.P. (1988) Measuring exposure to environmental tobacco smoke. Report prepared
for the U. S. Environmental Protection Agency.
Leaderer, B.P.; Cain, W.S.; Isseroff, R.; Berglund, L.G. (1984) Ventilation requirements in
buildings II. Particulate matter and carbon monoxide from cigarette smoking. Atmos.
Environ., 18, 99-106.
Leaderer, B. (1990) Assessing exposure to environmental tobacco smoke. Risk Analysis, Vol.
10, No. 1.
R-14 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Lebowitz, M.D.; Knudson, R.J.; Burrows, B. (1975) The Tucson epidemiology study of chronic
obstructive lung disease, I. Methodology and prevalence of disease. Am J. Epidemiol.
102:137-152.
Lebowitz, M.D.; Burrows, B. (1976) Respiratory symptoms related to smoking habits of family
adults. Chest 69(1):48-50.
Lebowitz, M.D.; Armet, D.B.; Knudson, R. (1982) The effect of passive smoking on pulmonary
function in children. Environment International 8(l-6):371-373.
Lebowitz, M.D.; Holberg, C.J.; Knudson, R.J.; Burrows, B. (1987) Longitudinal study of
pulmonary function development in childhood, adolescence, and early adulthood. Am.
Rev. Resp. Dis. 136:69-75.
Lebowitz, M.D.; Holberg, CJ. (1987) Physiological changes in pulmonary development related
to passive smoking and its interaction with active smoking. Toxicology Letters 35:101-105.
Lebowitz, M.D.; Holberg, C.J. (1988) Effects of parental smoking and other risk factors on the
development of pulmonary function in children and adolescents. American Journal of
Epidemiology 128(3) :589-597.
Lee, B.L.; Benowitz, N.L.; Jacob P. (1987) Influence of tobacco abstinence of the disposition
kinetics and effects of nicotine. Clin Pharm Ther 41:474-479.
Lee, P.N.; Chamberlain, J.; Alderson, M.R. (1986) Relationship of passive smoking to risk of
lung cancer and other smoking-associated diseases. Br J Cancer 54:97-105.
Lee, P.N. (1987a) Lung cancer and passive smoking: Association an artefact due to
misclassification of smoking habits. Toxicology Letters 35:157-162.
Lee, P.N. (1987b) Passive smoking and lung cancer association: a result of bias? Human
Toxicol 6:517-524.
Lee, P.N. (1988) Misclassification of smoking habits and passive smoking. Berlin: Springer.
Leeder, S.R.; Corkhill, R.T.; Irwig, L.M.; Holland, W.W. (1976a) Influence of family factors on
asthma and wheezing during the first five years of life. In: British Journal of Preventive
and Social Medicine 30(4):213-218.
Leeder, S.R.; Corkhill, R.T.; Irwig, L.M.; Holland, W.W. (1976b) Influence of family factors on
the incidence of lower respiratory illness during the first year of life. British Journal of
Preventive and Social Medicine 30(4):203-212.
Lehnert, G. (1984) Roundtable discussion from Symposium "Medical Perspectives on Passive
Smoking." Prev Med 13:732- 733.
Leung, H.M.; Kupper, L.L. (1981) Comparisons of confidence intervals for attributable risk.
Biometrics 37:293-302.
R-15 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Lindblad, A.; Marsal, K.; Andersson, K.-E. (1988) Effect of nicotine on human fetal blood flow.
Obstet. Gynecol. 72(3)1:371-382.
Lubin, J.H., Blot, W.J., Berrino, F., Flamant, R., Gillis, C.R., Kunze, M., Schmahl, D., Visco, G.
(1984) Modifying risk of developing lung cancer by changing habits of cigarette smoking.
BrMed 1288:1953-1956.
Luchsinger, P.C.; LaGarde, B.; Kilfeathers, J.E. (1968) Particle clearance from the human
tracheobronchial tree. Am. Rev Resp Dis 97:1046.
Mantel, N. (1963) Chi-square tests with one degree of freedom: extensions of the
Mantel-Haenszel procedure. J Am Statist Assoc 58:690-700.
Mantel, N. (1987a) Lung cancer and passive smoking. (Letter) Br J Med 294:440-441.
Mantel, N. (1987b) Re: Passive smoking and lung cancer in Swedish women (Letter) Am J
Epidemiol 126(6): 1214-1216.
Mantel, N.; Haenszel, W. (1959) Statistical aspects of the analysis of data from retrospective
studies of disease. J Natl Cancer Inst 22:719-748.
Maritz, G.S. (1988) Effect of maternal nicotine exposure on growth in vivo of lung tissue of
neonatal rats. Biol. Neonate, Foetal Neonatal Res.53(3):163-170.
Marks, B.E. (1988) Respiratory illness and home environment. British Medical Journal
296(6638):1740.
Martin, T.R.; Bracken, M.B. (1986) Association of low birth weight with passive smoke exposure
in pregnancy. American Journal of Epidemiology 124(4):633-642.
Martinez, F.D.; Antognoni, G.; Macri, F.; Bonci, E.; Midulla, F.; De Castro, G.; Ronchetti, R.
(1988) Parental smoking enhances bronchial responsiveness in nine-year-old children. Am.
Rev. Respir. Dis. 138(3):518-523.
Martonen, T.; Hofmann, W. (1986) Factors to be considered in a dosimetry model for risk
assessment of inhaled particles. Radiation Protection Dosimetry 15:225-232.
Masi, M.A.; Hanley, J.A.; Ernst, P.; Becklake, M.R. (1988) Environmental exposure to tobacco
smoke and lung function in young adults. Am. Rev Resp Dis 138:296-299.
Matsuki, H.; Kasuga, H.; Osaka, F. et al. (1985) A comparative study on the health effects of
smoking and indoor air pollution in summer and winter. In: Tokai Journal of
Experimental and Clinical Medicine 10(4):427-437.
Matsukura,S.; Taminato, T.; Kitano, N. et al. (1984) Effects of environmental tobacco smoke on
urinary cotinine excretion in nonsmokers. N Engl J Med 311:828-832.
McConnochie, K.M.; Roghmann, K.J. (1986) Breast feeding and maternal smoking as
predictors of wheezing in children age 6 to 10 years. Pediatr Pulmonol 2:260-268.
R-16 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
McDowell, E.M.; Trump, B.F. (1983) Histogenesis of preneoplastic and neoplastic lesions in
tracheobronchial epithelium. Survey and Synthesis of Path. Res., 2, 235-279.
Medina, E.; Medina, R.; Kaempffer, A.M. (1988) Smoking and incidence of respiratory diseases
in children. Rev. Chil. Pediatr. 59(l):60-64.
Meyer, M.B., Tonascia J.A., Buck, C. (1974) The interrelationship of maternal smoking and
increased perinatal mortality with other risk factors. Further analysis of the Ontario
perinatal mortality study, 1960-1961. American Journal of Epidemiology 100:443-452.
Miettinen, O.S. (1974) Proportion of disease caused or prevented by a given exposure, trait or
intervention. Am J Epidemiol 99: 325-332.
Miller, G.H. (1984) Cancer, passive smoking and nonemployed and employed wives. Western
J Med 140(4):632-635.
Mitchell, R.I. (1962) Controlled measurement of smoke particle retention in the respiratory
tract. Am Rev Resp Dis 85:526-533.
Moghissi, A.A.; Seiler, F.A.; Dellarco, M.J. (1987) Enhancement of exposure to radon progeny
as a consequence of passive smoking. Proceedings of the 4th International Conference on
Indoor Air Quality and Climate, West Berlin, Volume 2, August 1987, pp. 360-364.
Moolgavkar, S.H.; Dewanji, A.; Luebeck, G. (1989) Cigarette smoking and lung cancer:
reanalysis of the British doctors data. J Am Cancer Inst. 81(6):415-420.
Moorhead, R. (1985) Passive smoking and young children's health. Australian Family Physician
14(10):1058-1062.
Muramatsu, M.; Umemura, S.; Okada, T.; Tomita, H. (1984) Estimation of personal exposure
to tobacco smoke with a newly developed nicotine personal monitor. Environ Res 35:218-
227.
Murray, A.B.; Morrison, B.J. (1986) The effect of cigarette smoke from the mother on bronchial
responsiveness and severity of symptoms in children with asthma. J Allergy Clin Immunol
77:575-581.
Murray, A.B.; Morrison, B.J. (1989) Passive smoking by asthmatics: Its greater effect on boys
than on girls and on older than on younger children. Pediatrics 84:451-459.
Murray, A.B.; Morrison, B.J. (1988) Passive smoking and the seasonal difference of severity of
asthma in children. Chest 94:701-708.
Muscati, S.K.; Mackey, M.A.; Newsom, B. (1988) The influence of smoking and stress on
prenatal weight gain and infant birth weight of teenage mothers. J. Nutr. Educ. 20(6):299-
302.
Nash, I.E.; Persaud, T.V.N. (1988) Influence of nicotine and caffeine on rat embryonic
development.
R-17 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
National Academy of Sciences. (1988) Health risks of radon and other internally deposited
alpha-emitters, BEIR IV, National Academy Press, Washington, D.C.
National Council on Radiation Protection and Measurement (NCRP) (1984) Evaluation of
occupational and environmental exposures to radon daughters in the United States.
NCRP Report No. 78, Bethesda, MD.
National Research Council (NRC) (1986a) Environmental tobacco smoke: Measuring
exposures and assessing health effects, National Academy Press, Washington, D.C.
National Research Council (NRC) (1986b) Air quality and safety: The airliner cabin
environment. National Academy Press, Washington, DC.
Nazaroff, W.W.; Cass, G.R. (1989) Mathematical modeling of indoor aerosol dynamics.
Environ Sci Technol 23:157-166.
Neal, A.D.; Wadden, R.A.; Rosenberg, S. (1978) Evaluation of indoor particulate
concentrations for an urban hospital. Am Ind Hyg Assoc J 29:578-582.
Nelson, P.A.; Qvant, F.R.; Sem, G.F. (1982) Experimental measurements of aerosol
concentrations in offices. Environ Int 8:223-227.
Neuspiel, D.R.; Rush, D.; Butler, N.R.; Golding, J.; Bijur, P.E.; Kurzon M. (1989) Parental
smoking and post-infancy wheezing in children: A prospective cohort study. Am J Public
Health 79(2):168-171.
O'Connor, G.; Weiss, ST.; Tager, I.; Speizer, F.E. (1986) The effect of passive smoking on
nonspecific bronchial responsiveness in a population sample of children and young adults.
Clinical Research 34(2):581A (abstract).
Ogston, S.A.; Florey, C; Du V; Walker, C.H.M. (1987) Association of infant alimentary and
respiratory illness with parental smoking and other environmental factors. Journal of
Epidemiology and Community Health 41:21-25.
Ostro, B.D. (1988) Estimating the risks of smoking, air pollution, and passive smoke on acute
respiratory conditions. Risk Analysis 9(2): 189-196.
Paoletti, P; Prediletto, R.; Carozzi, L.; Viegi, G.; Di Pede, F.; Carmignani, G.; Mammini, U.;
Giuntini, C.; Lebowitz, M.D. (1989) Effects of childhood and adolescence-adulthood
respiratory infections in a general population. Eur Respir J 2:428-436.
Park, J.K.; Kim, I.S. (1986) Effects of family smoking on acute respiratory disease in children.
Yonsei Med J 27(4):261-270.
Pattishall, E.N.; Strope, G.L.; Etzel, R.A.; Helms, R.W.; Haley, N.J.; Denny, F.W. (1985)
Serum cotinine as a measure of tobacco smoke exposure in children. AJDC 139:1101-
1104.
R-18 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Pedreira, F.A.; Guandolo, V.L.; Feroli, E.J.; Mella, G.W.; Weiss, IP. (1985) Involuntary smoking
and incidence of respiratory illness during the first year of life. Pediatrics 75(3):594-597.
Pelech, L. (1989) Cooperation between the German Democratic Republic and the Czechoslovak
socialist Republic in the study of long-term air pollution effects in selected groups of
school-age children. Z. Gesamyte Hyg. Ihre Grenzgeb. 35(1):50.
Perry, J. (1973) Fasten your seat belts: no smoking. Br Columbia Med J 15:304-305.
Pershagen, G.; Hrubec, Z.; Svensson, C. (1987) Passive smoking and lung cancer in Swedish
women. Am J Epidemiol 125(1): 17-24.
Phillips, R.L.; Garfinkel, L.; Kuzma, J.W. et al. (1980a) Mortality among California Seventh-Day
Adventists for selected cancer sites. J Natl Cancer Inst 65:1097-1107.
Phillips, R.L.; Kuzma, J.W.; Beeson, W.L. et al. (1980b) Influence of selection versus lifestyle on
risk of fatal cancer and cardiovascular disease among Seventh-Day Adventists. Am J
Epidemiol 112: 296-314.
Pimm, P.E.; Silverman, F.; Shephard, R.J. (1078) Physiological effects of accute passive exposure
to cigarette smoke. Archives of Environmental Health 33(4):201-213.
Plant, M.L.; Plant, M.A. (1988) Maternal use of alcohol and other drugs during pregnancy and
birth abnormalities: Further results from a prospective study. Alcohol Alcolhol. 23(3):229-
233.
Polydorva, M. (1961) An attempt to determine the retention of tobacco smoke by means of
membrane filters. In: Davies, C.N. ed. Inhaled Particles and Vapours (Davies, C. N.,
ed.) Oxford: Pergammon Press, pp. 142-144.
Portheine, F. (1971) To the problem of passive smoking. Munch Med Wochenschr 18:707-709.
Pritchard, J.N.; Black, A.; McAughey, J.S. (1988) Physical behavior of sidestream smoke under
ambient conditions. Arch Environ Letters 9(6): 545-552.
Pullan, C.R.; Hey, E.N. (1982) Wheezing, asthma, and pulmonary dysfunction 10 years after
infection with respiratory syncytial virus in infancy. British Medical Journal
284(6330): 1665-1669.
Pulkander, J.; Luotonen, J.; Timonen, M.; Karma, P. (1985) Risk factors affecting the occurence
of acute otitis media among 2-3 year-old urban children. Acta Otolaryngolica 100(3-
4):260-265.
Raabe, O.G., (1969) Concerning the interactions that occur between radon decay products and
aerosols. Health Phys., 17, 177-187.
Radsel, Z.; Kambic, V. (1978) The influence of cigarette smoke on the pharyngeal mucosa.
Acta. Otolaryngol, 85:128-134.
R-19 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Randerath, E.; Avitts, T.A.; Reddy, M.V.; Miller, R.H.; Everson, R.B.; Randerath, K. (1986)
Comparative 32P-analysis of cigarette smoke-induced DNA damage in human tissues and
mouse skin. Cancer Res 46:5869-5877.
Rantakallio, P. (1978) Relationship of maternal smoking to morbidity and mortiality of the child
up to the age of five. Acta Paediatrica Scandinavica 67(5):621-631.
Redmond, C.K.; Ciocco, A.; Lloyd, J.W.; Rush, H.W. (1972) Long term mortality study of
steelworkers. VI. Mortality from malignant neoplasms among coke oven workers. J
Occup Med 14:621-629.
Reed, B.D.; Lutz, L.J. (1988) Household smoking exposure-association with middle ear
effusions. Fam Med 20(6):426-430.
Remmer, H. (1987) Passively inhaled tobacco smoke: a challenge to toxicology and preventive
medicine. Arch Toxicol 61:89-104.
Repace, J.L. (1987) Indoor concentrations of environmental tobacco smoke: field surveys. In:
O'Neill, I.K.; Brunnemann, K.D.; Dodet B.; Hoffmann, D. eds. Environmental
Carcinogens: Methods of Analysis and Exposure Measurements Vol.9 : Passive Smoking.
Lyon, France: IARC Scientific Pub. No. 81, pp. 141-162.
Repace, J.L.; Lowrey, A.H. (1980) Indoor air pollution, tobacco smoke and public health.
Science 208:464-472.
Repace, J. L., Lowrey, A. H. (1982) Tobacco smoke, ventilation and indoor air quality. Am
Soc Heat Refrig Air Cond Eng Trans 88:894-914.
Repace, J.L.; Lowrey, A.H. (1985) A quantitative estimate of nonsmokers' lung cancer risk
from passive smoking. Environ Int'l. 11:3-22.
Repace, J.L.; Lowrey, A.H. (1987) Predicting the lung cancer risk of domestic passive smoking
(Letter). Amer Rev Resp Dis 136:1308.
Repace, J.L. (1989) Workplace restrictions on passive smoking: Justification on the basis of
cancer risk (in press).
Repace, J.L.; Lowrey, A.H. (1990) Risk assessment methodologies in passive smoking. Journal
of Risk Analysis Vol. 10, No. 1.
Riboli, E. (1987) Passive smoking and lung cancer: Epidemiological evidence and ongoing
international collaborative studies. Toxicology Letters 35:19-27.
Rickert, W.S.; Robinson, J.C.; Collinshaw, N.E. (1984) Sidestream yields of tar, nicotine and
carbon monoxide from 15 brands of Canadian cigarettes. Am J Pub Health 74:228-231.
Robins, J.M.; Blevins, D.; Schneiderman, M. (1989) The effective number of cigarettes inhaled
daily by passive smokers: Are epidemiologic and dosimetric estimates consistent? Journal
of Hazardous Materials 21:215-238.
R-20 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Rogot, E. (1974) Smoking and mortality among U.S. veterans. J Chronic Dis 27:189-203.
Rona, R.J.; Chinn, S.; Du Ve Florey, C. (1985) Exposure to cigarette smoking and children's
growth. In: International Journal of Epidemiology 14(3):402-408.
Rose, K.J.; Strand, F.L. (1987) Response of the developing neuromuscular system of the rat to
nicotine and the neurotropic peptide fragment ACTH/MSH 4 - 10. Ann. New York
Acad. Sci. 494:319-322.
Rubin, B.K. (1986) Summer camp and passive smoking: Effects on nutrition, lung function and
self-image in children with cystic fibrosis. Chest 89(6):480s.
Rubin, D.H.; Damus, K. (1988) The relationship between passive smoking and child health:
Methodologic criteria applied to prior studies. Yale J Biol Med 61(5) :401-411.
Rubin, D.H.; Krasilnikoff, P.A.; Leventhal, J.M.; Weile, B.; Berget, A. (1986) Effect of passive
smoking on birth-weight. The Lancet Aug. 23.
Rudnick, S.N.; Maher, E.F., (1986) Surface deposition of Rn-222 decay products with and
without enhanced air motion. Health Phys. 51:283-293.
Rush, D.; Cassano, P. (1970) Relationship of cigarette smoking and social class to birth weight
and perinatal mortality among all births in Britain. Journal of Epidemiology and
Community Health 37:249-55.
Rush, D.; Cassano, P.; Harlap, S. (1988) Perinatal outcome, maternal weight gain, cigarette
smoking and social status in Jerusalem. Rev. Epidemiol. Sante Publ. 36(3): 186-195.
Rush, D.; Callahan, K.R. (1989) Exposure to passive cigarette smoking and child development.
A critical review. Ann N Y Acad Sci 562:74-100.
Russell, M.A.H.; Jarvis, M.J.; West, R.J. (1986) Use of urinary nicotine concentrations to
estimate exposure and mortality from passive smoking in non-smokers. Br. J. Addiction
81: 275-281.
Rylander, E.; Eriksson, M; Freyschuss, U. (1988) Risk factors from occasional and recurrent
wheezing after RSV infection in infancy. Acta Paediatr. Scand. 77(5):711-715.
Sackett, D.L. (1979) Bias in analytic research. J Chronic Dis 32(l-2):51-63.
Said,G.; Lellouch, J.; Patois, E.; Zalokar, J. (1978) Parental smoking related to adenoidectomy
and tonsillectomy in children. Journal of Epidemiology and Community Healt 32(2):97-
101.
Samet, J.M.; Tager, I.E.; Speizer, F.E. (1983) The relationship between respiratory illness in
childhood and chronic air-flow obstruction in adulthood. American Review of Respiratory
Disease 127(4):508-523.
R-21 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Samet, J.M. (1988a) Involuntary exposure to tobacco smoke. Annals of Sports Medicine
Samet, J.M. (1988b) Environmental tobacco smoke and cancer. Report prepared for the U.S.
Environmental Protection Agency.
Sandier, D.P.; Everson, R.B.; Wilcox, AJ. (1985) Passive smoking in adulthood and cancer risk.
Am J Epidemiol 121(1)37-48.
Saracci, R. (1989) Passive smoking and cancer risk. IARC report of panel of experts. Prepared
at the request of the European School of Oncology through the Europe Against Cancer
program of the European Economic Community.
Saracci, R.; Riboli, E. (1989) Passive smoking and lung cancer: Current evidence and ongoing
studies at the International Agency for Research on Cancer. Mutation Research
222:117-127.
Schenker, M.B.; Samet, J.M.; Speizer, F.E. (1983) Risk Factors for childhood respiratory
disease: The effect of host factors and home environmental exposures. American Review
of Respiratory Disease 128(6): 1038- 1043.
Schesselman, JJ. (1982) Case-control studies: design, conduct, analysis. Monographs in
Epidemiology and Biostatistics. New York: Oxford University Press, p. 190.
Schilling, R.S.F.; Letai, A.D., Hui, S.L.; Beck, G.J.; Schoenberg, J.B.; Bouhuys, A.H. (1977) Lung
function, respiratory disease, and smoking in families. American Journal of Epidemiology
106(4):274-283.
Schwartz-Bickenbach, D.; Schulte-Hobein, B.; Abt, S.; Plum, C; Nau, H. (1987) Smoking and
passive smoking during pregnancy and early infancy: Effects on birth weight, lactation
period, and cotinine concentrations in mother's milk and infant's urine. Toxicology
Letters 35:73-81.
Sebben, J.; Pimm, P; Shephard, R. J. (1977) Cigarette smoke in enclosed public facilities. Arch
Envir Health 32:53-58.
Sepkrovic, D.W.; Haley, N.J.; Hoffman, D. (1986) Elimination from the body of tobacco
products by smokers and passive smokers. J Am Med Assoc 256:863.
Shamsuddin, A.K.M., Sinopoli, N.T., Hemminki, K, Boesch, R.R., Harris, C.C. (1985)
Detection of benzo[a]pyrene: DNA adducts in human white blood cells. Cancer Res
45:66-68.
Shimizu, H.; Morishita, M.; Mizuno, K.; Masuda, T.; Ogura, Y.; Santo, M.; Nishimura, M.;
Kunishima, K.; Karasawa, K.; Nishiwaki, K.; Yamamoto, M.; Hisamichi, S.; Tominaga, S.
A case-control study of lung cancer in nonsmoking women. (1988) Toboku J Exp. Med
154:389-397.
R-22 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Sims, D.G.; Downham, M.A.P.S.; Gardner, P.S.; Webb, J.K.G.; Weightman, D. (1978) Study of
8-year-old children with a history of respiratory syncytial virus bronchiolitis in infancy.
British Medical Journal 1(6104):11-14.
Slavin, R.G.; Hertz M. (1975) Indoor air pollution: a study of the 30th annual meeting of the
American Academy of Allergy. Proceedings of the 30th Annual Academy of Allergy
Conference, p. 4.
Somerville, S.M.; Rona, R.J.; Chinn, S. (1988) Passive smoking and respiratory conditions in
primary school children. Journal of Epidemiology And Community Health 42(2): 105-110.
Speizer, F.E.; Ferris, B.,Jr. Bishop, Y.M.M.; Spengler, J. (1980) Respiratory disease rates and
pulmonary function in children associated with NO2 exposure. American Review of
Respiratory Disease 121(1):3-10.
Speizer, F.E. (1989) Diagnosis of chronic obstructive pulmonary disease. Am Rev of Respir Dis
140:S106-S107.
Spengler, J.D.; Dockery, D.W.; Turner, W.A.; Wolfson, J.M.; Ferris, E.G. (1981) Long-term
measurements of respirable sulphates and particles inside and outside homes. Atmos
Environ 15:23-30.
Spinacci, S.; Arossa, W.; Burgiani, M.; Nortale, P.; Bucca, C; DeCondission, E. (1985) The
effects of air pollution on the respiratory health of children: A cross-sectional study. In:
Pediatric Pulmonology l(5):262-266.
Stanton, M.F.; Miller, E.; Wrench, C.; Blackwell, R. (1972) Experimental induction of
epidermoid carcinoma in the lungs of rats by cigarette smoke condensate. J Natl Cancer
Inst 49:867-877.
Stehlik, G.; Rihter, O; Altmann, H. (1982) Concentration of dimethylnitrosoamine in the air of
smoke-filled rooms. Ecotoxical Environ Safety 6:495-500.
Sterling, T.D. (1983) Possible effects on occupational lung cancer from smoking related changes
in the mucus content of the lung. J Chronic Dis 36:669-676.
Stern, B.; Raizenne, M.; Burnett, R.; Kearney, J. (1987) Respiratory effects of early childhood
exposure to passive smoke. Indoor Air '87 Proceedings of the 4th International
Conference on Indoor Air Quality and Climates, Springer-Verlag New York, Inc.
Stevens, R.G. and Moolgavkar, S.H. (1984) A cohort analysis of lung cancer and smoking in
British males. Am J Epidemiol 119:624-641.
Stober, W. (1984) Lung dynamics and uptake of smoke constituents by nonsmokers: a survey.
Prev Med 13:589-601.
Strachan, D.P.; Jarvis, M.J.; Feyerabend, C. (1989) Passive smoking, salivary cotinine
concentrations, and middle ear effusion in 7 year old children. British Medical Journal
298:1549-1552.
R-23 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Svensson, C.; Pershagen, G.; Klominek, J. (1988) Smoking and passive smoking in relation to
lung cancer in women. Department of Epidemiology, National Institute of Environmental
Medicine, Stockholm, Sweden.
Swift, D.L. (1989) Seminar presented at the University of North Carolina at Chapel Hill
(unpublished).
Szadkoski, D.; Harke, H.P.; Angerer, J. (1976) Burden of carbon monoxide from passive
smoking in offices. Int Med 3:310-313.
Tachi, N.; Aoyama, M. (1989) Effects of cigarette smoke exposure on early stage embryos in the
rat. Bull. Environ. Contam. Toxicol. 43(3):467-472.
Tager, I.E.; Munoz, A.; Rosner, B.; Weiss, S.T.; Carey, V.; Speizer, F.E. (1985) Effect of
cigarette smoking on the pulmonary function of children and adolescents. American
Review of Respiratory Disease 131(5):752-759.
Tager, I.E.; Weiss, S.T.; Munoz, A.; Rosner, B.; Speizer, F.E. (1983) Longitudinal study of the
effects of maternal smoking on pulmonary function in children. New England Journal of
Medicine 309(12):699-703.
Tager, I.E.; Weiss, ST.; Rosner, B.; Speizer, F.E. (1979) Effect of parental cigarette smoking on
the pulmonary function of children. American Journal of Epidemiology 110(l):15-26.
Tager, I.E.; Segal, M.R.; Munoz, A.; Weiss, ST.; Speizer, F.E. (1987) The effect of maternal
cigarette smoking on the Pulmonary function of children and adolescents. Am Rev Respir
Dis 136:1366-1370.
Tager, I.E. (1986) Passive smoking and respiratory health in children: Sophistry or cause for
concern? American Review of Respiratory Disease 133(6):959-961.
Tainio, V.M.; Savilahti, E.; Salmenpera, L; Arjomaa, P. (1988) Risk facors for infantile
recurrent otitis media: Atopy but not type of feeding. Pediatr. Res. 23(5):509-512.
Tashkin, D.; Clark, V.A.; Simmons, M.; Reems, C.; Coulson, A.M.; Bourque, L.B.; Sayre, J.W.;
Detels, R.; Rokaw, S. (1984) The UCLA population studies of chronic obstructive
respiratory disease: 7. Relationship between parental smoking and children's lung function.
American Review of Respiratory Disease 129(6):891-897.
Teculescu, D.B.; Pham, Q.T.; Varona-Lopez, W.; Deschamps J.P.; Marchand, M.; Henquel, J.C.;
Manciaux, M. (1986) The single-breath nitrogen test does not detect functional
impairment in children with passive exposure to tobacco smoke. Bull. Eur. Physiopathol.
Respir. 22:605-607.
Tenovuo, A.H.; Kero, P.O.; Korvenranta, H.J.; Erkkola, R.U.; Klemi, P.J.; Tuominen, J. (1988)
Risk factors associated with severely small for gestational age neonates. Am. J. Perinatol.
5(3):267-271.
R-24 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Thorslund, T.W.; Charnley, G. (1987) Use of the multistage model to predict the carcinogenic
response associated with time-dependent exposure to multiple agents. Proceedings of the
ASA/EPA Conference on Current Assessment of Combined Toxicants. EPA-230-03-87-
027. Washington, D.C., Environmental Protection Agency, pp. 9-15.
Thorslund, T.W. (1990) Task 115- Obtaining unit risk estimates for inhaled B[a]P. Task 116 -
Development of a systematic approach for estimating PAH relative potencies. Prepared
for the Office of Health and Environmental Assessment. Under EPA contract no. 68-02-
4403. Preliminary Report.
Thyssen, J.; Altoff, J.; Kimmerle, G.; Mohr, U. (1981) Inhalation studies with benzo[a]pyrene in
Syrian golden hamsters. J Natl Cancer Inst 66:575-577.
Tredaniel, J; Hill, C; Chastang C; Hirsch A. (1989) Passive smoking and respiratory diseases.
Current data. Rev Mai Respir 6(2): 109-120.
Trichopoulos, D. (1988) Passive smoking and lung cancer. Scand J Soc Med 16:75-79.
Trichopoulos, D.; Kalandidi, A.; Sparros, L. (1983) Lung cancer and passive smoking:
conclusion of Greek study. (Letter) Lancet 667-668.
Tzoumaka-Bakoula, C.; Lekea-Karanika, V.; Matsaniotis, N.S.; Goldin, J. (1989) The Greek
national perinatal survey. II: Socioeconomic factors and perinatal mortality in Greece.
Paediatr. Perinat. Epidemiol. 3(l):41-52.
Uberia, K. (1987) Lung cancer from passive smoking: hypothesis or convincing evidence? Int
Arch Envir Occup Health 59:421- 437.
U.S. Department of Health and Human Services. (1984) The health consequences of smoking:
Chronic obstructive lung disease. A Report of the Surgeon General. DHHS Pub. No.
(PHS)84-50205. U.S. Department of Health and Human Services, Public Health Service,
Office of the Assistant Secretary for Health, Office on Smoking and Health, 1984.
U.S. Department of Transportation. (1971) Health aspects of smoking in transport aircraft.
U.S. DOT, Washington, DC, National Technical Information Service.
U.S. Environmental Protection Agency (EPA). (1984) Summary of the issues with regard to the
carcinogenicity, mutagenicity, and exposure of polycyclic organic matter. NTIS No. PB88-
220371.
U.S. Surgeon General (U.S. SG) (1986) The health consequences of involuntary smoking. U.S.
Department of Health and Human Services, Public Health Service.
U. S. Surgeon General (U.S. SG). (1988) The health consequences of smoking: nicotine
addiction. U.S. Department of Health and Human Services, Public Health Service.
• U.S. Surgeon General (U.S. SG) (1989) Reducing the health consequences of smoking. U.S.
~" Department of Health and Human Services, Public Health Service.
R-25 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Vainio, H.; Partanen, T. (1989) Population burden of lung cancer to environmental tobacco
smoke. Mutation Research 222:137-140.
Vandenbroucke, J.P. (1988) Passive smoking and lung cancer: a publication bias? Br J Med
296:391-392.
Varela, L.R. (1987) Assessment of the association between passive smoking and lung cancer.
Dissertation to Yale University in candidacy for the degree of Doctor of Philosophy.
Vedal, S.; Schenker, M.B.; Samet, J.M.; Speizer, F.E. (1984) Risk factors for childhood
respiratory disease: Analysis of pulmonary function. American Review of Respiratory
Disease 130(2): 187-192.
Vosmar, M.B.J.G.; Jongsma, H.W.; Van Dongen, P.W.J. (1989) The value of ultrasonic placental
grading: No correlation with intrauterine growth retardation or with maternal smoking. J.
Perinat. Med. 17(2): 137-143.
Vutuc, C. (1984) Quantitative aspects of passive smoking and lung cancer. Prev Med 13:
698-704.
Wald, N.J.; Idle, M.; Boreham, J.; Bailey, A.; Van Vonatcis, H. (1981) Serum cotinine levels in
pipe smokers: evidence against nicotine as a cause of coronary heart disease. Lancet,
2:775-777.
Wald, N.J.; Nanchahal, K.; Thompson, S.G.; Cuckle, H.S. (1986) Does breathing other people's
tobacco smoke cause lung cancer? Br J Med 293:1217-1222.
Wald, N.J., Ritchie, C. (1984) Validation of studies on lung cancer in nonsmokers married to
smokers. Lancet 1:1067.
Wald, N.; Doll, R.; Darby, S.; Kiryluk, S.; Peto, R.; Pike, M. eds. (1988) U.K. Smoking
Statistics. Oxford University Press.
Walter, S.D. (1975) The distribution of Levin's measure of attributable risk. Biometrika 62:
371-374.
Walter, S.D. (1976) The estimation and interpretation of attributable risk in health research.
Biometrics 32:829-849.
Wanner, W.A.; Hirsch, J.A.; Greeneltch, D.E.; Swenson, E.W.; Fore T. (1973) Tracheal mucus
velocity in beagles after chronic exposure to cigarette smoke. Arch Environ Health
27:370-371.
Ware, J.H.; Dockery, D.W.; Spiro, A.III; Speizer, F.E.; Ferris, B.G.,Jr.(1984) Passive smoking,
gas cooking, and respiratory health of children living in six cities. American Review of
Respiratory Disease 129(3):366-374.
Weber, A.; Fischer, T. (1980) Passive smoking at work. Int Arch Occup Environ Health
47:209-221.
R-26 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Weber A.; Fischer T.; Grandjean, A. (1979) Passive smoking in experimental and field
conditions. Environ Resp. 20:205-216.
Weibel, E.R. (1963) Morphometry of the human lung. New York, NY: Academic Press.
Weiss, ST.; Tager, I.E.; Speizer, F.E.; Rosner, B. (1980) Persistent wheeze: Its relation to
respiratory illness, cigarette smoking, and level of pulmonary function in a population
sample of children. American Review of Respiratory Disease 122(5):697-707.
Weiss, ST.; Tager, I.E.; Schenker, M.; Speizer, F.E. (1983) The health effects of involuntary
smoking. Am Rev Respir Dis 128:933-942.
Weiss, T.; Dorow, P.; Felix, R. (1983) Regional mucociliary removal of inhaled particles in
smokers with small airway disease. Respiration 44:338-345.
Weiss, ST. (1986) Passive smoking and lung cancer: what is the risk? (Editorial) Am Rev
Respir Dis 33:1-3.
Weitzman, M.; Gortmaker, S.; Walker, D.K.; Sobol, A. (1990) Maternal smoking and childhood
asthma. Pediatrics, April 1990, 505-511.
Wells, A.J. (1988a) Re: Passive smoking and lung cancer: a publication bias? (Letter) Br Med
1296:1128.
Wells, A.J. (1988) An estimate of adult mortality in the United States from passive smoking.
Environ. Int'l 14:249-265.
Wells, A.J. (1986) Misclassification as a factor in passive smoking risk. (Letter) Lancet 2:638.
Wells, A.J. (1990) An estimate of adult mortality in the United States from passive smoking: A
response to criticism. Environment International, Vol. 16.
White, J.R.; Froeb, H.F. (1980) Small-airways dysfunction in nonsmokers chronically exposed to
tobacco smoke. New England Journal of Medicine 302(13):720-723.
Whittemore, A.S. (1982) Statistical methods for estimating attributable risks from retrospective
data. Stat Med 1:229-243.
Whittemore, A.S. (1988) Effect of cigarette smoking in epidemiological studies of lung cancer.
Stat Med 7:223-238.
Wiedemann, H.P.; Mahler, D.A.; Loke, J.; Virgulto, J.A.; Snyder, P.; Matthay, R.A. (1986)
Acute effects of passive smoking on lung function and airway reactivity in asthmatic
subjects. Chest 89(2): 180-185.
Wigle, D.T.; Collishaw, N.E.; Kirkbride, J.; Mao, Y. (1987) Deaths in Canada from lung cancer
due to involuntary smoking. J. Can. Med. Assoc. 136: 945-.
R-27 05/17/90
-------
DRAFT-DO NOT QUOTE OR CITE
Willatt, D.J. (1986) Children's sore throats related to parental smoking. Clinical Otolaryngol
11:417- 321.
Witorsch, R.I.; Witorsch, P. (1989) A critical analysis of the relationship between parental
smoking and pulmonary performance in children. Off Gesundheitswes 51(2):78-83.
Woolf, B. (1955) On estimating the relationship between blood group and disease. Ann Human
Genet 19:251-253.
World Health Organization (WHO). (1986) Principles for evaluating health risks from
chemicals during infancy and early childhood: The need for a special approach.
Environmental Health Criteria 59.
Wu, A.H.; Henderson, B.E.; Pike, M.D.; Yu, M.C. (1985) Smoking and other risk factors for
lung cancer in women. J Natl Cancer Inst 74(4):747-751.
Wu, A.; Samet, J. Environmental tobacco smoke: Exposure-response relationships in
epidemilogic studies. Risk Analysis Vol. 10, No. 1.
Wynder, E.L.; Hoffman, D. (1967) Tobacco and tobacco smoke: studies in experimental
carcinogenesis. New York: Academic Press.
Yarnell, J.W.G.; St. Leger, A.S. (1979) Respiratory illness, maternal smoking habit and lung
function in children. British Journal of Diseases of the Chest 73(3):230-236.
Yusuf, S.; Peto, S.R.; Lewis, J.; Collins, R.; Sleight, P. (1985) Beta blockade during and after
myocardial infarction: an overview of the randomized trials. J Prog Cardiovasc Dis
27:335-371.
Zielhuis, G.A.; Heuvelmans-Heinen, E.W.; Rach, G.H.; Van den Broek, P. (1989)
Environmental risk factors for otitis media with effusion in preschool children. Scand. J.
Prim. Health Care 7(l):33-38.
R-28 05/17/90
------- |