<pubnumber>
600884003F
</pubnumber>
<title>Airborne Asbestos Health Assessment Update</title>
<pages>216</pages>
<pubyear>1984</pubyear>
<provider>NEPIS</provider>
<access>online</access>
<operator>mja</operator>
<scandate>20150415</scandate>
<origin>PDF</origin>
<type>single page tiff</type>
<keyword>asbestos exposure cancer lung risk mesothelioma fiber chrysotile fibers workers studies data mortality exposures years dose deaths occupational mcdonald crocidolite</keyword>
<author></author>
<publisher></publisher>
<subject></subject>
<abstract></abstract>
"United States
Environmenial Prelection
Agency
Office of Health and
Environmental Assessment
Weshington DC 2C460
EPA/SOO/8-84/003F
June 1986
Research and Development
EPA
Airborne Asbestos
Health Assessment
Update
image:
EPA/600/8-84/003F
June 1986
Airborne Asbestos
Health Assessment Update
Environmental Criteria and Assessment Office
Office of Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
image:
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental
Protection Agency policy and approved for publication. Mention of trade names
of commercial products does not constitute endorsement or recommendation for
use.
n
image:
TABLE OF CONTENTS
LIST OF TABLES vi
LIST OF FIGURES x
PREFACE xi
ABSTRACT xi i
AUTHORS, CONTRIBUTORS, AND REVIEWERS xiii
SCIENCE ADVISORY BOARD ENVIRONMENTAL HEALTH COMMITTEE xv
1. SUMMARY 1
2. INTRODUCTION 4
2.1 SUMMARY OF ASBESTOS HEALTH EFFECTS THROUGH 1972 6
2.1.1 Occupational Exposure 6
2.1.2 Environmental and Indirect Occupational Exposure
Circumstances , 9
2.1.3 Analytical Methodology 10
2.1.4 Experimental Studies 10
2.2 CURRENT ASBESTOS STANDARDS .., 11
3. HUMAN HEALTH EFFECTS ASSOCIATED WITH OCCUPATIONAL EXPOSURE
TO ASBESTOS 13
3.1 INTRODUCTION 13
3.2 MORTALITY ASSOCIATED WITH ASBESTOS EXPOSURE 13
3,2.1 Accuracy of Cause of Death Ascertainment 16
3.3 EPIDEMIOLOGICAL STUDIES OF ASBESTOS HEALTH EFFECTS:
STRENGTH OF THE EVIDENCE 16
3.4 MATHEMATICAL MODELS OF HUMAN CARCINOGENESIS 20
3.5 LINEARITY OF EXPOSURE-RESPONSE RELATIONSHIPS 23
3. 6 TIME AND AGE DEPENDENCE OF LUNG CANCER 32
3.7 MULTIPLE FACTOR INTERACTION WITH CIGARETTE SMOKING 40
3.8 METHODOLOGICAL LIMITATIONS IN ESTABLISHING DOSE-RESPONSE
RELATIONSHIPS 42
3.9 QUANTITATIVE DOSE-RESPONSE RELATIONSHIPS FOR LUNG CANCER ... 46
3.9.1 Textile Products Manufacturing, United States
(Chrysotile); Dement et al. (1982, 1983a, 1983b) ... 51
3.9.2 Textile Products Manufacturing, United States
(Chrysotile); McDonald et al. (1983a) 55
3.9.3 Textile Products Manufacturing, Rochdale, England
(Chrysotile); Peto (1980) 56
3.9.4 Textile and Friction Products Manufacturing:
United States (Chrysotile, Amosite, and
Crocidolite); McDonald et al. (1983b); Robinson
et al. (1979) 60
3.9.5 Friction Products Manufacturing, Great Britain
(Chrysotile and Crocidolite); Berry and Newhouse
(1983) 61
3.9.6 Friction Products Manufacturing, United States
(Chrysotile); McDonald et al..(1984) 63
iii
image:
TABLE OF CONTENTS (continued)
3.9.7 Mining and Milling, Quebec, Canada (Chrysotile);
Liddell et al. (1977); McDonald et al. (1980) 65
3.9.8 Mining and Milling, Thetford Mines, Canada
(Chrysotile); Nicholson (1976b); Nicholson et al.
(1979) 67
3.9.9 Mining and Milling, Italy (Chrysotile); Rubino
et al. (1979) 68
3.9.10 Insulation Manufacturing, Paterson, NJ
(Amosite); Seidman et al. (1979) 69
3.9.11 Insulation Application, United States (Chrysotile
and Amosite) 71
3,9.12 Asbestos Products Manufacturing, United States
(Chrysotile and Crocidolite); Henderson and
Enterline (1979) 74
3.9.13 Asbestos Cement Products, United States (Chrysotile
and Crocidolite); Wei 11 et al. (1979); Hughes and
Weill (1980) , 75
3.9.14 Asbestos Cement Products, Ontario, Canada
(Chrysotile and Crocidolite); Finkelstein (1983) ... 76
3.9.15 Lung Cancer Risks Estimated in Other Reviews 78
3.9.16 Summary of Lung Cancer Dose-Response Relationships . 80
3.10 TIME AND AGE DEPENDENCE OF MESOTHELIOMA 82
3.11 QUANTITATIVE DOSE-RESPONSE RELATIONSHIPS FOR MESOTHELIOMA .. 86
3.11.1 Insulation Application; Selikoff et al (1979);
Peto et al. (1982) 90
3.11.2 Amosite Insulation Manufacturing; Seidman et al.
(1979) 90
3.11.3 Textile Products Manufacturing; Peto (1980); Peto
et al. (1982) 90
3.11.4 Asbestos Cement Products, Ontario, Canada;
Finkelstein (1983) 91
3.11.5 Other Studies 91
3.11.6 Summary of Mesothelioma Dose-Response
Relationships , 91
3.12 ASBESTOS CANCERS AT EXTRATHORACIC SITES 96
3.13 ASBESTOSIS 99
3,14 MANIFESTATIONS OF OTHER OCCUPATIONAL EXPOSURES TO ASBESTOS . 102
3,15 DEPOSITION AND CLEARANCE 102
3.15.1 Models of Deposition and Clearance 104
3.16 EFFECTS OF INTERMITTENT VERSUS CONTINUOUS EXPOSURES 104
3.17 RELATIVE CARCINOGENICITY OF DIFFERENT ASBESTOS VARIETIES ... 106
3.17.1 Lung Cancer 107
3.17.1.1 Occupational Studies 107
3.17.1.2 Environmental Exposures 108
3.17.2 Mesothel ioma 110
3.17.2.1 Occupational Exposures 110
3.17.2.2 Environmental Exposures 116
3.18 SUMMARY 117
IV
image:
TABLE OF CONTENTS (continued)
Page
4. EXPERIMENTAL STUDIES 119
4.1 INTRODUCTION . 119
4.2 FIBER 'DEPOSITION AND CLEARANCE 119
4.3 CELLULAR ALTERATIONS 125
4.4 MUTAGENICITY 125
4. 5 INHALATION STUDIES 126
4.6 INTRAPLEURAL ADMINISTRATION 132
4.7 INTRATRACHEAL INJECTION 137
4.8 INTRAPERITONEAL ADMINISTRATION 137
4. 9 TERATOGENICITY '. 141
4.10 SUMMARY 141
5. ENVIRONMENTAL EXPOSURES TO ASBESTOS 142
5.1 INTRODUCTION 142
5. 2 GENERAL ENVIRONMENT 146
5.3 CHRYSOTILE ASBESTOS CONCENTRATIONS NEAR CONSTRUCTION
SITES 147
5.4 ASBESTOS CONCENTRATIONS IN BUILDINGS IN THE UNITED
STATES AND FRANCE 148
5.5 ASBESTOS CONCENTRATIONS IN U.S. SCHOOL BUILDINGS 152
5.6 CHRY' 1NCENTRATIONS IN THE HOMES OF WORKERS 155
5.7 SUMMARY UF ENVIRONMENTAL SAMPLING 156
5.8 OTHER EMISSION SOURCES 159
5.9 INTERCONVERTIBILITY OF FIBER AND MASS CONCENTRATIONS 159
5.10 SUMMARY 161
6. RISK EXTRAPOLATIONS AND HUMAN EFFECTS OF LOW EXPOSURES 162
6.1 RISK EXTRAPOLATIONS FOR LUNG CANCER AND MESOTHELIOMA 162
6.1.1 Alternative Analyses 167
6. 2 OBSERVED ENVIRONMENTAL ASBESTOS DISEASE 168
6.3 COMPARISON OF OBSERVED MORTALITY WITH EXTRAPOLATED DATA .... 170
6.4 COMPARISON OF ESTIMATED MESOTHELIOMA WITH SEER DATA 171
6.5 ' LIMITATIONS TO EXTRAPOLATIONS AND ESTIMATIONS 171
7. OTHER REVIEWS OF ASBESTOS HEALTH EFFECTS 172
7.1 INTRODUCTION 172
7.2 THE SPECTRUM OF ASBESTOS-RELATED MORTALITY AND FIBER
TYPE EFFECTS 172
7.3 MODELS FOR LUNG CANCER AND MESOTHELIOMA 173
7.4 EXTRAPOLATIONS TO LOW EXPOSURE CIRCUMSTANCES 174
7.5 RELATIVE CARCINOGENICITY OF DIFFERENT FIBER TYPES 176
7.6 NON-MALIGNANT EFFECTS 177
8. REFERENCES 178
image:
LIST OF TABLES
Table Page
3-1 Deaths among 17,800 asbestos insulation workers In the
United States and Canada, January 1, 1967 - December
31, 1976, number of men 17,800, man-years of observation
166,853 15
3-2 Observed and expected deaths from all causes, lung cancer,
gastrointestinal cancer, and mesothelioma in 41 asbestos-
exposed cohorts ,. 17
3-3 The risk of death from mesothelioma according to the time of
asbestos exposure, i n four studies 26
3-4 Analysis of residuals in polynomial fit to observed
mesothelioma dose-response data 29
3-5 Increasing risk of mesothelioma with increasing duration and
intensity of exposure 29
3-6 Comparison of linear weighted regression equations for lung
cancer and GI cancer in six cohorts of asbestos-exposed
workers 31
3-7 Relative -risk of lung cancer during 10-year intervals at
di fferent times from onset of exposure 38
3-8 Estimates of the percentage of the maximum expressed excess
risk of death from lung cancer for a 25-year exposure to
asbestos beginning at age 20 39
3-9 Age-standardized Tung cancer death rates for cigarette smoking
and/or occupational exposure to asbestos dust compared with no
smoking and no occupational exposure to asbestos dust 41
3-10 Estimates of the percentage increase in lung cancer per
f-y/ml of exposure (100 x «.), according to different
procedures in 14 epidemiclogical studies 52
3-11 Lung cancer risks, by dose, among South Carolina asbestos
textile workers (Dement et al. , 1983b) 54
3-12 Lung cancer risks, by dose, among South Carolina asbestos
textile workers (McDonald et al., 1983a) 56
3-13 Mortality experience of 679 male asbestos textile workers 57
3-14 Previous and revised estimates of mean dust levels in
f/ml (weighted by the number of workers at each level in
selected years) 59
3-15 Dust levels: Rochdale asbestos textile factory, 1971 59
3-16 Lung cancer risks, by dose, among Pennsylvania asbestos
textile and friction products workers 62
3-17 Lung cancer risks, by dose, among British asbestos friction
products workers 63
3-18 Lung cancer risks, by dose, among asbestos friction products
production workers 64
3-19 Lung cancer risks, by dose, among Canadian chrysotile asbestos
mi ners 66
3-20 Lung canqer incidence rates in urban and rural areas of
Quebec Provi nee, 1969-1973 67
vi
image:
LIST OF TABLES (continued)
Table Page
3-21 Expected and observed mortality among 544 Quebec asbestos
mi ne and mi 11 empl oyees, 1961-1973 68
3-22 Cumulative observed and expected deaths from lung cancer 5 to
40 elapsed years since onset of work in an amosite asbestos
factory, 1941-1945, by estimated fiber exposure 70
3-23 Summary of average asbestos air concentration during
i nsul ati on work 72
3-24 Lung cancer risks, by dose, among retirees of U.S.
asbestos products manufacturers 74
3-25 Lung cancer risks, .by dose, among asbestos cement production
workers 76
3-26 Lung cancer risks, by dose, among Ontario asbestos cement
workers ..-.•. 78
3-27 Comparison of estimated lung cancer risks by various groups or
individuals in studies of asbestos-exposed workers 79
3-28 Weighted geometric mean values and estimated 95% confidence
limits on K. for the various asbestos exposure circumstances
depicted in Table 3-10 and Figure 3-7 81
3-29 Mesothelioma incidence by years from onset of exposure, in
four studies 87
3-30 Summary of the data K.,, the measure of mesothelioma risk per
fiber exposure, in four studies of asbestos workers 90
3-31 Estimate of a measure of mesothelioma risk relative to lung
cancer risk, in 14 studies 92
3-32 Estimated geometric mean values of the relative mesothelioma
hazard for the various asbestos exposure circumstances
listed in Table 3-31 94
3-33 Observed and expected deaths from various causes in selected
mortality studies 97
3-34 Mortality from selected causes in Asbestos and Thetford Mines
compared to Quebec Province, females, 1966-1977 109
3-35 Risk of death from mesothelioma as a percentage of excess
1ung cancer, according to fiber exposure Ill
3-36 Mesothelioma from family contact in three occupational
circumstances 116
4-1 Distribution of fiber at the termination of 30-minute
inhalation exposures in rats (percent of total deposited) ...... 120
4-2 Summary of experiments on the effects of inhalation of
asbestos 128
4-3 Experimental inhalation carcinogenesis in rats and mice 129
4-4 Number of rats with lung tumors or mesotheliomas after
exposure to various forms of asbestos through inhalation 130
4-5 Number of rats with lung tumors or mesotheliomas after
various lengths of exposure to various forms of asbestos
through inhalation 130
4-6 Experimental inhalation carcinogenesis in rats 131
4-7 Summary of 72 experiments with different fibrous materials 134
vii
image:
LIST OF TABLES (continued)
Tab1e Page
4-8 Percentage of rats developing mesotheliomas after
intrapleural administration of various materials 136
4-9 Dose-response data following intrapleural administration
of asbestos to rats 136
4-10 Tumors in abdomen and/or thorax of rats after intraperitoneal
injection of glass fibers, crocidolite, or corundum 138
5-1 Cumulative distribution of 24-hour chrysotile asbestos
concentrations in the ambient air of U.S. cities and
Paris, France 147
5-2 Distribution of 4- to 8-hour daytime chrysotile asbestos
concentrations in. the ambient air of New York City, 1969-1970 .. 148
5-3 Distribution of 6- to 8-hour chrysotile asbestos concen-
trations within one-half mile of the spraying of asbestos
materials on building steelwork, 1969-1970 149
5-4 Cumulative distribution of 8- to 16-hour chrysotile asbestos
concentrations in buildings with asbestos-containing surfacing
materials in rooms or air plenums 150
5-5 Cumulative distribution of 5-day asbestos concentrations in
Paris buildings with asbestos-containing surfacing materials ... 151
5-6 Distribution of chrysotile asbestos concentrations in 4- to
8-hour samples taken in public schools with damaged asbestos
surfaces 153
5-7 Cumulative distribution of 5-day chrysotile asbestos concen-
trations in 25 schools having asbestos surfacing materials,
1980-1981 154
5-8 Airborne asbestos in buildings having friable asbestos
material s , 155
5-9 Distribution of 4-hour chrysotile asbestos concentrations in
the air of homes of asbestos mine and mill employees 156
5-10 Summary of environmental asbestos sampling 157
5-11 Measured relationships between optical fiber counts and mass
airborne chrysotile 160
6-1 Lifetime risks per 100,000 females of death from mesothelioma
and lung cancer from continuous asbestos exposures of 0.0001
and 0.01 f/ml according to age at first exposure, duration
of exposure, and smoking 163
6-2 Lifetime risks per 100,000 males of death from mesothelioma
and lung cancer from continuous asbestos exposures of 0.0001
and 0.01 f/ml according to age at first exposure, duration
of exposure, and smoki ng 164
6-3 Lifetime risks per 100,000 persons of death from mesothelioma
and lung cancer from continuous asbestos exposures of 0.0001
and 0.01 f/ml according to age and duration of exposure, U.S.
general population death rates were used and smoking habits
were not considered 165
vm
image:
LIST OF TABLES (continued)
Table
6-4 Comparison of the effect of different models for the time
course of mesothelioma risk for a five-year exposure to
0.01 f/ml 167
6-5 Prevalence of radiographic abnormalities associated with
asbestos exposure among household members of amosite
asbestos workers 169
6-6 Chest X-ray abnormalities among 685 household contacts of
amosite asbestos workers and 326 individual residents in
urban New Jersey, a matched comparison group • 169
6-7 Mesothelioma following onset of factory asbestos exposure,
1941-1945 170
7-1 The risks of death from mesothelioma and lung cancer from a
lifetime asbestos exposure to 0.01 f/ml 175
1x
image:
LIST OF FIGURES
Page
2-1 Dose-response relationship for prevalence of basal rales in a
chrysotile asbestos factory ,. 8
3-1 Exposure-response relationships for lung cancer observed in
seven studi es 24
3-2 Exposure-response relationships for mesothelioma observed
in four studies 28
3-3 Relative risk of death from lung cancer among insulation
workers according to age 33
3-4 Relative risk of death from lung cancer among insulation
workers according to time from onset of exposure 34
3-5 Relative risk of death from lung cancer (RR) among amosite
factory workers according to time from onset of exposure ., 36
3-6 Plot of membrane filter and midget impinger counts 44
3-7 Values of K, , the fractional increase in lung cancer per
f-y/ml of exposure in 14 asbestos-exposed cohorts 53
3-8 Risk of death from mesothelioma among insulation workers
according to age and years from onset of exposure 84
3-9 Match of curves calculated using Equation 3-6 to data on
the incidence of mesothelioma in two studies , 88
3-10 Match of curves calculated using Equation 3-6 to data
on the incidence of mesothelioma in two studies , 89
3-11 Ratio of observed to expected mortality from lung cancer
versus the ratio of observed to expected mortality from
gastrointestinal cancer . 98
3-12 Aerosol deposition in the respiratory tract 105
4-1 Measurements of animal radioactivity (corrected for decay) at
various times after inhalation exposure to synthetic
fluoramphibole 121
4-2 Correlation between the alveolar deposition of a range of
fibrous and non-fibrous particles inhaled by rats and
the corresponding activity median aerodynamic diameters 123
4-3 Mean weight of dust in the lungs of rats in relation to dose
and time .124
4-4 Regression curve relating probability of tumor to logarithm
of the number of particles per microgram with diameter
<Q.25 urn and 1 ength >8 urn 135
4-5 Hypothesis concerning the carcinogenic potency of a fiber
as a function of its length and width using data on tumor
incidence from injection and implantation studies 140
5-1 Fiber concentrations by optical microscopy versus asbestos
mass concentrations by electron microscopy 145
5-2 Cumulative distribution, on a log-probability plot, of
urban, school, and building asbestos samples 158
image:
PREFACE
This Asbestos Health Assessment Update document has been prepared by the
Environmental Criteria and Assessment Office of the U.S. Environmental Protection
Agency (EPA), Office of Health and Environmental Assessment (OHEA). The document
was developed to serve as the scientific basis for EPA review and revision, as
appropriate, of the National Emission Standards for Asbestos as a hazardous
air pollutant.
The document was reviewed and critiqued in July, 1984, by the Environmental
Health Committee (EHC) of the U.S. EPA Science Advisory Board (SAB) and subse-
quently revised to take into account the peer-review comments of that SAB
committee. The Science Advisory Board provides advice on scientific matters
to the Administrator of the U.S. Environmental Protection Agency.
In the development of this assessment document, pertinent scientific
literature has been critically evaluated and conclusions are presented in such
a manner that the toxicity of asbestos and related characteristics are identi-
fied. Estimates of the fractional increased risk of lung cancer and mesothe-
Tioma per unit exposure of asbestos are also discussed, in an attempt to
quantify adverse health effects associated with exposure to asbestos via
inhalation.
image:
ABSTRACT
Data developed since the early 1970s, from large population studies with
long follow-up, have added to our knowledge of asbestos-related diseases and
strengthened the evidence for associations between asbestos and specific types
of health effects. Lung cancer and mesothelioma are the most important asbestos-
related causes of death among exposed individuals. Cancer at other sites also
has been associated with asbestos exposure. The accumulated data suggest that
the excess risk of lung cancer from asbestos exposure is proportional to the
cumulative exposure (the duration times the intensity) and the underlying risk
in the absence of exposure. The risk of death from mesothelioma is approxi-
mately proportional to the cumulative exposure to asbestos and increases
sharply with time since onset of exposure. Animal studies confirm the human
epidemiological results and indicate that all major asbestos varieties produce
lung cancer and mesothelioma, with only limited differences in carcinogenic
potency. Some measurements demonstrate that asbestos exposures exceeding 100
times background occur in non-occupational environments. Currently, the most
important of these non-occupational exposures is the release of fibers from
asbestos-containing surfacing materials in schools, auditoriums, and other
public buildings, or from sprayed asbestos fireproofing in high-rise office
buildings. Extrapolations of risks of asbestos cancers from occupational
circumstances can be made, although numerical estimates in a specific exposure
circumstance have a large (approximately tenfold) uncertainty. Because of
this uncertainty, calculations of unit risk values for asbestos at low concen-
trations must be viewed with caution. They are subjective, to some extent,
and are also subject to the following limitations in data: 1) variability in
the exposure-response relationship at high exposures; 2) uncertainty in extra-
polating to exposures 1/100 as much; and 3) uncertainties in conversion of
optical fiber counts to electron microscopic fiber counts or mass determina-
tions.
Xll
image:
AUTHORS, CONTRIBUTORS, AND REVIEWERS
This health assessment update document for asbestos was prepared by
Dr. William J. Nicholson, Ph.D. (Mt, Sinai School of Medicine, New York, N.Y.-)
under contract with the U.S. EPA Environmental Criteria and Assessment Office
in Research Triangle Park, NC (Dr. Dennis J. Kotchmar, M.D., Project Manager).
The following individuals reviewed earlier drafts of this document during
its preparation and their valuable comments are appreciated.
Dr. Steven Bayard, Office of Health and Environmental Assessment (RD-689),
U.S. Environmental Protection Agency, 401 M Street, SW, Washington, DC 20460
Mr. Michael Beard, Environmental Monitoring Systems Laboratory (MD-77), U.S.
Environmental Protection Agency, Research Triangle Park, NC 27711
Or. David L. Coffin, Health Effects Research Laboratory (MD-70), U.S.
Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Devra Davis, Environmental Law Institute, 1346 Connecticut Avenue, NW,
Suite 600, Washington, DC 20036
Professor Sir Ri'chard Doll, ICRF Cancer Epidemiology and Clinical, Trials
Unit, Gibson Laboratory, Radcliffe Infirmary, Oxford, 0X2 6HE, England
Dr. Philip Enter-line, Graduate School of Public Health, Department of
Biostatistics, University of Pittsburgh, 130 Desoto Street, Pittsburgh, PA
15261
Dr. Lester D. Grant, Director, Environmental Criteria and Assessment Office
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
Dr. Robert E. McGaughy, Office of Health and Environmental Assessment (RD-689),
U.S. Environmental Protection Agency, 401 M Street, SW, Washington, DC 20460.
Dr. Paul Kotin, Mansville Corporation, Ken-Caryl Ranch, Denver, CO 80271
Dr. James R. Millette, Health Effects Research Laboratory, U.S. Environmental
Protection Agency, 26 West St. Clair, Cincinnati, OH 45268
Dr. Charles H. Nauman, U.S. Environmental Protection Agency (RD-689), 401 M
Street, SW, Washington, DC 20460
Dr. William Nelson, Health Effects Research Laboratory (MD-55), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
image:
Professor Julian Peto, Section of Epidemiology, Institute of Cancer Research,
Sutton, Surrey SM2 5PX, England
Dr. James N. Row, Office of Toxic Substances (TS-796), U.S. Environmental
Protection Agency, 401 M Street, SW
Dr. Marvin A. Schneiderman, Clement Associates, Inc., 1515 Wilson Boulevard,
Arlington, VA 22209
Mr. Ralph Zumwalde, c/o Chief Criteria Document Section, National Institute of
Occupational, Safety and Health, 46-76 Columbia Parkway, Cincinnati, OH 45226
xiv
image:
SCIENCE ADVISORY BOARD
ENVIRONMENTAL HEALTH COMMITTEE
This document was independently peer-reviewed in public session by the
Environmental Health Committee (EHC), Environmental Protection Agency Science
Advisory Board. Members and consultants of the EHC participating in the
review included:
Chairmarv, Environmental Health Committee (EHC)
Dr. Richard A. Griesemer, Directors Biology Division, Oakridge National Labora-
tory, Martin Marietta Energy Systems, Inc. P.O. Box Y, Oakridge, Tennessee
37831
EHC
Dr. Herschel E. Griffin, Professor of Epidemiology, Graduate School of Public
Health, 6505 Alvarado Road, San Diego State University, San Diego, California
92182-0405
Exec ut j ve Sec re tary , EHC
Dr, Daniel Byrd III, Executive Secretary, Science Advisory Board, A- 101 F,
U.S. Environmental Protection Agency, Washington, D.C. 20460
Members
Dr. Herman E. Collier, Jr., President, Moravian College, Bethlehem, Pennsylvania
18018
Dr. Morton Corn, Professor and Director, Division of Environmental Health
Engineering, School of Hygiene and Public Health, The Johns Hopkins University,
615 N. Wolfe Street, Baltimore, Maryland 21205
Dr. John Doull, Professor of Pharmacology and Toxicology, University of Kansas
Medical Center, Kansas City, Kansas 66103
Dr. Jack D. Hackney, Chief, Environmental Health Laboratories, Professor of
Medicine, Rancho Los Amigos Hospital Campus of the University of Southern
California, 7601 Imperial Highway, Downey, California 90242
Dr. Marvin Kuschner, Dean, School of Medicine, Health Science Center, Level 4,
State University of New York, Stony Brook, New York 11794
Dr. Daniel Menzel , Director and Professor, Pharmacology and Medicine, Director,
Cancer Toxicology & Chemical Carcinogenesis Program, Duke University Medical
Center, Durham, North Carolina 27710
Dr. D. Warner North, Principal, Decision Focus Inc., Los Altos Office Center,
Suite 200, 4984 El Camino Real, Los Altos, California 94022
xv
image:
Dr. William J. Schull, Director and Professor of Population Genetics, Center
for Demographic and Population Genetics, School of Public Health, University
of Texas Health Science Center at Houston, Houston, Texas 77030
Dr. Michael J. Symons, Professor, Department of Biostatisties, School of
Public Health, Un'iversity of North Carolina, Chapel Hill, North Carolina
27711
Dr. Seymour Abrahamson, Professor of Zoology and Genetics, Department of
Zoology, University of Wisconsin, Madison, Wisconsin 53706
Dr. Edward F. Ferrand, Assistant Commissioner for Science and Technology, New
York City Department of Environmental Protection, 51 Astor Place, New York,
New York 10003
Dr. Ronald D. Hood, Professor, Development Biology Section, Department of
Biology, The University of Alabama, and Principal Associate, R.D. Hood and
Associates, Consulting "Toxicologists, P.O. Box 1927, University, Alabama
35486
Dr. Bernard Weiss, Professor, Divison of Toxicology, P.O. Box RBB, University
of Rochester, School of Medicine, Rochester, New York 14642
Consultants
Dr. Brooke T. Mossman, Associate Professor of Pathology; Department of Pathology,
University of Vermont, Burlington, Vermont 05405-Q068
Dr. J, Corbett McDonald, Professor, Dust Disease Research Unit, McGill University,
1110 Pine Avenue, West, Montreal, PQ, Canada H3A1A3
xvi
image:
1. SUMMARY
Data developed since the early 1970s, from large population studies with
long follow-up, have added to our knowledge of asbestos disease. These data
strengthen and quantitatively define the association of asbestos exposure with
disease. Lung cancer and mesothelioma are the most important asbestos-related
causes of death among exposed individuals. Gastrointestinal cancers are also
increased in most studies cf occupationally exposed workers. Cancer at other
sites (larynx, kidney, ovary) has also been shown to be associated with asbes-
tos exposure in some studies, but the degree of excess risk and the strength
of the association are less for these and the gastrointestinal cancers than
for lung cancer or mesothelioma. The International Agency for Research on
Cancer (1982) lists asbestos as a group 1 carcinogen, meaning that exposure to
asbestos is carcinogenic to humans. EPA's proposed guidelines would categorize
asbestos as Group A, human carcinogen (Federal Register, 1984b).
Data from a study of U.S. insulation workers allow models to be developed
for the time and age dependence of lung cancer and mesothelioma risk. Thirteen
other studies provide exposure-response information. The accumulated data
suggest that the excess risk of death from lung cancer from asbestos exposure
is proportional to the cumulative exposure (the duration times the intensity)
and the underlying risk in the absence of exposure. The time course of lung
cancer is determined primarily by the time course of the underlying risk.
However, the risk of death from mesothelioma increases very rapidly after the
onset of exposure and is independent of age and cigarette smoking. As with
lung cancer, the risk appears to be proportional to the cumulative exposure to
asbestos in a given period. The dose and time relationships for other asbestos
cancers are uncertain.
Fourteen studies provide data for a best estimate fractional increased
risk of lung cancer per unit exposure. The values characterizing the lung
cancer risk obtained from different studies vary widely. Some of the varia-
bility can be attributed to specific processes. Chrysotile mining and milling,
and perhaps friction product manufacture, appear to have lower unit exposure
risks than chrysotile textile production and other uses of asbestos. Other
variability can be associated with the uncertainties of small numbers in
epidemiological studies and misestimates of the exposures of earlier years.
Finally, some differences between studies may be related to differences in
image:
fiber type, but these are much less than those associated with specific processes.
Four studies provide similar quantitative data on the unit exposure risk
for mesothelioma and six additional studies provide corroborative, but less
accurate, quantitative data. The same factors that affect the lung cancer
unit exposure risk appear to affect that of mesothelioma as the ratio of a
measure of raesothelioma risk to excess lung cancer risk is roughly constant
across the ten studies. However, In other studies the ratio of number of
mesothelioma deaths to lung cancer deaths among groups exposed to substantial
quantities of crocidolite is two to four times higher than among groups exposed
predominantly to other fibers. Further, the risk of peritoneal mesothelioma
appears to be less from exposure to chrysotile than to either crocidolite or
amosite, but this suggestion is tempered by uncertainties associated with the
greater possibility of misdiagnosis of the disease.
Animal studies confirm the human epidemiological results. All major
asbestos varieties produce lung cancer and mesothelioma with only limited
differences in carcinogenic potency. Implantation and injection studies show
that fiber dimensionality, not chemistry, is the most important factor in
fiber-induced carcinogenicity. Long (>4 urn) and thin (<1 urn) fibers are the
most carcinogenic at a cancer-inducible site. However, the size dependence of
the deposition and migration of fibers also affects their carcinogenic action
in humans.
Measurements demonstrate that asbestos exposures exceeding 100 times the
background occur to individuals in some non-occupational settings. Currently,
the most important of these non-occupational exposures is from the release of
fibers from asbestos-containing surfacing materials in schools, auditoriums,
and other public buildings, or from sprayed asbestos-containing fireproofing
in high-rise office buildings. A high potential exists for future exposure
from the maintenance, repair, and removal of these materials.
Extrapolations of risks of asbestos cancers from occupational circum-
stances can be made, although numerical estimates in a specific exposure
circumstance have a large (approximately tenfold) uncertainty. Because of
this uncertainty, calculations of unit risk values for asbestos at the low
concentrations measured in the environment must be viewed with caution. The
best estimate of risk to the United States general population for a lifetime
continuous exposure to 0.0001 f/ml is 2.8 mesothelioma deaths and 0.5 excess
lung cancer deaths per 100,000 females. Corresponding numbers for males are
image:
1.9 mesothelioma deaths and 1.7 excess lung cancer deaths per 100,000 individ-
uals. Excess GI- cancer mortality is approximately 10-30 percent that of
excess lung cancer mortality. These risks are subjective, to some extent, and
are also subject to the following limitations in data: 1) variability in the
exposure-response relationship at high exposures; 2) uncertainty in extrapo-
lating to exposures 1/100 as much; and 3) uncertainties in conversion of
optical fiber counts to electron microscopic fiber counts or mass determina-
tions.
Recently several government agencies in different countries reviewed
asbestos health effects. Areas of agreement and disagreement between these
other reviews and those of this document are presented, A comparison of the
different risk estimates is provided.
image:
2. INTRODUCTION
The principal objective of this "Airborne Asbestos Health Assessment
Update" document is to provide the U.S. Environmental Protection Agency (EPA)
with a sound scientific basis for review and revision, as appropriate, of the
national emission standard for asbestos, 40 CFR 61, subpart B, as required by
the 1977 Clean Air Act Amendments, Sections 111 and 112. The health effects
basis for designating asbestos as a hazardous pollutant and minimizing emis-
sions via the original 1973 National Emissions Standard for Hazardous Air
Pollutants (NESHAP) was scrutinized, at that time, during two public hearings
and a public comment period. Once a pollutant has been designated as a "hazar-
dous" air pollutant, the burden of proof is placed on proving that designation
wrong. The original health effects basis for designating asbestos as a hazard-
ous air pollutant was qualitative evidence establishing asbestos-associated
carcinogenic effects. However, insufficient bases then existed by which to
define pertinent quantitative dose-response relationships; i.e., unit risk
values could not be credibly estimated. The main focus of this update document
is to describe asbestos-related health effects developments since 1972, and to
determine if new data warrant the specification of unit risk values for asbes-
tos. This report forms part of the basis to perform a risk assessment. The
National Academy of Sciences (NAS) in 1983 suggested a definition of risk
assessment as the use of the factual data base to define the health effects of
exposure of individuals or populations to hazardous materials, such as asbestos
in this case (National Academy of Sciences, 1983). This update document is
not meant to characterize the status of asbestos measurement techniques or
mineralogical characterization, although they are presented briefly as back-
ground information. Because this document is concerned only with the excess
risk of cancer from inhalation of asbestos fibers, consideration of the risk
posed from ingesting asbestos fibers also is outside its scope. A separate
criteria document for asbestos in water is being prepared by the EPA.
Thus, emphasis is placed on the literature published after 1972 and on
those papers that provide information on the risk from low-level exposures,
such as those encountered in the non-occupational environment. Specifically,
this report addresses the following issues:
image:
1. Are there models that illustrate the age, time, and exposure
dependence of asbestos diseases that can be used satisfactorily
in a quantitative risk assessment?
2, Is there consistency among studies and sufficiently good esti-
mates of exposure in occupational circumstances so that useful
exposure-response relationships can be established?
3. Do these studies indicate any significant differences in the
carcinogenic potency of different asbestos minerals or of
fibers of different dimensionality?
4. What additional or confirmatory information relating to human
carcinogenicity is provided by animal studies?
5, What are the non-occupational concentrations of asbestos to
which populations are exposed?
6. Is there a basis for making numerical estimates of risks of
asbestos disease that might result from non-occupational expo-
sures?
Two documents provide good reviews of the status of knowledge of the
health effects of asbestos in the early 1970s. One is the criteria document
for occupational exposure to asbestos produced by the National Institute of
Occupational Safety and Health as part of the Occupational Safety and Health
Administration's consideration of an asbestos standard in early 1972 (National
Institute for Occupational Safety and Health, 1972). The second is the proceed-
ings of a conference sponsored by the International Agency for Research on
Cancer (IARC), which was convened in October 1972 with the stated purpose of
reviewing the knowledge of the biological effects of asbestos (Bogovski et
al., 1973), and included a report by an Advisory Committee on Asbestos Cancers
appointed by the IARC to review evidence relating exposures to asbestos dust
to cancers.
image:
2.1 SUMMARY OF ASBESTOS HEALTH EFFECTS THROUGH 1972
This section relies heavily on review articles found in the proceedings
of the October 1972 IARC meeting and in the report of the IARC Advisory Commit-
tee published therein (Bogovski et al., 1973) for a summary of health effects
knowledge as of 1973.
2.1.1 OccupationalExposure
Diseases considered to be associated with asbestos exposure in 1972
included asbestosis, mesothelioma, bronchogenic carcinoma, and cancers of the
gastrointestinal (GI) tract, including the esophagus, stomach, colon, and
rectum. Lung cancer was associated with exposure to all principal commercial
varieties of asbestos fiber: amosite, anthophyllite, crocidolite, and chryso-
tile. Excess risks of bronchogenic carcinoma were documented in mining and
milling, manufacturing, and end product use (application of insulation mater-
ials). Mesothelioma was a cause of death among factory employees, insulation
applicators, and workmen employed in the mining and milling of crocidolite. A
much lower risk of death from mesothelioma was observed among chrysotile or
amosite mine and mill employees, and no cases were associated with anthoph-
yllite exposure. The IARC Advisory Committee suggested that the risk of death
from mesothelioma was greatest with crocidolite, less with amosite, and still
less with chrysotile. This suggestion was based on the association of disease
with exposures. No unit exposure risk information existed.
Information on exposure-response relationships for lung cancer risk among
various exposed groups was scanty. Data from Canadian mine and mill employees
clearly indicated an increasing risk with increasing exposure, measured in
terms of millions of particles per cubic foot-years (mppcf-y), but data on the
risk at minimal exposure were uncertain because the number of expected deaths
calculated using adjacent county rates suggested that all exposure categories
were at elevated risk (McDonald et al., 1971). A study of retirees of the
largest U.S. asbestos manufacturer showed lung cancer risks ranging from 1.7
times that expected in the lowest exposure category to 5.6 times that expected
in the highest (Enterline and Henderson, 1973). Exposures were expressed in
mppcf-y and information on conversion of mppcf to fibers per milliliter was
available only for textile production. Despite the paucity of data, the
report of the Advisory Committee on Asbestos Cancers to the IARC (Bogovski et
al., 1973) stated, "The evidence ... suggests that an excess lung carcinoma
risk is not detectable when the occupational exposure has been low. These low
image:
occupational exposures have almost certainly been much greater than that to
the public from general air pollution." Limited data existed on the assoc-
iation of GI cancer with asbestos exposure, but the "excess is relatively
small compared with that for bronchial cancer."
The prevalence of asbestosis, particularly as manifested by X-ray abnor-
malities of the pleura or parenchymal tissue, had been documented more exten-
sively than the risk of the asbestos-related malignancies. In part, this
documentation resulted from knowledge of this disease extending back to the
turn of the century, whereas the malignant potential of asbestos was not
suggested until 1935 (Lynch and Smith, 1935; Gloyne, 1936) and not widely
appreciated until the 1940s (Merewether, 1949). Asbestosis had been docu-
mented in a wide variety of work circumstances and associated with all commer-
cial types of asbestos fibers. Among some heavily exposed groups, 50 to
80 percent of individuals employed for 20 or more years were found to have
abnormal X-rays characteristic of asbestos exposure (Selikoff et al., 1965;
Lewinsohn, 1972), A lower percentage of abnormal X-rays was present in
lesser exposed groups. Company data supplied to the British Occupational
Hygiene Society (British Occupational Hygiene Society, 1968) on X-ray and
clinical abnormalities among 290 employees of a large textile production
facility in Great Britain were analyzed by Berry (1973) in terms of a fiber
exposure-response relationship. The results were utilized in establishing the
1969 British regulation on asbestos. These data, shown in Figure 2-1, sug-
gested that the risk of developing the earliest signs of asbestosis (rales)
was less than 1 percent for accumulated fiber exposure of 100 fiber-years/ml
(f-y/ml), e.g., 2 fibers/milliliter (f/ml) for 50 years. However, shortly
after the establishment of the British Standard, additional data from the same
factory population suggested a much greater prevalance of X-ray abnormalities
than was believed to exist at the time the British Standard was set (Lewinsohn,
1972). These data resulted from use of the new International Labour Office
(ILO) U/C standard classification of X-rays (International Labour Office,
1971) and the longer time from onset of employment. Of the 290 employees
whose clinical data were reviewed by the BOHS, only 13 had been employed for
30 or more years; 172 had less than 20 years of employment. The progression
of asbestosis depends on both cumulative exposure and time from exposure;
therefore, analysis in terms of only one variable (as in Figure 2-1) can be
misleading.
image:
cc
O
9.
20
58 io
£>
z*
LU
o
cc
1U
a.
• BASAL RALES
O X-RAY ABNORMALITIES
0 100 200 300 400 500
CUMULATIVE EXPOSURE, years x fibres/cm3
Figure 2-1. Dose-response relationship for prevalence
of basal rales in a chrysotile asbestos factory.
Source: Berry (1973); x-ray data added from British
Occupational Hygiene Society (1968).
image:
2.1.2 Environmental and Indirect Occupational Exposure Circumstances
Several research groups had shown that asbestos disease risk could develop
from other than direct occupational exposures. Wagner, Sleggs, and Marchand
(1960) showed that a mesothelioma risk in environmental circumstances existed
in the mining areas of the Northwest Cape Province of South Africa. Of 33
mesotheliomas reported over a 5-year period, roughly half were from occupa-
tional exposure. However, all but one of the remainder resulted from exposure
occasioned by living or working in the area of the mining activity. A second
study that showed an extra-occupational risk was that of Newhouse and Thompson
(1965) who investigated the occupational and residential background of 76
individuals deceased of mesothelioma in the London hospital .< Forty-five of
the decedents had been employed in an asbestos industry; of the remaining 31,
9 lived with someone employed in asbestos work and 11 were individuals who
resided within half a mile of an asbestos factory. Bohlig and Hain (1973)
identified environmental asbestos exposure in 38 mesothelioma cases without
occupational exposure who resided near an asbestos factory, further defining
residential risk. A final study, which is particularly important because of
the size of the population implied to be at risk, was that of Harries (1968),
who pointed to a risk of asbestos disease from indirect occupational exposure
in the shipbuilding industry. He described thu presence of asbestosis in 13
individuals and mesothelioma in 5 others who were employed in a shipyard, but
were not members of trades that regularly used asbestos. Rather, they were
exposed to the dust created by other employees placing or removing insulation.
Evidence of ubiquitous general population exposure and environmental
contamination from the spraying of asbestos on the steel-work of high rise
buildings was established by 1972. Data by Nicholson and Pundsack (1973)
showed that asbestos was commonly found at concentrations of nanograms per
cubic meter (ng/m ) in virtually all United States cities, and at concentra-
tions of tnicrograms per liter (pg/l) in river systems of the United States.
Concentrations of hundreds of nanograms per cubic meter were documented at
distances up to one-quarter of a mile from fireproofing sites. Mesothelioma
was acknowledged by the Advisory Committee to be associated with environmental
exposures, but they suggested that "the evidence relates to conditions many
years ago .... There is no evidence of a risk to the general public at present.
Further, their report stated that, "There is at present no evidence of lung
damage by asbestos to the general public," and "Such evidence as there is does
not indicate any risk" from asbestos fibers in water, beverages, food, or
image:
parenteral drugs. No mention was made in the report of risks from indirect
occupational asbestos exposures.
2.1,3 Analytical Methodology
During the late 1960s and early 1970s, significantly improved methods
were developed for assessing asbestos disease and quantifying asbestos in the
environment. In 1971, a standardized methodology was established for the
identification of pneumoconiosis: the ILO U/C Classification of Pneumoconioses
(International Labour Office, 1971). This methodology provided a uniform cri-
terion for assessing the prevalence of asbestos-related X-ray abnormalities.
Significant advances were also achieved in the quantification of asbestos
aerosols. In the late 1960s, the membrane filter technique was developed for
the measurement of asbestos fibers in workplace aerosols. While this procedure
has som*1 limitations, it did establish a standardized method, using simple
instrumentation, that was far superior to any that existed previously. This
method subsequently allowed epidemiclogical studies to be done that based
exposure estimates on a standardized criterion, Experimental techniques in
the quantification of asbestos at concentrations of tenths of ng/m of air and
tenths of ug/1 of water were also developed, extending the sensitivity of
exposure estimates approximately three orders of magnitude below those of
occupational aerosols and allowing assessment of general population exposures.
Finally, techniques for the analysis of asbestos in lung and other body tissues
were developed. Digestion techniques and the use of electron microscopy to
analyze fibers contained in the digest and in thin sections of lung tissue
showed that asbestos fibers were commonly present in the lung tissue of gene-
ral population residents as well as individuals exposed in occupational circum-
stances.
2,1.4 Experimental Studies
Experimental animal studies using asbestos fibers confirmed the risks of
lung cancer and mesothelioma from amosite, crocidolite, and chrysotile. In
each case, the establishment of a risk in animals followed the association of
the malignancy with human exposure. For example, a causal relationship be-
tween lung cancer and asbestos exposure in humans was suggested in 1935 and
confirmed in the late 1940's, but was not described in the open literature in
animals until 1967 (Gross et al., 1967). Mesothelioma, reported in an asbestos
10
image:
worker in 1953 (Weiss, 1953), was produced in animal experimentation in 1965
(Smith et al., 1965). Other animal experimentation showed that combinations
of asbestos and other carcinogenic materials produced an enhanced risk of
asbestos cancer. Asbestos exposure combined with exposure to benz(a)pyrene
was demonstrably more carcinogenic than exposure to either agent alone.
Additionally, organic and metal compounds associated with asbestos fibers were
ruled out as important factors in the carcinogenicity of fibers. Lastly,
animal experimentation involving the application of fibers onto the pleura of
animals indicated that the important factor in the carcinogenicity was the
length and width of the fibers rather than their chemical properties (Stanton,
1973). The greatest carcinogenicity was related to fibers that were less than
2.5 urn in diameter and longer than 10 Mm-
2.2 CURRENT ASBESTOS STANDARDS
The current Occupational Safety and Health Administration (OSHA) stand-
ards for an 8-hour time-weighted average (TWA) occupational exposure to asbestos
is 2 fibers longer than 5 urn in length per milliliter of air (2 f/ml or
2,000,000 f/m ). Peak exposures of up to 10 f/ml are permitted for no more
than 10 rain (Code of Federal Regulations, 1984a), This standard has been in
effect since July 1, 1976, when it replaced an earlier one of 5 f/ml (TWA).
In Great Britain, a value of 0.5 f/ml is now the accepted level for chrysotile.
This standard has evolved from recommendations made in 1979 by the Advisory
Committee on Asbestos (1979a), which also recommended a TWA of 0.5 f/ml for
amosite and 0.2 f/ml for crocidolite. From 1969 to 1983, 2 f/ml (TWA) was the
standard for chrysotile (British Occupational Hygiene Society, 1968). This
earlier British standard served as a guide for the OSHA standard (National
Institute for Occupational Safety and Health, 1972).
The 1969 British standard was developed specifically to prevent asbestosis
among working populations; data that would allow a determination of a standard
for cancer (British Occupational Hygiene Society, 1968) were felt to be lacking.
Unfortunately, among occupational groups, cancer is the primary cause of
excess death among workers (see Chapter 3). Three-fourths or more of asbestos-
related deaths are from malignancy. This fact led OSHA to propose a lowered
TWA standard to 0.5 f/ml (500,000 f/m3) in October, 1975 (Federal Register,
1975). The National Institute for Occupational Safety and Health anticipated
11
image:
hearings on a new standard and proposed a value of 0,1 f/ml (National Institute
for Occupational Safety and Health, 1976) in an update of their 1972 criteria
document. In the discussion of the NIOSH proposal, it was stated that the
value was selected on the basis of the practical limitations of analytical
techniques using optical microscopy, and that 0.1 f/ml may not necessarily
protect against cancer. The preamble to the OSHA proposal acknowledges that
no information exists by which to define a threshold for asbestos carcino-
genesis. The OSHA proposal has been withdrawn, and a new proposal was submit-
ted on April 10, 1984 (Federal Register, 1984a). In it, OSHA proposed a TWA
standard of either 0.2 or 0.5 f/ml, depending upon information to be obtained
in hearings (held during the summer of 1984). NIOSH reaffirmed its position
on a 0.1 f/ml TWA standard (Occupational Safety and Health Administration,
1984).
The existing Federal national emission standards for asbestos are pub-
lished in Part 61, Title 40, Code of Federal Regulations (1984b). In summary,
these apply to milling, manufacturing, and fabrication sources, and to demoli-
tion, renovation, and waste disposal, and include other limitations. In
general, the standards allow compliance alternatives, either (1) no visible
emissions, or (2) employment of specified control techniques. The standards
do not include any mass or fiber count emission limitations. However, some
local governmental agencies have numerical standards (e.g., New York: 27
3
ng/m ), but these have little regulatory relevance.
12
image:
3. HUMAN HEALTH EFFECTS ASSOCIATED WITH OCCUPATIONAL EXPOSURE TO ASBESTOS
3.1 INTRODUCTION
The evidence that asbestos is a human carcinogen Is overwhelming. Studies
on more than 30 cohorts of workers exposed to asbestos have demonstrated an
elevated risk of cancer at the 5% level of significance. All four major
commercial varieties have been linked to excess cancer and asbestosls. The
question is not so much what disease, but how much disease. Our concerns are
now more quantitative than qualitative. What are the dose, time, and age
relationships for the different asbestos cancers? Are there differences in
the carcinogenic potencies of the different asbestos minerals? What are the
cancer risks at low exposures? What are the estimates of uncertainty?
This chapter Is largely concerned with those studies that provide quanti-
tative exposure-response relationships for asbestos diseases. While lung
cancer and mesothelioma are the most dominant asbestos-related malignancies,
the strength of the evidence and the relative excess of cancers at other sites
are discussed. "Models for assessment of the risk of lung cancer and mesothelioma
are reviewed. Unit exposure risks are estimated from 14 studies that provide
Information on exposure-response relationships. These estimates Illustrate
considerable variation in the calculated unit exposure risks for mesothelioma
and lung cancer in the different studies. The magnitude and possible sources
of these different unit risks are discussed. The extent to which the varia-
tion is the result of methodological or statistical uncertainties (i.e., on
the estimates of exposure or of the magnitude of disease) or of differences 1n
the character of the exposure 1n terms of fiber size and mineraloglcal species
1s considered 1n detail.
3.2 MORTALITY ASSOCIATED WITH ASBESTOS EXPOSURE
The study of U.S. and Canadian insulation workers by Sellkoff et al.
(1979) contains the largest number of asbestos-related deaths among any group
of asbestos workers studied. Thus, 1t best demonstrates the full spectrum of
disease from asbestos exposure. The mortality experience of 17,800 asbestos
Insulation workers was studied prospectively from January 1, 1967 through
13
image:
December 31, 1976.. .These workers were exposed primarily to chrysotile prior
to 1940, to chrysotile and amosite from 1940 through 1965, and largely to
chrysotile thereafter. No crocidolite is known to have been used in the U. S.
insulation material (Selikoff et al. 1970). The workers mainly applied new
insulation; removal of old materials would have constituted less than 5% of
their activities.
In this group, 2271 deaths occurred, and their analysis provides impor-
tant insights into the nature of asbestos disease. Table 3-1 lists the expected
and observed deaths by cause, and includes data on tumors less frequently
found. Lung tumors were common and accounted for approximately 21 percent of
the deaths; 8 percent were from mesothelioma of the pleura or peritoneum, and
7 percent died from asbestosis. Considering all cancers, 675 excess malig-
nancies occurred, constituting 30 percent of all deaths. In addition to lung
cancer and mesothelioma, the incidences of cancers of the gastrointestinal
tract, larynx, pharynx and buccal cavity, and kidney were significantly ele-
vated.
Other tumors were also increased, but not to a statistically significant
degree for individual sites. However, these other cancers, as a group, were
significantly in excess: 184 observed (using best available evidence for
classification) versus 131.8 expected (p<0.001). Some of this excess, however,
may be the result of misclassification of asbestos-related lung cancer or
peritoneal mesothelioma. Rather than 184 deaths, certificate of death classi-
fication attributed 252 cancers to these other sites. After a review of
pathological material and available medical records, pancreatic, liver, and
unspecified abdominal cancers are found to be commonly misclassified. Indivi-
duals certified as dying of cancers of the pancreas and the abdomen were often
found to have peritoneal mesotheliomas, and several liver cancers were the
result of a primary malignancy in the lung. As it was not possible to review
all cases, some additional misclassification may still exist. However, its
magnitude would not be great compared to the remaining excess of 52 cases.
The excess at extra-thoracic sites may reflect mortality from the dissemination
of asbestos fibers to various organs (Langer, 1974). Alternatively, it has
been suggested that asbestos could exert a systemic effect, perhaps on the
immune system, that leads to a general increased risk of cancer (Goldsmith,
1982).
14
image:
TABLE 3-1, DEATHS AMONG 17,800 ASBESTOS INSULATION WORKERS IN THE UNITED
STATES AND CANADA, JANUARY 1, 1967 - DECEMBER 31, 1976,
NUMBER OF MEN 17,800,
MAN-YEARS OF OBSERVATION 166,853
Number of Deaths
Underlying cause of death
Total deaths, all causes
Total cancer, all sites
Cancer of lung
Pleura! mesothelioma
Peritoneal mesothelioma
Mesothelioma, n.o.s.
Cancer of esophagus
Cancer of stomach
Cancer of colon-rectum
Cancer of larynx
Cancer of pharynx, buccal cavity
Cancer of kidney
Cancer of pancreas
Cancer of liver and biliary
passages
Cancer of brain
Cancer of lymphatic and
hematopoietic system
All other cancer
Noninfectious pulmonary
diseases, total
Asbestosis
All other causes
Expected3
1658.9
319.7
105.6
_b
_b
_b
7.1
14.2
38.1
4.7
10.1
8.1
17.5
7.2
10.4
33.2
63.5
59.0
_b
1280,2
Observed
BE
2271
995
486
63
112
0
18
22
59
11
21
19
23
5
14
34
108
212
168
1064
DC
2271
922
429
25
24
55
18
18
58
9
16
18
49
19
17
31
136
188
78
1161
Ratio of
observed
to expected
BE
1.37
3.11
4.60
_b
_b
_b
2,53
1.54
1.55
2.34
2.08
2.36
1.32
0.70
1.35
1.02
1.65
3.59
_b
0.83
DC
1.37
2.88
4.06
_b
_b
_b
2.53
1.26
1.52
1.91
1.59
2.23
2.81
2.65
1.63
0.93
2.16
3.19
_b
0.91
BE = Best evidence. Number of deaths categorized after review of best
available information (autopsy, surgical, clinical).
DC = Number of deaths as recorded from death certificate information only.
aExpected deaths are based upon white male age-specific U.S. death rates of
the U.S. National Center for Health Statistics, 1967-1976. (National Center
for Health Statistics, 1977).
Rates and thus ratios are not available, but these have been rare causes of
death in the general population.
Source: Selikoff et al. (19790.
15
image:
3,2.1 Accuracy ofCause of Death Ascertainment
Table 3-1 lists the observed deaths according to the cause recorded on
the certificate of death (DC) and according to the best evidence (BE) available
from medical records, surgical specimens, and autopsy protocols. In comparing
occupational mortality with that of the general population, one usually uti-
lizes information as recorded on death certificates since such information,
without verification, serves as the basis for "expected rates." However,
since mesotheHoma and asbestosis are virtually unseen in the general popula-
tion, their misdiagnosls (which has been common) is of little importance. In
contrast, their misdiagnosis among asbestos workers can cause serious distort-
ions in cause-specific mortality. Not only are asbestos-related causes under-
stated, but others, such as pancreatic cancer, might wrongly appear to be
significantly elevated (Selikoff and Seidman, 1981). While substantial dif-
ferencps exist in the DC and BE characterization of deaths from mesothelioma,
asbestosis, pancreatic cancer, and liver cancer, the numbers of BE and DC
deaths from cancer of other specific sites agree reasonably well.
Mesothelioma is best described by an absolute risk model and lung cancer
by a relative risk model. Thus, risks for mesothelioma are expressed in
absolute rates (e.g., deaths/1000 person-years), and the best medical evidence
is used, when available, to establish the number of cases. Deaths from asbes-
tosis are treated similarly. Risks for lung cancer are quantified by the ratio
of observed to expected deaths. Here, it is expected that misclassification
of lung cancer deaths would occur as frequently in asbestos workers as in the
general population (in terms of the percentage of lung cancer cases). Therefore,
the certificate of death cause is used to establish the relative risks of lung
cancer in asbestos-exposed groups. However, when possible, account is taken
of deaths from mesothelioma and asbestosis. The treatment of other malig-
nancies also uses DC causes of death.
3.3 EPIDEMIC-LOGICAL STUDIES OF ASBESTOS HEALTH EFFECTS: STRENGTH OF THE
EVIDENCE
Many epidemiological studies have documented the presence of asbestos
disease among occupationally-exposed workers. The larger and more recent
studies are listed in Table 3-2 according to the type of fiber exposure and
16
image:
TABLE 3-2, OBSERVED AND EXPECTED DEATHS FROM ALL CAUSES, LUNG CANCER, GASTROINTESTINAL
CANCER, AND MESOTHELIOMA IN 41 ASBESTOS-EXPOSED COHORTS
ftcteuo »t ll. 119821
Dml il il. USEO]" b
NcDonaKi it »1. (iW3a.br
McDDMld «L il. (1980)
(Uontlll It •!. (1980) ,
Kkholwn it il. (197»c
McDonald it ll. 11984)
Rubin >l il. (1379)
•Win (1977)
fndoiluntlx chiytotlli (>9aX}
McDonild it il. (1981blk
Bobtnun it il. (1919)"
tooinun «t il. (1979)" .
MUKUIO i El-llUr (1*))
fmla (1977)' j
•h«.^. ,1. (1U1)
Athiion it ll. (19B4Ik
Slldun it ll. (1919)
PndiBlninll* croc'doltu1
Achtton it il. (1982)
Hobbi it ll. (1980).
Jowl It ll. (1980T*_
ylgnlll i FtJ. (1981)'
NcDmuld 4 NcOonild (1*78)
AnUMSlalllM
Mturman tl ll, (1974)
UlC (TPMDlIU)
Klllnlild it li (»f*i
IroB.l li. (UN)
M«d Enpoiurai
AlDtn «l il. (I9S4)
tarry i HMTWUM (1983)
tarry i NHMHIH (1983) '
Flntclftlln (1S81)
fenfcnon i Entirllre (MOT)
Selllurr <t ll, (1978)
Sdlkarr <t il, (1979)
HelnMd ml «1. (1967)
lolonil «t l), (1380)
•MraiM I tarry (1979)
Nntuiu i Barry (1979)
Mcho)un (19761)
•unlonl it l). (1)7$)
Botttur 4 Coin (1980)
Mill (1984)
InduiLry
Git nit
T.illl.l
Mining
Mining
Mining
Frlcl. prod.
Mining
HinulictMrlng
1». Irlct.
TM. Irtfl.
Tu. Irlcl.
lu. Irlcl.
laltlllft
CMIt prod.
Minuficturlng
Huufictiirlna
Gil BUt
Mining
df Mf h
bf ••f.t
(•u feuk
Mining
Mining
Mining
Cacnl prnd.
Frlct. pnd.
Frlct. prod.
biMiilllon
Cewnt prod.
HmuficturlaB
Intulatlan
Intulitlan
Inxilittan
Shipyard
Hnmr^ctiirlna
Htaaltf turino
Muurac luring
Shipyard
Shipyard
Ctwit prod,
Country
UX
US*
US*
daMU
Ciaidi
bind.
US*
Italy
USA
USA
USA
USA
US*
UK
IK
Ul
USA
1*
Ainlril.
IK
Ul
CiMdl
Finland
iiin
USA
bfldin
UK
UK
UK
Caudi
USA
USA
USA
USA
USA
Ul
UK
USA
luly
Ul
USA
Su
m
n
N
N
N
F
N
F
F
MF
•r
rf
N
II
M
F
II
H
M
N
H
N
H
M
F
NF
n
H
N
OHM!
lotil S Vim of
NurfMr Unlwid follov-up
570
1261
2M1
9767
440
S44
3D)
9S2
2S4
4117
2722
S&4
1493
822
11S2
4820
UO
7S7
6200
951
S23
199
1091
ns
Ul
S98
7474
3708
170
241
10/1
17800
612
152
4779
4(00
922
tea
42647
6292
S64S
0.9
2.1
10.0
1.0
0,0
1.5
2.0
6.1
2.7
2.1
3.1
1,2
1.1
8.S
4,6
2.4
20.0
39 Z
6 S
11.6
4.7
4.0
i.a
2.9
3,3
8.0
23.
0.0
1,4
29.
1951-1980
1940- 7S
1938-77
1926-7S
1926-76
1961-77
1938-77
1946-75
1945-74
1938-77
1940-7S
1940- 75
1940-64
1933-74
193S-77
1947-78
1961-16
1951-80
1918-18
1941-18
1951-77
ISM-IS
19M-65
•*r4 SS
1947-75
19S7-80
1942-80
1942-80
1940-75
1963-00
1941-73
1967-76
1943-76
1945-65
19SO-IO
1916-15
19M-7S
WS9-71
1960-15
1947-78
1940-14
frac
DIIMt
ID*
IS*
20*
70*
20*
2ft*
20*
20*
0
]0«
u*
i«
lot
»*,
10*
15 •
10*
10*
10*
IS*
R.I.
20*
20*
IS*
20>
10*
10*
20*
20*
Ota.
177
sro
1291
84
178
Bfil
220
66
BM
912
UB
110
291
261
333
S28
21»
526
166
131
U
248
lit*
74
172
1339
299
122
72
781
1946
469
46
112
S4S
200
199
1070
1041
601
U.L K»)HS
Up. SM
148. S
1W 5
447.
1011.1
1S9 9
740.1
160.2
U8.0
eni
741.3
B9.1
139.7
224 9
HI. 2
298.0
M7.2
ZOJ.1
5B7.Z
13S.O
bl.l
1S7.4
1161.8
m.o
K ft
4M
640.7
1176.0
119.9
116.9
418.0
118. 0
134.3
BS3.7
998 4
B90 1
119
161
127
109
111
109
111
61
109
123
145
236
110
107
111
113
107
ga
gg
110
109
98
91
220
169
1»
141
147
96
124
169
144
US
104
6f
L
Obi.
6
11
210
n
71
9
4
S]
49
14
33
49
22
S7
81
11
60
12
10
1
21
13
9
12
14]
6
27
20
63
390n
91"
10
1]
103
J7
2J
123
84
SI
UK CJMCfB*
[J*. SM
4.S
9 B
29.6
L&4.0
1.2
11.1
49,1
8,7
4.1
50.5
16.1
i.r
14.8
12.9
25 8
29.1
21.9
64
38.2
6.3
3.1.
2.4"
12.6
4.S
3.3
6.6
13*. S
11.1
S
1.3
23.3
91.7
13.1
1.4"
7.5
43.2
3.2
8.4
54.9
100.1
49.2
133
336*
200*
US'
83
22S»
149*
103
93
UTS
13S*
§24'
221*
214-
85
196*
JW
M7-
157"
140*
273"
875*
W7*
289*
170'
iu«
103
S3
S40*
60S1
270*
416'
710*
702*
173
238'
Ml*
321'
224'
84
104
M CMCIB*
Dot. Exp. SM
4
18
26
2D9
10
19
IS
4
54
SO
a
16
16
M
19
28
S
10
1
7
7
3
19
1D3
29
13
4
ii
89
41q
S
40
10
11
66
61
»
4.9
B.I
17.1
203.7
9,5
SI 6
14.9
1.8
47.9
41.4
60
8,9
IS. 7
14.1
17.1
4.0
20.1
Ifl.I
14.9
t.D
1.0
10. e
107.2
27.4
I
2.S
39-9
S3. 2
14.8
i.a
M
10.2
s.o
41.6
76 Z
SO 1
82
124
1ST
101
IDS
114
103
109
113
121
131
IBfl-
102
111
U3
US
49
65
47
101
100
176*
96
UK
1300*
UD
138*
167"
*£91*
278*
118
196"
260*
136
83
50
BUUll
PI
1
0
0
10
1
1.
of
l'
0
ID
4
1
1
9
t
1
1
3
11
1]
3
0
n
0
4
e
2
8
6
S
61
11
|
Or
19
11
B
fl'
r of
Pir
0
1
1
0
0
D
0
0
4
S
1
B
0
0
1
7
2
a
4
1
S
0
1
0
0
0
0
S
S
0
112
27
2
0
27
B
1
D
Other unurt
Altai- In nctu it
Uniixc U»li «Jl lion Iml
17
Id
46
0
»
11.1 Uiyiu*
U
4 H.5
2 13.1
31
20.6
0
8.9
H
Oviry
14
1 0««ry
4
13
28
1" 1.1
10.1
ll" M
S.S
M
1EO Klaiwy
41
2
16.3
6.7
24
Sa.G Kidney, urinary
orgim, tnd laryiu
11 9
image:
Footnotes for Tab_1_e_3-2
a. The deaths from lung cancer and gastrointestinal cancer are those desig-
nated on the certificate of death. The cases of mesothelioma are those
determined from the review of all available evidence. Such cases will
not be included with the lung cancers. The asbestosis cases will be
those specifically listed, when provided. Otherwise, the number will be
the difference between the observed and expected for non-infectious
respiratory disease. The latter can be identified by the use of the
decimal point notation.
b. Two studies of the same plant but with different cohort definitions.
c. The majority of this cohort would also be included in that of McDonald
et al. (1980).
d. No mesotheliomas were identified in the defined cohort. However, three
mesotheliomas, two in women and one in an individual terminated prior to
1937, from this plant have been identified in the Tumor Registry of
Connecticut (Teta et al., 1983).
e. Twelve cases of pneumoconiosis were identified in this cohort. However,
these were all in individuals who had previous exposure to anthracite
coal containing silica.
f. Death certificate diagnosis of mesothelioma based upon clinical findings
and analysis of pleural fluid. No histological material was available
for review.
g. Significant at the 5 percent level in the entire cohort.
h. Three studies of the same plant at different periods of time and with
different cohort definitions. Between 3000 and 6000 tons of chrysotile
were used annually. Amosite constituted less than 1 percent of the
asbestos used except for a 3-year period, 1942-1944, where an average of
375 tons per year were used. Crocidolite usage was approximately 3-5
tons per year (Robinson et al., 1979).
i. Between 1931 and 1970 an average of 60 tons of crocidolite per year were
used (Berry et al., 1979). This would probably constitute about 1 percent
of the total fiber usage.
j. The factory operated between 1932 and 1980. Between 1932 and 1935 croci-
dolite and chrysotile asbestos were used; thereafter, only chrysotile.
The two mesothel iomas in this study were in the group exposed to both
chrysotile and crocidolite.
k. Amosite was the predominant fiber used. However, chrysotile was also
used between 1946 and 1973.
1. All of the groups in this category had a high exposure to crocidolite.
In some cases, however, there was also a substantial exposure to chrysotile
as wel1.
18
image:
m. Two cohorts at the same facility with different definitions and follow-up
periods.
n. Estimated as a proportion of deaths.
o. May have had exposure to asbestos in the construction industry,
p. Pleural mesothelioma or lung cancer.
q. Number of deaths based upon a review of all medical evidence.
r. No cases observed through the period of follow-up. Three cases have
occurred subsequently.
s. No cases occurred in the cohort as defined during the period of observa-
tion. Two occurred in individuals prior to 20 years from onset of employ-
ment and nine cases (8 pleural and 1 peritoneal) have developed subsequent
to termination of follow-up (Weill, 1984).
*p <0.05.
19
image:
work circumstance. Of the 41 groups listed, significantly increased (at the
5% level with a one-sided test) lung cancer is found in 32. Gastrointestinal
cancers are elevated at a significant level in 10. Moreover, strong exposure-
response relationships are seen for lung cancer and mesothelioma. They are
also seen for gastrointestinal cancer, but to a lesser extent.
The follow-up period was relatively long in most of the studies listed in
Table 3-2. However, in many cohorts, individuals continued to enter the
studies through the follow-up years, particularly in the period after World
War II. Thus, many individuals in some groups are just now reaching a time of
high potential risk for mesothelioma (30 or more years from onset of exposure).
In some cases, this can be seen in the finding of substantially increased
risks of mesothelioma subsequent to the termination of follow-up (see Table
3-2 footnotes).
3.4 MATHEMATICAL MODELS OF HUMAN CARCINOGENESIS
The quantitative determination of cancer risk in an occupational group
can be used to predict risks in similar exposure circumstances in the absence
of any model of action; observations in one group would apply to identically
exposed workers. If, however, a risk determination fits within the framework
of a general mathematical model for cancer, then predictions outside the range
of measurement can be made within the range of validity of a model. Validation
of a mathematical model, of course, requires the testing of such predictions.
If a mathematical model has a mechanistic basis, e.g., at a molecular level of
action, its use is considerably strengthened. To the extent that a model is
applicable, it strengthens risk estimates made for exposures and times different
from those directly observed. To the extent that a model may be applicable,
it points to issues that must be considered in any general risk assessment.
In the case of human carcinogenesis, a variety of multistage models have
been proposed to describe a number of observations, most notably the power law
dependence of human cancer risk with age and the time and dose dependence of
induced malignancy in some animal experiments. The models were initially
suggested to explain the observation that site-specific cancer mortality
increases as the fifth or sixth power of age (e.g., Cook et al., 1969;
Armitage and Doll, 1954). The models suggested ranged from proposals that
multiple (up to six or seven) mutations (or carcinogenic events) occur in the
20
image:
same or adjacent cells (Muller, 1951; Fisher and Holloman, 1951; Nordling,
1953) to models that involve preferential clonal development of altered cell
lines (Fisher, 1958; Armitage and Doll, 1957, 1961). Depending on the model,
some or all of the states are capable of being affected by an external carci-
nogen. For those susceptible states, it is expected that the probability of
progression to the next stage would be proportional to the time that a car-
cinogenic agent, or its active metabolite, is at a reaction site. A constant
exposure to environmental carcinogens would then introduce a power of time for
each state that is affected by a particular external carcinogen. Powers of
time also arise from exposure-independent processes. It is important to note,
however, that a power of dose is introduced for each exposure-dependent step
(for short-term exposures). Motivated by the experimental demonstration of
initiation and promotion in skin cancer (Berenblum and Shubik, 1949), Armitage
and Doll (1957) discuss a two-state model with an intermediate time-dependent
growth phase that is compatible with the observed age dependence of cancer
incidence.
In its generalized form, the model suggests that the time dependence of
site-specific cancer incidence in the general population is
I(t) = CX^ ... X^t-w)^1 (3-1)
where the A. are the transition probabilities of each state, k is the number
of stages and w 1s the growth time for a fully transformed cell to become
clinically detectable. One, or several, of the X^ can be influenced by the
application of an external carcinogen. There would be a power of dose (or
intensity of exposure) for each stage so affected. To account for this, the
most general form of the multistage model can be written
Kt) = C(qQ + I,q1d1)(t-w)k~1 (3-2)
Within this model, one can consider carcinogenic action on specific stages at
different times in the carcinogenic process.
Whlttemore (1977a, 1977b) and Day and Brown (1980) have explored some of
the time courses of cancer risk that are predicted by the model. The important
aspects of these analyses are:
21
image:
1. The effects of early stage carcinogens are most Important early
1n life (the cells or cell lines that are started 1n the car-
cinogenic process are available for a long time for further
alteration). In addition, their effect diminishes slowly after
cessation of exposures relative to continuous exposure.
2. The effects of late-stage carcinogens are most Important late
1n life when many altered cells are available to be acted upon.
The effects of exposure to late-stage carcinogens diminish
rapidly after cessation of exposure.
3. For each stage that an externally applied carcinogen acts,
there 1s a power of intensity of exposure (or dose for short-
term exposures).
Thus, the. predicted time dependence of cancer risk can be highly varied
depending on the stage affected, and subllnear, as well as linear, dose-
response relationships can be Incorporated within the model. Here, sublinear
refers to a relationship that contains a power of dose greater than unity. A
supralinear relationship Is not contained within the framework of the model.
The multistage model has provided a basis for dose-Incidence extrapo-
lation procedures. These have been formulated by Guess, Crump, and others
(Guess and Crump, 1976, 1978; Guess et al.f 1977). The procedure makes no a
priori assumptions on the dose-response relationship, but utilizes a maximum
likelihood procedure to calculate the q. values along with their 95 percent
confidence limits. In practice, 1t is found that most experimental carcinoge-
nesls data cannot rule out a linear dose term. Thus, the 95 percent confi-
dence limit on the risk at low exposure is dominated by the uncertainty on the
linear term (Guess et al., 1977).
It should be noted that the exposure 1n the multistage model 1s to the
site of action of an alterable cell. Significant non-linearities can be
introduced Into an exposure-response relationship by non-linearities 1n the
metabolism of a chemical to an active species or 1n the detoxification of an
active chemical. Such non-Hnearit1es have been observed in the case of vinyl
chloride (Gehring et al., 1978). A general discussion of activation non-
linearities 1n dose-response relationships has been published by Hoel et al.
(1983).
22
image:
Human data supporting a multistage model are Hun"ted because of lack of
information on the age, time, and dose dependence of cancer risk from exposure
to external agents. Recent data from the study of smoking effects among
British doctors (Doll and Peto, 1978) suggest that the dose-response relation-
ship is quadratic and that cigarette smoke may act at two stages, one early
and one late, in the carcinogenic process. This concept is supported by the
partial reduction 1n lung cancer risk after smoking cessation (relative to
continued smoking). On the other hand, U.S. smoking data suggest a linear
dose-response relationship (Hammond, 1966; Kahn, 1966). In the case of radia-
tion, the long lasting Increased risk of solid tumors among residents of
Hiroshima and Nagasaki (Beebe et a!., 1978) suggests an early stage action for
radiation. However, the age dependence of risk demonstrates a risk that 1s
proportional to the risk In the absence of radiation exposure, suggesting a
late-stage action. The dose-response relationship, however, does not suggest
a supra-linear relationship, which would be the case if two stages were affected.
In contrast to a somewhat equivocal application to human data, the model de-
scribes very well the time and dose dependence of skin tumors 1n benzo(a)pyrene
painted mice (Lee and O'Neill, 1971; Peto et al., 1975).
In summary, the multistage model provides a useful conceptual framework
for considering the age, time and dose dependence of site specific cancer
incidence. However, 1t 1s so general that it can be made to fit virtually any
animal or human carcinogenesls dose-response data. The requirements are more
stringent for fitting t1me-to-tumor data. Here, however, few human data are
available for validation. At this time, the model cannot predict a priori
either the dose or time dependence of human cancer. Nevertheless, the concepts
of the model are plausible and warrant consideration when the data on the age,
time, and dose dependence of asbestos cancers are reviewed.
3.5 LINEARITY OF EXPOSURE-RESPONSE RELATIONSHIPS
Direct evidence for linearity of response with asbestos exposure is
available from seven studies (two of the same plant) that compared lung cancer
mortality to the cumulative total dust exposure 1n asbestos workplaces (Dement
et al.,.1982; Henderson and Enterline, 1979; McDonald et al., 1980, 1983a,
1983b; Finkelsteln, 1983; Seidman, 1984). Figure 3-1 plots the exposure-
response data in these studies as the ratio of observed to expected lung
23
image:
10
i
ff-i 1 1 r
/ O DEMENT etBl,(1983bl
I Q MCDONALD at ai, (19B3«|
I A MCDONALD «t ai. <iW3b)
I O HENDERSON AND
/ ENTERLINE, (19791
/ • MCDONALD at ai. (1930)
200 400 600 BOO 1000
ESTIMATED ASBESTOS DOSE, mppcf-y
20
16
1200 0
O FINKELSTEIN (1983)
OSEIDMAN, (1984)
I L__
100 200 300 400
ESTIMATED ASBESTOS DOSE, f-Y/ml
Figure 3-1. Exposure response relationships for lung cancer observed in seven studies.
Cumulative exposures are measured in terms of millions of particles per cubic foot-
years (mppcf-y) or fibers per milliliter-years (f-y/ml).
24
image:
cancer mortality against the measured cumulative dust exposure 1n millions of
particles per cubic foot-years (mppcf-y) or cumulative asbestos exposure In
fiber-years per m1H1l1ter (f-y/tnl). (Henceforth, the term "dose" will be
used to designate cumulative exposure.) While different exposure-response
relationships appear to exist for the five studies of Figure 3-la, each demon-
strates a very good linear relationship over the entire range of observation.
The differences 1n the slopes of the relationships may relate to differences
1n the quantity of the other dust present, the fiber size distribution, the
fiber type, the age of the population under observation, the representative-
ness of the dust sampling programs and possibly other factors. .These factors
are discussed later, when the exposure-response relationships1 of all available
studies are compared (see Section 3.9). In the case of the two studies 1n
Figure 3-lb, the form of the dose-response relationship 1s less clear, particu-
larly for the group studied by F1nkelste1n (1983). The data from three other
studies that provide dose-response information are not shown. In one (Wei 11
et al., 1979), the dose-response relationship was affected by the large number
of untraced individuals in the study; 1n two others of friction products manu-
facturing (Berry and Newhouse, 1983; McDonald et a.l., 1984), the relationship
was too weak to provide any guidance as to its form. (These three studies are
considered later, 1n Section 3.9.) In one case, when exposure-response rela-
tionships were analyzed according to both duration and intensity of exposure
(McDonald et al., 1980), the results were less dramatic than shown 1n Figure
3-la. However, this may be the result of small numbers; only 46 excess lung
cancer deaths are reported in all exposure categories.
In the discussion of the time relationship of lung cancer risk and asbestos
exposure, the data can be Interpreted in terms of a multistage model of cancer
1n which asbestos appears to act at a single late stage. The continued high
risk following cessation of exposure results from the continued presence of
asbestos in the lungs. This model 1s compatible with a linear dose-response
relationship and with the synerglstic Interaction between asbestos and cigar-
ette smoking.
Fewer data are available on the exposure-response relationship for meso-
thelioma. Table 3-3 lists the mesothelloma mortality from four studies (Seidman,
1984; Hobbs et al., 1980; Jones et al., 1980; Finkelsteln, 1983) in terms of
cases per 1000 person-years of observation or percentage of mesothelioma
deaths. The data of Seidman are presented both 1n terms of duration of employ-
ment and estimated cumulative fiber exposure. The exposure circumstances of
25
image:
TABLE 3-3. THE RISK OF DEATH FROM MESOTHELIOMA ACCORDING TO THE TIME
OF ASBESTOS EXPOSURE, IN FOUR STUDIES
Study
Hobbs et
Jones et
Seidman
Exposure
period
(months
unless
noted)
al. (1980)
<3
3 - 11
12+
al. (1980)
<5
5 - 10
10 - 20
20 - 30
30+
(1984)
2.2
7.1
15.4
57
8.8?
37.5a
TC
125a
200a
375a
Number
of
deaths
0
10
16
0
3
4
4
5
1
5
4
7
2
5
6
2
1
1
Estimated
person-years
(10+ years
from first
exposure)
21,213
19,548
14,833
3,700
1,203
1,263
1,248
4,104
1,162
1,053
420
425
250
Deaths/
1000
person-
years
0
0.5
1.1
2.7
4.2
3.2
5.6
0.5
4.3
5.7
4.8
2.4
4.0
Percent
Number of
exposed deaths
314 0
116 2.6
145 2. 8
101 4.0
51 9.8
Finkelsteln (1983)
44
92
180
1
2
6
1.9
4.9
11.9
Exposure 1n fiber-years/ml.
26
image:
the groups studied by Jones et al. (1980) and Seldman (1984) offer the Ideal
circumstances for. studying the effects of cumulative exposure on risk. The
average exposure duration of each group was short (less then two years) and
all Individuals began exposure at approximately the same time during World War
II. Thus, the confounding effect of time on the observed risk 20 or more
years from onset of exposure 1s largely removed, To the extent that the
distributions 1n duration and time from onset of «mployment are similar 1n the
different exposure categories of F1nkelste1n (1983) and Hobbs et al. (1980),
the data would reflect an exposure-response relationship. This 1s likely to
be approximately correct, but direct Information 1s not available.
Figure 3-2 displays the data of Table 3-3. To the extent that duration
of employment 1s related to dose, the studies of Jones et al. (1980) and Hobbs
et al. (1980) are compatible with a linear dose-response relationship, as Is
that of F1nkelste1n (1983). The study of Seldman (1984) is highly non-linear,
especially when mesothelloma risk 1s plotted against estimated dose 1n f-y/ral.
The relationship, however, 1s supralfriear (I.e., one Involving fractional
powers of dose). This 1s likely to be the result of statistical uncertainties
associated with small numbers rather than exposure m1sclass1f1cat1on; 1n the
case of lung cancer a highly linear dose-response relationship was observed,
albeit one that suggested a zero dose Intercept at an SMR (standard mortality
ratio) greater than 100.
Polynomials of degree one and two were fitted to the data of Jones et al.
(1980), Hobbs et al. (1980), and F1nkelste1n (1983). The effect of Including a
quadratic term 1s shown 1n Table 3-4. In no case 1s a quadratic term re-
quired; 1n one case Its coefficient 1s negative, Indicating a suprallnear
relationship, and 1n the case where the effect 1s greatest (Flnkelstein,
1983), the effect on the slope at zero dose 1s only a factor of 1.76. A
quadratic term for the data of Seldman (1984) 1s clearly unwarranted.
A final study which provides some dose-response Information 1s that of
Newhouse and Berry (1979), which shows an Increasing risk of mesothelloma with
Increasing duration and Intensity of exposure (Table 3-5). However, a quanti-
tative relationship cannot be determined.
Because of the limited dose-response data, the model for mesothelloma 1s
not as well established as that for lung cancer. As will be seen, the time
course of mesothelloma appears to be related only to the asbestos exposure.
At this time, no Interactive effects have been observed between asbestos and
27
image:
I r i
D FINKELSTEIN (1983)
OSEIDMAN (19841
I
L
1
0 100 200 300 400 500 600
CONCENTRATION, f • y/ml
OHOBBS atai.d980(
QSEIDMAN (1984)
A JONES atfll, (1990)
30
TIME, months
40
60 60
Figure 3-2. Exposure-response relationships for mesothelioma observed in four studies.
Exposures are measured in terms of fiber per rmlliliter-years (f-y/ml) or duration of em-
ployment.
28
image:
TABLE 3-4. ANALYSIS OF RESIDUALS IN POLYNOMIAL FIT TO OBSERVED
MESOTHELIOMA DOSE-RESPONSE DATA
Sum of Squares
Accounted for by
Study
Hobbs et al . , 1980
Jones et al . , 1980
Flnkel stein, 1983
Linear
term
0.8133
77.64
78.50
Quadratic
term
0. 0015
0.51
1.19
Residual
0.0067
2.92
0.27
Prob-
ability3
0.72
0.39
0.28
Ratio offa
slopes
0.85C
1.38
1.76
The probability that the observed deviation from linearity 1s by chance alone.
The ratio of the slope of the dose-response function at zero dose without and
with Inclusion of a quadratic term.
GThe sign of the quadratic term is negative Indicating a supralinear relation-
ship (I.e., one containing fractional powers of dose).
TABLE 3-5. RISK OF MESOTHELIOMA/100,000 PERSON-YEARS WITH INCREASING
DURATION AND INTENSITY OF EXPOSURE (Newhouse and Berry, 1979)
Duration of
exposure
Males <2 yrs
>2 yrs
Females <2 yrs
>2 yrs
Deaths/100,000 Person- Years
Intensity of Exposure
Low- moderate3
33
93
{48}
combined
Severe
104
243
136
360
*5-10 f/ml.
3>20 f/ml.
other agents 1n the etiology of the disease. The steep power law dependence
of risk on time from asbestos exposure suggests that mesothelloma can be
described within the framework of the multistage model (see Peto et al., 1982)
and that asbestos may act early 1n the carcinogenic process. However, because
asbestos has been shown to act late 1n the carcinogenic process 1n the case of
lung cancer, It could do so also 1n the case of mesothelloma. If so, the
dose-response relationship would Involve higher than linear powers of dose.
29
image:
While a quadratic component in the dose-response relationship has plausibility,
the existing data provide no support for it. Further, the finding of mesothe-
lioma among family contacts of workers suggests that a substantial risk exists
at much less than occupational exposures among family contacts of chrysotile
miners and millers and amosite factory workers. Among the miners and millers,
3 family member contact cases are known (McDonald and McDonald, 1980) compared
to 12 among the miners and millers. For the amosite factory workers, there
are 4 cases of family member contact mesothelioma compared to 15 cases due to
occupational exposure (Anderson et al., 1976).
Even more limited data are available on a dose-response relationship for
gastrointestinal (GI) cancer. As seen in Table 3-2, the strength of the
evidence relating asbestos exposure to GI malignancy is less than that from
lung cancer and mesothelioma; the excess relative risk, when present, is lower
than that for lung cancer. Of the seven studies providing a clear dose-response
relationship for lung cancer, information is available from six of them on a
dose-response relationship for GI cancer. Weighted least squares regression
analyses were run on the data of the studies. Table 3-6 lists the coefficients
of these analyses, along with the standard errors of the slopes. As can be
seen, five of the six studies which demonstrated a fairly steep dose-response
relationship for lung cancer demonstrate a consistent and positive trend with
exposure for GI cancer, but less strong than that for lung cancer. However,
while indicating a positive trend with exposure, the data on GI cancer dose-
response relationships are inadequate to establish the functional relationship
between dose and risk.
This document uses a linear exposure-response relationship for estimating
unit exposure risks for lung cancer and mesothelioma and for calculating risks
at cumulative exposures 1/10 to 1/100 of those of the occupational circum-
stances of past years. It is a plausible relationship, and for lung cancer is
strongly indicated by the existing evidence. While more limited data exist
for mesothelioma, they also indicate a linear relationship. Its use has three
distinct advantages: 1) point estimates of exposure-response can be made with-
out knowledge of individual exposures, i.e., the excess mortality of an entire
group can be related to the average exposure of the group; 2) extrapolation to
various exposure circumstances can be made easily; and 3) it is likely to be a
conservative extrapolation procedure from the point of view of human health.
It is emphasized that linearity of exposure-response obtains only for similar
times of exposure and observation among similarly aged individuals with similar
personal habits.
30
image:
TABLE 3-6. COMPARISON OF LINEAR WEIGHTED REGRESSION EQUATIONS FOR LUNG CANCER
AND GI CANCER IN SIX COHORTS OF ASBESTOS-EXPOSED WORKERS
Study
Regression Equation
Lung cancer
GI cancer
Dement et al., 1983b
McDonald et al., 1983a
McDonald et al., 1983b
McDonald et al., 1980
Seldman, 1984
Finkelsteln, 1983
Textiles
SMR = 151 + 4.19(±0.84)f-y/mlb
SMR = 110 + 2.07(±0.25)f-y/ml
%RR - 61 + 2.27(±0.63)f-y/mlc
SMR = 53 + 0.86(10.15)f-y/ral
%RR = 70 + 1.20(±0.33)f-y/ml
Mining
SMR = 92 + O.D43(±0.008)f-y/ml
Manufacturing
SMR = 325 + 2.72(±0.54)f-y/ml
XRR = 100 * 4.79(±2.70)f-y/ml
SMR = 34 + 1.18(±0.62)f-y/ml
SMR = 113 + O.S9(±0.37)f-y/rol
%RR = 82 + 1.19(±0.42)f-y/ral
SMR = 82 + 0.42(±0.19)f-y/ml
XRR = 84 + 0.38(±0.32)f-y/ml
SMR = 88 + 0.011(±0.010)f-y/ml
SMR = 110 + 0.084(±0.43)f-y/ml
%RR = 100 + 3.11(±0.16)f-y/ml
Equations are calculated for the Increased risk per f-y/nl of exposure. Data of McDonald et al., given in mppcf-y,
were converted to f-y/ml using the relationship: 1 mppcf = 3 f/ml.
± standard error of the coefficient of f-y/ml.
1s relative risk x 100.
image:
3.6 TIME'AND AGE DEPENDENCE OF LUNG CANCER
A relative risk model has long been assumed to be applicable for the
description of the incidence of lung cancer induced by occupational asbestos
exposure. Such a model is tacitly assumed in the description of mortality in
terms of observed and expected deaths. Virtually every study of asbestos
workers is described in these terms. Early suggestive evidence supporting it
is found in the synergistic action between asbestos exposure and cigarette
smoking (Selikoff et al., 1968), in which the lung cancer risk from asbestos
exposure depended on the underlying risk in the absence of exposure. Relative
risk models were discussed previously by Enterline (1976) and Peto (1977) and
utilized in projections of lung cancer from past asbestos exposure by Nicholson
et al. (1982). They were adopted in the risk analyses of the Advisory Committee
on Asbestos (1979a,b), the U.S. Consumer Product Safety Commission (1983),
and the National Academy of Sciences (1984). Information on lung cancer risk
from exposures at different ages is now available from two studies (Selikoff
et al., 1979; Seidman, 1984), The analyses of these data, along with the
observations of linear dose-response relationships, provide substantial sup-
port for the use of such a formulation for lung cancer.
Information from the insulation workers study by Selikoff et al. (1979)
on the time course of asbestos cancer risk is given in Figure 3-3, which shows
the relative risk (here taken to be the ratio of observed to expected deaths)
of death from lung cancer according to age for individuals first employed
between ages 15 and 24 and for those employed between ages 25 and 34. The two
curves rise with the same slope and are separated by the 10 years of difference
in age at first exposure. This suggests that the relative risk of developing
asbestos-related lung cancer according to time from onset of exposure is
independent of age and of the pre-existing risk at the time of exposure. In
contrast, both the slope and the value of the excess risk of lung cancer are
two to four times greater for the group first exposed at older ages compared
to those exposed at younger ages. The similarity of the data for each group
in Figure 3-3 suggests that the data be combined and plotted according to time
from onset of exposure. The result, shown in Figure 3-4, plots the data
according to years from onset of exposure. However, because of the great
stability of union insulation work, the curve also reflects effects according
to duration of exposure up to at least 25 years from onset of exposure. A
linear increase with time from onset of exposure occurs for about 35 years
32
image:
IS
a»
•o
93
Q.
X
09
1
I 4
J3
O
*
CO
ee
u 3
AGE AT ONSET
• 15 24 YEARS
O 25-34 YEARS
30
40
50
60
AGE
70
80
90
Figure 3-3. The relative risk of death from lung cancer
among insulation workers according to age. Data supplied
by I.J. Selikoff and H. Seidman.
Source: Nicholson (1982).
33
image:
(/)
»rf
<0
X
e
I
0) _
> 3
(A
E 2
LU
1
LU .
DC 1
• ALL WORKERS
O WORKERS WHO SMOKE CIGARETTES
10
20
30
40
50
60
YEARS FROM ONSET OF EXPOSURE
Figure 3-4, The relative risk of death from lung cancer among
insulation workers according to time from onset of exposure
( • all insulators; O indicates insulators who were smoking
cigarettes at the start of follow-up in 1967.) Data supplied by
I.J Selikoff and H, Seidman.
Source: Nicholson (1982),
34
image:
(to about the time when many Insulation workers would have terminated employ-
ment), after which the relative risk falls substantially. The decrease 1s, 1n
part, the result of the earlier deaths of smokers from the group under study
due to their higher mortality from lung cancer and cardiovascular disease.
However, the decrease 1s not solely the result of the deaths of smokers since
a similar rise and fall occurs among those Individuals who were smokers at the
start of the study compared to smokers 1n the general population. Part of
the decrease may relate to the elimination of asbestos, particularly chrysotlle,
from the lung; selection processes, such as differing exposure patterns (e.g.,
the survivors may have avoided Intense exposures); or differing Individual
biological susceptibilities. While the exact reason for the effect 1s not
understood, 1t 1s a general phenomenon seen 1n other mortality studies of
asbestos workers (Nicholson, et al., 1979; 1985).
The early portions of the curves of Figures 3-3 and 3-4 have three Impor-
tant features. After a short delay, ttjey show a linear Increase 1n the relative
risk of asbestos lung cancer according to time from onset of exposure. Figure
3-4 shows that this Increased relative'risk 1s proportional to the time worked,
and, thus, to the cumulative asbestos exposure. However, the linear rise can
occur only 1f the Increased relative risk that 1s created by a given cumulative
exposure of asbestos continues to multiply the underlying risk for several de-
cades thereafter. Finally, an extrapolated linear line through the observed
data points crosses the line of relative risk equal to one (that expected in
an unexposed population) at between five and ten years from onset of exposure.
This means that the increased relative risk appropriate to a given exposure 1s
achieved soon after the exposure takes place. However, 1f there 1s a low
underlying risk at the time of the asbestos exposure (as for Individuals aged
20-30), most of the cancers that will arise from any Increased risk attribu-
table to asbestos will not occur for many years or even decades until the
underlying risk becomes substantially greater.
The data of Seldman (1984) also show that exposure to asbestos multiplies
the pre-existing risk of lung cancer and that the multiplied risk becomes
manifest 1n a relatively short time. Figure 3-5 depicts the time course of
lung cancer mortality beginning five years after onset of exposure of a group
exposed for short periods of time. The average duration of exposure was 1.46
years; 77 percent of the population was employed for less than 2 years. Thus,
35
image:
I
e 5
I
v.
J 4
I
V)
1
§
10
20
30
40
YEARS FROM ONSET OF EXPOSURE
Figure 3-5. The relative risk of death from lung
cancer (BE) among amosite factory workers
according to time from onset of exposure.
Source: Seidman (1984).
36
image:
exposure had largely ceased prior to the beginning of the follow-up period. A
rise to a significantly elevated relative risk occurred within ten year* and
remained constant throughout the observation period of the study. Furthermore,
the relative risk from a specific exposure 1s independent of the age at which
exposure began, whereas the excess risk would have increased considerably with
the age of exposure. Table 3-7 shows the relative risk of death from lung
cancer for individuals exposed for less than and greater than 25 f-y/ml accord-
Ing to age at time of entrance Into a ten-year observation period. Within a
given age category, relative risk was similar during different decades from
onset of exposure, as previously shown in Figure 3-5 with the overall data.
However, relative risk also was independent of the age decade at entry into a
ten-year observation period (see rows labeled "AVI" in each exposure category
of Table 3-7). There is some reduction in the oldest, most heavily exposed
group. This may be attributed to the same selection effects manifest at older
ages in insulation workers.
In terms of carcinogenic mechanisms, it appears that asbestos acts largely
like a lung cancer-promoting agent. However, because of the continued resi-
dence of the fibers in the lung, the promotional effect does not diminish with
time after cessation of exposure as it may with chemical or tobacco promoters.
Further, inhalation of the fibers can precede Initiating events because many
fibers remain continuously available in the lung to act after other necessary
carcinogenic processes occur.
A feature of Figure 3-4 important in the assessment of asbestos carcino-
genic risk is the decrease in relative risk after 40 years from onset of
exposure, or 60 years of age. As mentioned previously, we do not have a full
understanding of this decrease, but it generally applies. A virtually identical
time course of lung cancer risk occurs in asbestos factory employees (Nicholson
et al., 1985) and in Canadian chrysotile miners and millers (Nicholson et al.,
1979), Because of the significant decrease at long times from onset of expo-
sure and older ages, observations on retiree populations can seriously under-
state the actual risk of asbestos-related death during earlier years. To the
extent that time periods between 25 and 40 years from onset of exposure are
omitted from observation, a study will underestimate the full impact of asbestos
exposure on death.
37
image:
TABLE 3-7. RELATIVE RISK OF LUNG CANCER DURING 10-YEAR INTERVALS
AT DIFFERENT TIMES FROM ONSET OF EXPOSURE
Years from
onset of
exposure
4T
- 49
Age
at
50 -
Less than
5
15
25
35
All
5
15
25
35
All
0.
12.
5.
—
6.
0.
7.
25.
--
8.
0
0
9
3
0
7
0
3
[0.7]a
(3)
(1)
(4)
Greater
U-7]
(2)
(3)
(5)
1
5
2
2
3
25
.4
.1
.3
.8
.0
start of
59
period
60
, years
- 69
70
- 79
f-y/ml exposure
(Db
(4)
(2)
(1)
(8)
than 25 f-y/ml
12
11
9
4
10
.9
.1
.7
.3
.5
(8)
(8)
(7)
(1)
(24)
0.0
2.2
6.4
8.1
3,9
[4.1]
(3)
(9)
(6)
(18)
0
4
28
1
3
.0
.9
.0
.9
.1
[0.7]
(5)
(3)
(1)
(9)
exposure
6.6
5.6
12.0
4.0
7.6
(5)
(6)
(13)
(2)
(26)
3
6
2
8
4
.7
.2
.1
.8
.5
(1)
(4)
(2)
(3)
(10)
a[] = no cases seen. Number of cases expected on the basis of the average
relative risk 1n the overall exposure category.
() = number of cases.
Source: Seidraan (1984).
To appreciate the effect of the observed lung cancer time dependence upon
the results of an ep1dem1olog1cal study, the excess risk of lung cancer was
calculated for different observation periods for a hypothetical group exposed
for 25 years beginning at age 20. The time course of the risk was set propor-
tional to that of Figure 3-4 and 1978 general population rates were used.
Table 3-8 lists the percent excess lung cancer mortality observed for three
follow-up periods, 10 years, 20 years, and lifetime, beginning at different
ages. As can be,seen, the percent excess risk from start of exposure at age
20 to the complete death of all cohort members 1s 55 percent of the maximum.
The percent excess risk Increases up to age 50 as the follow-up period starts
later, reflecting the Increased relative risk concomitant with Increased
exposure. For observations starting after age 50 1t falls substantially;
follow-up begun at age 65 observes only 38 percent of the full risk. To the
extent that a group under observation has an age distribution that Is similar
38
image:
TABLE 3-8. ESTIMATES OF THE PERCENTAGE OF THE MAXIMUM EXPRESSED EXCESS
RISK OF DEATH FROM LUNG CANCER FOR A 25-YEAR EXPOSURE
TO ASBESTOS BEGINNING AT AGE 20a
Age at start of
observation,
years
20
30
40
50
60
65
70
10
2
34
69
97
73
55
37
Period of follow-up,
20
32
65
91
81
55
41
29
years
Lifetime
55
55
56
55
46
38
29
Years from
onset of
exposure
0
10
20
30
40
45
50
aThe maximum expressed risk is that manifest 7.5 years after the conclusion
of the 25-year exposure.
to the number alive 1n each quinquennium 1n a lifetime follow-up, an observation
for any period of time would reflect the same mortality ratio as an observation
from onset of exposure to the death of the total cohort.
The data in Table 3-8 came from observations on long-term exposures to
high concentrations of asbestos (>10 f/ml) where preferential death of suscep-
tible individuals occurred. Thus, appropriate comparisons between heavily
exposed groups could be made on the basis of lifetime risk (i.e. 55 percent of
the maximum), as well as on the maximum risk. However, in groups exposed to
low levels (<0.1 f/ml), even for many years, selection effects may be much
less important. A minimal excess risk would barely affect the pool of suscep-
tible individuals. A lesser effect would also be expected from short-term
exposures (to less than extreme concentrations). If selection effects are
largely the cause of the disease, the maximum expressed relative risk would be
most appropriate for estimating risks associated with low-level exposures.
However, 1f the decrease is largely the result of elimination of asbestos from
the lung or the biological neutralization of deposited fibers, a decrease in
relative risk beginning at about 35 years from onset of exposure should be
considered. This Is discussed in Chapter 6.
The above discussion supports a general model for lung cancer 1n which
the asbestos-related risk, t years from onset of exposure, is proportional to
the cumulative exposure to asbestos at time t-10 years multiplied by the age
39
image:
and the calendar year risk of lung cancer in the absence of exposure. The
incidence of lung cancer can be expressed formally by
IL(a,y,t,d,f) = IE(a,y) [1 + K,_-f-d(t-lO)] (3-3a)
Here, I.(a,y,t,d,f) is the lung cancer incidence observed or projected in a
population of age, a, observed in calendar period, y, at t years from onset of
an asbestos exposure of duration, d, and average intensity, f. Ir(a,y) is the
age and calendar year lung cancer incidence expected in the absence of exposure.
If smoking data are available, IL and 1^ can be smoking-specific incidences.
f is the intensity of asbestos exposure to fibers longer than 5 urn/ml (f/ml),
d is the duration of exposure up to 10 years from observation, and K. is a
proportionality constant that is a measure of the carcinogenic potency of the
asbestos exposure. A delay in manifestation of risk is based on the data of
Seidman (1984) and Selikoff et al. (1979); in neither study was any excess
lung cancer seen prior to 10 years from onset of exposure. From Equation 3-3a,
the relative risk of lung cancer, Ii/Iri is independent of age and depends
only on the cumulative exposure to asbestos.
Different asbestos varieties have different size distributions, and the
fraction greater than 5 urn depends on fiber type and the production process
(Nicholson et al., 1972; Gibbs and Hwang, 1975). Animal data demonstrate that
dimensions (length and width) are important variables in fiber carcinogeni-
city. Thus, K. would be expected to depend on fiber type and fiber dimension.
In practice, however, uncertainties in establishing quantitative dose-response
relations, through the application of Equation 3-3a to observed data, may
preclude the determination of K. by fiber type (see Section 3.17).
3.7 MULTIPLE FACTOR INTERACTION WITH CIGARETTE SMOKING
The multiplicative interaction between asbestos and the underlying risk
of lung cancer is seen in the synergism between cigarette smoking and asbestos
exposure, first identified by Selikoff et al. (1968). Later data on U.S.
insulation workers confirm and extend the initial findings (Hammond et al.,
1979a): In this larger study, 12,051 asbestos workers, 20 or more years from
onset of their exposure, were followed from 1967 through 1976. At the outset,
6841 volunteered a history of cigarette smoking, 1379 said they had not smoked
40
image:
cigarettes, and the rest provided no Information. By January 1, 1977, 299
deaths had occurred among the cigarette smokers and 8 among those not reported
as smokers.
This experience was compared to an age- and calendar year-specific basis
with that of like men with the same smoking habits 1n the American Cancer
Society's prospective Cancer Prevention Study (Hammond, 1966). For the control
group, 73,763 white males who were exposed to dusts, fumes, gases, or chemicals
at non-farming work were selected. The age standardized rates per 100,000
person-years for each group are shown 1n Table 3-9. The results show that
both the smoking and non-smoking lung cancer risks are multiplied five times
by the worker's asbestos exposure. However, since the risk 1s low for non-
smokers, multiplying 1t five times does not result 1n many cases, although any
excess 1s clearly undesirable. On the other hand, smoking by Itself causes a
major Increase and when that high risk 1s then multiplied five times, an
Immense Increase Is found. Corroborative data on the multiplicative smoklng-
asbestos Interaction are seen 1n studies by Berry et al. (1972), McDonald et
al. (1980), and Sellkoff et al. (1980). However, these do not show as exact a
multiplicative effect as that of Hammond et al, (1979a).
TABLE 3-9. AGE-STANDARDIZED LUNG CANCER DEATH RATES FOR CIGARETTE SMOKING
AND/OR OCCUPATIONAL EXPOSURE TO ASBESTOS DUST COMPARED WITH NO
SMOKING AND NO OCCUPATIONAL EXPOSURE TO ASBESTOS DUST
Group
Control
Asbestos Workers
Control
Asbestos Workers
Exposure
to
asbestos?
No
Yes
No
Yes
History
cigarette
smokl ng?
No
No
Yes
Yes
Death
rate3
11.3
58.4
122.6
601.6
Mortality
difference
0.0
+47.1
+111.3
+590.3
Mortality
ratio
1.00
5.17
10.85
53.24
Rate per 100,000 person-years standardized for age on the distribution of
the person-years of all the asbestos workers. Number of lung cancer deaths
based on death certificate Information.
Source: Hammond et al. (1979a).
41
image:
The study by Hammond et al. (1979a) also carried the asbestos-smoking
interaction a step further, to show increased risk of death from asbestosis.
As noted previously, .insulation work carries a risk of fatal progressive
pulmonary fibrosls, and some of those who never smoked cigarettes died of
asbestos!s. Nevertheless, asbestosis mortality for men who smoked a pack or
more a day was 2.8 times higher than for men who never smoked regularly.
Cigarette smoking, with resulting bronchitis and emphysema, adds an undesirable
and sometimes unsupportable burden to the asbestos-induced pneumoconiosis.
Interactive effects between cigarette smoking and the prevalence of X-ray
abnormalities were reported previously (Weiss, 1971). However, no relation-
ship was found 1n the Hammond et al. (1979a) study (Seldman, quoted in Frank,
1979) between cigarette smoking and the risk of death from mesothelioma or gas-
trointestinal cancer.
3.8 METHODOLOGICAL LIMITATIONS IN ESTABLISHING DOSE-RESPONSE RELATIONSHIPS
There are substantial difficulties in establishing dose-response relation-
ships for human exposure to asbestos, perhaps the most important being that
current health effects are the result of exposures to dust 1n previous decades
when few and imperfect measurements of fiber concentrations were made. Current
estimates of what such concentrations might have been can be inaccurate, since
individual exposures were highly variable. Further, while disease response
now can be established through epidemiological studies, these, too, can be
misleading because of methodological limitations. Despite this difficulty,
useful estimates of risk can be made to provide an approximate measure of
asbestos disease potential in environmental circumstances. Limitations of
existing data can be taken into account and their recognition can stimulate
appropriate research to fill identified gaps.
One of the important limits on the accuracy of exposure-response data for
asbestos diseases is our lack of information concerning past fiber exposures
of those populations whose mortality or morbidity have been evaluated. Few
measurements were made in facilities using asbestos fibers prior to 1965, and
those measurements that were done quantified all dust (both fibers and particles)
present in the workplace air. Current techniques, using membrane filters and
phase contrast microscopy for the~ enumeration of fibers longer than 5 Mm> naye
been utilized in Great Britain and the United States only since 1964 (Ayer et
42
image:
al., 1965; Holmes, 1965), They have been standardized in the United States
only since 1972 (National Institute for Occupational Safety and Health, 1972;
Leidel et al., 1979), and even later in Great Britain.
Modern counting techniques may be utilized to evaluate work practices and
ventilation conditions believed to be typical of earlier activities. However,
it is always difficult to duplicate materials and conditions of earlier decades
so that such retrospective estimates are necessarily uncertain. Alternatively,
fiber counting techniques using the particle counting instrumentation of
earlier years can be used now to evaluate a variety of asbestos-containing
aerosols. The comparative readings would then serve as a "calibration" of the
historic instrument in terms of fiber concentrations. Unfortunately, the
calibration depends on the type and size distribution of the asbestos used in
the process under evaluation and on the quantity of other dust present in the
aerosol. Thus, no universal conversion has been found between earlier dust
measurements and current fiber counts!
In the United States and Canada, those few data that were obtained on
asbestos workers' exposures prior to 1965 are based largely upon total dust
concentrations measured using a midget impinger. Fibers were inefficiently
counted with this instrument because of the use of bright field microscopy.
Attempts to compare fiber concentrations with midget impinger particle counts
generally showed poor correlations (Ayer et al.,, 1965; Gibbs and LaChance,
1974) (e.g., see Figure 3-6). In the United Kingdom, the thermal precipitator
was used from 1951 through 1964 in one plant for which environmental data have
been published. This instrument, too, does not allow accurate evaluation of
fiber concentrations. The variability in the correlation between fiber measure-
ments and thermal precipitator data is reported to be large (Steel, 1979), but
no specific data are given. Finally, both the midget impinger and the konimeter
were often used as area rather than personal samplers. Sources of dust were
often sampled for control purposes, even though no personnel were directly
exposed.
Even with the advances in fiber counting techniques, significant errors
may be introduced into attempts to formulate general fiber exposure-response
relationships. The convention now in use, that only fibers longer than 5 urn
be counted, was chosen solely for the convenience of optical microscopic
evaluation (since surveillance agencies are generally limited to such instru-
mentation). It does not necessarily correspond to any sharp demarcation of
effect for asbestosis, lung cancer, or mesothelioma. While it is readily
43
image:
140
120
J
™«.
z 100
o
(J
CC 80
j£ 60
CC
5 40
Ui
5
20
o <1 FIBER PER FIELD
• 2*1 FIBER PER FIELD
•• '
*. • •.. '
• •
*•••'•
% '••* • **
*o<Pa*
I
O 01
1
10 15 20 25
MIDGET IMPINGER COUNT, MPPCF
30
35
Figure 3*6. A plot of membrane filter and midget impinger counts;
MPPCF represents millions of particles per cubic foot.
Source: Gibbs and LaChance (1974).
44
image:
understood that counting only fibers longer than 5 urn enumerates just a fraction
of the total number of fibers present, there 1s Incomplete awareness that the
fraction counted 1s highly variable, depending upon the fiber type, the pro-
cess or products used, and even the past history of the asbestos material
(e.g., old versus new Insulation material), among other factors. For example,
the fraction of chrysotlle fibers longer than 5 urn 1n an aerosol can vary by a
factor of 10 (from as little as 0.5 percent of the total number to more than
5 percent). When amoslte aerosols are counted, the fraction longer than 5 um
may be 30 percent, extending the variability of the fraction counted to two
orders of magnitude (Nicholson et al., 1972; Nicholson, 1976a; Winer and
Cossette, 1979).
Even 1f consideration Is restricted to fibers longer than 5 um, many
fibers are missed by optical microscopy. Using electron microscopy, Rendall
and Sklkne (1980) measured the percentage of fibers with a diameter less than
0.4 um (the approximate limit of resolution of an optical microscope) 1n
various asbestos dust samples. In general, they found that more than 50 percent
of the 5 um or longer fibers are less than 0.4 urn 1n diameter and, thus, are
not visible using a standard phase contrast optical microscope. Moreover, as
with length distribution, diameter distribution varies with activity and fiber
type. As a result, the fraction of fibers longer than 5 um visible by light
microscopy varies from about 22 percent 1n chryscitlle and croddollte mining
and amoslte/chrysotlle Insulation manufacturing to 53 percent 1n amoslte
mining. Intermediate values of 40 percent are measured 1n chrysotlle brake
lining manufacturing and 33 percent 1n amoslte mill operations. Thus, even
perfect measurement of workplace air, with accurate enumeration of fibers ac-
cording to currently accepted methods, would be expected to lead to different
exposure-response relationships for any specific asbestos disease when dif-
ferent work environments are studied. Conversely, risks estimated for a given
exposure circumstance must have a large range of uncertainty to allow for the
variability resulting from fiber size effects.
Those uncertainties 1n the physical determinations of past fiber concen-
trations and the difficulty 1n evaluating the exposure parameter of Importance
1n current measurements are exacerbated by sampling limitations 1n determining
individual or even average exposures of working populations; only few workmen
at-a worksite are monitored, and then only occasionally. Variability 1n work
practices, ventilation controls, use of protective equipment, personal habits,
45
image:
and sampling circumstances add considerable uncertainty to our knowledge of
exposure.
Statistical variability associated with small numbers and methodological
difficulties 1n the estimation of disease also are Important contributions to
the variability 1n exposure-response relationships. Studies can be signifi-
cantly biased by inclusion of recently employed workers 1n study cohorts, use
of short follow-up periods, and Improper treatment of the various time factors
that are important in defining asbestos cancer. Particularly, inadequacies of
tracing, can lead to significant misestimates of disease. Generally, 10 percent
to 30 percent of an observation cohort will be deceased (sometimes even less).
If 10 percent of the group is untraced and most are deceased, very large
errors 1n the determination of mortality could result, even if no person-years
are attributed to the lost-to-follow-up group. Finally, the choice of compari-
son mortality rates can introduce substantial errors. Local rates are gener-
ally the most desirable to use, but these may be unstable because of small
numbers, or they may be affected by special circumstances (e.g., other Industry).
Data on general population worker mortality rates are not available, and
existing general population rates may overstate the expected total mortality
due to a "healthy worker effect" (Fox and Collier, 1976). Proper consideration
of smoking habits is important 1n the determination of lung cancer risks.
Unfortunately, full Information on the smoking patterns of all Individuals in
a cohort is often not available.
3.9 QUANTITATIVE DOSE-RESPONSE RELATIONSHIPS FOR LUNG CANCER
In concept, exposure-response relationships can best be determined from
studies 1n which individual exposures are estimated for each cohort member,
subgroups are established according to cumulative exposure (with proper con-
sideration of time factors), and an exposure-response relationship 1s deter-
mined from effects observed in all exposure categories. Consistencies 1n the
observed exposure-response relationships, and an appropriate intercept at zero
exposure, strengthen the risk estimates made from such studies. Dose-response
relationships are commonly obtained by two methods. One method utilizes
mortality rates in a comparison population (usually the general population of
the same area) with standard mortality ratio (SMR) calculated for each exposed
subgroup by multiplying the ratio of observed to expected deaths by 100.
Crucial to the validity of the calculation is the choice of comparison rates.
46
image:
Ideally, exposures to confounding factors, such as from cigarettes, should be
the same in the study and comparison populations. The second method generates
a relative risk (RR) factor at each exposure by a case-control anaylsis, where
the number of cause-specific deaths is compared with the number of internal
controls in each dose category. Such analysis is less subject to confounding
factors in the comparison population, but has greater statistical variability.
In calculating a dose-response relationship, a weighted, rather than
unweighted, least square analysis is most appropriate because there are large
differences in the statistical validity of the individual SMRs or RRs in a
given study. Values of K., the fractional increase in risk per unit exposure,
can be calculated directly from the slopes of the regression lines of SMR or
RR on dose (with a conversion, if necessary, from mppcf-y to f-y/ml).
Ideally, regression lines should pass through zero dose at an SMR of 100
or an RR of 1. The chances of this occurring are minimal. Statistical vari-
ability, even in the most ideal circumstances, will lead to intercepts differ-
ent from that expected; in the case of SMRs, the comparison population may not
be completely appropriate; incomplete tracing of a cohort can distort both
SMRs and RRs; the comparison group in a relative r.isk analysis usually has
some exposure; and finally, dose-response relationships can be affected by
improper estimates of dose. It is important to identify the factor which may
have led to an abnormal intercept, because it would indicate what adjustments
might be made to the observed slope. For example, if improper comparison
rates were used for the calculation of SMRs, and they were the sole cause of a
higher or lower than expected intercept, it would be appropriate to divide
both the slope and the intercept by the intercept/100 because the same percen-
tage misestimate would be expected to exist in each exposure category. However,
if the deviation from 100 were simply random, such division would compound
what is already a statistical misestimate of the true slope. For example, if
statistical variability led to an SMR intercept higher than 100, the observed
slope would be less than the true slope. To divide by the intercept/100 would
reduce it even further.
It may be difficult to identify misestimates of dose, especially within a
single study. However, comparisons between estimates in similar exposure
circumstances by different groups are useful in establishing the reasonable-
ness of stated exposure estimates. In analyses of the available data on lung
cancer risk for several studies, the uncertainties associated with response are
47
image:
greater than those associated with dose. This 1s particularly true in groups
demonstrating low risks, where the difference between observed and expected
deaths has an extremely large uncertainty relative to the difference.
Dose-response data can also be obtained using the overall SMR for a group
and the average exposure for all cohort members. This calculation assumes
that a linear dose-response relationship exists throughout the range of exposure
and that the comparison population rates are appropriate to the study popula-
tion. The first assumption would appear to be generally valid for lung cancer,
but the second must be considered carefully 1n the analysis of each study.
Such calculations will generally use Equation 3-3a, which 1s simplified as
IL = IE(1 + KL-f-d) (3-3b)
Rearranging, one obtains
KL = [(IL - IE)/IEJ/f-d (3-3c)
or
(3-3d)
= (Relative Risk -l)/Cumulat1ve Exposure
Two approaches are possible 1n developing an exposure-response relation-
ship for asbestos. One is to select the study or studies with the best exposure
data, assuming an adequate measure of effect. The exposure-response relation-
ship developed certainly would apply to similar exposure circumstances and may
apply to others as well. Alternatively, all studies for which exposure-response
information 1s available can be utilized along with estimates of the uncertainty
of such data. An appropriate weighted average of the relationships found in
different studies, taking Into account observable differences In exposure
circumstances, yields an overall exposure-response relationship. The former
procedure has particular merit in evaluating the risk from an agent whose
exposure can be well characterized, such as that from a single chemical species.
However, this 1s not the case with asbestos where we are generally concerned
with exposures to mixtures of different asbestos minerals. Even exposures to
a single mineral species can Involve substantially different fiber-size distri-
butions which would strongly affect the carcinogenic potentials of the expo-
sures. As mentioned above, a large fraction (usually greater than 50 percent)
48
image:
of the fibers longer than 5 pm are too thin to be visible by light microscopy.
These thin and long fibers are the most carcinogenic in experimental studies
(see Chapter 4) and are believed to be so in humans. The fraction of these
uncounted fibers will vary with the particular process and a study or studies
selected on the basis of the "best exposure measurements" may not be typical
of most exposure circumstances in terms of its fiber-size distribution, even
for one asbestos mineral. Thus, the quality of "good" exposure data for
carcinogenic risk assessment may be illusionary.
The advantages of considering all studies for which exposure-response
data can be developed are
1. any bias in the choice of studies selected for analysis 1s largely
removed,
2. information can be obtained on the uncertainty of the estimate of an
average value of K,,
3. estimates of the effect of fiber type differences or process differ-
ences can be estimated better. Such information is of crucial
importance and efforts to obtain it are warranted.
Primary among the disadvantages of the use of all exposure-response data
is the fact that the quality of some of the data can only be estimated subjec-
tively. The statistical variability in measures of response can be established
quantitatively. However, biases 1n epidemiological studies may not be perceived
and, of most importance, evaluations of the quality of exposure estimates are
highly subjective, as are the estimates themselves.
Because of the above advantages, in the analysis that follows, all studies
that provide exposure-response information are utilized. This procedure was
also followed in the asbestos health effects reviews of the Consumer Products
Safety Commission (1983) and the National Academy of Sciences (1983). In
contrast, the recently published review by Doll and Peto (1985) for the British
Health and Safety Commission selected two studies for analysis, based upon the
quality of exposure measurements. These were the study by McDonald et al.
(1983) of South Carolina textile workers and Peto et al.'s (1985) update of
the mortality of Rochdale textile workers. As will be .seen, their results are
virtually identical to those obtained using all available studies.
In this document estimates of K, are made from all sources of data within
each study. If the data indicate that the results of a study are substantially
49
image:
affected by possible misestimates of exposure, that non-local rates are used
for the expected mortality, or that Inadequate tracing exists, an adjustment
and Its magnitude are clearly indicated. Consideration 1s made for deviations
of the Intercept of SMR regression lines from 100. However, 1f the source of
the deviation cannot be Identified, the slope as calculated 1s used.
For nine studies, values of K, are estimated from a weighted linear
regression analysis of the relationship between lung cancer risk and cumulative
exposure. The weighting is the reciprocal of the variance of a particular
data point. Perceived biases are taken into account and adjustments for them
described 1n the text. Generally, the adjustment accounts for the difference
in local lung cancer rates compared to those used in the published study. A
value for K. is calculated for each study using the slope of observed dose-
response data, the slope of the odds ratios at different doses 1n case control
analyses, or an average of the two procedures when both are done. In three
studies, K, is estimated from the difference in risk between heavily and
lightly exposed groups (using individual exposure estimates) and/or the risk
estimated from the ratio of overall excess lung cancer to the average exposure
for the group. Finally, in one study, the relationship between SMR and dura-
tion of employment 1s used, assuming average group exposure per year of
employment.
Table 3-10 shows the results of a variety of analytical procedures using
the published data in 14 studies, along with 95 percent confidence limits
calculated from the variance of the observed number of lung cancer cases and
the slope of weighted regression lines. Adjustments for potential biases are
shown as well as alternate regression analysis which either forces the regres-
sion line through an SMR of 100 at 0 dose or adjusts for a non-zero intercept
by dividing by the intercept/100. It 1s emphasized that these two procedures
can lead to misestimates of the actual exposure and increased uncertainty
estimates. They are included, however, to provide a measure of the uncertainty
that may be associated with regression analysis. Further, an analysis is
shown 1n which the overall SMR and average exposure of the group was utilized
to estimate the value of K,. This analysis 1s' particularly useful in estimat-
ing the range of uncertainty that may be present 1n given studies. For example,
consider the study of Peto (1980). In the cohort exposed after 1950, 11 lung
cancers were observed and 3.35 expected in the group followed 15 years after
first employment and deemed to have a cumulative exposure of 200 f-y/ml. The
50
image:
excess risk 1s 7.65 cases, using Equation 3-3c, and K, - (11 - 3.35)/3.35/200
li L
= 0.0114 (f-y/ml) . Assuming the number of deaths 1s an expression of a
Poisson varlate, the 95 percent confidence limit (from statistical considera-
tions) will be from KL = [0.0114 (5.4 - 3.35)]/7.75 to KL = [0,0114 (19.7 -
3.35)]/7.75; I.e., from 0.0030 to 0.024.
The method for estimating K, and the 95 percent confidence limit for each
study is described in the text that follows. These data are listed in Table 3-10
and displayed in Figure 3-7. In addition to the statistical uncertainty
listed in Table 3-10, the effect of a ± two-fold range of uncertainty in
cumulative exposure is indicated in Figure 3-7 for most studies. This twofold
range 1s a subjective choice, but is felt to be a realistic representation of
the uncertainty in the cumulative exposure estimates from all the sampling
problems mentioned previously. In some cases, for specific reasons listed, a
greater exposure uncertainty is indicated. Even though response uncertainties
and exposure uncertainties are unlikely to be correlated, the overall 95 percent
confidence limit on a study is considered to be the sum of the listed exposure
and response uncertainties.
3.9.1 Textile Products Manufacturing, United States (Chrysotile); Dement at al.
(1982. 1983at 1983b)
Mortality data from a chrysotile textile plant studied by Dement et al.
(1982, 1983a, 1983b) allow a direct estimate of lung cancer risk per fiber
exposure. Here, data from impinger measurements of total dust in terms of
mppcf were available, characterizing dust concentrations since 1930. Further,
1106 paired and concurrent impinger-membrane filter measurements allow conver-
sion of earlier dust measurements to fiber concentrations, suggesting that 3
f/ml is equivalent to 1 mppcf for all operations except fiber preparation.
(The 95 percent confidence interval is 2-3.5 f/ml/mppcf.) A value of 8 f/ml/
mppcf characterizes fiber preparation work (confidence Interval, 5-9). Subse-
quent to 1940, average fiber concentrations in most operations are estimated
to range from 5 to 10 f/ml, with the exception of fiber preparation and waste
recovery where mean concentrations are 10-80 f/ml.
The study cohort consisted of all 1261 white males employed one or more
months between January 1, 1940 and December 31, 1965. Vital status was deter-
mined for all but 26 individuals who were considered alive for purposes of
analysis. SMRs for lung cancer were presented for five exposure categories in
terms of cumulative fiber exposure (Table 3-11). A weighted regression line
51
image:
TABLE 3-10. ESTIMATES OF THE PERCENTAGE INCREASE IN LUNG CANCER PER f-y/nl OF EXPOSURE (100 x K,),
ACCORDING TO DIFFERENT PROCEDURES IN 14 EPIDEMIOLOGICAL STUDIES
Study
Oemenl el al., 19S3b
McDonald el al., 19B3o
Peto. 1980
McDonald et al., ]983b
Berry & Henhouse. 1583
McDonald et al., 1984
McDonald et al., 1980
Nicholson et al., 1979
lublno ct al., 1979
SeidBan. 1984
Sell toff et al., 1979
Henderson « Enterltne,
1979
velll et al.. 1979
FlnkelsUIn, 1383
Tears
fro*
onset
15
20
15
20
10
20
20
20
20
5
ZO
ret.'
20
20
Directly froa
weighted SKR
regression
4. 19(11. 65)'
2.07(10.50)
0.86(10.29)
Negative
0.043(10.015)
0.23(--)*
0.51(-)e
2.72(11.06)
1.10(10.097)
0.34(10.17)
0.31(10.31)
Negative
Adjusted for
local rates or
other factors
(see text)
2. 79(tl.lO)
1.38(10.33)
1.06(±fl.3S)
0.064(10.022)
0.30(-)
0.75(10 066)
0.49(10.25)
0.53(10.54)
Adjusted
to SMR = 100
at zero dose
2.77(11.08)
1.88(10.45)
1.62(10.55,
0.047(10.016)
0.84(10.33)
0.24(10.12)
0.42(10.44)
SHR
regression
forced
through 100
at rero dose
4.48(11.10)
2.21(10.39)
0.41(10,71)
0.13(11. S3)
0.035(10.014)
0.30(-)
4.28(12.27)
0.43(10.13)
0.22(10.31)
RR
regression
adjusted to
RR = 1
at zero dose
3.72U2.04)
1.71(10.93)
Negative
0.003(10.95)
0.057(10.009)
0.89 (")
0.35(10.26)
4.80(15.29)
Overall SHR-100
divided by
average exposure
5.37 (Z. 94-8,45)
3.22 (1.46-4.95)
1.14 (0.30-2.41)
0.10 (0.0-0.66)
0.87 (0.29-1, 79)c
0.068 (D.O-D.52)
0.79 (0.017-1 74)
O.DB5 (0.0-0.55)
0.045 (0.016-0 074)
0.011 (0.043-0.21)
0.013 (O.D-O.U)
5.92 (4.49-7.36)
0.86 (0. 75-0. 97)
0.046 (0.27-0.63)
0.041 (0.0-0.36)
O.D34 (0.13-0.63)fl
6.70 (3.53-11.25)
Adjusted for
local rates or
other factors
(see text)
3.58 (1.99-5.63)
2.15 (0.97-3.30)
0.12 (0.0-0.81)
1.07 (0.36-2.21)
0.064 (0.023-0.11
0.17 (0.064-0.32)
0.69 (0.60-0.78)
0.67 (0.39-0.91)
0.64 (0.0-1.1)
0.38 (0.14-0.70)
Adopted values
and range
2.8 (1.7-5.6)
2.5 (1.0-3.7)
1.1 (0.30-Z.4)b
1.4 (0.36-1.7)
O.OSfl (0.010-D.BO)
0.010 (0.010-0.55)
0.060 (0.023-0.11)
0.17 (0.064-0.32)
0.075 (0.010-0.89)
4.3 (0.84-7.4)
0.75 (0.60-1.1)
0.49 (0.24-0.91)
0.53 (0.14-1.1)
6.7 (3.5-11.2)
"() = 95S confidence Halts
bDoll and Petu (1985) refer to in update of this study (Peto et al. 1985). They calculate values of
1.5 and O.S4 for 100 x RL for workers first exposed after 1950 and after 1932, respectively.
'Calculated frn highest exposure category.
'Calculated Hitting lowest exposure category.
eOnly lira values.
'Retirees.
"calculated froa highest two exposure categories.
image:
01
LO
0.5000
0.2000
0.1000
0.0500
0.0200
0.0100
•
0.0050
0.0020
0.0010
0.0005
0.0002
-
-
-
_
-
_
«
"j
f
g
r
1
t
i-
z
u
5
|
!
!
i
a
•
i
«
- Q
_
—
0.0001 -
mm
^
't\
to
at
—^
I
4
Z
i
2
I
i
d
H
•oT
;;
S
w~
-,
D
_l
Z
O
a
u
TEXTILE PRODUCTION
CO
^
M
3
i S
s ^
1 5
1 E
:* o
> -i
a i
o o
MM
2
mm
••
0
T"
~J
E
Q
_l
i
u
2
M
s
-i
z
a
O
0
z
••
••
5i
F*.
™
_.•
D
O
z
ffl
=
•I
I
CD
i
UJ
CO
O
r
i
t
1
>
«
1
D
Z g
ERSON
ILINE.I
Z ™
liJ E
MM
^
MM
0)
r-
OJ
"5
ts
_i
£
' -
-
-
-
_
_
n
S ~
z'
at
iv — -
CO
ui
Z —
IL
£
u. —
u
O
ui
w —
FRICTION PRODUCTS MINING AND MILLING INSULATION
PRODUCTS
PREDOMINANTLY CHRYS
rf^TII C
Vl 1 IL& -
_ - « »
1^' ^ Mil
IOSITE •»
INSU- MIXED
LATION PROD-
CEMENT
PRODUCTS
WORKERS UCTS
-m MIXED FIBERS
Figure 3-7. Values of K|_, the fractional increase in lung cancer per f-y/ml of exposure in 14 asbestos exposed cohorts. The open bar
reflects the estimated 95% confidence limits associated with measures of response. The line represents the uncertainties associated
with measures of exposure, generally ± a factor of two.
image:
TABLE 3-11. LUNG CANCER RISKS, BY DOSE, AMONG SOUTH CAROLINA
ASBESTOS TEXTILE WORKERS
(Dement et al., 1983b)
Exposure in f-y/ml SMR
1.4 (<2.74) 140 (5)a
15.1 (2.74-27.4) 279 (9)
68.5 (27.4-109.6) 352 (7)
191.8 (109.6-274.0) 1099 (10)
411.0 (>274.0) 1818 (2)
Complete cohort: 336 (33)
Estimated average cumulative exposure: 43.9 f-y/ml
a( ) = number of deaths.
Regression equations
SMR = 150 + 4.19(±0.84) x f-y/ml weighted
SMR = 169 + 4.13(±0.32) x f-y/ml unweighted
Weighted regression equation forced through an SMR of 100
SMR = 100 + 4.48(±0.56) x f-y/ml
yields SMR = 150 + 4.19 x f-y/ml, for a KL of 0.042. The standard error of
the estimate of the slope Is ± 0.84.
Dement et al. (1983b) uses U.S. rates for calculating expected deaths.
Age-adjusted county rates are 75 percent higher, 1.e 66.5/10 versus 38.0/10
(Mason and McKay, 1974). Dement et al. presents arguments for using national
rates. Local rates are probably Influenced by nearby shipyard employment (and
perhaps by the study plant) and the smoking habits of the study population
reflect those of the U.S. general population. Blot et al. (1979) found that
World War II shipyard employment leads to a 60 percent increased risk of lung
cancer. This increase, however, would be substantially diluted in county
rates. Across the United States these rates are 11 percent higher in shipyard
counties compared with control counties. Further, Acheson and Gardner (1983)
point out that the rates for women 1n the county are equally high and they
suggested an exposure to some unknown carcinogen in the population. The
age-adjusted rates of contiguous counties are only 16 percent greater than
54
image:
those of the United States; those of the State of South Carolina are virtually
identical to the United States rates.
It is unlikely that the origin of the high local rates will ever be
resolved. As seen above, the SMR at zero exposure is calculated to be 150
from the weighted regression analysis, We will use this value as a measure of
possible overestimates of the SMRs at all exposures, and we will divide the
value of K, above by 1.5. This brings the SMR at zero exposure to 100 and
allows virtually full consideration that higher local rates are the appropriate
comparison. (The remainder would be accounted for by shipyard employment.)
The adjusted KL is 0.028.
3.9.2 Textile Products Manufacturing, United States (Chrysotile); McDonald
et al. (198lal
Exposure-related mortality data at this same plant have recently been
published by McDonald et al. (1983a). Their cohort consisted of all individuals
employed for one or more months prior to January 1, 1959 and for whom a Social
Security Administration (SSA) record existed. This eliminated from considera-
tion individuals who began and ended their employment prior to mid-1937, when
SSA numbers were first assigned. The same data used by Dement on past exposures
were utilized to assign cumulative dust exposures, in mppcf-y, to each study
participant. Male deaths, by cause, 20 years after first employment, are
related to dust exposure accumulated to 10 years prior to death. Data for
lung cancer are shown in Table 3-12. A weighted regression analysis yields
the relation SMR = 110 + 6.22 mppcf-y. No data are given by McDonald et al.
(1983a) on cumulative fiber exposures. If we use the average relationship
found by Dement et al., 1 mppcf = 3 f/ml, we obtain a K, of 0.021. Adjusting
by the value 1.5, as above, to account for the higher local rates, yields a K,
of 0.014. (McDonald et al. (1983a) used South Carolina rates rather than local
rates).
McDonald et al. (1983a) also made estimates of risk using a Mantel and
Haenszel (1959) case-control analysis, as in Table 3-12. A weighted regression
line yields a slope of 0.068. Because the RR regression was obtained using
internal controls, no adjustment for local rates is necessary. However, since
the controls were exposed, the zero dose intercept should be used as the
measure of risk in an unexposed group. This requires dividing the slope by
the intercept to obtain an adjusted regression line. Dividing by the zero
exposure intercept, 0.61, and by 3 to convert to fiber exposures, gives a
55
image:
TABLE 3-12. LUNG CANCER RISKS, BY DOSE, AMONG SOUTH CAROLINA ASBESTOS
TEXTILE WORKERS (McDonald et al., 1983a)
Exposure
in mppcf-y
5 (<10)
15 (10-19)
30 (20-39)
60 (40-79)
120 (>80)
SMR
143.1 (31)c
182.7 (5)
304.2 (8)
419.5 (7)
1031.9 (8)
RRb
1.00 (25)
0.98 (3)
2.95 (8)
4.32 (7)
15.00 (6)
Complete cohort: 199.5 (53)
Estimated average cumulative exposure: 10.3 mppcf-y.
Exposure accumulated to 10 years before death.
Relative risk from an internal case-control analysis.
c( ) = number of deaths.
Regression equations
SMR = 110 + 6.22(±0.76) x mppcf-y weighted
SMR = 63 + 7.68(10.76) x mppcf-y unweighted
RR = 0.61 + 0.068(±0.019) x mppcf-y weighted
RR = -0.80 + 0.123(10.017) x mppcf-y unweighted
Weighted regression equation forced through an SMR of 100:
SMR = 100 + 6.63 (10.61) x mppcf-y
value of KL = 0.037. We will use 0.025, the average of 0.014 and 0.037, to
represent this study. The agreement with the results of Dement et al. (1982,
1983a,b) is very good.
3.9.3 Textile ProductsManufacturing, Rochdale, England (Chrysotije);
Peto (19801
Table 3-13 shows the lung cancer and mesothelioma mortality experience
from an often-studied British textile plant (Doll, 1955; British Occupational
Hygiene Society, 1968; Berry et al., 1979; Knox et al., 1968; Peto, 1980;
British Occupational Hygiene Society, 1983). The data are difficult to inter-
pret because dust concentrations have changed fairly dramatically over the
past five decades of plant operations, and so have subsequent estimates of
56
image:
TABLE 3-13. MORTALITY EXPERIENCE OF 679 MALE ASBESTOS TEXTILE WORKERS
(Peto, 1980)
Year
first
exposed
1933-1950
N = 424
1951 or later
N = 255
Period since
first exposure
(yrs)
10-14
15-19
20-24
25-29
30-34
35-39
Total
10-14
15-19
20-24
25-29
Total
Man-years
1633
1860
1760
1496
837
507
8093
1123
1022
556
96
2797
Lung
0
2
4
3
10
8
1
28
1
3
7
1
12
cancer
E
1.80
2.98
3.97
4.54 •
3.14
2.20
18.63
1.30
1.74
1.31
0.31
4.65
Mesothelioma
0
0
0
1
2
2
2
7
0
0
0
0
0
rate per
103 p-y
0.0
0.0
0.6
1.3
2.4
3.9
-
0.0
0.0
0.0
0.0
-
those concentrations. No measurements of dust concentrations were made prior
to 1951. Between 1951 and 1964, thermal predpltators were used to evaluate
total dust levels; thereafter, filter techniques similar, but not Identical,
to those 1n the United States were used. Average fiber concentrations are
published for earlier years based on a comparison of fiber counting with ther-
mal predpltator techniques (Berry, 1973). Later these estimates were stated
to be Inaccurate; Berry et al. (1979) reported that a re-evaluation of the
work histories indicated that some men had spent more time in less dusty jobs
than previously believed and that previous average cumulative doses to 1966
had been overestimated by 50 percent.
Recently, as part of the British Government's review of Its asbestos
standard, the hygiene officers of the plant re-evaluated previously reported
exposure data. It 1s now suggested that earlier static sampling methods
underestimated personal exposures by a factor of about 2, and that whole
field, rather than graticule field, microscopic counting understated fiber
concentrations by another factor of 2 to 2.5 (Steel, 1979). In 1983, the
57
image:
British Occupational Hygiene Society (1983) reported information on the
differences between personal and static sampling. Data were presented for
thirty-one simultaneous samples comparing the two techniques, the personal
samplers indicating a greater fiber concentration in 22 cases. Using these
data, the BOHS committee evaluated the cumulative fiber exposure (as of
approximately 1976) for 284 individuals employed for 10 or more years sub-
sequent to 1951. The overall average of the entire group was 182 f-y/ml.
This is slightly less than the estimate of Peto (1980), who suggested that the
exposure of 10+ years employees was 200-300 f-y/ml. However, Peto's estimate
was based on preliminary data on only 126 men first employed between 1951 and
1955 (see Table 3-14).
These most recent estimates are clouded by questions concerning the
appropriateness of multiplying static sampler concentrations by a factor
approaching two. The BOHS data are directly contradicted by published data
(See Table 3-15) from the factory on other comparisons of static and personal
sampling results by job (Smither and Lewinsohn, 1973). Dr. Lewinsohn (1983)
confirmed these results. He stated that the static sampler concentrations
were generally higher than those of the personal samplers of men working at
the monitored job. The company placed the static samplers to best reflect the
breathing zone dust concentrations of machine operators while tending machines.
Dr. Lewinsohn (1983) stated that if a machine were running smoothly, a worker
would move away to the aisle adjacent to the machine from where he or she could
continue to observe the operation and experience a lower dust concentration.
The difference between static and personal sampling data appears to be greater
in the dustier jobs. In the Rochdale factory, the average of the ratios of
static to personal sample concentrations at the same work station is 1.8 (1.5
if the fiberizing operation is not considered). The recent comparison may not
reflect the movement of a worker from his machine.
We will use a value of 200 f-y/ml to represent cumulative exposure of the
post-1951 group fifteen or more years from onset of exposure, which probably
overestimates the effective exposure of the group. While 200 f-y/ml, the
average dose of all men employed 10 or more-years, may underestimate the
average total dose of men employed 15 or more years, it certainly overestimates
the effective dose that accumulates to about 10 years prior to end of follow-up
or death. As was shown above, this yields a K. of 0.011, To reflect what could
be a twofold lesser exposure, the upper exposure-related uncertainty in risk
was increased from 2 to 4 in Figure 3-7.
58
image:
TABLE 3-14. PREVIOUS AND REVISED ESTIMATES OF MEAN DUST LEVELS IN f/ml
(WEIGHTED BY THE NUMBER OF WORKERS AT EACH LEVEL IN SELECTED YEARS)
1936 1941 1946 1951 1956 1961 1966 1977 1974
Previous estimates
corresponding to
early fiber counts
13.3 14.5 13.2 10.8 5.3 5.2 5.4 3.4
Revised estimates
corresponding to
modern counting »
of static samples
No measurements 32.4 23.9 12.2 12.7 4.7 1.1
prior to 1951
aThese estimates are based on preliminary data on 126 workers first employed
between 1951 and 1955, and should be regarded as provisional.
Source: Peto (1980).
TABLE 3-15. DUST LEVELS: ROCHDALE ASBESTOS TEXTILE FACTORY, 1971
Department
F1ber1z1ng
Carding
Spinning
Weaving
Process
Bag slitting
Mechanical bagging
Fine cards
Medium cards
Coarse cards
Electrical sliver cards
Fine spinning
Roving frames
Intermediate frames
Beaming
Pirn weaving
Cloth weaving
Listing weaving
Static
3
4
3.5
4.5
8
1.5
2.5
6
5.5
0.5
1.5
2
0.5
Personal
1
1
2
3.5
6
1
3
3
3
0.5
1
1
0.5
Plaiting
Medium plaiting
Source: Smlther and Lewinsohn (1973).
59
image:
A second difficulty of the British textile factory study is that the
dose-response data calculated from groups exposed before and after 1950 differ
considerably. While no cumulative exposure data are published for the pre-1951
group, it is surprising that more disease is seen in the latter group, as the
average Intensity of exposure was certainly greater for the earlier group,
perhaps by a factor of three. It is difficult to reconcile the differences
between the two subcohorts employed in this facility. The data are severely
limited by the relatively small size of the cohort and the few deaths available
for analysis. Nevertheless, what would appear to be a nearly tenfold differ-
ence in the estimated risk of death from lung cancer suggests the possible
existence of some unidentified bias in the pre-1951 group. The post-1950
group's mortality experience is more in accord with U.S. textiTe plants. The
finding of only a 50 percent increase in lung cancer in exposure circumstances
leading to 5.3 percent of deaths being from asbestos is is certainly unusual,
as is the finding that there are as many mesotheliomas as excess lung cancers.
Doll and Peto (1985) recently reviewed the new information on the health
effects of asbestos for the British Health and Safety Commission. Many of the
above uncertainties, particularly that of the ratio of personal to static
sampling counts, are discussed. A regression analysis of the ratio of personal
to static counts against mean concentration indicated that the ratio is greater
than one for concentrations less than 2 f/ml, but less than one for higher
concentrations. Doll and Peto (1985) estimate values of K, from the mortality
in an expanded and updated study of the Rochdale cohort. Their results indicate
K, is 0.015 for workers first employed after 1950 and 0.0054 for all workers
first employed after 1932.
3.9.4 Textile and Friction Products Manufacturing, United States (Chrysotile,
Amosite, and Crocidolite); McDonald et a!. (1983b>; Robinson et al. (1979)
A plant located near Lancaster, Pennsylvania, which produced mainly
textiles but also friction products and packings, was studied by Robinson et
al. (1979), McDonald et al. (1983b), and earlier by Mancuso and Coulter (1963)
and Mancuso and El-attar (1967). The plant, which began operations in the
early 1900s, used between 3000 and 6000 tons of chrysotile per year over most
of the period of its operation. Amosite constituted less than 1 percent of
the fiber used, except for a three-year period, 1942 - 1944, when 375-600 tons
of amosite were used in insulation blankets and mattresses. Crocidolite usage
was approximately 3-5 tons per year (Robinson et al. , 1979). The reports of
60
image:
Robinson et al. (1979), Mancuso and Coulter (1963), and Mancuso and El-attar
(1967) provide no information on the exposure of the cohort members to asbestos;
so they cannot be used in establishing exposure-response relationships. In
the study of McDonald et al. (1983b), dust concentrations, measured in mppcf,
available from the 1930s through 1970 were used. However, no attempt was made
to relate particle exposures to fiber exposures. The study cohort of McDonald
et al. (1983b) comprised all individuals employed for one or more months prior
to January 1, 1959 with their Social Security file identifiable in the Social
Security Administration offices. These individuals were traced through December
31, 1977, and cause-specific mortality ratios, based on state rates, were
related to cumulative dust exposure.
The results for lung cancer are shown in Table 3-16. The regression of
SMR on dose has an unusually low intercept of 53. The overall SMR for lung
cancer is also low. The low local rates (30.1 versus 37.7 for the state)
(Mason and McKay, 1974) do not fully account for these deficits. Smoking
histories are reported for only 36 individuals and indicate no unusual pattern.
Because the full deficit cannot be explained, we have adjusted the slope by
the ratio of the local to state lung cancer rates (0.81) rather than by 0.53,
resulting in a slope of 0.032. The adjusted slope of the RR regression is
0.051. If these two values are averaged and a factor of 3 is used to convert
from mppcf to f/ml, the exposure-response relationships give average K, =
0.014. The factor of 3 was previously measured in textile manufacturing, the
predominant activity in this plant. Calculating K. using the overall SMR of
the study suggests that the lower confidence limit of K. is 0, but the SMR and
RR regression lines strongly contradict this. Thus, for the lower confidence
limit we will use a value calculated from the highest exposure relationship,
*
where the uncertainty in comparison rates has less of an effect.
3.9.5 Friction Products Manufacturing, Great Britain (Chrysgti^e
Crqcidolite); Berry and Newhouse (1983)
Berry and Newhouse analyzed the mortality of a large workforce manufac-
turing friction products. All individuals employed in 1941 or later were
included in the study, and the mortality experience through 1979 was determined.
Exposure estimates were made by reconstructing the work and ventilation con-
ditions of earlier years. Fiber measurements from these reconstructed condi-
tions suggested that exposures prior to 1931 exceeded 20 f/ml but those after-
wards seldom exceeded 5 f/ml. From 1970, exposures were less than 1 f/ml.
61
image:
TABLE 3-16. LUNG CANCER RISKS, BY DOSE, AMONG PENNSYLVANIA ASBESTOS
TEXTILE AND FRICTION PRODUCTS WORKERS
(McDonald et al., 1983b)
Exposure
in mppcf-y
5 (<1Q)
15 (10-19)
30 (20-39)
60 (40-79)
120 (>80)
SMR
66.9 (21)C
83.6 (5)
156,0 (10)
160.0 (6)
416.1 (11)
RRb
1.00 (20)
0.83 (4)
1.54 (10)
2.90 (6)
6.82 (11)
Complete cohort: 105.0 (53)
Estimated average cumulative exposure: 16.9 mppcf-y.
aExposure accumulated to 10 years before death.
Relative risk from an internal case-control analysis.
c( ) = number of deaths.
Regression equations
SMR = 53 + 2.58(10.45) x mppcf-y weighted
SMR = 41 + 2.94(±0.42) x mppcf-y unweighted
RR = 0.70 + 0,036(±0.010) x mppcf-y weighted
RR = 0.24 + 0.050(±0.005) x mppcf-y unweighted
Weighted regression equation forced through an SMR of 100:
SMR = 100 + 1.22 (±1.07) x mppcf-y
These relatively low intensities of exposure kept the average cumulative
exposure for the group to less than 40 f-y/ml.
The overall mortality of all study participants, 10 years and more after
onset of exposure, was no greater than expected for all causes. Data for lung
cancer are shown in Table 3-17. Cancer of the lung and pleura was slightly
elevated in men (151 observed versus 139.5), but the excess was largely ac-
counted for by eight mesothelioma deaths. No unusual mortality was found in
those employed 10 or more years. Using a case-control analysis according to
cumulative exposure, Berry and Newhouse (1983) estimated that the lung cancer
increased risk- was 0.06 percent per f-y/ml (K^ = 0.00058), with an upper 90
percent confidence limit of 0.8 percent per f-y/ml. Table 3-17 lists the
results of the case control analysis. The weighted regression of RR on dose
62
image:
has a negative slope. The ratio of excess lung cancer to average group expo-
sure yields a value of KL = 0.00068 = [(143/139.5)-l]/37.1. We will use the
value published by Berry and Newhouse, 0.00058, and their confidence limits
for KL.
TABLE 3-17. LUNG CANCER RISKS, BY DOSE, AMONG BRITISH ASBESTOS
FRICTION PRODUCTS WORKERS
(Berry and Newhouse, 1983)
Exposure in mppcf-y RRa
5 (0-90 1.00 (50)b
30 (10-49) 0.79 (37)
75 (50-99) 0.86 (13)
200 (100-356) 0.88 (5)
Estimated average cumulative exposure: 31.7 f-y/ml.
af
b,
aRelat1ve risk from an internal case-control analysis.
( ) = number of deaths.
Regression equations
RR = 0.91 - 0.00076(10.0016) x f-y/ml weighted
RR = 0.90 - 0.00019(10.00070) x f-y/ml unweighted
3.9.6 Frlct1on ProductsHanufacturingtUnited States(Chrysotl1e);
McDonald et al. (1984f
McDonald et al. (1984) analyzed the mortality of the workforce employed
in friction products production in the United States and attempted to relate
it to cumulative dust exposure. However, a highly unusual mortality experience
is observed. The overall mortality shows an elevated risk of death in the
complete cohort for virtually all causes, largely confined to individuals
employed for less than one year. The correlation of respiratory cancer SMR
with cumulative dust exposure of those employed for more than one year shows
little, 1f any, trend with increasing dust exposure, even though the overall
SMR for lung cancer (see Table 3-18) is 137 for these individuals. The slopes
of the regression equations of SMR on dose are slightly negative and those of
relative risk are slightly positive. As with the McDonald et al. (1983b)
Pennsylvania textile study, we will use the dose-response regression relation-
ship for the measure of risk and set KL = 0.0001 for this group. In Figure
3-7, this represents "zero" for the purpose of calculating geometric means.
63
image:
TABLE 3-18. LUNG CANCER RISKS, BY DOSE, AMONG ASBESTOS
FRICTION PRODUCTS PRODUCTION WORKERS
(McDonald et al., 1984)
Exposure
in mppcf-y
5 (<10)
15 (10-19)
30 (20-39)
60 (40-79)
120 (>80)
SMR
167.4 (55)b
101.7 (6)
105.4 (5)
162.8 (6)
55.2 (1)
RRa
1.00 (54)
0.40 (4)
0.91 (5)
1.40 (16)
1.13 (1)
Complete cohort: 148,7 (73)
1+ yrs employment: 136.8 (49)
Estimated average cumulative exposure: 10.3 mppcf-y.
Estimated average exposure for
those employed more than 1 year: 15.5 mppcf-y.
aRelative risk from an internal case-control analysis.
( ) - number of deaths.
Regression equations
SMR = 160 - 0.85(±0.52) x mppcf-y weighted
SMR = 147 - 0.62(10.46) x mppcf-y unweighted
RR = 0.69 + 0.00006(10.01) x mppcf-y weighted
RR = 0.78 + 0.0041(±0.0039) x mppcf-y unweighted
Weighted regression equation forced through an SMR of 100:
SMR = 100 + 0.13 (+0.83) x mppcf-y
The low value, however, is qualified by the overall high lung cancer mortality.
As the origin of this elevated lung cancer mortality is workers employed for
more than one year (where total mortality is close to that expected) is unknown,
the upper limit of uncertainty will be given by the upper confidence limit on
the ratio of lung cancer excess risk to average exposure in the 10-19 mppcf-y
exposure groups. This procedure is similar to that used to estimate the lower
confidence limit in the Pennsylvania textile cohort.
64
image:
3.9.7 Mining and Milling. Quebec. Canada (Chrysotile); Llddell et al.
(1977); McDonald et. al. (1980)
The results reported by Llddell et al, (1977) and McDonald et al. (1980)
on mortality (Table 3-19) according to total dust exposure 1n Canadian mines
and mills can be converted to relationships expressed In terms of fiber expo-
sures. SMR values are provided by McDonald et al. (1980) for various exposure
categories 1n four different duration-of-employroent categories. A weighted
regression analysis of these data yields a relationship, SMR = 92 + 0.13 x
mppcf-y. Using a value of 3 f/ml/mppcf for the particle fiber conversion
factor yields a K. of 0.00043. The factor of 3 f/ml/mppcf 1s the midpoint of
the range of 1-5 f/ml/mppcf suggested by McDonald et al. as being applicable
to most jobs 1n mining and milling. However, since McDonald et al. used the
rates of the Province of Quebec for their comparison data, K, 1s likely to be
underestimated. In an earlier paper, McDonald et al. (1971) suggested that
the lung cancer rates 1n the counties adjacent to the asbestos mining counties
were about two-thirds those of the Province. This Is substantiated by lung
cancer Incidence rates, 1n the Province of Quebec, published by Graham et al.
(1977). These data for the years 1969-1973 are shown In Table 3-20 and confirm
the earlier statement of McDonald et al. (1971). Thus, the above K. will be
multiplied by a factor of 1.5. Llddell et al. (1977) performed a case control
analysis of the relative risk of lung cancer In this same period. Their
regression equation suggests a K. of 0.00057. We will use the average of
these two estimates, 0.00060, for «L.
The overall SMR of 125 based upon Quebec rates, for lung cancer mortality
among all miners 1s surprising. In studies of the mortality of male residents
of Thetford, 1n the midst of the Canadian asbestos mining area (Toft et al.,
1981; Wigle, 1977), an SMR of 184 was seen for lung cancer and 230 for cancer
of the stomach. Because no corresponding Increases were seen in female cancer
rates, Toft et al. (1981) and Wigle (1977) attributed the excesses to occupa-
tional exposure 1n the mines. S1em1atyck1 (1982) presented data on the mor-
tality of male residents of Asbestos and Thetford Mines, Quebec, that indicated
an SMR for lung cancer of 148 compared to Quebec rates. The origin of a
lower SMR for those employed 1n mining and ml'lling compared to all male resi-
dents could result from the departure of most short-term workers from the
area, but data on this possibility are lacking. While the risk appears low
compared to town mortality, the agreement between the SMR and RR analyses 1s
very good.
65
image:
TABLE 3-19. LUNG CANCER RISKS, BY DOSE, AMONG
CANADIAN CHRYSOTILE ASBESTOS MINERS
McDonald et al.
In mppcf-y
1980
SMR
Liddell et al.
Exposure
in mppcf-y
1977
Complete cohort:
Estimated average cumulative exposure:
185 mppcf-y.
Relative risk from an Internal case-control analysis.
) - number of deaths,
SMR
SMR
RR
RR
Regression equations
92 + 0.13(±0.024) x mppcf-y
93 + 0,13(±0.024) x mppcf-y
weighted
unweighted
RRe
< 1 year of employment
1 to
5 to
.5
1.7
5.8
39.0
4. 9 years of employment
3.3
13.6
59.0
231.3
19.9 years of employment
16.0
58.2
178.5
704.0
117 (19)b
91 (12)
88 (9)
80 (7)
66 (5)
95 (13)
82 (6)
78 (5)
141 (13)
122 (14)
83 (7)
217 (16)
3 (<6)
8 (6-10)
20 (10-30)
65 (30-100)
200 (100-300)
450 (300-600)
800 (600-1000)
1250 (1000-1500)
1750 (1500-2000)
3000 (2000+)
1.00 (43)
1.07 (10)
0.95 (24)
1.16 (37)
1.22 (31)
1.88 (27)
2.39 (18)
3,49 (10)
4.97 (6)
5.42 (9)
20+ years of employment
104.6
261.3
549.1
1141.4
121 (28)
108 (20)
220 (24)
265 (32)
125 (230)
0.99 + 0.0017(10.00013) x mppcf-y weighted
1.10 + 0.0017(10.00013) x mppcf-y unweighted
Weighted regression equation forced through an SMR of 100:
SMR = 100 + 0.12 (±0.02) x mppcf-y
66
image:
TABLE 3-20. LUNG CANCER INCIDENCE RATES IN URBAN AND
RURAL AREAS OF QUEBEC PROVINCE,
1969-1973
Region
Asbestos counties
Peripheral counties
Other rural
Montreal
Quebec City
Province
Ratio: Rural /Province
Ratio: Peripheral/Province
Rate
33.59
23.71
27.29
48.67
50.53
37.47
•
MALES
Population
57,1585
209,320
1,295,395
1,222,245
204,435
2,989,580
728
633
Rate
4.39
4.64
3.87
8.70
6.96
6.20
FEMALES
Population
57,630
210,180
1,264,795
1,281,865
218,745
3,033,215
.624
.748
From: Graham et al. (1977).
3.9.8 Mining and Milling, Thetford Mjnes_A Canada (Chrysotilejj Nicholson
(1976b); Nicholson et al. (1979)
Somewhat higher risks in the mining industry were obtained by Nicholson
(1976b) and Nicholson et al. (1979) from the mortality experience of a smaller
group of miners and millers employed 20 or more years at Thetford Mines,
Quebec. In this study, 178 deaths occurred among 544 men who were employed
during 1961 in 1 of 4 mining companies. In the ensuing 16 years of follow-up,
26 deaths occurred from asbestosis, 28 (25 on DC) from lung cancer (11.1
expected), and 1 from mesothelioma.
Fiber measurements were made during 1974 in five mines and mills, and
data on particle counts from 1948 were supplied by the Canadian Government.
From these data, exposure estimates were made for each of the 544 individuals
according to their job histories. Fiber exposures for earlier years were
estimated by adjusting current measurements by changes in particle counts
observed since 1950. The 20-year cumulative exposure for the entire group was
estimated to be 1080 f-y/ml.
The mortality experience of the whole group from an earlier follow-up was
reported by two exposure categories (Nicholson, 1976b) (see Table 3-21). The
difference in lung cancer SMRs in these two exposure groups suggests that
KL = 0,0023 [(333-55)/(1760-560)/100]. However, Canada rates were used to esti-
mate expected deaths and these overestimated mortality. As with the McDonald
67
image:
TABLE 3-21. EXPECTED AND OBSERVED MORTALITY AMONG 544 QUEBEC ASBESTOS
MINE AND MILL EMPLOYEES, 1961-1973
Average Exposure
Causes of death
All causes of death
All cancers
Lung
Mesothelloma
Gastrointestinal
Other cancers
Respiratory diseases
Pneumonia
Asbestosls
Other respiratory
All other causes
Exp.
68.29
15.45
4.52
4.18
6,75
4.79
2.01
--
2.79
48.05
560 f-v/ml
Obs.
65
15
7
1
3
4
10
1
7
2
40
Ratio
0.95
0.97
1.55
0.72
0.59
2.09
0.50
--
0.72
0.83
Cumulative Exposure
Exp.
44.56
10.11
3.00
2.71
4.40
3.02
1.27
—
1.76
31.43
1760 f-v/ml
Obs.
67
18
13
0
3
2
15
1
11
3
34
Ratio
1.50
1.78
4.33
1.11
0.45
4.24
0.78
—
1.70
1.08
Best estimate cause of death.
et al. (1980) study, KL will be multiplied by a factor of 1.5 to 0.0034 and
then reduced to 0.0030 to convert to DC lung cancer diagnosis. An analysis,
adjusted to local rates, using the overall SMR and average group exposure,
yields a value of K, = 0.0017. Because there 1s likely to be greater uncer-
tainty associated with the regression analysis than with the use of average
values, we will use the estimate of K, = 0.0017 for this study.
3.9.9 Mining and Milling, Italy (ChrysotHe); Rublno et al. (1979)
A final study of chrysotlle mining and milling 1s that of Rubino et al.
(1979) of the Balangero Mine and Mill, northwest of Turin. A cohort was
established of 952 workers, each with at least 30 calendar days of employment
between January 1, 1930 and December 31, 1965, who were alive on January 1,
1946. Ninety-eight percent of the cohort was traced and their mortality
experience through 1975 was ascertained. Overall, an exceptionally high
mortality was seen compared to that expected; 332 deaths were observed versus
214.4 expected. The excess mortality, however, was largely confined to non-
malignant respiratory diseases, cardiovascular diseases, and accidents. The
overall SMR for all malignant neoplasms was 106, with only cancer of the
68
image:
larynx found to be significantly in excess in the whole group. While the
overall data were relatively unremarkable, the age standardized rates of lung
cancer according to cumulative dust exposure showed a relative risk of 2.29
(2.54 based upon cancer of the lung and pleura) for a high exposure group (376
f-y/ml) compared to a low exposure group (75 f-y/ml) [K. = 1.29/(376-75) =
0.0043)]. A case-control analysis of lung cancer according to cumulative dust
exposure showed a relative risk of 2.61. Adjusting to a relative risk of 1 at
zero exposure gives a K. of 0.089. However, the characterization of the
exposures in the study may have created an artificially steeper dose-response
relationship than actually exists. Rubino et al. (1979) calculated the person-
years at risk in two exposure categories (±100 f-y/url). A person contributed
to the lower category until his exposure exceeded 100 f-y/ml. However, in
Section 3.6 it is shown that there is a 5-10 year lag before the risk is
manifest from a given exposure. Thus, the transition should be delayed by
5-10 years after achievement of 100 f-y/ml. Deaths and person-years at risk
occurring in this delay period should be attributed to the lower exposure
category. If lung cancer deaths occurred in the delay period, the dose-
response relationship is probably artificially steeper than it should be; if
no lung cancer deaths occurred, it is artificially shallower. The overall SMR
of those 20 years from onset yields a KL of 0.00013 [(103.4 - 100)7100/273 f-y/
ml]. The uncertainty in the estimate of K. is enormous. We will use the
geometric mean of 0.0043 and 0.00013, 0.00075, to represent KL.
3.9.10 Insulation Manufacturing, Paterson, NJ (Amosite); Seidman et al.
(1979)
The study by Seidman et al. (1979) also can be used for quantitative risk
estimates". The study was recently updated and the new mortality results were
submitted for the OSHA hearings record on a revised standard for asbestos
(Seidman, 1984). In this update, dose-response data, based upon estimates of
individual exposures for each cohort number, are available. Data for lung
cancer are listed in Table 3-22.
Because no data exist on air concentrations for the Paterson factory, the
data in terms of fiber counts were estimated from air concentrations in two
other plants manufacturing the same products with the same fiber and machinery.
One of these plants, in Tyler, Texas, opened in 1954 and operated until 1971;
the other, in Port Allegany, Pennsylvania, opened in 1964 and closed in 1972.
As in the Paterson factory, efforts to control dust in these newer plants were
69
image:
TABLE 3-22. CUMULATIVE OBSERVED AND EXPECTED DEATHS FROM LUNG CANCER
5 TO 40 ELAPSED YEARS SINCE ONSET OF WORK IN AN AMOSITE ASBESTOS FACTORY,
1941-1945, BY ESTIMATED FIBER EXPOSURE
(Seidman, 1984)
Cumulative
exposure
(f-y/nii)
<6.0
6.0 -
12.0 -
25.0 -
50,0 -
100.0 -
150.0 -
250+
Total
11.9
24.9
49.9
99.9
149.9
249.9
Number
of men
177
109
139
123
104
57
58
53
820
Number of deaths
(BE) (DC)
15
12
15
13
17
9
15
15
111
14
12
15
12
17
9
12
11
102
Expected
deaths3
5.31
2.89
3.39
2.78
2.38
1.49
1.32
0.94
20.51
SMR
(BE) (DC)
282
415
442
468
714
604
1136
1596
541
264
415
442
432
714
604
909
1170
497
Estimated average cumulative exposure: 67.1 f-y/ml.
BE = best estimate of cause of death based on all medical evidence.
DC = Death certificate cause of death.
aExpected deaths based on New Jersey white male quinquennial age and calendar
year period specific death rates.
Regression equations
SMR = 325 + 2.72(±0.54) x f-y/ml weighted
SMR = 330 + 2.45(±0.37) x f-y/ml unweighted
Weighted regression equation forced through an SMR of 100:
SMR = 100 + 4.28 (±1.17) x f-y/ml
limited. One, in fact, was housed in a low Quonset-type building where the
confined space exacerbated dust conditions. During 1967, 1970, and 1971,
asbestos fiber concentrations in these plants were measured by the U.S. Public
Health Service and the results published in the Asbestos Criteria Document of
the National Institute for Occupational Safety and Health (1972). These data
were supplemented by company data in one plant and individual worker estimates
of dustiness (which were used for some jobs not sampled).
The zero dose SMR intercept of 325 is highly anomalous and difficult to
understand. The use of New Jersey rates for calculating expected deaths is
70
image:
appropriate for the Paterson area (the age standardized county rates are 46.8
versus 46,3 for the state). The high Intercept 1s largely the result of a
disproportionately high risk observed in individuals employed for less then 6
months, whose SMR is 295 (32 observed, 10.86 exDosed), Certainly, new employees
usually get the dustiest jobs and if there are effects of intensity of exposure
separate from those of dose, very dusty environments may have contributed a
disproportionately greater risk. However, longer term employees also would
have had such jobs at one time and intensity effects are not seen 1n other
asbestos-exposed groups. Another possibility is that the short-term group
Includes many men exposed to carcinogens at work elsewhere or they are unusu-
ally heavy smokers. Abnormally high risks were also seen 1n the short-term
employees of a friction products plant studied by McDonald et al. (1984). A
third possibility is that there could have been misestimates of exposure for
the short-term employees who would have the extremely dusty jobs. However,
the dose-response relationship for death from asbestos is a reasonable one and
there is no unusual mesothelioma risk among those employed less than 6 months.
Finally, part of the excess may simply be the result of statistical fluctua-
tions.
The values of K, estimated by different treatments of the data range from
0.0084, obtained by adjusting the slope of the weighted regression line by the
intercept (2.72/325), to 0.059, obtained by dividing the excess overall lung
cancer SMR by the average group exposure [(49!i-100)/67.1/100]. If Inappro-
priate underlying rates (because of other exposures) .apply only to the short-
term group, an adjustment can be made by forcing the dose-response line through
the origin. This yields a value of KL = 0.043. Because this is most likely
to be the case, this value will be used for KL-
The uncertainty in the value extends from 0.0084 to 0.074 to account for
the statistical variability on the number of deaths and different values of K.
obtained from different analysis procedures.
3.9.11 Insulation Application, United States (Chrysotile and Amosite)
The previously discussed mortality study of Selikoff et al. (1979) can be
combined with published information on asbestos exposures measured for members
of this cohort to obtain an exposure-risk estimate. The data on insulation
workers' exposure were reviewed by Nicholson (1976a) and are summarized in
Table 3-23. Using the standard membrane filter technique of the U.S. Public
Health Service for counting asbestos fibers (Leldel et al., 1979), three
71
image:
TABLE 3-23. SUMMARY OF AVERAGE ASBESTOS AIR CONCENTRATION
DURING INSULATION WORK3
(Selikoff et al., 1979)
Averagefiber concentration, f/rnl
Light and heavy
Research group construction . Marine work
Nicholson (1975) 6.3
Cooper and Balzer (1973) 2.7 6.6
Ferris et al. (1971) 2,9
Harries (1971) 8.9
Average concentrations of all visible fibers counted with a konimeter
and bright-field microscope.
Murphy et al. (1971) 8.0
Fleischer et al. (1946) 30-40
Estimates of past exposure based on current membrane-filter data.
Nicholson (1976a) 10-15
aAverage concentrations of fibers longer than 5 urn evaluated by membrane
filter techniques and phase-contrast microscopy.
Source: Nicholson (1976a).
different laboratories in the United States found that the average fiber
concentration of asbestos dust in insulation work, between 1968 and 1971,
ranged from about 3 to 6 f/ml. A similar study in the Devonport Naval Dock-
yard in Great Britain, with the same techniques, obtained 8.9 f/ml for the
average of long-term sampling of asbestos concentrations measured during
application of insulating materials aboard ship (Harries, 1971). In the
research that led to these data, it was reported that peak exposures could be
extremely high. It was not uncommon, for example, to get 2- to 5-minute
concentrations of asbestos exceeding 100 f/ml during the mixing of cement.
This mixing, however, would only be done perhaps once an hour, so that exposures
measured during that hour, including the mixing, would seldom average more
than 10 f/ml. Similar experiences were subsequently reported by Cooper and
Miederna (1973), who stated, "Peak concentrations may be high for brief periods,
while time-weighted averages are often deceptively low."
72
image:
Direct Information on asbestos fiber concentration, measured by the
currently prescribed analysis procedures, has been available only since 1966.
Although Insulation materials have changed from earlier years (fiber glass has
found extensive use, and work with cork 1s seldom done today) and changes 1n
the asbestos composition of Insulating products have taken place (pipe cover-
Ings and Insulation blocks may have had twice the asbestos content 1n earlier
years), work practices are virtually Identical and few controls of consequence
were In use. Therefore, dust concentrations measured under these conditions
have relevance for estimating the levels of past years. Considering the
possible doubling of the asbestos content of older Insulation materials, the
data from the studies listed 1n Table 3-23 suggest that the aVerage exposures
of Insulation workers 1n the United States during past years could have ranged
from 10-15 f/ml for commercial and Industrial construction. In marine construc-
tion, 1t may have been between 15 and 20 f/ml. We will use a value of 15 f/ml
as an overall average. Because of the great variability In work activities of
this group, the range of uncertainty 1n the exposure 1s estimated to be from
7.5 to 45 f/ml, and this range 1s Indicated 1n Figure 3-7.
This information and the data 1n Figure 3-4 allow one to calculate a lung
cancer risk per unit of asbestos exposure (in f-y/ml) from the linearly rising
portion of the curve, the slope of which 1s 0.16 per year or 0.07 per f-yr/ml
(for an exposure Intensity of 15 f/ml).. However, the data of Figure 3-4
utilized BE (best estimates) 1n establishing lung cancer mortality. Adjusting
to DC (death certificate) diagnosis reduces the value of K. from 0.011 to
0.0094 (0.011 x 3.06/3.60). The statistical uncertainty on the estimate of
risk 1s very low. However, there 1s no independent Indication that the use of
U.S. mortality rates 1s appropriate. Hammond et al. (1979a) reported that
53.5 percent of Insulation workers were current cigarette smokers, 27.3 percent
were past smokers, and 17.2 percent never smoked cigarettes. The corresponding
data for the 1967 U.S. population were 49.1 percent current smokers, 23.6 per-
cent past smokers, and 27.3 percent non-cigarette smokers (Harris, 1979).
This difference would only affect the underlying rates by about 10 percent.
However, because insulation workers may have smoked more cigarettes, we will
reduce the value of K. by 20 percent to 0.0075.
73
image:
3.9.12 Asbestos Products Manufacturing, United States (Chrysotile and
Croddollte); Henderson and Enterline (1979)
The data of Henderson and Enter!ine (1979) (Figure 3-1 and Table 3-24)
can also be used to establish fiber dose-response data even though their data
were presented 1n terms of total dust concentrations measured in millions of
particles per cubic foot (mppcf). No data exist on the conversion between
mppcf and f/ml for most of the plants studied. However, there are data on the
relationship between fiber and total dust concentrations in textile operations
and asbestos cement production. Dement et al. (1982) found that conversion of
3 f/ml/mppcf was appropriate to most textile operations, although Ayer et al.
(1965) had earlier suggested a value of 6 f/ml/mppcf. In a plant making
asbestos cement pipe and sheets, Hammad et al. (1979) determined the conversion
value to be 1.4. It would be expected that the cement products value would be
most applicable to the Henderson and Enterline circumstance because of the
extensive use of cement and other mineral particles (e.g., calcium silicate,
talc, SiO-, MgO) in asbestos products manufacturing. The least squares weighted
regression line of SMR on dose is SMR = 143 + 0.51 x mppcf-y (see Table 3-24).
Using a value of 1.5 f/ml/mppcf to represent the conversion relationship, the
estimate of KL is 0.0034 (0.51/100/1.5).
TABLE 3-24. LUNG CANCER RISKS, BY DOSE, AMONG RETIREES
OF U.S. ASBESTOS PRODUCTS MANUFACTURERS
(Henderson and Enterline, 1979)
Exposure in mppcf-y SMR
62 (<10) 197.9 (19)a
182 (10-19) 180.0 (9)
352 (20-39) 327.6 (19)
606 (40-79) 450.0 (9)
976 (>80) 777.8 (7)
Complete cohort: 270.4 (63)
Estimated average cumulative exposure: 249 mppcf-y.
a() = number of deaths.
Regression equations
SMR = 143 + 0.51(10.13) x mppcf-y weighted
SMR = 100 + 0.66(±0.07) x mppcf-y unweighted
Weighted regression equation forced through an SMR of 100:
SMR = 100 + 0.64 (±0.097) x mppcf-y
74
image:
As described previously, observing a cohort beginning at age 65 may
seriously understate the full Impact of asbestos exposure. Most of the workers
1n this cohort began employment prior to age 25. To partially account for
selection effects among retirees, we will multiply the above value by 1.45.
[This adjustment 1s the ratio of the lifetime mortality from age 25 to lifetime
mortality at age 65 (see Table 3-8)]. Thus, K, 1s adjusted to a value of
0.0049.
3.9.13 Asbestos Cement Products. United States (Chrysotlle and Croc1do11te);
Welll et al. (1979); Hughes and Well! (11580)
A study of an asbestos cement production facility also provides exposure-
response Information (Welll et al., 1979; Hughes and Welll, 1980), as shown 1n
Table 3-25. Although the experience of 5645 Individuals was reported, 1791 of
whom had been employed for longer than two years, the dose-response Information
1s uncertain because of limitations 1n the mortality data. Of even greater
significance, tracing was accomplished*through Information supplied on vital
status by the Social Security Administration, and this Information only allowed
the vital status of 75 percent of the group to be determined. Those Individuals
untraced were considered alive 1n the analyses, which assumption may have led
to serious misestimates of mortality because prior to 1970, many deaths,
particularly of blacks, were not reported to the Social Security Administra-
tion. The percentage of unreported deaths of both sexes ranged from nearly
80 percent 1n 1950 to 15 percent 1n 1967 (Aziz and Buckler, 1980). Thus, many
cohort members could be deceased, a fact unknown to the researchers. This
could likely be the source of the extraordinarily low overall reported mortality
of the cohort, which allowed deficits of about 40 percent 1n several exposure
categories. (The overall SMR is 68.)
Two methods of adjustment for Incomplete trace can be made. In one, the
overall SMR for lung cancer 1s divided by the SMR for causes other than lung
and gastrointestinal cancer (66). This yields a value of K, = 0.0064, using a
value of 64 mppcf for the group exposure and a fiber-particle conversion
factor of 1.4 (Hammad et al., 1979) [((104/66)-l)/64/1.4]. Alternatively, a
regression of SMR on dose yields SMR = 70 + 0.43 x mppcf-y. The low value of
SMR 1s probably the result of missing deaths. If the percent missing 1s
similar in each category then KL = 0.0042 (0.43/100/1.4/0.70). We will use
the average of these values, 0.0053, for the point estimate of K.. The assump-
tion that there 1s an equal percentage of missing deaths In each category 1s
75
image:
TABLE 3-25. LUNG CANCER RISKS, BY DOSE, AMONG ASBESTOS CEMENT
PRODUCTION WORKERS (Wei 11 et al., 1979)
Exposure
in mppcf-y
5 (<10)
25 (11-50)
75 (51-100)
150 (101-200)
400 (>200)
SMR
77 (19)C
70 (8)
26 (1)
290 (9)
226 (14)
RRb
1.00
1.14
0.52
2.85
2.75
104 (51)
Estimated average cumulative exposure: 63.6 mppcf-y
Accumulated during first 20 years from initial employment.
Relative risk from an internal case-control analysis.
c( ) = number of deaths.
Regression equations
SMR = 70 + 0.43(10.22) x mppcf-y weighted
SMR = 77 + 0.46(10.31) x mppcf-y unweightea
RR = .96 + 0.47(10.18) x mppcf-y weighted
RR = .99 + 0.50(10.26) x mppcf-y unweighted
Weighted regression equation forced through an SMR of 100:
SMR = 100 + 0.31(10.22) x mppcf-y
uncertain. There are more untraced in the lowest category but a greater per-
centage of those untraced in the most exposed group may be deceased. If one
considers all of the untraced deaths to be in the lowest exposure categories
and forces a regression line through the origin, its slope is 0.0040. These
uncertainties in possible methods of adjusting for untraced deaths are indicated
in Figure 3-7.
3.9.14 AsbestosCement Products, Ontario. Canada (Chrysotile and Crocidolite);
Flnkelstein (1983)
A recent study by Finkelstein (1983) also relates mortality in an asbestos
cement products facility to measured exposures. He established a cohort of
241 production and maintenance employees from records of an Ontario asbestos
cement factory, consisting of all individuals who had nine or more years of
75
image:
employment beginning prior to 1960. Their mortality experience was followed
through October 1980. Impinger particle counts of varying degrees of compre-
hensiveness were available from various sources (government, insurance com-
pany, employer) from 1949 until the 1970s. After 1973, membrane fiber counts
were taken. Individual exposure estimates were constructed based on recent
fiber concentrations at a particular Job. They were modified for earlier
years due to changes in dustlness of the job, as determined by the impinger
particle counts. These counts were thought to be accurate to within a factor
of 3-5. Examples of exposure estimates for the years 1948-1954 for willow
operators, forming machine operators, and lathe operators were 40 f/ml, 16
f/ml, and 8 f/ml, respectively.
The lung cancer mortality data are shown in Table 3-26. The dose-response
relationship is anomalous. The first two exposure categories show the risk
increasing steeply with exposure, but 1n the last category it falls signifi-
cantly. Both GI cancer and mesothelioma show a strong positive trend with
exposure, suggesting that the exposure rankings are correct. The only regres-
sion line that makes sense is one forced through an RR of 1 at zero exposure.
This yields a K, of 0.048, which is close to that calculated from the overall
mortality excess and average group exposure. The average cumulative 18-year
exposure for the production group in the asbestos cement work was 112.5 f-y/ml.
Lung cancer deaths observed 1n this group were 17 versus 2.0 expected from
Ontario rates for an SMR of 850. This yields a value of KL = 0.067 [(850-100)/
112.5/100] which will be used as the estimate from this study.
We do not know the reasons for the very significant difference in risk
seen in two plants (of the same company) producing the same product. The
point estimate of risk from Finkelsteln (1983) (KL = 0.067) is 13 times that
of Wei 11 et al. (1979) («L = 0.0053) even after attempting to correct for the
Incomplete trace of the latter study. Data on the duration of exposure are
not given by Finkelsteln (1983), but it would appear that the estimated average
fiber exposure of his cohort was between 7 f/ml and 12 f/ml. (The average
cumulative exposure over 18 years was 112 f-y/ml; all cohort members were
employed for at least 9 years, one of which must have been 1n an asbestos work
area.) This average concentration is about half of that estimated by Welll
et al. (1979), using the particle-to-fiber conversion of Hammad et al. (1979).
It is not possible to evaluate the accuracy of either set of exposure estimates.
The exposure estimates of Finkelstein (1983) w«re submitted to company offi-
cials who thought they were reasonable; but worker descriptions of plant
77
image:
TABLE 3-26. LUNG CANCER RISKS, BY DOSE, AMONG
ONTARIO ASBESTOS CEMENT WORKERS
(Flnkelstein, 1983)
Standardized mortality deaths/1000 p-y
Exposure 1n f-y/ml Lung Cancer
Ontario
44
92
180
1.6
13.6 (5)a
92.1 (7)
11.9 (6)
Complete cohort: 850 (17)
Estimated average cumulative exposure: 112 f-y/ml.
a() = number of deaths.
Regression equations
(Forced through the value 1.6 at zero exposure)
Lung cancer RR = 1.60 + 0.077 x f-y/ml weighted
Lung cancer RR = 1.60 + 0.108 x f-y/ml unweighted
conditions suggest that very high exposures occurred periodically (Ontario
Royal Commission, 1984). In a study of asbestosis in the Ontario plant
(Flnkelstein, 1982), data comparable to that of Berry et al. (1979) were
obtained. Finkelstein observed prevalence rates of asbestosis of 4 percent
and 6 percent at 50-99 f-y/ml and 100-149 f-y/ml versus 2.5 percent and
8.5 percent by Berry et al. (1979). Henderson and Enterline (1979) observed
SMRs of 231 and 522 among retirees of cement sheet and shingle workers and
cement pipe workers, respectively. These values are more consistent with the
higher risk of Flnkelstein (1983) than the lower one of Weill et al. (1979).
In Figure 3-7, a fivefold downward uncertainty is indicated in K. to reflect
the maximum stated uncertainty 1n the exposure estimates of Finkelstein (1983).
3.9.15 Lung Cancer Risks Estimated in Other Reviews
A number of other individuals or groups have also estimated unit exposure
risks for lung cancer from these same epidemlological studies. These are
shown in Table 3-27. Because of general agreement on the appropriate model
for lung cancer, the unit exposure risks estimated in this document are very
78
image:
TABLE 3-27. COMPARISON OF ESTIMATED LUNG CANCER RISKS BY VARIOUS GROUPS
OR INDIVIDUALS IN STUDIES OF ASBESTOS-EXPOSED WORKERS
Percent increase
Study
Dement et al, (1983b)
McDonald et al. (1983a)
Peto (1980) after 1950
before 1951
McDonald et al. (1983b)
Berry and Newhouse (1983)
McDonald et al. (1984)
McDonald et al. (1980)
Nicholson et al. (1979)
Rubino et al . (1979)
Seidman (1984)
Selikoff et al. (1979)
Henderson and Enterline (1979)
Wei 11 et al. (1979)
Finkelstein (1983)
Newhouse and Berry (1979) Males
Females
This
Document
2
2
1
1
0
0
0
0
0
4
0
0
0
6
.8
.5
.1
.4
.058
.010
.06
.17
.075
.3
.75
.49
.53
.7
CPSC3
2.
1.
0.
0.
0.
0.
6.
1.
0.
0.
4.
3
0
06
06
12
17
8e
0
50
31
8
in lung cancer per f-y/ml
NASb
5.
0.
0.
0.
0.
9.
1.
0.
1.
8.
3
8
07
06
15
le
7
3
3
4
Ontario
Royal
Commission
4.2
1.0
0,058
0.020-0.046
1.0
0.069
f
4.2f
of exposure (100
Liddel
and Hanley
mppcf-y
6.
5.
5.
0,
0.
0.
3.
3.
0.
0.
9
9
1
,00
00
16
3e
7
35
66
1
(1985)
f-y/ml
2.4
2.0
1.7
0.00
0.00
0.05
1.1
1.2
0.23
0.47
x K )
Doll and
Peto (1985)
f-y/ml
1.25
1.5 ,
0.54d
Values used for risk extrapolation
Geometric mean of all studies
Geometric mean excluding
mining and mil 1 ing
0.3-3.0
2.0
0.02-4.2
0.65
1.0
1.0
U.S. Consumer Product Safety Commission (1983).
National Academy of Sciences (1984).
C0ntario Royal Commission (1984).
All men employed after 1932.
eData from Seidman et al. (1979).
Unpublished data supplied to the Commission.
image:
similar to those estimated by others. The differences in the values lie in
the choice of the method to obtain a dose-response relationship and the treat-
ment of potential biases in a study.
3.9.16 Summary of LungCancer Dose-Response Relationships
The results of all the determinations of K, , the fractional increases in
lung cancer risk per f-y/ml exposure, are displayed in Figure 3-7, along with
estimates of statistical variation, adjustments for possible biases, and
estimates of uncertainties associated with exposure determinations. The
details of the calculations of statistical uncertainty are provided in Table
3-10, which also shows that the confidence limits associated with an individual
value of KL are large. The uncertainties are largely the result of statis-
tical variations associated with small numbers and uncertainties in exposure
measurements. However, statistical variabilities appear to be more important.
In 9 of the 14 studies, uncertainties in the measure of response contribute
more to the overall uncertainties than do uncertainties in the measure of
exposure. Three studies have 95 percent confidence limits of about two orders
of magnitude.
Figure 3-7 displays the unit exposure risks in 14 studies, by predominant
fiber type in the exposure and by industrial process. Table 3-28 lists the
geometric mean of the unit exposure risks, estimated for the different indus-
trial processes, showing substantial differences in the risks observed, even
between processes using predominantly the same asbestos mineral. Significantly
lower unit exposure risks (p <0.05) are associated with chrysotile mining and
milling and friction product manufacturing compared to the other three processes
studied. However, because of the great uncertainty associated with the unit
exposure risks in friction products manufacturing, the level of significance
of the difference is less than for mining and milling.
There is reasonable agreement between the unit risks observed in different
studies within a given industrial process. In the case of textile production,
even though the cohorts studied by Peto (1980) and McDonald et al. (1983b)
were exposed to some quantities of crocidolite, the unit risks were very
similar to that of the plant studied by Dement et al. (1983b) and McDonald
et al..(1983a). The only substantial difference in the four groups exposed to
mixed fibers in manufacturing processes is the high unit risk observed in the
study of Finkelstein (1983). Whether this is real or the result of uncertain-
ties in the study cannot be established at this time. There is no statistical
80
image:
TABLE 3-28. WEIGHTED GEOMETRIC MEAN VALUES AND ESTIMATED 95 PERCENT
CONFIDENCE LIMITS ON K, FOR THE VARIOUS ASBESTOS EXPOSURE CIRCUMSTANCES
DEPICTED IN TABLE 3-10 AND FIGURE 3-7.
Asbestos process
or use
Textile production
Friction products
Fiber exposure
Predominantly
Chrysotlle
Chrysotile
Geometric
mean.
value of K.
0.020
0.00023
95% confidence
Interval
(0.0096 - 0.042)
(0.00010 - 0.0051)
manufacturing
Mining and milling
Amoslte Insulation
production
Mixed product
manufacturing
or use
All processes
All processes
except mining
and milling
Textile production
and mixed product
manufacturing or
use
Chrysotlle
Amos He
Amoslte
Chrysotlle
CrocidolHe
Amoslte
Chrysotlle
Crocidolite
Amoslte
Chrysotlle
CrocidolHe
Amoslte
Chrysotlle
Crocidollte
0.00098
0. 043
0. 0068
0.0065
0.010
0.013
(0.00028
(0.0084 -
(0.0035 -
(0.0025 -
(0.0040 -
(0.0074 -
- 0.0034)
0.074)
0.013)
0.017)
0.027)
0.024)
difference in the unit exposure risk seen In the group exposed only to amosite
asbestos compared to those exposed predominantly to Chrysotlle in textile
production or to mixed fibers in manufacturing.
The origin of the differences in unit exposure risks between mining and
milling and other Chrysotlle exposure circumstances is not completely clear.
It was suggested by many individuals, Including McDonald et al. (1984), that
the differences between mining and milling and various production processes
may be related to differences in the fiber size distributions. As in the review
of experimental studies (Chapter 4), fiber length and diameter strongly affect
the potential for fibers to produce mesothelloma. Corresponding data are not
81
image:
available for lung cancer, but It would be expected that different fiber,size
distributions would produce different responses. There are many long and
curly fibers present 1n the environment of miners and millers which are easily
counted, but not easily Inspired because of their large equivalent dtameter.
In asbestos-using industries, as fibers are broken apart a greater percentage
are deposited in the lung. Many of these will remain within a carcinogenic
size range. However, the number counted by the membrane filter procedure
compared to the number that are potentially carcinogenic may substantially
decrease in such circumstances.
As shown in Table 3-28, the geometric mean value of K, , using data from
all studies, is 0.0065, and that for all studies exclusive of mining and
milling is 0.010, Because the mining and milling exposures (long and curly
fibers, preprocessed) are likely to be less typical of those experienced in
the environment (processed, see also Sections 3-8, 3-9, 3-17, 4-2, and 5-1 to
5-8), the best estimate for the fractional increased risk of lung cancer, K. ,
for environmental asbestos exposures appears to be 0.010. This value is the
same as that used by the Occupational Safety and Health Administration in
their risk assessment for the proposed revision to the asbestos standard
(OSHA, 1983). OSHA's analysis also was based on risks in studies other than
chrysotile mining and milling. The value is one-half that which was adopted
by the National Academy of Sciences in their risk analysis (National Academy
of Sciences, 1984). The NAS value was based on rounding upward, to 0.02, a
median risk of 0.011 estimated in a group of 11 epidemiological studies.
The 95 percent confidence limits on the value 0.010 for KL are from
0.0040 to 0.027 (a factor of 2.5). This is the result of the analysis of
variance in 11 separate estimates. The 95 percent confidence limits on the
value of K, that might be measured in any unstudied exposure circumstance is
estimated to be a factor of 10 (8.3 by calculation). The range of uncertainty
may, in fact, be greater than the 10 fold factor estimated here, but insuffi-
cient information exists by which to make any more precise or definite estimate.
3.10 TIME AND AGE DEPENDENCE OF MESOTHELIOMA
In contrast to lung cancer, for which a relative risk model well explained
the data, mesothelioma is best described by an absolute risk model in which
the incidence is independent of the age at first exposure and increases accord-
ing to a power of time from onset of exposure. The rationale for such a model
82
image:
describing human cardnogenesls was discussed by several authors (e.g., ArrrH-
tage and Doll, 1961; P1ke, 1966; Cook et al., 1969). Such a model was utilized
by Newhouse and Berry (1976) 1n predicting raesothelloma mortality among a
cohort of factory workers In England. Specifically, they matched the Incidence
of mesothelloma to the relationship
IM = c(t - w)k (3-4)
where !„ 1s the mesothelloma incidence at time t from onset of exposure, w is
a delay 1n the expression of the risk, and c and k are empirically derived
constants. The incidence of asbestos-Induced mesothelioma in rats (Berry and
Wagner, 1969) followed this time course. In the case of the analysis of
Newhouse and Berry (1976), the data suggested that the value of k was between
1.4 and 2 and w between 9 and 11 years. However, the relatively small number
of cases available for analysis led to a large uncertainty in the values
estimated for either k or w. Peto et al. (1982) recently analyzed mesothelloma
incidence in five groups of asbestos-exposed workers. In one study analyzed,
that of Sellkoff et al. (1979), the number of cases of mesothelloma were
sufficiently large that the age dependence of the mesothelioma risk could be
Investigated. Peto et al. (1982) showed that the absolute Incidence of meso-
thelloma was Independent of the age at first exposure and that a function, IM
32
= ct (see Equation 3-4), fit the data well between 20 and 45 years from
onset of exposure. However, observed incidence rates for earlier times were
less than those projected, and the authors suggested that an expression propor-
2
tlonal to (t - 10) better fit the data up to 45 years from onset of exposure.
The analysis of Peto et al. (1982) excluded individuals first employed before
1922 and after 1946 and over the age of 80; the fit to the mortality of the
entire group suggested a value of k of about 5,
Figure 3-8 shows the risk of death of mesothelloma, according to age, for
Individuals first exposed between ages 15 and 24 and between ages 25 and 34.
As can be seen, these data, although somewhat uncertain because of small
numbers, are roughly parallel and separated by 10 years, as was the relative
risk for lung cancer. Thus, the absolute risk of death from mesothelioma
appears to be directly related to onset of exposure and is independent of the
age at which the exposure occurs. The risk of death from mesothelloma among
the insulation workers is plotted, according to time from onset of exposure,
on the right side of Figure 3-8. It increases to 40 years from onset of
exposure. Thereafter, the increase 1s less. There is even a decrease 1n the
83
image:
1000
500
CO
oc
<
LU
o
W
cc
UJ
0.
oc
UJ
0.
W
UJ
Q
200
100
50
20
10
AGE AT ONSET
A< AGE25yr.
• > AGE25yr.
I I
1000
500
.00
100
50
20
10
J L
10
20
40 60 80
10
20
40
60
AGE, years
YEARS FROM ONSET
OF EXPOSURE
Figure 3-8. The risk of death from mesothelioma
among insulation workers according to age and
years from onset of exposure. The risk of death
according to age is shown separately for insulators
first employed before age 25 and after age 25.
Data supplied by I.J. Selikoff and H. Seidman.
Source: Nicholson et al. (1982).
84
image:
risk at 50+ years from onset. This can be the result of misdiagnosis of the
disease in individuals age 75 and older, statistical fluctuations associated
with small numbers, or selection factors also seen in the risk of lung cancer
(e.g., those who lived to age 80 may have had jobs with much lower exposure),.
The graph of Figure 3-8 is also represented by an equation of the form
IM = c-f(t-w)k+1 (3-5)
The data of Figure 3-8, however, are not sufficient to separately specify w
and k. If w is 0, k lies between 4 and 5, If w is 10, k lies between 2 and
3. To estimate the risk from long-term exposures; consider an exposure of
duration d that began T years ago. The incidence of mesothelioma at time t
from the entire exposure is
IM = c<f'/}-d (t-10)kdt (3-6a)
assuming a delay of 10 years. The choice of a delay of 10 years is indicated
by the data on lung cancer risk, where a delay of from 5 to 10 years was
juserved beu^ -Bastes exposure and the manifestation of risk, f is the
intensity of the asbestos exposure, and as used in Equation 3-6, assumes a
linear relationship between intensity of exposure and risk (see Figures 3-4
and 3-5). Equation 3-6 is also linear in dose for short duration exposures.
Equation 3-6 yields
' f •
c k+l k+l C3"6b)
c
tt • f • t(T-io)1
Using a value of k = 2 (which best fits the workers' data) and letting c/k+1
KM leads to the following relations for varying times of exposure:
IM(t,d,f) = KM • f[(T-10)3 - (T-10-d)3] for: T > 10+d (3-6c)
= KM • f(T-10)3 for: 10+d > T > 10 (3-6d)
= 0 for: 10 > T (3-6e)
85
image:
Here I,. Is the mesothelioma Incidence at t years from onset of exposure
to asbestos for duration d at a concentration f. KM is carcinogenic potency
and may depend on fiber type and dimensionality. Note that I., depends only
upon exposure variables and not upon age or calendar year period.
1C. 1s the measure of the mesothelioma risk per year. In order to calculate
the full effect of an asbestos exposure on an exposed population over time,
the calculated incidence per year must be summed for each interval from onset
of exposure. In such a calculation, it is necessary to take account of the
mortality that occurs in the exposed population as it ages. In practice, such
calculations, are carried out by 5-year age and onset of exposure intervals.
3.11 QUANTITATIVE DOSE-RESPONSE RELATIONSHIPS FOR MESOTHELIOMA
Four studies provide information on the incidence of mesothelioma (pleura!
and peritoneal combined) according to time from onset of exposure, and contain
data that allow estimates to be made of the duration and intensity of asbestos
exposure. These data are given in Table 3-29. Values for 1C., the potency
factor for mesothelioma risk, can be estimated using Equations 3-6c, 3-6d, and
3-6e. Other studies reported cases of mesothelioma, but incidence data are
lacking or simply not provided. In others, the data were not given because
very few mesothelioma deaths were seen. Thus, some studies with missing data
could have a lower value of K... Note that we are estimating values of Ku from
a biased sample of those studies 1n which K, was estimated. A measure of the
bias can be estimated by comparing the values of 1C. and K, obtained in each
analysis with an analysis of the percentage of deaths from mesothelioma compared
to excess lung cancer in other studies. The estimate of 1C, for each of the
four studies was made by calculating a relative mesothelioma incidence using
Equation 3-6 and data on duration and intensity of asbestos exposure. The
relative Incidence curves were then superimposed on the observed Incidence
data 1n each study to obtain the value of KM> These fits are depicted on
Figures 3-9 and 3-10. The four studies are described below and summary data
are listed in Table 3-30.
86
image:
TABLE 3-29. MESOTHELIOMA INCIDENCE BY YEARS FROM ONSET OF EXPOSURE,
IN FOUR STUDIES
Incidence (cases/10,000 person-years)
Years from onset Insulation workers Textile workers
of exposure Peto et al. (1982) Peto (1980)
15-19
20 - 24
25 - 29
30 - 34
35 - 39
40 - 44
45 - 49
50+
1.2 (2,3)a
3.2 (7,6)
15.4 (18,29)
28.9 (16,34)
52.6 (20,26)
56.9 (6,19)
108.1 (14,18)
66.4 (4,14)
0.0
5.7 (1,0)
13.4 (2,0)
23.9 (2,0)
39.4 (2,0)
Amoslte factory Asbestos cement
workers workers
Seidman (1984) Finkelstein (1983)
15 -
20 -
25 -
30 -
35 -
40 -
45 -
50+
19
24
29
34
39
44
49
0.
7.
26.
50.
18.
0
4
2
8
4
(1
(3
(4
(0
,D
,2)
,4)
,2)
8.
37.
90.
96.
5
7
9
2
(1)
(4)
(5)
(1)
a( , ) = number of pleural and peritoneal deaths, respectively.
87
image:
100
50
1 I F~.TT
UJ
>r 20
o
u>
cc
10
"*• r-
0 5
SELIKOFFETAL. (1979)
PETO ETAL. (1982)
INSULATION WORKERS
20
— 100
T I TTT
50
20
10
1
SEIDMAN (1984)
AMOSITE FACTORY WORKERS
I I I I
40 60 20
YEARS FROM ONSET OF EXPOSURE
40
60
Figure 3-9. The match of curves calculated using Equation 3-6 data
on the incidence of mesothelioma in two studies. The fit is achieved
for Kjyj = 1.5 x 10"8 for insulators data and KM =3.2x10"B for the
amosite workers data.
Source; Petoetal. (1982); Selikoff et al. (1979); Seidman (1984).
88
image:
100
50
I T
CO
20
10
UJ
Q
PETO (1980)
TEXTILE WORKERS
I I I
100
BO
20
10
1
FINKELSTEIN 11983)
CEMENT WORKERS
I I
20
40 60 20
YEARS FROM ONSET OF EXPOSURE
40
60
Figure 3-10. The match of curves calculated using Equation 36 to
data on the incidence of mesothelioma in two studies. The fit is
achieved for KM = 1.0 x 10*a for the textile workers data and KM =
1.2 x 10~7 for the cement workers data,
Source: Peto (1980); Finkelstein (1983).
89
image:
TABLE 3-30. SUMMARY OF THE DATA KM, THE MEASURE OF MESOTHELIOMA RISK PER
FIBER EXPOSURE, IN FOUR STUDIES OF ASBESTOS WORKERS
Study
Average Average
employment exposure,
duration f/ml K,
H
VKL
Insulation workers 25
(Sellkoff et al.t 1979;
Peto et al., 1982)
Textile workers 25
(Peto, 1980;
Peto et al., 1982)
Amosite factory workers 1.5
(Seldman, 1984)
Cement factory workers 12
(Finkelstein, 1983)
15 1,5 x 10~8 2.0 x 10"6
20 1.0 x 10"8 0.9 x 10~6
35 3.2 x 10"8 0.7 x 10"6
9 1.2 x ID"' 1.8 x 10"6
3.11.1 Insulation Application; Selikoff et al. (1979); Peto et al. (1982)
A follow-up through 1979 of the cohort of insulation workers provides
data on the incidence of mesothelioma with time from onset of exposure (Peto
et al., 1982). It was estimated that their time-weighted average exposure was
15 f/ml (Nicholson, 1976a). Using these data and 25 years for their average
_Q
duration of exposure, a value of K,. - 1.5 x 10 is estimated.
3.11.2 Amosite Insulation Manufacturing; Seidman et al. (1979)
The average employment time of all Individuals in this factory was 1.46
years. This value and the previously used value of 46 f/ml for the average
exposure yields an estimate for KM of 3.2 x 10 .
3.11.3 Textile Products Manufacturing; Peto (1980); Peto et al. (1982)
A 20-30 f/ml value for exposure intensity is suggested by data presented
by Peto (1980). However, some uncertainty exists regarding this value because
of discrepancies in relative exposures measured by personal samplers and
static samplers. If exposures measured by personal samplers are less than
static samplers, as suggested by the data of Smither and Lewinsohn (1973), the
average exposure could be about 15 f/ml. Using 20 f/ml and an employment
_Q
period of 25 years, a value of KM = 1.0 x 10 is estimated.
'M
90
image:
3.11.4 Asbestos Cement Products, Ontario. Canada; FlnkeUtein (1983)
The cumulative exposure of the cohort over 18 years was 112 f/yr. Only
men with nine or more years of employment were Included'in the cohort. Although
data on the exact duration and Intensity of exposure are unavailable, we will
use a value of 12 years for duration of exposure and 9 f/ml for the Intensity
of exposure. This yields a value of K^ = 1.2 x 10 .
3.11.5 Other Studies
A note on the friction product studies 1s appropriate. In the study of
Berry and Newhouse (1983) little excess lung cancer risk was observed (see
Section 3.9.5). Eleven deaths from mesotheliona occurred. A comparison of
the work histories of the cases and 40 controls matched for sex, age, and date
of hire showed an Increased probability of croddollte exposure among the
cases (eight had such exposure) and an Increased probability of heavy chrysotlle
exposure. In the study of McDonald et al. (1984), an elevated risk of lung
cancer was observed but no trend with increasing exposures (see Section 3.9.6).
McDonald et al. (1984) did not find any mesothelioma deaths among the cohort
members. However, three mesothelioma deaths among former plant employees were
reported to the Connecticut Tumor Registry (Teta et al., 1983). Two were in
women and one 1n a male who terminated employment prior to receiving a Social
Security number and, thus, all were excluded from the cohort of McDonald et
al. (1984). Mention of the mesothellomas is Important because 1t Illustrates
that cases can occur from chrysotlle exposures in friction products manufacture.
Because of the low observed lung cancer dose-response relationship in both the
studies of McDonald et al. (1984) and Berry and Newhouse (1983), no meaningful
data on mesothelioma risk relative to lung cancer can be obtained.
3.11.6 Summary of Mesothelioma Dose-Response Relationships
A review of the four studies for which values of 1C. were obtained indicate
that three are very similar while 1C. from the study of Finkelstein (1983) 1s
much higher. This was also found in the value of K. estimated in that study.
Much closer agreement exists in the ratio of KM/K,. While it 1s not possible
to make an accurate estimate of the value of 1C. In the 10 other studies used
to estimate K, , a rough measure of mesothelloma risk can be obtained by calcu-
lating the ratio of the number of mesothelioma deaths to total deaths and
dividing by the cumulative exposures of the groups. This is done in Table 3-31.
91
image:
TABLE 3-31. ESTIMATE OF A MEASURE OF MESQTHELIQMA SJSK RELATIVE TO LUNG CANCER RISK, IN 14 STUDIES
Column 1
Calculated
Study KH(xlQ8)
Textile Production
Dement et al. , 19836
McDonald et al. , 1983a
Peto, 1980 1.0
McDonald et al. , 19835
Friction Products
Berry & Newhouse, 1983
McDonald et al., 1984
fti m'ng_and Mi 1 1 ing
McDonald et al., 1980
Nicholson et al, , 1979
Rubino el al., 1979
Anosite Insulation Manufacturing
Seldman, 1984 3,2
Insulation Application
Selikoff et al., 1979 1.5
Asbestos Products Manufacturing
Henderson & Enter! ine, 1979
Wei 11 et al. , 1979; tfeill, 1984
Finkelstein, 1983 12
Column Z
\
0,028
0.025
0.011
0.014
0,00058
0,00010
0.00060
0,0017
0.00081
0.043
0. 0075
0. 0049
0.0053
0.048
Column 3
Cumulative
exposure
(f-y/ml)
43.9
30.9
500
50.7
37.1
30.9
555
1070
258
67.1
375
373
89
112
Column 4 Column 5
Hesothelioma
deaths Col. 4 x 104
lota) deaths Col. 3 £
0.0041 0.91
0.0018 0.58
0.040 0.80
0.016 3.16
0.0050 1.62
0.0030* 0.97
0.0030 0.05
0.0056 0.05
0.0045 0.17
0.029 4.26
0.087 2.32
0.0064 0.17
0.0046° 0.652
0.153 13,66
Geometric means
excluding friction products
excluding friction products
and studies of Dement and
Nichalson
Column 6 Column 7
Col. 5
al, 2 x Iff2 K,^
0.33
0.23
0.73 0.91 x 10"5
2.25
27.9
97
0.83
0.29
2.10
0.99 0.74 x 10"
3.09 2.0 x 10"
0.35
0.98
2.85 1.8 * l<f
0.87
1.07
3No mesolheliomas were reported in the male cohort studied. However, three raesothelionas (two in women) were reported
from the workforce of the plant studied (Teta et al., 1983). The rough mesotheliona risk calculation uses these three
cases and a value of 1000 for the total mortality in the plant work force.
bln 1984 testimony before OSHA, Weill reported 9 mesotheliomas among 1953 deaths in his cohort of cement workers.
92
image:
Column 5 of Table 3-31 indicates this rough mesothelioma risk in all 14 studies,
and Column 6 shows the ratio of this risk to 100 x K, . Note that the two
measures of risk are not commensurate. To make this explicit the ratio will
be designated as the "relative mesothelioma hazard." The geometric mean of
the relative mesothelioma hazard in all studies except friction products
manufacturing is 0.87. The ratios in the two friction products studies are
very uncertain because of the great uncertainties in the lung cancer risks,
and they are not included in the average. Table 3-32 lists the geometric
means, by process, of the relative mesothelioma hazards in all studies except
Dement et al. (1983b) and Nicholson et al. (1979) (whose mesothelioma cases
are included in the larger studies of McDonald et al., 1980, 1983a,b).
The geometric means of the relative mesothelioma hazards, by process,
differ very little (excluding consideration of friction products because of
the large uncertainties in lung cancer risk.) Textile production, including
studies of plants that used some crocidolite and amosite have the lowest
average hazard. Product manufacture and use has the highest relative mesothe-
lioma hazard. This is largely the result of the high hazard found among
insulation workmen who were exposed only to amosite and chrysotile, but where
a review was made of all available pathological material to identify cases.
The geometric average of the manufacturing plant studies is 0.99, coincidentally
the same as found in amosite insulation manufacture. Chrysotile mining also
demonstrated a high relative mesothelioma hazard (although in absolute terms
the unit exposure risks for both mesothelioma and lung cancer are lower than
other asbestos exposure circumstances). The high relative hazard was, in
part, the result of a high relative hazard found in the study of Rubino.
Nevertheless, the hazard found in the large study of McDonald et al. (1980),
0.83, is higher than that of textile production (predominantly chrysotile but
with some crocidolite and amosite) and little different from all product
manufacturing, 0.99, using all types of asbestos, Thus the geometric mean of
all studies, 1.07, fairly represents all exposure circumstances, except perhaps,
insulation work.
There is no evidence in those studies listed in Table 3-31 and 3-32 that
would suggest a substantially different relative mesothelioma hazard for the
different types of asbestos varieties. However, this conclusion is limited by
the fact that crocidolite was not the dominant fiber exposure in any of the
study groups. In an analysis of the risk of pleural and peritoneal mesothe-
lioma relative to excess lung cancer in all published cohorts, including those
93
image:
TABLE 3-32. ESTIMATED GEOMETRIC MEAN VALUES OF THE RELATIVE
MESOTHELIOMA HAZARD (COL. 6 OF TABLE 3-31) FOR THE VARIOUS
ASBESTOS EXPOSURE CIRCUMSTANCES LISTED IN TABLE 3-31
Geometric mean value
of relative hazard
(Col 6, Table 3-31)
Textiles (except Dement et al., 1983b)a 0,72
Friction products 52
Mining and milling 1.32
(except Nicholson et al., 1979)a
Araosite manufacturing 0.99
Asbestos product manufacturing 1.32C
and use (crocidolite 0% of insulation,
15% of two factories; 5% of Manville plant)
Geometric mean'of all except 1.07
friction products (excluding Dement et al.,
1983b, and Nicholson et al., 1979)
Geometric mean of all except friction 1.02
products and mining and milling
aA single mesothelioma case is included in the larger study of McDonald
et al.
b
An unreasonably high value because of low lung cancer risk.
Crocidolite contribu
tion of crocidolite.
cCrocidolite contribution very small and can't extract out relative contribu-
with only crocidolite exposures, it would appear that the ratio of the cases
of pleural mesothelioma to excess lung cancers is two to three times greater
than that from amosite, chrysotile or mixed fiber exposures. (See Section 3-17,
Relative Carcinogenicity of Different Asbestos Varieties.) Considering both
pleural and peritoneal sites this ratio increases to three or four times for
pure crocidolite exposures. There are no estimates of the relative exposures
to crocidolite in those cohorts where such exposure was possible. However, to
estimate the possible effect, the relative mesothelioma hazard for the studies
of Peto (1980) and McDonald et al. (1983b) were reduced by 20 percent to
account for effects of a 2 percent crocidolite usage and those of asbestos
94
image:
products manufacturing by 50 percent. This yields a geometric mean of 0.85
rather than 1.07, This 26 percent difference for an assumed effect of croci-
dolite in five studies is far less than the tenfold uncertainty in the estimated
values of K, or K., for an unstudied exposure circumstance. Because of the
absence of any evident effect of crocidolite in the values of relative mesothe-
lioma risk in the Table 3-32 and small estimated crocidolite correction to the
relative mesothelioma hazard, no adjustment will be made to the final estimated
value of KH (which have associated with it a twentyfold uncertainty in estimating
an unknown exposure risk).
The'relative mesothelioma hazard in the four studies for which the geometric
mean of K.. was calculated is 1.59. The geometric mean of the relative mesothe-
lioma hazard in all studies (excluding friction products) is 1.07. This
suggests that the value of KjVK. in the four studies is 49 percent higher than
the average for all studies. As the geometric mean of the calculated values
of KL/K. in the four studies is 1.25 x 10 , the above data suggest a value of
KM/KL for all studies of 0.84 x 10 . However, this is certainly a lower
limit on the value of the ratio. Firstly, inclusion of the friction products
studies would raise it by some (unknown) amount. Secondly, 3 of the 4 studies
for which K^/K, was calculated used data from all available pathological
materials and medical records to identify mesothelioma cases, while those not
analyzed generally did not. Had all studies done so, the relative mesothe-
lioma hazard would be higher (in the Seidman, 1984 and Selikoff et al., 1979
studies such review increased the number of mesothelioma cases by 75 percent).
To partially account for these factors we will use a value of 1.0 x 10 for
~8
the ratio of KM/KL- The average value of KM is thus 1.0 x 10 .
The 95 percent confidence limits on the estimated value of K. was a
factor of 2.5 and a factor of 10 on its application to any unknown exposure
circumstance. Larger uncertainty factors would apply to KM because the data
from which it was estimated are more uncertain than those from which K. was
estimated. While it is not possible to estimate the 95 percent confidence
limit directly, a factor of 5 would appear to be reasonable for the average
value of KM and a factor of 20 on its application to any unknown exposure
circumstance.
The range of uncertainty may in fact be greater than that suggested.
While this 20-fold factor provides a range of 400 (i.e., estimates are divided
by 20 and multiplied by 20 to determine the range), the range could be greater
95
image:
yet. However, insufficient information exists by which to make any more
precise or definite estimate of uncertainty.
3.12 ASBESTOS CANCERS AT EXTRATHORACIC SITES
The consistency of an increased cancer risk and its magnitude, either in
absolute (observed-expected deaths) or relative (observed/expected deaths)
terms is less for cancer at other sites. Nevertheless, many studies document
significant cancer risks at various gastrointestinal (GI) sites. Cancer of
the kidney and urinary organs was also found to be significantly elevated in
two large studies (Selikoff et al., 1979; PuntolH et al., 1979). Among female
workers, ovarian cancer was found in excess (Newhouse et al., 1972; Wignall
and Fox, 1982; Acheson et al., 1982). While no other specific sites were
shown to be elevated at the 0.05 level of significance, the category of all
cancers other than the lung, GI tract, or mesothelioma is significantly
elevated (e.g., Selikoff et al., 1979).
Table 3-33 lists all studies in which more than 10 GI cancers were expec-
ted or observed and in which the overall lung cancer risk was elevated at the
0.05 level of significance. Some studies having statistically uncertain data
were eliminated from consideration, as were several larger studies demon-
strating a low risk of lung cancer because of exposure or follow-up circum-
stances. Because the excess risk of GI cancer is less than that of lung
cancer, significantly elevated risks are unlikely to be seen in studies that
demonstrate little risk of lung cancer; therefore, negative data in such
studies do not have much significance. In considering Table 3-33, note that
all but 3 of the 23 listed studies show an excess GI cancer risk, even though
the risk is small in several studies. However, 10 of the 23 studies demonstrate
risk at a 0.05 level of significance. Figure 3-11 displays the relationship
between the relative risk of lung cancer and relative risk of GI cancer in the
23 studies in Table 3-33. Figure 3-11 shows there is a consistent relationship
between increased GI cancer risk and increased lung cancer risk. Fiber exposure
to the GI tract is probable because the majority of fibers inhaled are brought
up from the respiratory tract and swallowed (Morgan et al. , 1975), and some
may become entrapped within the gut wall (Storeygard and Brown, 1977).
Additionally, fibers may be swallowed directly. Nevertheless, the magnitude
96
image:
TABLE 3-33. OBSERVED AND EXPECTED DEATHS FROM VARIOUS CHUSES IN SELECTED MORTALITY STUDIES
Respiratory cancer
ICD 162-164
1.
2.
3.
4.
5.
6.
7.
8.
9.
10-
11-
12.
13.
14,
15.
16,
17.
IB,
19.
20.
21.
22.
23.
Henderson and Enterline (1979)
McDonald et al . (1980)
Newhouse and Berry (197g) (male)
NewhOLse and Berry (197S) (female)
Selikcff et al. (1979) (NY-NJ)
Selikoff et al. (1979) (U.S.)
Nicholson eL al. (1979)
Peto (1977)
Mancuso and El-attar (1967)
Punloni et al. (1979)
Seidmar et al . (1979)
Dement et al. (1983b)
Jones el al . (1980)
McDonald et al. (1983a)
McDonald et al . (1984)b
Robinson et al. (1979)
Acheson et al. (1984)
Wignall & Fox (1982)
Mejrman et al. (1974)
Albin et al. (1984)
Elmes 4 Simpson (1977)
Nicholson (197€a)
Clenmesen & Hjalgrim-Jensen (1981)
a
63
230
103
-fr-
93a
390
25
51
30
123
83
33
12
59
73
49
57
10
21
U
24
27a
44
E
23.3
184,0
43.2
""'3.2
13.1
93.7
11.1
23.8
9.8
54.9
21.9
9.8
6.3
29,6
49.1
36.1
29.1
3.7
12.5
6.6
5
8-4
27.3
0-E
39.7
46.0
59, a
23.8
79.9
295.3
13.9
17.2
20.2
68.1
61.1
23.2
5.7
29.4
23.9
12.3
27,9
6,3
a, 4
5.4
19
18.6
16.7
Digestive cancer
ICD 150-159
0
55
276
40
20
43a
89
10
16
15
94
28
10
10
26
59
50
19 -
7
7
19
13
13a
31
E
39.9
272.4
34.0
10.2
14.8
53.2
9,5
15 7'
7 1
76.6
22.7
8.1
20.3
17.1
51,6
41.4
17.1
1C. 7.
H.9
1(1.8
5.0
J'S.9
0-E
15.1
3.6
6.0
9.8
28.2
35. 8
0.5
0.3
7.9
17.4
5.3
1,9
(10.3)
a. 9
7.4
8.6
1.9
(3-7)
C.9)
8.2
12
8.0
1.1
(OT)r
0.380
0.078
0.100
0.412
0.353
0,121
0.036
0.319
0.527
0.255
0.087
O.OB2
def.
0.302
0.309
0.667
0.068
def.
def.
1.519
0.632
0.430
0.066
Other cancers
ICD except 150-59, 162-4, ireso
0
55
237
38
33
28a
184
14
18
20
88
39
11
35
35
70
69
33
35
21
10
17a
89
E
45,5
217.4
27.4
20.4
24 5
131,8
16.1
24.8
6.8
81.3
35.9
14.1
39.5
27.7
60,4
51,2
28,2
21,6
no
20.4
no
14.4
93.9
0-E
9.4
19.6
10.6
12.6
3.5
52.2
(2.1)
(6.8)
13.2
6.7
3.1
(3.1)
(4.5)
7,4
9.6
17.8
4.8
13.4
data
0,6
data
2.6
(4,9)
%$
0.237
0.426
0.177
0.529
0.044
0.176
def.
def.
0.653
0.098
0-037
def.
def,
0-252
0.402
0,380
0.172
2.127
0.111
0.140
def
0 = observed deaths.
E = expected deaths.
d = digestive cancer.
r = respiratory cancer,
o = other cancer.
ICD = International Classification of Diseases.
def. = no ratio when deficient in 0-E.
*Best estimate data on causes of death.
Excess risk may not be asbestos-related; see Section 3.9.6.
97.
image:
cc
01
o
<
o
2
o
oc
u.
V)
LU
Q
Q
LU
LU
Q.
X
01
\
O
LU
>
QC
LU
V)
to
O
18
13
19
23 •
•41-
14
21
22
20
0.5
1.5
2.0
2.5
3.0
-$s-
13
OBSERVED/EXPECTED DEATHS
FROM GASTROINTESTINAL CANCER
Figure 3-11. The ratio of observed to expected mortality from lung cancer
versus the ratio of observed to expected mortality from gastrointestinal
cancer.
Source; Table 3-33, reference numbers 1 through 23.
image:
of the excess at GI sites is much less than for the lung. In recent studies,
the GI excess is about 10-30 percent of the lung excess.
The number of studies demonstrating a statistically significant excess
risk of gastrointestinal cancer in asbestos-exposed groups and the correlation
of the relative risk of gastrointestinal with the relative risk of lung cancer
are highly suggestive of a causal relationship between asbestos exposure and
gastrointestinal cancer. However, alternative interpretations of the above
data are possible. Doll and Peto (1985) have suggested that many of the
excess cancers attributed to gastrointestinal sites may be misdiagnosed lung
cancers or mesotheliomas. They also cite the absence of confirmatory animal
data showing a risk of cancer at extrapulmonary sites as weighing against a
causal relationship. However, it is difficult to accept that all excess
gastrointestinal cancers are the result of misdiagnosis. While cancers of
some of the gastrointestinal sites, particularly the pancreas and the stomach
to some extent, are often misdiagnosed jnesothellomas, cancers of the colon and
rectum are usually correctly certified and the excesses at these sites across
studies are unlikely to be the result of misdiagnosis.
The U. S. , Environmental Protection Agency Cancer Assessment Group has
reviewed studies with GI cancer excess, They have concluded that the associa-
tion between GI cancer excess and asbestos exposure is strong.
Table 3-33 also lists the observed and expected mortality for cancers
other than mesothelioma, the GI, or respiratory tract. The elevation is not
as consistent as for GI cancer. Only six studies have elevated risks that are
significant at a 0,05 level, and deficits are observed in five. The analysis
is further complicated by the possibility that misattribution of lung cancer
or mesothelioma may have occurred for some cases. For example, brain or liver
cancers could be metastatic lung cancers in which the primary site was not
properly identified. In the study of insulation workers, Selikoff et al.
(1979) found that 26 of 49 pancreatic cancers were misclassified; most of
those misclassified were peritoneal mesotheliomas. The excess at other sites
is much less than lung cancer and roughly similar to that of GI cancer.
3.13 ASBESTOSIS
Asbestosis, a Icuig-term disease entity resulting from the inhalation of
asbestos fibers, is a chronic, progressive pneumoconiosis. It is character-
ized by fibrosis of the lung parenchyma, usually radiologically evident only
99
image:
after ten years from first exposure, although changes can occur earlier follow-
ing more severe exposures. Shortness of breath is the primary symptom, cough
is less common, and signs such as rales, finger clubbing, and weight loss in
later stages of the disease appear in a proportion of cases. The disease was
first reported eight decades ago (Murray, 1907) and has occurred frequently
among workers occupationally exposed to the fiber in ensuing years. Charac-
teristic X-ray changes are small irregular opacities, usually in the lower and
middle lung fields, often accompanied by evidence of pleural fibrosis or
thickening, and/or pleural calcification, Both the visceral and, more commonly,
the parietal pleura may be involved.
Currently, 50-80 percent of individuals in groups with heavy occupational
exposures beginning more than 20 years earlier are found to have abnormal
X-rays. These include asbestos insulation workers (Selikoff et al., 1965),
miners and millers (Nicholson, 1976b), and asbestos factory employees
(Lewinsohn, 1972). In many circumstances, fibrosis progresses following
cessation of exposure. The prevalence of abnormal X-rays is much less in
groups exposed to lesser quantities of asbestos, such as shipyard or con-
struction workers or workers exposed recently. Berry et al. (1979) have
analyzed the development of clinical and x-ray signs of asbestosis according
to accumulated exposure among workers of the Rochdale factory studied by Doll
and Peto and others (see Section 3.9.3). The results suggest that the risk of
developing possible asbestosis is less than 1 percent from an exposure to 0.7
f/ml for forty years. However, these results must be interpreted cautiously
because all individuals studied began work with asbestos after 1950. The
possibility of an increasing prevalence of abnormalities with progression of
time, even with no further exposure, must be considered.
The British Occupational Hygiene Society (1983) evaluated the clinical,
physiological, and X-ray findings among groups of workers exposed in two
factories in Great Britain. From an analysis of the data they conclude that
the probability of developing any one of seven pulmonary or radiographic
abnormalities associated with asbestos exposure is less than 2 percent at
cumulative exposures of 25 f-y/ml. As with Berry's analysis, the progression
of abnormalities with time must be considered. Findings of abnormal X-rays,
predominantly of the pleura, among family contacts of asbestos workers
(Anderson and Selikoff, 1979) suggest that radiographic stigmata of asbestos
exposure may occur at very low exposures if a long enough time elapses between
100
image:
the exposure and the observation. The significance of pleural X-ray abnormal-
ities is uncertain. They may or may not be associated with deficits in pul-
monary function, and no information exists on whether the presence of pleural
plaques or pleural thickening implies a greater risk of cancer separate from
that associated with cumulative asbestos exposure,
Liddell and McDonald (1980) have correlated cause-specific mortality,
1951-1975, with the readings of the last available employment X-ray of a group
of Canadian miners and millers. They found that significantly increased risks
of death from pneumoccniosis, pulmonary TB, lung cancer, "other" respiratory
disease, and diseases of the heart were associated with a previous abnormal
X-ray. However, increased lung cancer risks were also found among individuals
with no detected parenchymal fibrosis, but who may have had pleural abnormal-
ities. Again, unknown progression of fibrosis could have occurred between the
last reading and death.
In addition to disease and disablement during life, asbestosis has ac-
counted for a large proportion of deaths among workers in some occupational
groups. The first reports of the disease (Auribault, 1906; Murray, 1907)
described complete eradication of workers in textile carding rooms. Much
improvement in dust control has taken place in the industry since the turn of
the century, but even recently those exposed to extremely dusty environments,
such as textile mills, may have as much as 40 percent of their deaths attribu-
table to this cause (Nicholson, 1976a). Groups with lesser exposures for 20
or more years, such as in mining and milling (Nicholson, 1976b) or insulation
work CSelikoff et al., 1979) may have 5 to 20 percent of their deaths attributed
to pneumoconiosis. All varieties of asbestos appear equally.capable of produc-
ing asbestosis (Irwig et al., 1979). In groups exposed at lower concentrations,
such as the families of workers, death from asbestosis has not been reported.
It is not clear what the dose-response relationship is for the most
minimal manifestations of asbestos exposure, such as a pleural or diaphragmatic
plaque or unilateral pleural thickening. The possibility exists that such
abnormalities may develop in some individuals long after exposure to very low
doses of asbestos (1-10 f-y/ml, for example.) This is suggested by the finding
of significant percentages of such abnormalities among family contacts of
asbestos workers. However, these x-ray abnormalities are unlikely to be
associated with any discernible pulmonary function deficit in individuals
exposed to less than 10 f-y/ml. At such exposures, the primary risk considera-
tion is cancer rather than non-malignant disease,
101
image:
3.14 MANIFESTATIONS OF OTHER OCCUPATIONAL EXPOSURES TO ASBESTOS
In the past decade, considerable evidence was developed on the prevalence
of asbestos disease in workers exposed to a variety of work activities. Workers
in shipyard trades (other than insulation work), in particular, were shown to
have had significant exposure. By 1975, Harries (1976) identified 55 mesothe-
liomas in the Devonport Dockyard, only two of which were in asbestos workers.
In a case-control study of four Atlantic Coast areas, an average relative risk
for lung cancer of 1.4 was determined (Blot et al., 1978). The study population
had an average employment time of only three years and no exposure data are
available. X-ray abnormalities among non-insulator shipyard employees are
also common. Among long-term (mostly 30+ year) shipyard workers, 86 percent
have X-ray abnormalities characteristic of asbestos exposure (Selikoff et al.,
1980). Maintenance personnel are also at risk from asbestos disease. Lilis
et al. (1979) reported finding X-ray abnormalities among 55 percent of 20+
year chemical plant workers.
These findings are important because they point to sources of environmen-
tal asbestos emissions in the future. Removal of asbestos from friable pro-
ducts, including insulation material, and installation of engineering controls
in factories have significantly reduced exposure and emissions from primary
manufacturing or new construction work. However, more than one million tons
of asbestos are in place as friable materials in ships, buildings, power
plants, chemical plants, refineries, and other locations of high temperature
equipment (Nicholson, 1976a). Maintenance, repair, and removal of this material
will continue to be an important source of future exposure to workers and of
emissions into the environment (both inside and outside buildings).
3.15 DEPOSITION AND CLEARANCE
Considerable data are available on the quantity of asbestos fibers in
lungs of individuals?with and without known exposures to asbestos (Sebastien
et al., 1979; Jones et al., 1980; Wagner et al., 1982). Most of the cases
analyzed were selected because of death from mesothelioma, often coupled with
an investigation of a specific work group (Wagner et al., 1982; Berry and
Newhouse, 1983). However, they have not been correlated with known cumulative
exposures. Generally, amphibole burdens of heavily exposed individuals range
7 8
from 10 to 10 fibers/gram dry weight; general population controls (in Great
102
image:
Britain) are usually less than 10 fibers/gram dry weight (Jones et al.,
1980). Similar concentrations of chrysotile are seen in exposed workers
(Wagner et al., 1982) and unexposed controls (Jones et al., 1980).
Very few data are available that provide a basis for establishing a model
for the deposition and clearance of fibers in humans. It is expected that
both short- and long-term clearance mechanisms exist in humans, as in animals
(see Chapter 4), If only long-term processes are considered (characterized by
months or years) the simplest model is one in which the change in lung burden
(N) is proportional to the rate of deposition of fibers (A) (assuming continuous
exposure-) diminished by a clearance that is proportional (by factor p) to the
number of fibers present.
^ = A - pN (3-7a)
This yields for the number of fibers present after a constant exposure of
duration, t,,
N = A-d-e^l) (3-7b)
and at a time, t? after cessation of a constant exposure of duration t..
N = -(l-e~ptl)e~pt2 (3-7c)
P
Such a model is applicable at times t.. and t? which are long compared to
any short-term clearance mechanisms. It is clearly a very simplistic model in
that it considers only one characteristic time for long-term removal pro-
cesses. Nevertheless, it illustrates the difficulty of. applying even the
simplest model. In order to systematize lung burdens, information is needed
on the duration and intensity of the exposure and the time from last exposure
in order to obtain a measure of the characteristic removal time for a given
fiber type. Such information is not yet available for the individuals whose
lungs have been analyzed.
Data have been presented by Bignon et al. (1978) on the number of amphi-
bole fibers detected in lung washings of seven asbestos insulation workers.
All were exposed between 10 and 16 years. While individual exposures are
103
image:
unknown, fewer fibers were found in the washings of those longest removed from
exposure. The data are consistent with a decrease of 50 percent in the number
of washable fibers at five to seven years after cessation of exposure. However,
it is noted that washable fibers may not be proportional to the residual lung
burden or to the number of fibers trapped within lung tissue. The lung wash-
ings were largely amphibole; no corresponding data are available for chrysotile
fibers.
Data or> the fiber dimensions from these studies show a decrease in the
average length and diameter of fibers found in the pleura compared with those
found in the parenchyma. However, no distinction was made between amphiboles
and chrysotile in this analysis, and the different length-width data could
simply be a reflection of the predominance of chrysotile in the pleura.
3-15.1 Models of Deposition and Clearance
The Task Group on Lung Dynamics of the International Commission on Radio-
logical Protection proposed a model for the deposition and retention of parti-
cles (see Brain and Valberg, 1974). The results of this model are shown in
Figure 3-12, which depicts the percentages of particles of different sizes
deposited in the various compartments of the respiratory tract. Figure 3-12
shows that alveolar deposition is dominant for particles with a mass median
diameter less than 0.1 urn. As the particle size increases, deposition in this
area decreases, falling to 25 percent at 1 urn and to 0 at 10 pm or above.
Nasal and pharyngeal surface deposition becomes important above 1 |jm and rises
rapidly to be the dominant deposition site for particles 10 urn in diameter or
greater. This model was developed for spherical particles. Timbrel 1 (1965)
has shown that the settling velocities of particles, and their aerodynamics,
are such that fibers with aspect ratios greater than three behave like particles
with a diameter three times as great, independent of the length of the fiber.
This was corroborated by calculations of Harris and Fraser (1976). Thus, few
fibers with diameters as large as 2 urn are likely to penetrate into the alveolar
spaces, although finer fibers, even as long as 200 urn, may do so.
3.16 EFFECTS OF INTERMITTENT VERSUS CONTINUOUS EXPOSURES
Two distinct kinds of exposure occurred to workers in the different
studies reviewed. In some production operations (e.g., textiles), workers are
104
image:
V)
2
LU
o
BRONCHO-ALVEOLAR
NASO-PHARYNGEAL
TRACHEO
BRONCHIAL
0.01
0.05 0.1 0,5 1.0 5 10
MASS MEDIAN DIAMETER, urn
Figure 3-12. Aerosol deposition in the respiratory
tract. Tidal volume is 1,450 ml; frequency, 15 breaths
per minute. Variability introduced by change of sigma,
geometric standard deviation, from 1.2 to 4.5. Particle
size equals diameter of mass median size.
Source: Brain and Valberg (1974).
105
image:
exposed to a relatively constant concentration of asbestos fiber throughout
their work day; in other production operations (e.g., insulation, maintenance,
and some production), workers are exposed to extremely variable concentrations
of asbestos, with most of their cumulative exposure resulting from short
duration, but intense, exposures. Implicit in the use of a linear dose-response
relationship and average exposures is the concept that the risk of cancer is
directly related to the cumulative asbestos exposure received in a period of
time, i.e., the effect of an exposure to 100 f/ml for 1 hour is the same as
that of 1 f/ml for 100 hours. (This equivalence applies only for short time
periods. Because of the time dependence of mesothelioma risk, 100 f/ml for
one year is not equivalent to 2 f/ml for 50 years.) Short, intense exposures
could have an effect different from longer and lower exposures if clearance
mechanisms are altered by very high concentrations of inspired asbestos.
Although there are no data that directly address this question, indirect
information suggests that the magnitude of the effect is less than the variabil-
ity between studies with continuous exposure. Henderson and Enterline (1979)
found that the excess lung cancer risk for plant-wide maintenance mechanics
was only slightly higher (21 percent) than that for production workers, on a
unit exposure basis. Curiously, the risk of pneumoconiosis was much less per
unit of cumulative exposure among maintenance workers. The similarity of unit
exposure risks of insulation workers compared to groups having continuous
exposure suggests that the character of their exposure is not important. How-
ever, both comparisons depend upon the exposure estimates of the groups in
question. Clearly, average exposures are difficult to estimate from episodic
exposures and the above numerical similarities may be fortuitous. The unusu-
ally low pneumoconiosis risk among mechanics in the Henderson and Enterline
(1979) study may be the result of exposure misestimates.
3.17 RELATIVE CARCINOGENICITY OF DIFFERENT ASBESTOS VARIETIES
Whether there is -a different carcinogenic response according to fiber
type or industrial process is an issue of increasing concern in the under-
standing of asbestos disease. Considerable controversy has developed as to
whether one variety of asbestos (crocidolite) or mineral class (amphibole) is
more carcinogenic than another (the serpentine mineral, chrysotile). Great
106
image:
Britain, Canada, and Sweden have imposed far more rigid standards for crocido-
lite than other varieties of asbestos. In contrast, the United States has no
special standard for any specific asbestos mineral.
Prior to the late 1960s the question was moot, because most epidemio-
logical studies did not accurately characterize the asbestos fiber types used
and measurements were not made of fiber concentration by mineral species.
Most measurements only characterized the total quantity of dust in the aerosol
(in terms of millions of particles per cubic foot) rather than in terms of
fiber concentration. This lack of information on fiber exposure by mineral
type was recognized at the time of the 1964 New York Academy of Sciences
Conference on Asbestos (Whipple and van Reyen, 1965), and a recommendation was
made that the importance of fiber type on the risk of developing asbestosis,
carcinoma of the lung, and mesothelial tumors be investigated. In the ensuing
years, considerable information was developed on the mortality experience of
different groups exposed to different varieties of asbestos in different work
processes. Unfortunately, the differential unit exposure risk associated with
different fiber types is still not completely understood.
3,17.1 Lung Cancer
3.17.1.1 Occupational Studies. Figure 3-7, Table 3-28 and Table 3-10 summar-
ize the information available on the unit expoisure risk for lung cancer in 14
different epidemiological studies. The range of the fractional increase in
lung cancer per unit asbestos exposure, expressed in terms of f-y/ml, varies
by more than two orders of magnitude. What is unique about this variation is
that exposures to a single fiber type yield results that differ by nearly
100-fold. One of the highest unit exposure risks was found in a textile plant
that used only chrysotile asbestos (Dement et al., 1983b; McDonald et al.,
1983a) and the lowest values were found in a large study of chrysotile mine
and mill employees (McDonald et al., 1980) and in groups exposed only to
chrysotile asbestos in friction products manufacturing (Berry and Newhouse,
1983; McDonald et al. , 1984). Similarly, large (10-fold) differences are
found in studies ostensibly of the same process, using the same mix and quality
of asbestos fibers in different plants of the same company. A study of asbestos
cement manufacturing shows one of the highest unit exposure risks (Finkelstein,
1983). Another study (Wei 11 et al., 1979) suggests a risk more than 1/10 as
much, while a 10-fold difference in risk appears to exist in two groups working
at different periods in a single British Textile facility (Peto, 1980).
107
image:
: There is only one study in which the exposure was solely to amosite
asbestos (Seidman, 1984), and the risk was comparable to the risk found in
chrysotile textile operations. However, in several groups exposed to a mixture
of chrysotile, amosite, and crocidolite in insulation work (Selikoff et al.}
1979), the risk was less than that experienced by either chrysotile textile
manufacturers or amosite factory workers.
No data exist, in any study, of unit exposure risks to workers exposed
solely to crocidolite asbestos. Enterline and Henderson (1973) and Wei 11
et al. (19.79) suggest that workers exposed to chrysotile and crocidolite may
have a greater lung cancer risk than those exposed to chrysotile alone, perhaps
by a factor of two. However, this suggestion is based on air concentrations
of total particles in the respective work environments (including much other
dust) and a significant amount of crocidolite could also have been present
without affecting the total particle count.
The wide divergence of risks according to fiber type, and even among
similar work processes, suggests that factors other than mineral type substan-
tially influenced the studies reviewed. These other factors could include
errors in the estimation of exposures that occurred decades previously, biases
or other limitations in epidemiological studies describing the disease experi-
ence, and statistical uncertainties associated with a limited number of deaths.
While the above factors undoubtedly contribute to some of the observed
variability in Figure 3-7, certain consistent differences are likely to be
real. Chrysotile textile production imparts a significantly higher risk per
fiber exposure than chrysotile mining or friction products manufacturing. The
data supporting this suggestion are very convincing for mining versus textiles.
They are less convincing for friction products versus textiles because of
greater uncertainties in the mortality experience of friction product workers
and estimates of their fiber exposure.
McDonald et al. (1984) and others suggested that differences in risk may
be caused by differences in fiber size and dynamics of penetration. As chryso-
tile is processed, the percentage of long curly fibers (which are easily
counted but not easily inspired) decreases and the percentage of shorter,
straighter, and narrower fibers increases.
3.17.1.2 Environmenta1 Exposures. Data on the risk of lung cancer by fiber
type from non-occupational exposures to asbestos are extremely scarce.
Siemiatycki (1982) reported on the mortality experience of the general popula-
tion of Asbestos and Thetford Mines, Quebec. These two areas account for the
108
image:
great preponderance of chrysotile mining in Canada. The female population in
these towns has experienced substantial exposure compared to that of individuals
in non-mining areas. Data from Gibbs et al. (1980) indicate that recent town
2
air concentrations range from 170 to 3500 ng/m , Additionally, home exposures
to the wives of workers in the plant also occurred. Table 3-34 lists the mor-
tality experience for selected causes among the female population of Asbestos
and Thetford Mines during the years 1966-1977. The observed mortality was
compared to the mortality experience of the entire Province of Quebec. There
is no statistically significant excess of lung cancer among the mining popula-
tion females compared to that expected. However,, the use of the entire Province
of Quebec as the reference population appears to be inappropriate, although
the degree of inappropriateness is difficult to ascertain. Lung cancer rates
in rural areas are considerably lower than those of urban centers. McDonald
et al. (1971) stated that the lung cancer rate for males in the counties
surrounding the mining area is two-thirds that of the Province as a whole.
Table 3-20 gives the regional lung cancer incidence rates in Quebec Province
for males and females for the years 1969-1973. The rate for males in rural
counties is 73 percent of the rate in the Province, in agreement with McDonald
et al. (1971); however, the relative rates for rural females is even lower, 62
percent of the Provincial rate. Thus, a female lung cancer relative risk of
1.06 compared to Quebec Province translates into a 70 percent increase compared
to all of Quebec except Montreal and Quebec City.
TABLE 3-34. MORTALITY FROM SELECTED CAUSES IN ASBESTOS AND THETFORD MINES
COMPARED TO QUEBEC PROVINCE, FEMALES, 1966-77.
Cau.se
All causes
All cancers
Digestive cancer
Respiratory cancer
Other respiratory
diseases
0
1130
289
117
23
30
E
1274.6
318.1
110.7
21.5
51.8
0-E
-144.6
-29.1
6.3
1.5
-21.8
L.C.L.a
0.84
0.81
0.88
0.68
0.39
0/E
0.89
0.91
1.06
1.07
0.58
U.C.L.9
0.94
1.02
1.28
1.61
0.83
95-percent confidence limits.
Source: Siemiatycki (1982).
109
image:
This increase is also compatible with data published by Wigle (1977) on
cancer mortality in relation to asbestos in municipal water supplies. He
compared the cancer risk, by site, for Asbestos and Thetford Mines with nearby
communities having moderate concentrations of asbestos in their water siupply,
and with various other communities throughout the Province of Quebec, including
some in populated and industrial areas. The relative cancer risk for females
was 1.3 for Asbestos and Thetford Mines, 0.7 for five nearby towns, and 0.8
for other communities (some urban or industrial).
The increases indicated by the adjusted relative risks in Siemiatycki's
(1982) study and those indicated by Wigle1s (1977) data are both statistically
significant. However, these data are only indicative and do not demonstrate
an increased lung cancer risk due to environmental asbestos exposure, because
the effect of confounding variables was not explored. Nevertheless, the data
show that population comparisons between residents of Asbestos and Thetford
Mines J other regions of Quebec cannot be used to indicate the absence of a
risk.
3.17.2 Mesothelioma
3.17.2,1 Occupational Exposures. Table 3-31 lists values characterizing the
risk of death from mesothelioma and lung cancer per f-y/ml in four studies,
along with cruder estimates of the mesothelioma risk compared to that of lung
cancer in 14 studies. One noticeable feature among all studies is that the
ratios of the unit exposure risks of mesothelioma and lung cancer are very
similar, irrespective of the type of exposure experienced. Thus, it appears
that the same factors affect the variability of mesothelioma risk as affect
lung cancer risk, and that mesothelioma risk can be estimated from values of
K, and an average ratio of K../K,. Again, it appears impossible to separate
the effect of mineral type from other factors contributing to the variability
of potency.
In order to make a broader comparison of mesothelioma according to expo-
sure by mineral type, the risk of pleural and peritoneal mesothelioma can be
compared with that of lung cancer in a variety of studies. Because the asbes-
tos risk of lung cancer is directly proportional to the underlying risk of
lung cancer, the comparisons are most appropriately made to a lung cancer risk
that is standardized to a similar background. In particular, one would expect
the ratio of mesothelioma to excess lung cancer among women to be several
110
image:
TABLE 3-35. RISK OF DEATH FROM MESOTHELIOHA AS A PERCENTAGE OF EXCESS LUNG CANCER. ACCORDING TO FIBER EXPOSURE
Study and fiber type
Chrysotlle
Acheson et al. (19BZ)
Dement et al. (1983a,b)«
McDonald et al, (1983a)
McDonald et al. (1980)
Nicholson et al. (1979)*
McDonald et al. (1964)
Rub 1 no et al. (1979)
Weiss (1977)
Totals (excluding * studies)
ToUTs (adj. for additional cases)
Predominantly chrysotile (>98*)
McDonald et al, (19S3b)
Robinson et al. (1979)
Robinson et al. (1979)
Hancuso & El-attar (1967)
Peto (1980)
Thous et al. (1982)
Obs.
0
6
33
59
230
25
73
9
4
53
49
14
33
30
22
Exp. Lung Cancer
E D-E Adj.
4.5
9.B
29.6
104.0
11.1
49.1
8.7
4.3
50.5
36.1
1.7
14.8
15.5
25.8
1.5
23.2
29.4
46.0
13.9
Z3.9
D.3
-0.3
2.5
12.9
12.3
18.2
14.5
-3.8
5.
18.
15.
126.
17.
24.
0.
-0.
147.
187
18.
28.
123.
28.
12.
-3.
5
5
4
2 (166)a
2 b
8 (0.0)"
3
3
1
0
4
0 (20)C
3
0
8
MesothelloHa
PI. Per. Tot.
1
fl
0
10(20+)"
1 b
0(3)b
1
0
12
25
10
4
1
1
7
2
0
1
1
0
0
0
0
0
1
1
4
s
1
a
0
0
i
1
i
10(20+)
1
0(3)
1
0
13
26
14
13
4
9
7
2
Mesothellona as a % of
excess of lung cancer
Pl./O-E Per./O-E Tol./O-E
18.2
0.0
0.0
7.9(12.0+)
5.8
0.0(very
high)
333.3
0.0
8.2
13.4+
55.6
14.1
5.0
35.3
58.3
--
0.0
5.4
6.5
0.0
0.0
0.0
0.0
0.0
0.7
0.5
22.2
17.6
5.0
28. 3
0.0
--
18.2
5.4
6.5
7.9(12.0+)
5.8
0. 0(very
nigh)
333.3
0.0
8.8
14.0+
77.8
45.8
20.0
31.8
58.3
--
Totals (so«e unknown
dupl(cations of deaths)
Anosite
Acheson et al. (1984)
Seidwin et al. (1979)
Totals
Predominantly crocldoHte
Acheson et al. (1982)
Hobbs et al. (1980)
Jones et al. (1980)
Wignall & Fox (1982)
McDonald & McDonald (1978)
57
83
13
60
12
10
7
29.1
21.9
6.6
38.2
6.3
3.7
2.4
27.9
61.1
6.4
21.8
5.7
6.3
4.6
102.9
25.4
61.1
86.5
24.0
21.8
21.0
23.2
16.8
25
4
7
11
3
17
13
9
3
18 49
5
14
19
5
17
17
12
9
24.3
15.7
11.5
12.7
12.5
78.0
61.9
38.8
17,9
17.5
3.9
11.5
8.3
0.0
19.0
12.9
35.7
47.6
19.7
Z2.9
9.2 22.0
20.8
78.0
81.0
57.7
53.6
Totals
1 106.8
45
13
68
42.1
12.2 63.7
image:
TABLE 3-35. (continued)
Study and fiber type
Anthophy 1 1 1 te
Heuntan et al. (1974)
Totals
Talc (Treaollte)
Kleinfeld et al. (1974)
Brown et al. (1979)
Totals
HUed exposures
Albin et al. (1984)
Berry A Henhouse (1983) (M)
Berry & Mewhou.se (1983) (F)
Cleuwsen & Hjalgrlv- Jensen (1981)
Elaes ft. Sfapson (1977)
Flnkelsteln (1983)
Henderson ft Enterllne ((1979)
Selikoff et al. (1979) (US)
SeHkolf et al. (1979) (HY-HJ)
Kleinfeld et al. (1967)
Koloiwl ei al. (1980)
Henhouse & Berry (1979) (M)
Newhouse & Berry (1979) (F)
Nicholson (1976a)
Puntonl et al. (1979)
floss Her 4 Coles (1980)*
Welll (1984)
Totals (except * study)
Obs.
0
21
13
9
12
143
6
47
27
20
63
390
93e
10
13
103
27
27*
123
84
188
fxp.
E
12.6
4.5
3.3
6.6
139.5
11.3
27.3
S.O
3.3
23.3
93.7
13.1
1.4
7.5
43.2
3.2
8.4
54.9
100.3
128.0
Lung Cancer
0-E Adj.
8.4
8.5
5.7
5.4
3.5
-5.3
19.7
22.0
16.7
39.7
296.3
79.9
8.6
5.5
58.8
23.8
IB. 6
68.1
-16.3
60.0
13
13,
16.
8,
24
12.
3.
-5.
26.
59.
15.
59.
259
106
14.
7.
69,
100
22.
79.
-16.
79.
892.
.4
.4
1
.6
7
2
5
3
2
4
9
6
4
3
1
6
]
3
5
2
Hesothelln
PI. Per.
0
0
0
0
0
4
8
2
3 ,
8(19)d
6
5
61
11
1
10
19
13
ft
0
—
8
168
0
0
1
0
1
0
0
0
0
5
5
0
109
27
2
0
27
8
T
0
—
1
191
•a
Tot.
0
a
i
i
2
4
B
2
3
24
11
170
170
38
3
46
46
n
0
0
--
9
359
Nesothellou as a X of
excess of lung cancer
PI. /0-E Per. /0-E Tot. /0-E
0.0
0.0
0.0
0.0
0.0
32.8
514.3
--
11.5
32.0
37.7
8.6
23.6
10.4
G.9
0.0
27.5
13.0
35.4
0.0
—
7.5
IB. 8
0.0
0.0
6.2
0.0
4.0
0.0
0.0
—
0.0
8.4
31.4
44.1
42.1
25.5
13.9
39.1
39.1
B.O
0.0
0.0
—
6.3
21.4
0.0
0.0
6.2
11.6
8.1
32. B
514.3
—
11.5
40.4
69.2
84
65.6
35.8
20.8
0.0
66.6
21.0
66.4
0.0
..
13.8
40. Z
•One •esathelloaa death Is Included in a larger study of McDonald et al. (19BO).
"Subsequent to termination of the study, many additional cases of mesotheIiona developed. Four occurred In 1976 and 1977 {McDonald and llddell. 1979) and
sin were found in one mining area In 90 consecutive autopsies during 1981-83. (Churg el al., 1984). To account for some of this Increase, the additional 10
nesotheHana cases were Included and the adjusted excess lung cancer deaths increased by 40 to account for mortality over the 5 additional years- The effect
ol considering these additional cases Is illustrated by data in parentheses.
bHo mesotheliona cases were found in the cohort. However, three deaths fro™ mesothelforaa were Identified In the Connecticut Tumor Registry fr-oa the
plant (Tela el aT., 1983). Ihese are Included In parentheses for the purposes of this analysis. While a high lung cancer risk, was noted In the cohort, the
absence of a dose-response relationship made attribution of the cause difficult and no lung cancer deaths were attributed to asbestos exposure.
cThe adjusted excess lung cancer risk Is unrealisticaTly high. A value of 20 will be used.
Eleven deaths were either from pleura! mesothellona or lung cancer.
Best estimate data on the cause of death.
In this analysis, all were considered nesolheliona.
image:
times higher than among men because of the greater background risk of lung
cancer among men. Table 3-35 lists the various studies from Table 3-2. In
each study, an attempt was made to estimate an excess lung cancer risk that
would have occurred if the U.S. male rates in 1970 had prevailed for the study
population. For example, the standardized number of deaths in women was calcu-
lated by multiplying the number of observed deaths minus the expected number
of deaths by the ratio of the age standardized male to female lung cancer
rate. Similar adjustments were made to the excess number of lung cancers of
cohorts followed for long periods of time, that would have had an average time
of death earlier than 1970. Adjustments were also made where the lung cancer
rates of other nations differed from those in the United Stages. The last two
adjustments led to only minor changes in most cohorts, while the adjustment
for gender was substantial and uncertain because of absence of information
about the smoking habits of the study group. Finally, adjustments to local
rates were made similar to those in Section 3.9. After all the adjustments
were made, the ratio of mesothelioma was calculated by type of fiber exposure
as a percentage of adjusted excess lung cancer, The results were summed and
the combined data for specific mineral exposures were obtained.
There are several limitations ta consider when reviewing these data.
Because of possible bias caused by underdiagnosis of peritoneal mesothelioma
in many cohorts, the principal focus should be on the ratios of pleura! meso-
thelioma to adjusted excess lung cancer. Tissue specimens of all abdominal
tumors were reviewed in only a few studies (Selikoff et al., 1979; Seidman,
1984; Newhouse and Berry, 1979; Finkelstein, 1S83) to determine if peritoneal
mesothelioma had been misdiagnosed. Because of the ongoing review of mesothe-
liomas in Canada by the McDonalds (McDonald and McDonald, 1978; McDonald et
al., 1970, 1971), the study of Canadian miners and gas mask workers can also
be considered to have benefited from review. These studies account for 194 of
236 identified peritoneal mesotheliomas. Substantial bias may also exist
because of studies in which the tracing of the cohort is limited; in some
studies as many as 39 percent of the exposed individuals were untraced. The
inadequacy of tracing was particularly high in studies of workers exposed to
crocidolite. The danger is that mesotheliomas were identified in registries
because of their uniqueness, but that lung cancers in untraced individuals
were not. Thus, it is likely that there is a substantial overestimate of the
number of mesotheliomas relative to lung cancer associated with crocidolite
113
image:
exposures. Also, the comparison of the ratio of mesothelioma to excess lung
cancer is uncertain because of substantially different time courses for the
two diseases. The time course for "lung cancer is determined by the time
course of the underlying risk, which is usually the time course of lung cancer
from cigarette smoking. On the other hand, the time course for mesothelioma
is strictly dependent upon the time from onset of exposure, rising at about
the fourth or fifth power of time from first exposure. The analysis utilized
in Table 3-35 does not fully account for such differences.
In comparing the different ratios of pleural mesothelioma to adjusted
lung cancer for all studies in which the major exposure was to one fiber type,
the ratios for chrysotile, amosite, and mixed exposures are roughly compar-
able. Crocidolite exposures have a twofold to threefold greater number of
pleural mesotheliomas relative to excess adjusted lung cancer. However, as
noted previously, the untraced individuals in the various crocidolite cohorts
may have led to an overestimate of this ratio. The possibility of underdiag-
nosis of mesothelioma notwithstanding, the risk of peritoneal mesothelioma is
much lower with pure chrysotile exposures than with amphiboles or mixed expo-
sure. Only one peritoneal mesothelioma has been identified among more than
25 mesotheliomas in chrysotile-exposed populations. Though a greater mesothe-
lioma potency may be considered for crocidolite (a factor of two or four
considering both pleural and peritoneal sites), the effect of other factors in
a given exposure circumstance leads to much greater differences, as for example
in the case of lung cancer, where different exposure circumstances with the
same fiber lead to nearly 100-fold differences in unit exposure risk. A
similar situation exists with mesothelioma where the manufacture of amosite
insulation is associated with a high risk of mesothelioma (see Table 3-34),
while amosite mining demonstrates little excess risk (Webster, 1978; Solomons,
1984). Also, great differences in risk appear to exist between the crocidolite
mines of the Transvaal and those of the Cape Province. Thus, any suggestion
that there are dramatic differences between asbestos varieties has to be
considered in the light of greater differences that appear to be related to
processing, fiber size distribution effects within a single asbestos variety
(e.g., the difference between textiles and mining), and to differences between
cohort studies of the same exposure circumstances (e.g., the asbestos cement
studies of Weil! et al. (1979) and of Finkelstein (1982, 1983), or the two
cohorts of Peto (1980).
114
image:
There was no evidence in Table 3-10 of a substantial difference in lung
cancer unit exposure risk attributable to fiber type. While a pure amosite
exposure had a unit exposure risk about twice that of chrysotile exposures,
the combination of amosite or crocidolite with chrysotile in other exposure
circumstances demonstrated lower unit exposure risks. The data from Tables
3-31 and 3-35 indicate the crocidolite mesothelioma to lung cancer risk ratio
is no more than four times that of other fibers;, and when crocidolite is used
with other fibers, the combined ratio differs little from non-crocidolite
exposures. These findings suggest that crocidolite or amphibole exposures
cannot be the explanation of most mesotheliomas found in some predominantly
chrysotile exposure circumstances (e.g., Canadian mining and milling and
Rochdale, England textile production). This conclusion is further supported
by the observation that all the mesotheliomas in the above circumstances were
of the pleura, whereas amphibole exposure generally produces comparable numbers
of pleural and peritoneal mesotheliomas (the study of Hobbs et al. (1980) is a
remarkable exception). Finally, in the case of the Rochdale factory, the risk
of mesothelioma in a factory using only 2.6 percent crocidolite from 1932-1968
(Doll and Peto, 1985) was as high as the risk in the London factory studied by
Newhouse and Berry (1979) in which large amounts of crocidolite and amosite
were used.
A careful consideration of the role of amphiboles in the production of
mesothelioma is important for control of asbestos disease. On the one hand,
it would be a mistake to minimize or ignore the mesothelioma risk of chrysotile.
Millions of tons of this fiber presently are in building materials and other
products. The potential for release in future years is substantial unless
proper work practices and care are utilized during repair and .maintenance
work. On the other hand, it should be recognized that crocidolite, particu-
larly, is a very dangerous asbestos material. This comes from two aspects of
the fiber. One is the above-mentioned 2-4 fold greater risk of mesothelioma
relative to lung cancer found in crocidolite exposure circumstances. This
certainly indicates a greater unit exposure risk for mesothelioma relative to
other asbestos fibers. Secondly, the large percentage of thin fibers in a
crocidolite aerosol (which may contribute to increased risk mentioned above)
also may contribute to a greater fiber exposure when crocidolite-containing
products are manufactured or used because these very thin fibers remain aloft
for longer periods of time. Considering all factors, the proscription on the
115
image:
use of crocidolite in several countries would appear to be justified. Fortu-
nately, few pure crocidolite exposure circumstances exist in the United States.
Subject to their uncertainties, the average values of K, and K^ reflect the
most important processes where crocidolite is a constituent of the material
being produced. Nevertheless, if a pure crocidolite exposure is encountered,
a mesothelioma risk greater than that estimated using the average value of 1C,
is likely to exist and correspondingly greater precautions should be exercised.
3.17.2.2 Environmental Exposures. Mesothelioma has been documented in a
variety of non-occupational circumstances, including family contacts of asbes-
tos-exposed individuals. Table 3-36 lists observed family contact mesothe-
liomas associated with three occupational exposure circumstances and mesothe-
liomas identified in the contact worker group (the observation periods are not
quite commensurate). It is important to note that family contact cases are
seen with exposure to chrysotile, amosite and crocidolite. By fiber type,
there appears to be little difference in the family contact risk relative to
the risk at work.
TABLE 3-36. MESOTHELIOMA FROM FAMILY CONTACT
IN THREE OCCUPATIONAL CIRCUMSTANCES
Mesothelioma
Occupation
Miners and millers
Insulation manufacturers
Mixed products
Country
Canada
U.S.A.
U.K.
Fiber type
Chrysotile
Amosite
Mixed
Family
members
3a
4C
9e
Workers
12b
14d
67f
McDonald and McDonald (1980). dSeidman et al. (1979).
DMcDonald et al. (1980). eNewhouse and Thomson (1965).
cAnderson (1976). Newhouse and Berry (1979).
Animal studies support this conclusion and suggest that all varieties of
asbestos should be considered equally potent with respect to the production of
either lung cancer or mesothelioma in both inhalation and implantation studies.
As discussed previously, many risk differences may be accounted for by
differences in fiber size distributions in different work environments, rather
than by fiber type. The greatest percentage of longer and thicker fibers
116
image:
occurs in the work environment of miners and millers. Asbestos used in manu-
facturing processes is broken apart while it is incorporated into the finished
product. During application or removal of insulation products it is further
manipulated and the fibers become further reduced in length and diameter with
many falling within the range of significant carcinogenic potency (see Section
4-6). Because these shorter and thinner fibers can readily be carried to the
periphery of the lung where they penetrate the visceral pleura and lodge in
the visceral or parietal pleura, they may be of importance in the etiology of
mesothelioma. Bignon, Sebastien, and their colleagues (1978) reported data
from a study of lungs and pleura of shipyard workers. Larger fibers, often
amphibole, were found in lung tissue. In the pleura, the fibers were generally
chrysotile, but shorter and thinner. The early association of mesothelioma
with crocidolite occurred because, even in mining, crocidolite is readily
broken apart, yielding many fibers in a respirable and carcinogenic size
range, and has been extensively used in Great Britain in extremely dusty
environments (e.g., spray insulation), creating high exposures for many indi-
vidual.s, with a concomitant high risk of death from mesothelioma. Thus the
&
disease came tn' attention (Wagner et al., 1960). The mining and milling of
chrysotile, on uie other hand, involves exposures to long and curly fibers
which are easily counted but not easily inspired.
Recent exposures in Turkey to the fibrous zeolite mineral, erionite, have
been associated with mesothelioma. Results reported by Baris et al. (1979)
demonstrate an extraordinary risk; annual incidence rates of nearly 1 percent
exist for mesothelioma. In 1974, 11 of 18 deaths in Karain, Turkey were from
this cause. The fiber lengths are highly variable; most erionite fibers are
shorter than 5 urn and 75 percent are less than 0.25 urn.
3.18 SUMMARY
Data are available that allow unit risks to be determined for lung cancer
and mesothelioma. The values for K, , the fractional risk per f-y/ml, vary
widely among the studies, largely because of the statistical variability
associated with small numbers but also because of uncertainties associated
with methodology and exposure estimates. Based on an analysis of the unit
exposure risk for lung cancer and mesothelioma in 11 studies (all studies for
which unit exposure risks can be estimated except chrysotile mining and milling),
117
image:
_Q
the best estimate for K. is 0.010, and for KH it is 1.0 x 10 , An analysis
of variability suggests that the 95 percent confidence limit on the estimate
of KL is generally from 0,0040 to 0.027 (a factor of 2.5), but for KL in an
unknown exposure circumstance it is a factor of 10. A greater range of uncer-
tainty applies to the best estimate for the value of KM, the uncertainty in a
given exposure circumstance is also greater, perhaps by a factor of 20.
Differences in asbestos type cannot explain the variability of K, observed in
different studies. However, the lower risk values found in chrysotile mining
and milling compared with chrysotile textile production suggest that fiber
length and width distribution is important. The unit exposure mesothelioma
risk also differs greatly in different exposure circumstances, but the ratio
of mesothelioma risk to excess lung cancer risk is relatively constant.
Peritoneal mesothelioma has largely been associated with amphibole exposure,
although this is qualified by the possibility of underdiagnosis in some studies.
Pleural mesothelioma is associated with exposure to chrysotile and crocidolite;
while differences in pleural mesothelioma risk attributable to fiber type may
exist, they are much less than differences attributable to other factors.
118
image:
4. EXPERIMENTAL STUDIES
4.1 INTRODUCTION
Most animal studies of asbestos health effect;; have been used to confirm
and extend previously established human data rather than to predict human
disease. This situation exists because asbestos usage predates the use of
animal studies for ascertainment of risk; because some animal models are rela-
tively resistant to the human diseases of concern; and because lung cancer,
the principal carcinogenic risk from asbestos, 1s the result of a multlfactorlal
Interaction between causal agents, principally cigarette smoking and asbestos
exposure, and 1s difficult to elicit 1n a single exposure circumstance.
Although all of the asbestos-related malignancies were first identified in
humans, experimental animal studies confirmed the identification of the diseases
and provided important information, not available from human studies, on the
deposition, clearance, and retention of fibers, as well as cellular changes at
short times after exposure. Unfortunately, one of the most important questions
raised by human studies, that of the role of fiber- type and size, is only
partially answered by animal research. Injection and Implantation studies in
animals have shown longer and thinner fibers to be more carcinogenic once in
place at a potential site of cancer. However, the size dependence of the
movement of fibers to mesothelial and other tissues is not fully elucidated,
and the questions raised by human studies concerning the relative carcino-
genidty of different asbestos varieties still remain.
4.2 FIBER DEPOSITION AND CLEARANCE
Deposition and clearance of fibers from the respiratory tract of rats
were studied directly by Morgan and his colleagues (Morgan et al., 1975; Evans
et al., 1973) using radioactive asbestos samples. Following 30-minute inhala-
tion exposures in a nose breathing apparatus, deposition and clearance from
the respiratory tract were followed. The distribution of fibers 1n various
organ systems was determined at the conclusion of Inhalation, showing that
31-68 percent of Inspired fibrous material is deposited in the respiratory
tract. The distribution of that deposited material is shown in Table 4-1.
Rapid clearance, primarily from the upper respiratory tract (airways above the
119
image:
TABLE 4-1. DISTRIBUTION OF FIBER AT THE TERMINATION OF 30-MINUTE INHALATION
EXPOSURES IN RATS (PERCENT OF TOTAL DEPOSITED)
Fiber
Chrysotile A
Chrysotlle B
Amoslte
Crocidolite
AnthophylUte
Fluoramphibole
Nasal
passages
9 ± 3
8 ± 2
6 ± 1
8 ± 3
1 ±2
3 ± 2
Esophagus
2 ± 1
2 ± 1
2 ± 1
2 + 1
2 ± 1
1 ± 1
Gastro-
intestinal
tract
51 ± 9
54 ± 5
57 ± 4
51 ± 9
61 ± 8
67 + 5
Lower
respiratory
tract
38 ± 8
36 ± 4
35 ± 5
39 + 5
30 ± 8
29 ± 4
Percent,
deposited
31 ± 6
43 ± 14
42 ± 14
41 ± 11
64 ± 24
68 ± 17
Mean and standard deviation.
Percent of total Inspired.
Source: Morgan et al. (1975).
trachea), occurs within 30 minutes; up to two-thirds of the fibers are
swallowed and found 1n the GI tract.
Clearance from the lower respiratory tract (trachea to alveoli) proceeds
more slowly and two distinct components of clearance are observed. The first,
believed to be caused by macrophage movement, leads to elimination of a consider-
able portion of the material deposited in the lower respiratory tract at a
half life of 6-10 hours. The slower component that follows has a half-life of
60-80 days and involves clearance from the alveolar spaces. Data for a synthe-
tic fluoramphibole (Figure 4-1) show one short-term and two long-term compo-
nents for clearance of fibers. Other data on the lung content of animals,
sacrificed at various times after exposure, show only a single long-term
clearance component (Morgan et al., 1978); however, the ratio of fibers in the
feces to those in the lung at the time of sacrifice is not a constant, as
would be expected from a single exponential clearance mechanism.
By extrapolating curves like that of Figure 4-1 to zero-time for a vari-
ety of fibers, it is possible to ascertain the relative amounts of fibers
120
image:
693t/0.38+5i3a-0.693t/8,8 2e-0,693t/118
20
40 60 »Q 100
TIME AFTER ADMINISTRATION, days
Figure 4-1. Measurements of animal radioactivity
(corrected for decay) at various times after inhalation
exposure to synthetic fluoramphibole. Mean result for
three animals expressed as a percentage of the counting
rate measured immediately after exposure.
Source: Morgan et al. (1977).
121
image:
deposited In the bronchiolar-alveolar spaces. These data are shown for dif-
ferent fibers in Figure 4-2. The relative similarity of the percentage depos-
ited in the lower bronchioles or alveoli for different fiber diameters is a
reflection of two competing processes: at lower fiber diameters, the fibers
can be inspired and then expired without impaction in the lower respiratory
tract, but as the fiber diameter increases, impaction in the upper respiratory
tract becomes important, leading to a lower percentage being carried to the
alveolar spaces.
Morgan et al. (1978) also studied the length distribution of fibers that
remain in the lungs of rats to determine the significance of fiber length on
clearance. They found that the shorter fibers are preferentially removed
within one week following inhalation and suggested that longer fibers reaching
the alveolar spaces are trapped.
The radioactive chrysotile used in the clearance experiments -allows auto-
radiography to demonstrate the location of fibers at different times after
exposure. At 48 hours after exposure, the distribution of fibers in the lung
is relatively uniform. However, at later times, there is a movement of fibers
to the periphery of the lung where they accumulate in subpleural foci con-
sisting of alveoli filled with fiber-containing cells.
Other data on the deposition and retention of inhaled asbestos were
reported by Wagner et al. (1974). Figure 4-3 shows the dust content of rat
lungs following exposures to different asbestos varieties. In the case of
amphibole exposures, a linear increase in the amount of retained fiber was
seen, whereas for chrysotile, the content of the lung rapidly reached an
equilibrium between removal or dissolution processes and deposition, and did
not increase thereafter. The long-term build-up of the amphiboles indicates
that, in addition to the clearance processes observed by Morgan et al. (1977),
there is a virtual permanent retention of some fibers. Using a minute volume
for the rat of 100 ml, it would appear that about 1 percent of the total
crocidolite or amosite inhaled is retained permanently in the lung.
The finding of a rapid movement from the upper respiratory tract and a
slower clearance from the lower respiratory tract to the GI tract demonstrates
a route of exposure that may be important for GI cancer. The observation in
humans of peritoneal mesothelioma, of excess cancers of the stomach, colon,
and rectum, and possibly of cancers at other non-respiratory sites, such as
122
image:
c
&
| 20
Ja
£
•g
I
£ 15
•s
£
D
a
S.
i 10
in
Q
5
o _
LU D
2
I i
Key
Glass fiber * 108 o
UICC Anthophyllrte •
Fluoramphibole o
UICC Chrysotile A *
_ UICC Chrysotile B A —
UICC Amosrte e
UICC Crocidolrte X
Kuruman CrocidoKte o
Mallpsdrift Crocidolite •
Cerium Oxide • T
1 „
x ., ^^
1 _^*^^^ ^^^^H|,^
1 ^^^1 * *^^V
s ° s\
t ^ ~
V
\
\
\
- \ -
\
1
0
1 1
ACTIVITY MEDIAN AERODYNAMIC DIAMETER, ^m
Figure 4-2. Correlation between the alveolar
deposition of a range of fibrous and non-fibrous
particles inhaled by the rat and the corresponding
activity median aerodynamic diameters.
Source: Morgan (1979).
123
image:
AFTER REMOVAL
FROM EXPOSURE
24 TIME (MONTHS)
10000 20000
CUMULATIVE DOSE, mg/m'/hr.
30000
Figure 4-3. Mean weight of dust in lungs of rats in
relation to dose and time.
Source; Wagner et al, (1974).
124
image:
the kidney, could result from the migration of such fibers to and across the
gastrointestinal mucosa. Additionally, fibers may reach organs 1n the peri-
toneal cavity by transdiaphragmatic migration or lymphatic-hematogenous trans-
port.
4.3 CELLULAR ALTERATIONS
Several studies describe cellular changes in animals following exposure
to asbestos. Holt et al. (1964) describe early (14-day) local inflammatory
lesions found in the terminal bronchioles of rats following inhalation of
asbestos fibers. These lesions consist of multi-nucleated giant calls, lympho-
cytes, and fibroblasts. Progressive fibrosls follows within a few weeks of
the first exposure to dust. Davis et al. (1978) describe similar early lesions
found 1n rats, consisting of a proliferation of macrophages and cell debris in
the terminal bronchioles and alveoli.
Jacobs et al. (1978) fed rats 0.5 mg or 50 mg of chrysotile daily for 1
week or 14 months and subsequently examined GI tract tissue by light and elec-
tron microscopy.' No effects were noted in the esophagus, stomach, or cecal
tissue, but structural changes in the ileum were seen, particularly cf the
villi. Considerable cellular debris was detected in the ileum, colon, and
rectal tissue by light microscopy. Electron microscopy data confirm the
light microscopy data and indicate that the observed changes are consistent
with a mineral-Induced cytotoxidty.
A single oral administration of 5-100 mg/kg of chrysotile to rats produces
a subsequent increase in thymidine in the stcmach, duodenum, and jejunum
(Amacher et al., 1975), suggesting that an immediate response of cellular
proliferation and DNA synthesis may be stimulated by chrysotile ingestion.
4.4 MUTAGENICITY
Many studies showed asbestos not to be mutagenlc, e.g., 1n Escherichla
co 11 and Salmonella typhlinurlum tester strains (Chamberlain and Tarmy, 1977).
Newman et al. (1980) reported that asbestos has no mutagenic ability in Syrian
hamster embryo cells, but may increase cell permeability and allow other
mutagens into the cell. Hossman et al. (1983) showed that UICC (Union Intrana-
tlonale Contra le Cancer) crocidollte and chrysotile do not produce DNA strand
125
image:
breaks 1n the alkyHne elutlon assay when applied to cultured hamster tracheal
cells. Similar negative results were obtained by Lechner et al. (1983) with
respect to induction of DNA strand breakage in human bronchial organ cultures
treated with UICC chrysotile, amosite, and crocidolite. Finally, Hart et al,
(1979) demonstrated that asbestos does not produce unscheduled DNA synthesis
in human fibroblasts or single or double strand breaks.
However, a few studies do show mutagenldty. Sincock (1977) used several
chrysotile, amosite, and crocidolite samples to show that an Increased frequency
of polyploids and cells with fragments results from passive Inclusion of
asbestos in the culture media of Chinese hamster ovary (CHO)-Kl cells. Similar-
ly, Lavappa et al. (1975) showed that chrysotile induced a significant and
exposure-related increase in chromosome aberrations in cultured Syrian hamster
embryo cells. Amosite, chrysotile, and crocldolite were found to be weakly
mutagenlc 1n Chinese hamster lung cells in the 6-th1oguanine-res1stance assay
(Huang, 1979). Livingston et al, (1980) showed that exposure to crocidolite
and amosite can increase the sister chromatid exchange rate In Chinese hamster
ovarian fibroblasts.
The evidence for chromosomal effects in human cells is contradictory.
Valerio et al. (1980) found that freshly Isolated lymphocytes undergo chromo-
somal changes when treated with UICC Rhodesian chrysotile. In contrast,
Sincock et al, (1982) found negative effects with lymphocytes exposed to UICC
crocldolite. Asbestos was shown to be highly cytotoxic in a variety of pre-
parations (e.g., Mossman et al., 1983; Chamberlln and Brown, 1978).
In summary, while some evidence exists for aneuploidy caused by asbestos,
most studies show that asbestos probably is not mutagenic in the classic sense
of causing gene mutations and/or chromosomal breakage.
4.5 INHALATION STUDIES
The first unequivocal data that showed a relationship between asbestos
inhalation and lung malignancy in laboratory animals were those of Gross et al,
(1967) who observed carcinomas in rats exposed to a mean concentration of
86 mg/m chrysotile for 30 hours a week from the age of 6 weeks. Of 72 rats
surviving for 16 months or longer, 19 developed adenocarcinomas, 4 developed
squamous cell carcinomas, and 1 developed a mesothelioma. No malignant tumors
were found in 39 control animals. A search was made for primaries at other
126
image:
sites which could have metastasized and none were found. These and other data
are summarized in Table 4-2.
Reeves et al. (1971) found two squamous cell carcinomas 1n 31 rats sacri-
ficed after 2 years following exposure to about 48 mg/m of croddoHte. No
malignant tumors were reported in rabbits, guinea pigs, or hamsters, or 1n
animals exposed to similar concentrations of chrysotile or amosite. No details
of the pathological examinations were given.
In a later study (Reeves et al., 1974), malignant tumors developed in 5
to 14 percent of the rats that survived 18 months after exposure. Lung cancer
and mesothelioma were produced by exposures to amosite and chrysotile, and
lung cancer was produced by crocidolHe Inhalation. Again, significant experi-
mental details were not provided; information on survival times and times of
sacrifice would have been useful. Available details of the exposures and
results are given In Table 4-3. While the relative carcinogenicity of the
fiber types was similar, the fibrogenlc potential of chrysotile, which had
been substantially reduced in length and possibly altered by milling (Langer
et al., 1978), was much less than that of the amphiboles. These results are
also discussed 1n a later paper by Reeves (1976).
The most Important series of animal Inhalation studies 1s that of Wagner
et al. (1974, 1977). Wagner exposed 849 Wistar SPF rats to the five UICC
asbestos samples at concentrations from 10.1 to 14.7 mg/m for times ranging
from 1 day to 24 months. These concentrations are typically 10 times those
measured 1n dusty asbestos workplaces during earlier decades. For all the
exposure times, 50 adenocarcinomas, 40 squamous-cell carcinomas, and 11 mesothe-
liomas were produced. All varieties of asbestos produced mesothelioma and
lung malignancies, 1n some cases from exposures as short as 1 day. Data from
these experiments are presented 1n Tables 4-4 and 4-5. These tumors follow a
reasonably good linear relationship for exposure times of 3 months or greater.
However, the Incidence in the 1-day exposure group is considerably greater
than expected. Exposure had a limited effect on length of life. Average
survival times varied from 669 to 857 days for exposed animals versus 754 to
803 days for controls. The development of asbestosis 1s also documented.
There are 17 lung tumors, 6 in rats with no evidence of asbestosis and 11 1n
rats with minimal or slight asbestosis. Cancers at extrapulmonary sites are
listed,' Seven malignancies of ovaries and eight malignancies of male genito-
urinary organs were observed 1n the exposed groups of approximately 350 male
127
image:
TABLE 4-2. SUMMARY OF EXPERIMENTS ON THE EFFECTS OF INHALATION OF ASBESTOS
Study
Animal species
Material administered
Dosage
Animals Examined
for tunors
Findings
(malignant tunors)
Average survival
time
Gross et al. (1967) 132 male white rats
Reeves et al. (1971)
Reeves et al, (1974)
IX)
CD
Wagner et al. (1974)
Wagner et al. (1977)
Davis et al. (1978)
55 male white rats
206 rats
106 rabbits
139 guinea pigs
214 boasters
219 rats
216 gerblls
100 nice
72 rabbits
lOfi guinea pigs
13 groups of approxi-
mately 50, and IS of
about 25 Wistar SPF
rats
CO Wistar nale and
female rats
CO Wistar male and
female rats
46 groups of approxi-
mately 20 Han SPF rats
and 20 Han SFP rats
Ball- and hammer-milled
Canadian chrysotlle
with/without 0.05 ml
Intratracheal 5 per-
cent NaOH
Controls with/without
5 percent NaOH
Ball-milled chrysotlle,
anoslte, and croddollte
Bal 1* and hammer--
milled chrysotlle,
anoslte, and
croc Idol He
Amoslte, anthophylllte,
croc Idol He. Canadian
chrysotlle, Rhodeslan
chrysotlle (U1CC sam-
ples)
Superfine chrysotlle
Nonfibrous cosmetic talc
UICC samples of anoslte,
chrysotlle, and
crocldgllte
42-14* mg/al
(mean concentra-
trailon, 86 mg/
m3) for 30 hours/
week
control
4B±2 mg/rn3 for
16 hours/week up
to 2 years
ng/m3 for
16 hours/Meek
up to 2 years
10.1 to 14.7
ng/m3 for 1 day
to 24 months,
35 hours/week.
10.8 mg/m3 37.5
hours/week for
3, 6. or 12 months
2 mg/n3 and
10 mg/m3 35
hours/week
for 224 days
72
39
not available
120 rats
116 gerblls
10 Bice
30 rabbits
43 guinea pigs
849
20B
17 adenocarclnomas
4 squamous-cell sarcomas
7 flbrosarcomas
1 mesothellaraa
none
2 squamous-cel1 carcino-
mas In 31 animals from
croc Idol He exposure
10 malignant tumors In
rats, 2 In nice
(Table 4-3)
(See Tables 4-4 and 4-5)
All asbestos varieties
produced nesothelloma and
lung cancer, some from ex-
posure as short as 1 day
1 adenocarclnoma of the
lung 1n 24 animals ex-
posed for 12 months
none
7 adenocarclnomas
3 squamous-cell
sarcomas, 1 pleural
mesothellona, 1
peritoneal mesothelioma
not available
not available
no Information
periodic sacri-
fices were made
no Information
periodic sacri-
fices were made
669 to 857 days
versus 754 to
803 for controls.
Survival times
not significant-
ly affected by
exposure.
not available
sacrificed at 29
months
20 Han SPF rats
control
control
20
none
image:
TABLE 4-3. EXPERIMENTAL INHALATION CARCINOGENESIS IN RATS AND MICE
Fiber
Chrysotile
Amos He
Crocidolite
Controls
Mass
mg/m
47.9
48.6
50.2
Exposure
Fiber
3 f/ml
54
864
1,105
Animals
examined
43
46
46
5
Rats
Malignant tumors
1 lung papillary carcinoma
1 lung squamous-cel 1 carcinoma
1 pleural mesothel ioma
2 pleural mesothel iomas
3 squamous-cel 1 carcinomas
1 adenocarcinoma
•1 papillary carcinoma - all of
the lung
Ndn.e
Animals
examined
19
17
18
r
D
Mice
Malignant tumors
None
None
2 papillary carcinomas
of bronchus
1 papillary carcinoma
of bronchus
The asbestos was comminuted by vigorous milling, after which 0,08 to 1.82% of the airborne mass was of fibrous
morphology (3:1 aspect ratio) by light microscopy.
Source: Reeves et al. (1974).
image:
TABLE 4-4. NUMBER OF RATS WITH LUNG TUMORS OR MESOTHELIOMAS AFTER EXPOSURE
TO VARIOUS FORMS OF ASBESTOS THROUGH INHALATION
Form of Asbestos
Amoslte
AnthophylUte
Croddolite
Chrysotlle
(Canadian)
Chrysotile
( Rhodes 1 an)
None
Number of
animals
146
145
141
137
144
126
Adenocardnomas
5
8
7
11
19
0
Squamous-cell
carcinomas
6
8
9
6
11
0
Mesothellomas
1
2
4
4
0
0
Source: Wagner et al. (1974)
TABLE 4-5. NUMBER OF RATS WITH LUNG TUMORS OR MESOTHELIOMAS AFTER VARIOUS
LENGTHS OF EXPOSURE TO VARIOUS FORMS OF ASBESTOS THROUGH INHALATION
Number Number of animals
Length of
exposure
None
1 day
3 months
6 months
12 months
24 months
of animals
tested
126
219
180
90
129
95
with lung
carcinomas
0
3a
8
7
35
37
Number of animals
with pleural
mesothellomas
0
2b
1
0
6
2
Percent
of animals
with tumors
0.0
2.3
5,0
7.8
31.8
41.0
Two rats exposed to chrysotlle and one to croddollte.
One rat exposed to amoslte and one to croddollte.
Source: Wagner et al. (1974).
130
image:
and female rats. No malignancies were observed In control groups of GO males
and females. The tnddence of malignancy at other sites varied little from
that of the controls. The authors note that 1f controls from other experiments
1n which ovarian and genitourinary tumors were present are Included, the
comparative Incidence 1n the exposed groups 1n the first study lacks statistical
significance. No data are provided on the variation of tumor Incidence at
extrapulmonary sites with asbestos dosage.
Wagner et al. (1977) also compared the effects of Inhalation of a super-
fine chrysotile to the effects of Inhalation of a pure nonflbrous talc. One
3
adenocarclnoma was found 1n 24 rats exposed to 10.6 mg/m of chrysotile for
37.5 hours a week for 12 months.
In a study similar to Wagner's, Davis et al. (1978) exposed rats to 2.0
or 10.0 mg/m of chrysotile, crocldollte, and amoslte (equivalent to 430 to
1950 f/ml). Adenocardnomas and squamous-cell carcinomas were observed 1n
chrysotile exposures, but not 1n crocldollte or amoslte exposures (Table 4-6).
One pleural mesothel1oma was observed wfth crocldollte exposure, and extrapulmo-
nary neoplasms Included a peritoneal mesothel1oma. A relatively large number
of peritoneal connective tissue malgnandes also were observed, these Including
a-leimyoflbroma on the wall of the small Intestine. The meaning of these
tumors Is unclear.
TABLE 4-6. EXPERIMENTAL INHALATION CARCINOGENESIS IN RATS
Exposure
Chrysotile
Chrysotile
Mass
mg/m3
10
2
Fiber
f> Sum/ml
1,950
390
Number of
animals
examined
40
42
Malignant tumors
6 adenocarclnomas
2 squamous-cell carcinomas
1 squamous-cell carcinoma
Source: Davis et al. (1978),
1 peritoneal mesothelloma
Amoslte
Crocldollte
Crocldollte
Control
10
10
5
550
860
430
43
40
43
20
None
None
1 pleural
None
mesothel 1oma
131
image:
Inhalation exposures result in concomitant GI exposures from the asbestos
that Is swallowed after clearance from the bronchial tree. Although all
inhalation experiments focus on thoracic tumors, those of Wagner et al. (1974),
Davis et al. (1978), and, to a limited extent, Gross et al. (1967) also in-
clude a search for tumors at extrathoracic sites. A limited number of these
tumors were found, but no association could be made with asbestos exposure.
One important aspect of the inhalation experiments is the number of
pulmonary neoplasms that can be produced by inhalation in the rat as compared
to other species (Reeves et al., 1971, 1974). This phenomenon illustrates the
variability of species response to asbestos and the need for an appropriate
model before extrapolations to man can be made with confidence. The absence
of significant GI malignancy from asbestos exposure in animals, In contrast to
t
that found in humans, may be the result of the use of Inappropriate animal
models.
4.6 INTRAPLEURAL ADMINISTRATION
Evidence that intrapleural administration of asbestos results in mesothe-
11oma was presented 1n 1970 when Donna (1970) produced mesothelionias in Sprague-
Dawley rats treated with a single dose of 67 mg of chrysotile, amoslte, or
crocidolite. Reeves et al. (1971) produced mesothellal tumors in rats (1 of 3
with croddollte and 2 of 12 with chrysotile) by Intrapleural injection of 10
mg of asbestos. Two of 13 rabbits Injected with 16 mg of crocidolite developed
mesotheliomas.
In a series of experiments, Stanton and Wrench (1972) demonstrated that
major commercial varieties of asbestos, as well as various other fibers,
produce mesotheliomas in as many as 75 percent of animals into which material
had been surgically Implanted onto the pleura! surface. The authors conclude
that the carcinogenicity of asbestos and other fibers is strongly related to
their physical size; fibers that have a diameter of less than 3 jjm are carcino-
genic and those that have a larger diameter are not carcinogenic. Further,
samples treated by grinding in a ball mill to produce shorter length fibers
are less likely to produce tumors. Although the authors attribute the reduced
cardnogenicity to a shorter fiber length, the question Vias raised of the
effect of the destruction of crystal Unity, and perhaps other changes 1n the
fibers, caused by the extensive ball milling (Langer et al., 1978).
132
image:
Since 1972, Stanton and his co-workers (Stanton et al., 1977, 1981) have
continued these investigations of the carcinogenic action of durable fibers.
Table 4-7 summarizes the results of 72 different experiments. In their analy-
ses, Stanton et al. (1981) suggest that the best measure of carcinogenic
potential is the number of fibers that measure <0,25 pm in diameter and >8 pm
in length, although a good correlation of carcinogenic}ty is also obtained for
fibers <1.5 pm in diameter and >4 \m in length. The logit distribution of
tumor incidence against the log of the number of particles having a diameter
<0.25 urn and length >8 pm Is shown in Figure 4-4. The regression equation for
the dotted line is
ln[p/(l-p)] = '2.62 + 0.93 log x (4-1)
where p is the tumor probability and x is the number of particles per pg that
are £0.25 urn diameter and >8 pm long. A reasonable relationship exists
between the equation and available data, but substantial discrepancies suggest
the possibility that other relationships may better fit the data. Bertrand
and Pezerat (1980) suggested that carcinogenicity may correlate as well with
the ratio of length to width (aspect ratio).
Another comprehensive set of experiments was conducted by Wagner et al.
(1973, 1977). Mesothelioma was produced from intrapleural administration of
asbestos to CD Wlstar rats, demonstrating that there is a strong dose-response
relationship. Tables 4-8 and 4-9 list the results of these experiments.
Pylev and Shabad (1973) and Shabad et al. (1974) reported mesotheliomas
in 18 of 48 rats and in 31 of 67 rats injected with 3 doses of 20 mg of Russian
chrysotlle. Other experiments by Smith and Hubert (1974) produced mesotheliomas
*
1n hamsters Injected with 10-25 mg of chrysotile, 10 mg of amosite or anthophyl-
lite, and 1-10 mg of croddollte.
Various suggestions have been made that the natural oils and waxes contam-
inating asbestos fibers might be related to the carcinogenicity of asbestos
fibers (Harington, 1962; Harington and Roe, 1965; Commins and Gibbs, 1969).
However, this theory was not substantiated in the experiments performed by
Wagner et al. (1973) or Stanton and Wrench (1972).
133
image:
TABLE 4-7. SUMMARY OF 172 EXPERIMENTS WITH DIFFERENT FIBROUS MATERIALS
OJ
-t.
Experiment
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
IB
19
ZO
21
22
23
24
25
26
27
28
29
30
31
32
33
34
3S
X
Compound
Tltanate 1
Tltanate 2
Sllcarblde
Dawson 5
Tremllte 1
TreaolIU 2
Dawson 1
Cracld 1
Crocld 2
Crocld 3
AMSlte
Crocld 4
Class 1
Crocld 5
Glass 2
Glass 3
Glass 4
Alwln 1
Glass 5
Dauson 7
Dawson 4
Dawson 3
Glass 6
Crocld 6
Crocld 7
Crocld fl
Alwln 2
Alwln 3
Crocld 9
Woll as ton 1
Alwln 4
Crocld 10
Alwln 5
Glass 20
Glass 7
tfollaston 3
Actual
tuBor
Incidence
21/29
20/29
,17/26 ,
26/29
22/28
21/28
20/25
18/27
17/24
15/23
14/25
15/24
9/17
14/29
12/31
20/29
18/29
15/24
16/25
16/30
11/26
9/24
7/22
9/27
11/26
8/25
8/27
9/27
8/27
5/20
4/25
6/29
4/22
4/25
5/28
3/21
Percent
tinor
probability
t SO
9514.7
100
100
100
100
100
95i4.B
94 ±6.0
9316.5
9316.9
9317. 1
8619.0
85±13. 2
7B110.B
77116. 6
7418.5
7119.1
70110.2
6919.6
68*9.8
66112.2
66H3.4
64117. 7
63113.9
56111. 7
53112.9
44111.7
41110.5
3319.8
31112.5
28112.0
37113.5
2219.8
22UO.O
2118.7
19110.5
Canon log
flbers/ug
<0.25 M" disaster x
>8 \im long
4.94
4.70
5.15
4.94
3.14
2.B4
4.66
5.21
4.30
5.01
3.53
5.13
5.16
3.29
4.29
3.59
4.02
3.63
3.00
4.71
4.01
5,73
4.01
4.60
2,65
0
2.95
2,47
4.25
0
2.60
3.09
3.73
0
2.50
0
Experiment
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
- - 55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
Compound
Hal lay 1
Hal lay 2
Glass 8
Crocld 11
Glass 19
Glass 9
Attain 6
Oavson G
Damon 2
Woll as ton 2
Crocld 12
Attapul 2
Glass 10
Glass 11
Tltanate 3
Attapul 1
Talc 1
Glass 12
Glass 13
Class 14
Glass 15
Alwln 7
Glass 16
Talc 3
Talc 2
Talc 4
Alwln 8
Glass 21
Glass 22
Glass 17
Glass 18
Crocld 13
Wo 11 as ton 4
Talc 5
Talc 6
Talc 7
Actual
tUBQr
Incidence
4/25
5/28
3/26
4/29
2/28
2/28
2/28
3/30
2/27
2/25
2/27
2/29
2/27
1/27
1/28
2/29
1/26
1/25
1/27
1/25
1/24
1/25
1/29
1/29
1/30
1/29
V28
2/47
1/45
0/28
0/115
0/29
0/24
0/30
0/30
0/29
Percent
twor
probability
i SO
2019.0
2319.3
19H0.3
1918.5
1519.0
1419.4
1316 8
1316.9
1217.9
1218.0
1017.0
1117.5
S15.6
815.5
818.0
815.3
716.9
715.4
615.7
615.5
615.9
515.1
514.4
414.3
413.8
514.9
313.4
614.4
212.3
0
0
0
0
0
0
0
CoDBon log
flbers/tig
<0.25 M dlaneter *
>8 |M long
0
0
3.01
0
0
1.84
0.82
0
0
0
3.73
0
0
0
0
0
0
0
0
0
1.30
0
0
0
0
0
0
0
0
0
0
0
0
0
3.30
0
SO = Standard deviation.
Source: Stanton et al. (1981).
image:
OJ
en
CC
o
D
O
m
O
I.U
0.9
n o
U.o
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
I
C
~~ G
D
L
~" S
A
— P
T
M
_ w
H
C O
—
-W
HG
_HCW
G
ADOW
.AGGPT .
LTGGGG
LTTGG
CWTTGG
0.0 0.5
I ! I
= crocJdolite
= glass
= dawsonfte
= aluminum oxide
= silicon carbide
= attapuigtte
= titanate
= talc
= tremolhe
= wollastontte
= halloysKe
= amosita
t
S
S
^**
^
^ G
+*L'
G
I I I
1.0 1.5 2.0
| M | M | | | 1 P D S |
D P c — •• "
o c c >-~
X.-"^CG
C ** G "" ~
G ^
G L,XG —
x' D
X DG D
X C
c x
s
s —
/ L
X. _ _
' c
_
L
G L _
**
C —
1 I T I ( 1 | |
2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.1
LOG NUMBER PARTICLES MEASURING < 0.25 fan x > tan PER MICROGRAM
Figure 4^4.Regression curve relating probability of tumor to logarithm of
number of particles per M§ with diameter < 0.25 ^m and length >8 ptn.
Source: Stanton et al. (1981).
image:
TABLE 4-8. PERCENTAGE OF RATS DEVELOPING MESOTHELIOMAS AFTER INTRAPLEURAL
ADMINISTRATION OF VARIOUS MATERIALS
Material
Percent of rats
with mesothe"Hernias
SFA chrysotile (superfine Canadian sample)
UICC crocidolite
UICC amosite
UICC anthophyllite
UICC chrysotile (Canadian)
UICC chrysotile (Rhodesian)
Fine glass fiber (code 100), median diameter =
0.12 Mm
Ceramic fiber, diameter = 0.5-1 pma
Glass powder
Coarse glass fiber (code 110), median diameter =
1.8 pro
66
61
36
34
30
19
12
10
3
0
From Wagner et al. (1973).
Source: Wagner et al. (1976).
TABLE 4-9. DOSE-RESPONSE DATA FOLLOWING INTRAPLEURAL ADMINISTRATION
OF ASBESTOS TO RATS
Material
SFA chrysotile
Crocidol ite
Dose
mg
0.5
1
2
4
8
0.5
1
2
4
8
Number of
rats with
mesothelioma
1
3
5
4
8
1
0
3
2
5
Total number
of rats
12
11
12
12
12
11
12
12
13
11
Percent
of rats
with tumors
8
27
42
33
62
9
0
25
15
45
Source: Wagner et al. (1973).
136
image:
4.7 INTRATRACHEAL INJECTION
Intratracheal Injection has been used to study the combined effect of the
administration of chrysotlle with benzo(a)pyrene 1n rats and hamsters. No
lung tumors were observed 1n rats given 3 doses of 2 mg of chrysotlle (Shabad
et al., 1974) and 1n hamsters given 12 mg of chrysotlle (Smith et al., 1970).
However, co-administration of benzo(a)pyrene resulted 1n lung tumors, which
suggests a co-carcinogenic or synerglstlc effect.
4.8 INTRAPERITONEAL ADMINISTRATION
Intraperltoneal Injections of 20 mg of croddollte or chrysotlle produced
3 peritoneal mesothellomas 1n 13 Charles River CD rats, but 20 mg of amosite
produced no tumors 1n a group of 11 rats (Maltonl and Annoscla, 1974). Maltonl
and Annoscla also Injected 25 mg of croddollte Into 50 male and 50 female
17-week-old Sprague-Dawley rats and observed 31 mesothelial tumors 1n males
and 34 1n females.
In an extensive series of experiments, Pott and Fr1edr1chs (1972) and
Pott et al. (1976) produced peritoneal mesothellomas 1n mice and rats that
were Injected with various commercial varieties of asbestos and other fibrous
material. These results are shown 1n Table 4-10, Using experiments with
Intrapleural administration, the malignant response was altered by ball-milling
the fibers for 4 hours. The rate of tumor production was reduced from 55 to
32 percent and the time from onset of exposure to the first tumor was length-
ened from 323 to 400 days following administration of 4 doses of 25 mg of UICC
Rhodeslan chrysotlle. In the case of the ball-milled fibers, 99 percent of
the fibers were reported to be smaller than 3 urn, 93 percent were smaller than
1 urn, and 60 percent were smaller than 0.3 urn.
Pott (1980) proposed a model for the relative carcinogen!city of mineral
fibers, according to their dimensionality, using the results of injection and
implantation data. Figure 4-5 shows the schematic features of this model.
The greatest carcinogenlclty is attributed to fiber lengths between 5 and 40
urn with diameters between 0.05 and 1 jam.
A strong conclusion that can be drawn from the above experimental data 1s
that long (>4 urn) and fine diameter (<1 urn) fibers are more carcinogenic than
short, thick fibers when they are implanted on the pleura or injected into the
peritoneum of animals. The origin of a reduction 1n carclnogeniclty for
137
image:
TABLE 4-10. TUMORS IN ABDOMEN AND/OR THORAX OF RATS AFTER INTRAPERITONEAL INJECTION OF GLASS FIBERS, CROCIDOLITE, OR CORUNDUM
Dust
Glass fibers
MM 104
Glass fibers
NN 104
Glass fibers
NN 104
CroddolHe
CorundiB
UICC Rhodes 1 an
chrysotlle
UICC Rhodes Ian
chrysotlle
UICC Rhodes 1 an
chrysotlle
UICC Rhodes 1 an
chrysotlle
UICC Rhodes Ian
chrysotlle
UICC Rhodes 1 an
milled
Paly goes cite
Fora*
f
f
f
f
g
f
f
r
f
f
f
f
IntrapeHtoneal
dose
•9
2
10
2 x 25
2
2 x 25
Z
6.25
25
4 x 25
3 K 25
s.c.
4 x 25
3 x 25
Effective
nuaber of
dissected
rats
73
77
77
39
37
37
35
31
33
33
37
34
Number of
days before
first tUKir
421
210
194
452
545
431
343
276
323
449
400
257
Average
survival tlw
of rats with
tumrs, days
after Injection
703
632
3G7
761
799
651
501
419
361
449
509
348
Rats
with
tuaors,
percent
27.4
53.2
71.4
38.5
8.1
16.2
77.1
0.6
54.5
3.0
32.4
76.5
1
17
36
47
12
1
4
24
21
16
-
9
24
Ti»or/typeb
23 456
3 - -1-1
4 - 13-
6 2 .
3 - - 2 1
222
2 - - 1 -
3 - ...
21 1
2 .
1 -
s.c.
3 - ---
2 -
image:
TABLE 4-10. (continued)
CO
UD
Dust
Glass fibers
S * S 106
Glass fibers
S * S 106
Glass fibers
S * S 106
GypsiM
Henatlte
ActlnoHU
Blotlte
HaeMtlte
(precipitation)
HaeMtlte
(Mineral)
Pectollte
Sanldlnc
Talc
NaCl (control)
For»a
f
1
f
f
f
g
9
g
g
g
g
g
-
Intraperltoneal
dose
•g
2
10
* x 25
4 * 25
4 K 25
4 K 25
4 * 25
4 x 25
4 N 25
4 x 25
4 x 25
4 x 25
4 x 2 •!
Effective
number of
dissected
rats
34
36
32
35
34
39
37
34
38
40
39
36
72
Niaber of
days before
first tuaor
692
350
197
579
249
-
-
„
-
569
579
587
-
Average
survival ttoe
of rats with
tiaors, days
after Injection
692
530
325
583
315
-
*^
«
-
569
579
587
-
Rats
with
tuaors, TuBor/type
percent 1 23 4 5 6 .
2,9 1 ...
11.1 2 2 - 1 -
71.9 20 3 - -
5.7 1 11-
73.5 17 8 ...
-
_
. -
... -
2.5 ... Ill
2.6 - 1 - -
2.B 1 - - -
-
f = fibrous; g = granular.
bTwor Types «•«: 1 Mesothcllow; 2 Spindle cell sarcoma. 3 Poly«-cell s»rco«a; 4 Carc1i»o»a; 5 Retlculua cell sarcoma;
6 Benign ~ not evaluated 1n tuaor rates.
Sources: Pott and Frledrlchs (1972); Pott et al. (1976),
image:
0.031 \
Figure 4-5. Hypothesis concerning the carcinogenic potency of a fiber
as a function of its length and width using data on tumor incidence
from injection and implantation studies.
Source: Pott (1980).
140
image:
shorter, ball-milled fibers 1s less clear because the relative contributions
of shorter fiber length and the significant alteration of the crystal structure
by Input of physical energy have not yet been defined. Extrapolation of data
on size-dependent effects obtained from intrapleural or intrapeHtoneal admin-
istration, to inhalation, where movement of the fibers in airways and subse-
quently through body tissues 1s strongly size-dependent, presents significant
difficulties. The number of shorter (<5 urn) fibers in an exposure circumstance
may be 100 times greater than the number of longer fibers; therefore, their
cardnogenicity must be 1/100 times as much before their contribution can be
neglected.
4.9 TERATOGENICITY
There 1s no evidence that asbestos 1s teratogenlc. Schneider and Maurer
(1977) fed pregnant CD-I mice doses of 4-400 mcj/kg body weight (1.43 to 143)
for gestation days 1 to 15, They also administered 1, 10, or 100 pg of asbes-
tos to 4-day blastocysts, which were transferred to pseudopregnant mice. No
positive effects were noted 1n either experiment.
4.10 SUMMARY
Animal data on the carcinogenlclty of asbestos fibers confirm and extend
epidemiologlcal human data. Mesothelioma and lung cancer are produced by all
the principal commercial asbestos varieties, chrysotile, amosUe, croddollte,
and anthophyllite, even by exposures as short as one day. The deposition and
clearance of fibers from the lung suggest that most inhaled fibers (~99 percent)
are eventually cleared from the lung by ciliary or phagocytic action. Chrysotile
appears to be more readily removed, and dissolution of the fibers occurs 1n
addition to other clearance processes. Implantation and Injection studies
suggest that the carcinogenlcity of durable mineral fibers is related to their
dimensionality and not to their chemical composition. Long (>4 urn) and thin
£<1 pm) fibers are most carcinogenic when they are in place at a potential
tumor site. However, deposition, clearance, and migration of fibers are also
size dependent, and the importance of all size-dependent effects in the cardno-
genicity of inhaled fibers is not fully established.
141
image:
5. ENVIRONMENTAL EXPOSURES TO ASBESTOS
5.1 INTRODUCTION
The analysis of ambient air samples for asbestos has utilized techniques
different from those used in occupational circumstances. This situation
occurred because typical urban air may contain up to 100 )jg/m of particulate
matter in which the researcher is attempting to quantify asbestos concentra-
3 3
tions from about 0.1 ng/m to perhaps 1000 ng/m . Thus, asbestos may con-
stitute only 0.0001 to 1 percent of the particulate matter in a given air
sample. Asbestos found in ambient air has a size distribution such that the
vast majority of fibers are too short or too thin to be seen with an optical
microscope. In many cases, these fibers and fibrils will be agglomerated with
a variety of other materials present in the air samples.
The only effective method of analysis uses electron microscopy to
enumerate and size all asbestos fibers (Nicholson and Pundsack, 1973; Samudra
et al., 1978).. Samples for such analysis are usually collected either on a
Nuclepore (polycarbonate) filter with a pore size of 0.4 urn or on a Millipore
(cellulose ester) filter with a pore size of 0.8 urn. In some cases the Millipore
is backed by nylon mesh. Samples collected on Nuclepore filters are prepared
for direct analysis by carbon coating the filter to entrap the collected
particles. A segment of the coated filter is then mounted on an electron
microscope grid, which is placed on a filter paper saturated with chloroform
so that the chloroform vapors dissolve the filter material. (Earlier electron
microscopic analysis utilized a rub-out technique in which the ash residue was
dispersed in a nitrocellulose film on a microscope slide and a portion of the
film was then mounted on an electron microscope grid for scanning.)
Samples collected on Millipore filters are prepared for indirect analysis
by ashing a portion of the filter in a low temperature oxygen furnace. This
removes the membrane filter material and all organic material collected in the
sample. The residue is recovered in a liquid phase, dispersed by ultrasonifi-
cation, and filtered on a Nuclepore filter. The refiltered material is coated
with carbon and mounted on a grid as above. The samples are then subjected to
analysis. Chrysotile asbestos is identified on the basis of its morphology in
the electron microscope and amphiboles are identified by their selected area
electron diffraction patterns, supplemented by energy-dispersive X-ray analy-
sis. Fiber concentrations in fibers per unit of volume (such as fibers/cm ,
142
image:
3
fibers/m , etc.) are calculated based on sample volume and filter area counted.
In some cases, mass concentrations are reported using fiber volume and density
relationships. However, mass concentrations may not be reliable if the sample
contains fibrous forms, such as clusters, bundles, and matrices, where fiber
volume is difficult to determine. These materials may constitute most of the
asbestos mass in some samples, particularly those reflecting emission sources.
Current fiber counting methods do not include those clumps. However, many of
them are respirable and to the extent that they are broken apart in the lungs
into individual fibers, they may add to the carcinogenic risk. On the other
hand, methods which break up fibers generally disperse the clumps as well. In
such analyses, the clumps, would contribute to tht> mass.
In much of the earlier analyses of chrysotile concentrations in the
United States the ashed material was either physically dispersed or disrupted
by ultrasonification. Thus, no information was obtained on the size distri-
bution of the fibers in the original aerosol. Air concentrations were given
only in terms of total mass of asbestos present In a given air volume, usually
in nanograms per cubic meter (ng/m ). (See Section 5-9 for data on the inter-
convertability of optical fiber counts and electron'microscopic mass determi-
nations.) With the use of Nuclepore filters and appropriate care in the
collection of samples and their processing, information on the fiber size
distribution can be obtained and concentrations of fibers of selected di-
mensions can be calculated. Samples collected on Millipore filters can be
®
ashed and the residue resuspended and filtered through Nuclepore filters.
However, some breakage of fibers during the process is likely. Direct pro-
is)
cessing of Millipore filters for electron microscopic analysis has been
reported by Burdett and Rood (1983) and is being tested by several labora-
tories. However, the utility and reliability of this technique is unknown at
present.
Ideally, one would like a measure of exposure that would be proportional
to the carcinogenic risk.. Unfortunately, this is not possible because of our
limited information on the carcinogenicity of fibers according to length and
width and the lack of information on the deposition, clearance, and movement
through the body of fibers of different sizes. Secondly, our epidemiological
evidence, of disease relates to fibers longer than 5 urn measured by optical
microscopy. It should be recognized that electron microscopic fiber counts of
fibers longer than 5 urn of length will differ considerably from optical micro-
scopy counts of the same sample because of the presence of a large number of
143
image:
fibers undetected by optical microscopy. Nevertheless, it would appear that
the best measure of risk would be electron microscopic fiber counts of fibers
greater than 5 urn in length and use of an empirically determined adjustment
for the increased resolving power of the electron microscope when such mea-
surements are used for risk assessment.
Two of the studies described below provide information on fiber as well
as mass concentrations. However, in one case (Constant, 1983) the fiber
concentrations were of fibers of all length, and thus are impossible to trans-
late into optical microscopic counts (other than by mass). While the other
studies are limited because of the absence of fiber concentrations, they are
sufficient to indicate exposure circumstances of concern or that warrant
further investigation. Further, using an empirical conversion factor (having
a very large uncertainty), estimates of environmental exposures can be made in
terms of optical fiber counts.
Unfortunately, few studies have been conducted which provide data relating
asbestos fiber concentrations and health effects. While estimates of asbestos
concentrations based on conversions from fiber-mass relationships have an
associated uncertainty, they are the best data available for such assessments.
Future studies will hopefully be designed to measure fiber number, size, and
type for correlation with health effects.
An analysis of 25 samples collected in buildings having asbestos surfac-
ing material (some buildings showing evidence of contamination) demonstrated
the inadequacy of phase contrast optical microscopic techniques for the quan-
tification of asbestos (Nicholson et al., 1975). Figure 5-1 shows the corre-
lation of optical fiber counts determined using National Institute for Occupa-
tional Safety and Health (1972) prescribed techniques and asbestos mass mea-
surements obtained on the same samples. In determining the fiber concen-
trations, all objects with an aspect ratio of three or greater were enumerated
using phase-contrast microscopy, Petrographic techniques were not utilized to
verify whether an object was an asbestos fiber. Figure 5-1 shows that the
optical microscopic data do not reflect the mass concentrations of asbestos
determined by electron microscopy, largely because of a considerable number of
nonasbestos fibers that were in the ambient air and were counted in the optical
microscopic analysis.
The available published asbestos exposure data are to a large extent
episodic in nature. The studies were not designed to provide measures of
ambient concentrations throughout the United States. The data presented here
144
image:
30
£
^
(0
«= 20
O
ill
U
z
O
U
10
m
I
1
50
100 150
ASBESTOS MASS CONCENTRATION, ng/m'
200
250
Figure 5-1. Fiber concentrations by optical microscopy versus asbestos mass concentrations by
electron microscopy.
Source: National Institute for Occupational Safety and Health (1972).
image:
represent the published data that are available. These data show what concen-
tration can occur 1n the circumstances given. When useful information (I.e.,
number of sites, frequency of samples) 1s available that helps characterize
the representativeness of exposure of the data, 1t 1s presented. But as can
be seen, these data generally do not represent the results of systematic
studies designed to characterize the ambient asbestos concentrations in the
United States or those 1n typical building circumstances.
5.2 GENERAL ENVIRONMENT
Asbestos of the chrysotlle variety has been found to be a ubiquitous
contaminant of ambient air. A study of 187 quarterly samples collected In 48
U.S. cities 1n 1969-1970 showed chrysotlle asbestos to be present 1n virtually
all metropolitan areas (Nicholson, 1971; Nicholson and Pundsack, 1973). Table
5-1 lists the distribution of values obtained in that study, along with similar
data obtained by the Battelle Memorial Institute (U.S. EPA, 1974). Each value
represents the chrysotlle concentration in a composite of from five to seven
24-hour samples, thus averaging possible peak concentrations which could occur
3
periodically or randomly. Of the three samples greater than 20 ng/m analyzed
by Mount Sinai School of Medicine, one sample was 1n a city that had a major
shipyard and another was 1n a city that had four brake manufacturing facilities
with no emission controls. Thus, these samples may have included a contribu-
tion from a specific source 1n addition to that of the general ambient air.
Also shown in Table 5-1 is the distribution of chrysotile concentrations from
five-day samples of the air in Paris (Sebastien et al., 1980). These values
were obtained during 1974 and 1975 and were generally lower than those measured
in the United States, perhaps reflecting a diminished use of asbestos in
construction compared to that of the United States during 1969-1970.
In a study of the ambient air of New York City, 1n which samples were
taken only during daytime working hours, higher values than those mentioned
above were obtained (Nicholson et al., 1971). These 4- to 8-hour samples were
collected between 8:00 A.M. and 5:00 P.M., and they reflect what could be
intermittently higher concentrations during those hours compared to nighttime
periods. Table 5-2 records the chrysotlle content of 22 samples collected in
the five boroughs of New York and their overall cumulative distribution. The
samples analyzed in all the studies discussed above were taken during a period
when f1reproofing of high rise buildings by spraying asbestos-containing
146
image:
TABLE 5-1. CUMULATIVE DISTRIBUTION OF 24-HOUR CHRYSOTILE ASBESTOS
CONCENTRATIONS IN THE AMBIENT AIR OF U.S. CITIES AND PARIS, FRANCE
Concentration
(ng/m3)
less than
1.0
2.0
5.0
10.0
20.0
50.0
100.0
Mount
School of
Number
of
samples
61
119
164
176
184
185
187
Electron
Sinai
Medicine3
Percentage
of
samples
32.6
63.6
87.7
94.2
98.5
99.0
100.0
Microscopy Analysis
Battelle
Memorial
Number
of
samples
27
60
102
124
125
127
127
Institute
Percentage
of
samples
21.3
47.2
80.1
97.6
98.5
100.0
100.0
Paris, France
Percentage
of
samples
70
85
98
100
.^urces:
(1971); bU.S. EPA (1974); C5ebast1en et al. (1980).
materials was permitted. The practice was especially common 1n New York City.
While no sampling station was known to be located adjacent to an active con-
struction site, unusually high levels could nevertheless have resulted from
this procedure. Other sources that may have contributed to these air concen-
trations Include automobile braking, other construction activities, consumer
use of asbestos products, and maintenance or repair of asbestos-containing
materials (e.g., thermal insulation).
5.3 CHRYSOTILE ASBESTOS CONCENTRATIONS NEAR CONSTRUCTION SITES
To determine if construction activities could be a significant source of
chrysotile fiber in the ambient air, 6- to 8-hour daytime sampling was conducted
in lower Manhattan in 1969 near sites where extensive spraying of asbestos-
containing fireproofing material was taking place. Eight sampling sites were
established near the World Trade Center construction site during the period
when asbestos material was sprayed on the steelwork of the first tower.
147
image:
TABLE 5-2. DISTRIBUTION OF 4- TO 8-HOUR DAYTIME CHRYSOTILE ASBESTOS
CONCENTRATIONS IN THE AMBIENT AIR OF NEW YORK CITY, 1969-1970
Asbestos concentration
(ng/m3) less than
1
2
5
10
20
50
100
Sampl ing locations
Manhattan
Brooklyn
Bronx
Queens
Staten Island
Cumulative number Cumulative percentage
of samples of samples
0
1
4
8
16
21
22
Distribution by
Number of samples
7
3
4
4
4
borough
Asbestos
Range
8-65
6-39
2-25
3-18
5-14
0.0
4.5
18.1
36.4
72.7
95.4
100.0
air level , ng/m3
Average
30
19
12
9
8
Source: Nicholson et al. (1971).
Table 5-3 shows the results of building-top air samples taken at sites within
one-half mile of the Trade Center site, demonstrating that spray fireproofing
did contribute significantly to asbestos air pollution (Nicholson et al.,
1971; Nicholson and Pundsack, 1973). In some instances, chrysotile asbestos
levels were observed that were approximately 100 times greater than the con-
centrations typically found in ambient air.
5.4 ASBESTOS CONCENTRATIONS IN BUILDINGS IN THE UNITED STATES AND FRANCE
During 1974, 116 samples of indoor and outdoor air were collected in 19
buildings (usually 4-6 Indoor samples and 1 ambient air control sample per
building) in 5 U.S. cities to assess whether contamination of the building air
resulted from the presence of asbestos-containing surfacing materials 1n rooms
or return air plenums (Nicholson et al., 1975). The asbestos materials in the
buildings were of two main types: 1) a cementitious or plaster-Uke material
that had been sprayed as a slurry onto steelwork or building surfaces, and
148
image:
TABLE 5-3. DISTRIBUTION OF 6- TO 8-HOUR CHRYSOTILE ASBESTOS
CONCENTRATIONS WITHIN ONE-HALF MILE OF THE SPRAYING OF ASBESTOS MATERIALS
ON BUILDING STEELWORK, 1969-1970
Asbestos concentration Cumulative number Cumulative percentage
(ng/m3) less than of samples of samples
5
10
20
50
100
200
500
0
3
8
14
16
16
17
0.0
17.6
47.1
82.3
94.1
94.1
100.0
Distribution of chrysotile air levels according to distance from
spray fireproofing sites
Sampling locations
1/8-1/4 mile
1/4-1/2 mile
1/2-1 mile
Number of samples
11
6
5
Asbestos air
Range
9-375
8-54
3.5-36
level, ng/m3
Average
60
25
18
Source: Nicholson et al. (1971).
2) a loosely bonded fibrous mat that had been applied by blowing a dry mixture
of fibers and binders through a water spray onto the desired surface. The
friability of the two types of materials differed considerably; the cemen-
titious spray surfaces were relatively impervious to damage while the fibrous
sprays were highly friable. The results of air sampling in these buildings
*
(Table 5-4) provide evidence that the air of buildings with fibrous asbestos-
containing materials may often be contaminated.
Similar data were obtained by Sebastien et al. (1980) in a survey of
asbestos concentration in buildings in Paris, France. Sebastien surveyed 21
asbestos-insulated buildings; 12 had at least one measurement higher than 7
ng/m , the upper limit of the outdoor asbestos concentrations measured by
these investigators. The distribution of 5-day asbestos concentrations in
these buildings, along with 19 outdoor samples taken at the same time, is
shown in Table 5-5. One particularly disturbing set of data by Sebastien et
al. is the concentrations of asbestos measured after surfacing material was
removed or repaired. The average of 22 such samples was 22.3 ng/m . However,
149
image:
en
O
TABLE 5-4. CUMULATIVE DISTRIBUTION OF 8- TO 16-HOUR CHRYSOTILE ASBESTOS
CONCENTRATIONS IN BUILDINGS WITH ASBESTOS-CONTAINING SURFACING MATERIALS
IN ROOMS OR IN AIR PLENUMS
Asbestos
concentration
ng/m3 less than
1
2
5
10
20
50
100
200
500
1000
Arithmetic average
concentration
Friable
Number of
samples
5
6
8
15
28
44
49
52
53
54
spray
Percentage
of samples
9.3
11.1
14.8
27.8
51.9
81.5
90.7
96.3
98.1
100.0
48 ng/m3
Cementitious spray
Number of
samples
3
6
10
17
26
27
27
28
Percentage
of samples
10.7
21.4
35.7
60.7
92.9
96.4
96.4
100.0
14.5 ng/m3
Control
Number
5
6
15
21
29
33
34
sampl es
Percentage
14.7
17.6
44.1
61.8
85.3
97.1
100.0
12.7 ng/m3
Source: Nicholson et al. (1975; 1976).
image:
TABLE 5-5. CUMULATIVE DISTRIBUTION OF 5-DAY ASBESTOS CONCENTRATIONS
IN PARIS BUILDINGS WITH ASBESTOS-CONTAINING SURFACING MATERIALS
Asbestos concentration
(ng/m3) less than
Bui1ding samp1es
Numberpercentage
Outdoor control samples
NumberPercentage
Chrysotlle
1
2
5
10
20
50
100
200
500
1000
Arithmetic average
concentration
57
70
92
104
117
128
129
130
132
135
42.2
51.9
68.1
77.0
86.7
94.8
,6
,3
95.
96.
97.8
100.0
25 ng/m3
14
16
17
19
73.7
84,2
89.5
100.0
ng/ma
AmpM boles'
1
2
5
10
20
50
100
200
500
Arithmetic average
concentration
112
115
122
125
129
131
132
133
135
83.0
85.
90,
92.
95.
97.
97.8
98.5
100.0
10 ng/m3
19
100.0
0.1 ng/m3
No value reported for 104 building samples. Some materials would have con-
tained no amphlbole asbestos.
Source: Sebastien et al. (1980).
151
image:
in two highly contaminated areas, significant reductions were measured (500 to
3 3
750 ng/m decreased to less than 1 ng/m ). The importance of proper removal
techniques and cleanup cannot be overemphasized.
i j
Sebastien et al. (1982) also measured concentrations of indoor airborne
asbestos up to 170 ng/m in a building with weathered asbestos floor tiles.
Asbestos flooring is used in a large number of buildings and is the third
largest use of asbestos fibers.
5.5 ASBESTOS CONCENTRATIONS IN U.S. SCHOOL BUILDINGS
Of concern was the discovery of extensive asbestos use in public school
buildings (Nicholson et al., 1978). Asbestos surfaces were found in more than
10 percent of pupil-use areas in New Jersey schools, with two-thirds of the
surfaces showing some evidence of damage. Because these values appear to be
typical of conditions in many other states, it was estimated that 2 to 6 million
pupils and 100,000 to 300,000 teachers may be exposed to released asbestos fibers
in schools across the nation. To obtain a measure of contamination for this use
of asbestos, 10 schools were sampled in the urban centers of New York and New
Jersey and in suburban areas of Massachusetts and New Jersey. Schools were
selected for sampling because of visible damage, 1n some cases extensive.
Table 5-6 lists the distribution of chrysotile concentrations found in
samples taken over 4 to 8 hours in these 10 schools (1-5 samples per school).
3 3
Chrysotile asbestos concentrations ranged from 9 ng/m to 1950 ng/m , with an
average of 217 ng/m . Outside air samples at 3 of the schools varied from 3
33 3
ng/m to 30 ng/m , with an average of 14 ng/m . In all samples but two (which
measured 320 ng/m ) no asbestos was visible on the floor of the sampled area,
although surface damage was generally present near the area. The highest
value (1950 ng/m ) was in a sample that followed routine sweeping of a hallway
in a school with water damage to the asbestos surface, although no visible
asbestos was seen on the hallway floor. It is emphasized that the schools
were selected in testing on the basis of the presence of visible damage.
Although the results cannot be considered typical of all schools having
asbestos surfaces, the results do illustrate the extent to which contamination
can exist.
A recent study suggests that the above school samples may not be atypical
(Constant et al., 1983). Concentrations similar to those Indicated above
were found in the analysis of samples collected during a 5-day period in 25
152
image:
TABLE 5-6. DISTRIBUTION OF CHRYSOTILE ASBESTOS CONCENTRATIONS IN
4- to 8-HOUR SAMPLES TAKEN IN PUBLIC SCHOOLS WITH DAMAGED ASBESTOS SURFACES
Asbestos
(ng/m3
concentration
) less than
5
10
20
50
100
200
500
1000
2000
Number of samples
0
1
1
6
12
19
25
26
27
Percentage of samples
0.0
3.7
3.7
22.2
44.4
70.4
92.6
96.3
100.0
Source: Nicholson et al. (1978).
schools that had asbestos surfacing materials. The schools were in a single
district and were selected by a random procedure, not because of the presence
or absence of damaged material. A population-weighted arithmetic mean concen-
3
tration of 179 ng/m was measured in 54 samples collected in rooms or areas
3
that had asbestos surfacing material. In contrast, a concentration of 6 ng/m
was measured in 31 samples of outdoor air taken at the same time. Of special
concern are 31 samples collected in the schools that used asbestos, but taken
in areas where asbestos was not used. These data showed an average concentra-
tion of 53 ng/m , indicating dispersal of asbestos from the source. The data
are summarized in Table 5-7. As published fiber counts were fibers of all
sizes, only the fiber mass data are listed in the table. Additionally, fiber
clumps were noted in many samples, but were not included in the tabulated
masses.
A study commissioned by the Ontario Royal Commission (1984) of asbestos
concentrations in buildings with asbestos insulation indicates levels comparable
to that of urban air. It is not clear whether "insulation" is thermal insula-
tion or sprayed surfacing material. Average concentrations (3-5 samples per
building) ranged from less than 1 to 11 ng/m . However, during very careful
maintenance and inspection work, concentrations substantially in excess of
background were observed.
Sawyer (1977, 1979) reviewed a variety of data on air concentrations,
measured by optical microscopy, for circumstances where asbestos materials in
schools and other buildings are disturbed by routine or abnormal activity.
153
image:
TABLE 5-7. CUMULATIVE DISTRIBUTION OF 5-DAY CHRYSOTILE ASBESTOS CONCENTRATIONS IN
25 SCHOOLS HAVING ASBESTOS SURFACING MATERIALS, 1980-1981
Asbestos
concentration
(ng/m3) less than
1
2
5
10
20
50
100
200
500
1000
Population weighted
mean concentration
1
2
5
10
20
50
100
500
Arithmetic mean
concentration
Rooms with
Number of
samples
5
6
7
in
19
26
39
45
52
54
44
45
49
50
52
52
54
asbestos
Percentage
of samples
9.2
11.1
13.0
25.9
35.2
48.1
72.2
83.3
96.3
100.0
179 ng/m3
81. 5
83.3
90.7
92.6
96.3
96.3
100.0
3.6 ng/m3
Rooms without
Number of
samples
Chrysotile
6
7
11
12
15
21
24
29
31
Amphiboles
21
22
26
27
27
29
31
asbestos
Percentage
of samples
19.4
22.6
35.5
38.7
48.4
67.7
87.1
93.5
100.0
53 ng/m3
67.7
-71.0
83.9
87.1
87.1
93.5
100.0
8.3 ng/m3
Outdoor
Number of
samples
17
22
27
28
30
31
26
29
31
controls
Percentage
of samples
54.8
71.0
87.1
90,3
96.8
100.0
6 ng/m3
83.9
93.5
100.0
0.5 ng/m3
Source: Constant et al. (1983).
image:
These results, shown in Table 5-8, demonstrate that a wide variety of activi-
ties can lead to" high asbestos concentrations during disturbance of asbestos
surfacing material. Maintenance and renovation work, particularly 1f performed
Improperly, can lead to substantially elevated asbestos levels.
TABLE 5-8. AIRBORNE ASBESTOS IN BUILDINGS HAVING
FRIABLE ASBESTOS MATERIALS
Classification
quiet, non-
specific,
routine
Maintenance
Custodial
Renovation
Vandal ism
Mean
count of
Main mode of Activity fibers per
contamination description cm3 n
Fallout None
reentralnment Dormitory
University, schools
Offices
Contact Re lamp ing
Plumbing
Cable movement
Mixed: contact
reentrainment Cleaning
Dry sweeping
Dry dusting
Bystander
Heavy dusting
Mixed: contact Ceiling repair
reentralnment Track light
Hanging light
Partition
Pipe lagging
Contact Celling damage
0.0
0.1
0.1
0.2
1.4
1.2
0.9
15.5
1.6
4.0
0.3
2.8
17.7
7.7
1.1
3.1
4.1
12.8
32
NA
47
14
2
6
4
3
5
6
3
8
3
6
5
4
8
5
Range
or SD
0.0
0.0-0.
0.1
0.1-0.
0.1
0.1-2.
0.2-3.
6.7
0.7
1.3
0.3
1.6
8.2
2.9
0.8
1.1
1.8-5.
8.0
8
6
4
2
8
Source: Sawyer (1979).
5.6 CHRYSOTILE CONCENTRATIONS IN THE HOMES OF WORKERS
The finding of asbestos disease in family contacts of individuals occupa-
tional^ exposed to chrysotile fibers directs attention to air concentrations
1n the homes of such workers. Thirteen samples were collected In the homes of
asbestos mine and mill employees and analyzed for chrysotile (Nicholson et
al. , 1980). The workers were employed at mine operations in California and
Newfoundland. At the time of sampling (1973 and 1976) they did not have
155
image:
access to shower facilities nor did they commonly change clothes before going
home. Table 5-9 lists the concentration ranges of the home samples. Three
samples taken in homes of non-miners in Newfoundland yielded concentrations of
3
32, 45, and 65 ng/m . In contrast, the concentrations In workers' homes were
much higher, pointing to the need for appropriate shower and change facilities
at asbestos workplaces. Because asbestos-generated cancers have' been documented
in family contacts of workers, concentrations such as those described in this
document should be viewed with particular concern.
TABLE 5-9. DISTRIBUTION OF 4-HOUR CHRYSOTILE ASBESTOS CONCENTRATIONS
IN THE AIR OF HOMES OF ASBESTOS MINE AND MILL EMPLOYEES
Asbestos concentration
(ng/m3) less than Number of samples Percentage of samples
50
100
200
500
1000
2000
5000
0
4
8
10
12
12
13
0.0
30.8
61.5
76.9
92.3
92.3
100.0
Source: Nicholson et al. (1980).
5.7 SUMMARY OF ENVIRONMENTAL SAMPLING
Table 5~10 summarizes those studies of the general ambient air or of
specific pollution circumstances that have a sufficient number of samples for
comparative analysis. The data are remarkably consistent. Average 24-hour
samples of general ambient air indicate asbestos concentrations of 1 to 2
ng/m (two U.S. samples that may have been affected by specific sources were
not included). Short-term daytime samples are generally higher, reflecting
the possible contributions of traffic, construction, and other human activi-
ties. In buildings having asbestos surfacing materials, average concentrations
100 times greater than ambient air are seen in some schools and concentra-
tions 5-30 times greater than ambient air are seen in some other buildings.
Figure 5-2 shows the cumulative distributions, on a log-probability plot,
of the urban, school, and building samples. The straight lines in the data of
Nicholson are suggestive of homogeneous sampling circumstances, but this may
be fortuitous. The sampling situation of Constant et al. appears not to be
homogeneous.
156
image:
TABLE 5-10. SUMMARY OF ENVIRONMENTAL ASBESTOS SAMPLING
Mean
Collection Number Concentration,
Sample set period of samples ng/m3
Quarterly composites of 5 to 7 1969-70 187 3.3 Ca
24-hour U.S. samples (Nicholson,
1971; Nicholson and Pundsack, 1973)
Quarterly composite of 5 to 7 1969-70 127 3.4C
24-hour U.S. samples
(U.S. EPA, 1974)
5-day samples of Paris, France 1974-75 161 0.96 C
(Sebastien et al., 1980)
6- to 8-hour samples of New York 1969 22 16 C
City (Nicholson et al., 1971)
5-day, 7-hour control samples 1980-81 31 6.5 (6C, 0.5Ab)
for U.S. school study (Constant
et al., 1983)
16-hour samples of 5 U.S. 1974 34 13 C
cites (U.S. EPA, 1974)
New Jersey schools with damaged 1977 27 217 C
asbestos surfacing materials 1n
pupil use areas (Nicholson et al.,
1978)
U.S. school rooms/areas with 1980-81 54 183 (179C, 4A)
asbestos surfacing material
(Constant, 1983)
U.S. school rooms/areas in 1980-81 31 61 (53C, 8A)
building with asbestos
surfacing material
(Constant, 1983)
Buildings with a'sbestos 1976-77 135 35 (25C, 10A)
materials in Paris, France
(Sebastien et al., 1980)
U.S. buildings with friable 1974 54 48 C
asbestos in plenums or as
surfacing materials (Nicholson
et al., 1975; Nicholson et al.,
1976)
U.S. buildings with cementi- 1974 28 15 C
tious asbestos material in
plenums or as surfacing materials
(Nicholson et al., 1975, 1976)
Ontario buildings with asbestos 1982 63 2.1
insulation (Ontario Royal
Commission, 1984)
aC = chrysotile. A = amphibole.
157
image:
z
o
O
Z
o
o
o
01
s
u
5
z
CO
CO
co
ui
a.
S
(O
u.
O
u
QC
UJ
a.
50 U.S. CITIES (NICHOLSON (1971)
CONSTANT
ETAL, (1983)
ASBESTOS
IN SCHOOL
ASBESTOS
IN ROOM
SCHOOLS WITH
DAMAGED
SURFACES
NICHOLSON
ETAL, (1978)
1234567
NATURAL LOGARITHM OF ASBESTOS CONCENTRATION, ng/m3
Figure 5-2. Cumulative distribution, on a log probability plot,
of urban, school, and building asbestos air concentrations.
158
image:
5.8 OTHER EMISSION SOURCES
Weathering of asbestos cement wall and roofing materials was shown to be
a source of asbestos air pollution by analyzing air samples taken 1n buildings
constructed of such material (Nicholson, 1978). Seven samples taken 1n a
school after a heavy rainfall showed asbestos concentrations from 20-4500
33 3
ng/m (arithmetic mean = 780 ng/m ); all but two samples exceeded 100 ng/m .
The source was attributed to asbestos washed from asbestos cement walkways and
asbestos cement roof panels. No significantly elevated concentrations were
observed 1n a concurrent study of houses constructed of asbestos cement mate-
rials. "Roof water runoff from the homes landed on the ground and was not
reentralned, while that of the schools fell to a smooth walkway, which allowed
easy reentrainment when dry. Contamination from asbestos cement siding has
also been documented by Spurny et al. (1980).
One of the more significant remaining contributions to environmental
asbestos concentrations may be emissions from braking of automobiles and other
vehicles. Measurements of brake and clutch emissions reveal that, annually,
2.5 tons of unaltered asbestos are released to the atmosphere and an addi-
tional 68 tons fall to roadways, where some of the'asbestos is dispersed by
passing traffic (Jacko et al., 1973).
5.9 INTERCONVERTIBILITY OF FIBER AND MASS CONCENTRATIONS
The limited data that relate asbestos disease to exposure are derived
from studies of workers exposed 1n occupational environments. In these studies,
concentrations of fibers longer than 5 [jm were determined using optical micros-
copy or they were estimated from optical microscopy measurements of total
particulate matter. All current measurements of low-level environmental pol-
lution utilize electron microscopy techniques, which determine the total mass
of asbestos present in a given volume of air. In order to extrapolate dose-
response data obtained in studies of working groups to environmental exposures,
it is necessary to establish a relationship between optical fiber counts and
the mass of asbestos determined by electron microscopy.
Data are available relating optical fiber counts (longer than 5 urn) to the
total mass of asbestos, as determined by electron microscopy techniques or
other weight determinations. These relationships (Table 5-11) provide crude
159
image:
TABLE 5-11. MEASURED RELATIONSHIPS BETWEEN OPTICAL FIBER COUNTS
AND MASS AIRBORNE CHRYSOTILE
Sampling situation
Fiber"
counts
f/ml
Mass
concentration
(jg/m3
Conversion factors
|jg/m3 or ug
fTiiiT ICFf 103 f/ng
Textile factory
British Occupational
Hygiene Society
(1968) (weight vs.
fiber count)
Air chamber monitoring
Davis et al. (1978)
Monitoring brake
repair work
Rohl et al. (1976)
Electron Microscopy
1950
120
10,000
60
5
16
200
(E.M. mass vs.
fiber count)
Textile mill
Lynch et al. (1970)
0.1 to 4.7 0.1 to 6.6
(7 samples)
Friction products manufacturing
Lynch et al. (1970)
Pipe manufacturing
Lynch et al. (1970)
0.7 to 24U
mean = 6
150C
70C
45C
170
6.7
13.9
22.5
All fiber counts used phase-contrast microscopy and enumerated fibers longer than 5 urn.
Conversion factor may be low due to losses in electron microscopy processing.
GConversion factor may be high because of overestimate of asbestos mass on the basis of
total magnesium.
image:
estimates of a conversion factor relating fiber concentration 1n fibers per
mill niter (f/ml) to airborne asbestos mass 1n mlcrograms per cubic meter
(ug/m ). The proposed standards for asbestos In Great Britain, set by the
British Occupational Hygiene Society (BOHS), states that a "respirable" asbestos
mass of 0.12 mg/m 1s equivalent to 2 f/ml (British Occupational Hygiene
Society, 1968). The standard does not state how this relationship was deter-
mined. If the relationship was obtained from magnesium determinations 1n an
aerosol, the weight determination would likely be high because of the presence
of other nonflbrous magnesium-containing compounds 1n the aerosol. Such was
the case 1n the work of Lynch et al. (1970), and their values for the conversion
factor are undoubtedly overestimates. The data of Rohl et al. (1976) are
likely to be underestimates because of possible losses In the determination of
mass by electron microscopy. No Information exists on the procedures used to
determine the mass of chrysotile in the data presented by Davis et al. (1978).
The range of 5 to 150 for the conversion factor relating mass concen-
tration to optical fiber concentration 1s large and any average value derived
from It has a large uncertainty. However, for the purpose of extrapolating to
low mass concentrations from fiber count, the geometric mean of the above
range of conversion factors, 30 ug/m /f/ml, will be used. The geometric
standard deviation of this value 1s 4, and this uncertainty severely limits
i
any extrapolation 1n which 1t 1s used. .In the case of amosite, the data of
Davis et al. (1978) suggest that a conversion factor of 18 1s appropriate.
However, these data yield lower chrysotile values than all other chrysotile
estimates; therefore, they may also be low for amosite.
5.10 SUMMARY
Measurements using electron microscopy techniques established the presence
i
of asbestos 1n the urban ambient air, usually at concentrations less than 10
ng/m . Concentrations of 100 ng/m to 1000 ng/m~ were measured near specific
asbestos emission sources, in schools where asbestos-containing materials are
used for sound control, and 1n office buildings where similar materials are
used for fire control. Excess concentrations in buildings have usually been
associated with visible damage or erosion of the asbestos materials. Many
buildings with Intact material have no Increased concentrations of asbestos.
Most ambient measurements were taken over ten years ago and it is very Important
to obtain more current data.
161
image:
6. RISK EXTRAPOLATIONS AND HUMAN EFFECTS OF LOW EXPOSURES
6.1 RISK EXTRAPOLATIONS FOR LUNG CANCER AND MESOTHELIOMA
To obtain dose-response estimates at current or projected environmental
asbestos concentrations, it is necessary to extrapolate from epidemiological
data on deaths that have resulted from exposures to the considerably higher
concentrations extant in occupational circumstances. As mentioned previously,
the available data are compatible with a linear exposure-response relation-
ship, with no evidence of a threshold. However, the limited data that indi-
cate the validity of this relationship are for exposures two or three orders
of magnitude higher than those of concern for environmental exposures.
The values determined for K, and KM in Chapter 3 are used to calculate
best estimate risks from continuous exposures to 0.0001 and 0.01 f/ml. The
values rfor continuous exposure were derived by multiplying 40 hr/wk risks,
obtained from occupational exposures, by 4.2 (the ratio of hours in a week to
40 hours.) The lower concentration is typical of urban ambient air and corres-
3
ponds to about 3 ng/m . The higher concentration, corresponding to about 300
ng/m ( was measured in several environmental exposure circumstances. These
two examples provide unit risks from which risk at other continuous exposures
can be calculated as needed.
Tables 6-1, 6-2, and 6-3 list the calculated lifetime risks of meso-
thelioma and lung cancer for continuous exposures to 0.0001 and 0.01 f/ml of
asbestos for various time periods. Risks from longer or shorter exposures can
be estimated by directly scaling the data in the tables, as can risks from
other concentrations (i.e., 0.1 f/ml). Equations 3-3a, 3-6c, 3-6d, and 3-6e
_ 9 _o
and values of K, = 1.0 x 10 and KH = 1.0 x 10 were used in these calcula-
tions. The calculation uses a lifetable approach, in which the hypothetical
population at risk is continuously decreased by its calculated mortality from
all causes. Different overall mortality rates for smokers and non-smokers, as
well as for males and females, lead to different estimated mesothelioma risks
by smoking and gender, in Tables 6-1, 6-2, and 6-3. In the calculation of lung
cancer risk it was assumed that the calculated asbestos-related risk continue
following cessation of any hypothetical exposure. U.S. 1977 mortality rates
(National Center for Health Statistics, 1977) are used as the basic data for
the calculation. The tables utilize both smoking specific (Tables 6-1 and
162
image:
TABLE 6-1. LIFETIME RISKS PER 100,000 FEMALES OF DEATH FROM
MESOTHELIOMA AND LUNG CANCER FROM CONTINUOUS ASBESTOS EXPOSURES OF 0.0001 AND Q.01 f/ml
ACCORDING TO AGE AT FIRST EXPOSURE, DURATION OF EXPOSURE, AND SMOKING3
Concentration - 0.0001 f/ml
years of exposure
Age at onset life-
of exposure 1 5 10 20 time
Mesothelioma
0
10
20
30
50
0
10
20
30
50
0
0
0
0
0
0
.1
.1
.1
,0
.0
.0
0.0
0
0
0
.0
.0
.0
0.6
0.4
0.2
0.1
0.0
0.1
0.1
0.1
0.1
0.1
1.2
0.7
0.4
0.2
0.0
Lung
0.3
0.3
0.3
0.3
0.2
1.9
1.1
0.6
0.3
0.0
Cancer
0.5
0.5
0.5
0.5
0.2
Mcsothel ioma
0
10
20
30
50
0
0
0
0
0
.1
.1
.1
.0
.0
0.7
0.4
0.3
0.1
0.0
1.2
0.8
0.4
0.2
0.0
2.0
1.2
0.7
0.3
0.0
1
Concentration = 0.01 f/ml
years of exposure
1 i f e-
5 10 20 time
in Female Smokers
2.5
1.4
0.7
0.3
0.0
13.9
9.0
5.3
2.8
0.6
64.0
40.3
23.5
12.3
2.0
115,1
71.4
40,7
20.6
2.9
186.2
112.0
61.3
29.4
3.5
252.0
142.8
72.8
32.8
3.5
in Female Smokers
1.5
1.2
1,0
0.'7
0.2
in Female
2.7
1.6
0.8
0.4
0.0
Lung Cancer in Female
0
10
20
30
50
0
0
0
0
0
.0
.0
.0
.0
.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.1
0.1
0.1
0.0
0.2
0.1
0.1
0.1
0.0
2.8
2.8
2.8
2.8
2.0
13,4
13.4
13.4
13.3
8.8
26.7
26.7
26.7
25.9
15.5
53.3
53.3
52.5
47.9
22.7
149.9
123.5
96.9
71.0
24.4
Nonsmokers
14.8
9.5
5.7
3.1
0.6
68.2
43.4
25.6
13.6
2.2
122.8
81.2
44.4
23.0
3.4
199.4
121.2
67.2
32.9
4.1
272.2
155.8
80.6
36.8
4.1
Nonsmokers
0.3
0.3
0.3
0.3
0,3
1,3
1.3
1.3
1.3
1.1
2.7
2.7
2.7
2.7
2.1
5.2
5.3
5.2
5.0
3.5
16.4
13.9
11.3
8.7
3.9
aThe 95% confidence limit on the risk values for lung cancer for an unstudied exposure cir-
cumstance is a factor of 10. The 95% confidence limit en the risk values for lung cancer on
the average determined from 11 unit exposure risk studies is a factor of 2.5. The 95% con-
fidence limit on the risk values for mesothelioma for an unstudied exposure circumstance is
a factor of 20. The 95% confidence limit on the risk values for mesothelioma for a studied
circumstance can be reasonably averaged as a factor of 5. The values for continuous expo-
sure were derived by multiplying 40 hr/wk risks, obtained from occupational exposures, by
4.2 (the ratio of hours in a week to 40 hours.)
163
image:
TABLE 6-2. LIFETIME RISKS PER lOO'.OOO MALES OF DEATH FROM
MESOTHELIOMA AND LUNG CANCER FROM CONTINUOUS ASBESTOS EXPOSURES OF 0.0001 AND 0.01 f/ml
ACCORDING TO AGE AT FIRST EXPOSURE, DURATION OF EXPOSURE, AND SMOKING
Concentration = 0.0001 f/ml
years of exposure
Age at onset life-
of exposure 1 5 ID 20 time
Mesothel ioma
0
10
20
30
50
0.1
0.1
0.0
0.0
0.0
0.5
0.3
0.2
0.1
0.0
0.
0.
0.
0.
0.
9
5
3
1
0
1.4
0,8
0.4
0.1
0.0
Lung Cancer
0
10
20
30
50
0.0
0.0
0.0
0.0
0.0
0.2
0.2
0.2
0.2
0.2
0.
0.
0.
0,
0.
4
4
4
4
3
0.8
0.8
0.8
0.8
0.4
Mesothel ioma
0
10
20
30
50
0
10
20
30
50
0.1
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.6
0.4
0.2
0.1
0.0
0.0
0.0
0.0
0.0
0.0
1.
0,
0,
0.
0,
0,
0,
0,
0
.0
,6
.4
.2
.0
Lung
.0
.0
.0
.0
0.0
1.6
1.0
0.5
0.2
0.0
in Male
1.8
1.0
0.5
0.2
0.0
in Male
2.4
2.0
1.6
1.2
0.4
in Male
2.2
1.2
0.6
0.3
0.0
1
Smokers
10.6
6.6
3.6
2.0
0.3
Smokers
4.2
4.2
4.2 .
4.2
3.6
Concentration = 0.01 f/ml
years of exposure
life-
5 10 20 time
48.3
29.4
16.4
8.1
1.1
20.9
21.0
21.3
21.3
16.2
85.5
51.5
28.0
13.4
1.5
41.9
42.0
42.3
42.0
28.4
137.5
77.8
41.2
18.5
1.8
83.4
83.9
83.4
79.2
40.3
181.0
98,3
47.9
20.2
1.8
238.1
197.8
157.5
117.6
42.0
Nonsmokers
12.5
7,8
4.5
2.4
0.4
57.0
35.3
20.4
10.5
1.5
102.3
62,6
35.1
17.5
2.2
164.5
97.3
52.4
24.6
2.7
220.1
122.6
61.7
26.9
2.7
Cancer in Male Nonsmokers
0.0
0.1
0.1
0.1
0.0
0.2
0.2
0.1
0.1
0.0
D.3
0.3
0.3
0.3
0.3
1.5
1.5
1.5
1.5
1.3
2.9
2.9
2.9
2.9
2.2
5.9
5.9
5.9
5.7
3.9
18.5
15.5
12.6
9.7
4.2
The 95% confidence limit on the risk values for lung cancer for an unstudied exposure cir-
cumstance is a factor of 10. The 95% confidence limit on the risk values for lung cancer on
the average determined from 11 unit exposure risk studies is a factor of 2.5. The 95% con-
fidence limit on the risk values for mesothelioma for an unstudied exposure circumstance is
a factor of 20. The 95% confidence limit on the risk values for mesothelioma for a studied
circumstance can be reasonably averaged as a factor of 5. The values for continuous expo-
sure were derived by multiplying 40 hr/wk risks, obtained from occupational exposures, by
4.2 (the ratio of hours in a week to 40 hours.)
164
image:
TABLE 6-3. LIFETIME RISKS PER 100,000 PERSONS OF DEATH FROM
MESOTHELIOMA AND LUNG CANCER FROM CONTINUOUS ASBESTOS EXPOSURES OF D.0001 AND 0.01 f/ml
ACCORDING TO AGE AND DURATION OF EXPOSURE. U.S. GENERAL POPULATION
DEATH RATES WERE USED AND SMOKING HABITS WERE NOT CONSIDERED3
Concentration = 0.0001
years of exposure
Age at onset
of exposure 1 5 10 20
f/ml
life-
time
1
Concentration = 0.01 f/ml
years of exposure
life-
5 10 20 time
Mesothelioma in Females
0
10
20
30
50
0.1
0.1
0.1
0.0
0.0 ,
0.7
0.4
0.3
0.1
0.0
1.2
0.8
0.4
0.2
0.0
2.0
1.2
0.7
0.3
0.0
2.8
1.5
0.8
0.4
0.0
Lung Cancer in
0
10
20
30
50
0.0
0.0
0.0
0.0
D.O
0.0
0.0
0,0
0.0
0.0
0.1
0.1
D.I
0.1
0.1
0.2
0.2
0.2
0.2
0.1
0.5
0.4
0.3
0.3
0.1
Mesothelioma
0
10
20
30
50
0
10
20
30
5D
0.1
0.1,
. . o'. o
0.0
D.D
0.0
0.0
0.0
0.0
0.0
0.5
0.3
0.2
0.1
0.0
0.1
0.1
0.2
0,1
D.I
, 0.9
0.6
0.3
0.1
0.0
0.3
0,3
0.3
0.3
0.2
1.5
0.8
0.4
0.2
0.0
Lung
0.6
0.6
0.6
0.6
0.3
1.9
1.1
0.5
D.2
O.D
14.6
9.4
5.6
3.1
0.6
Females
1.0
1.0
1.0
1.0
0.7
in Males
11.2
7.0
4.1
2.1
0.3
67.1
42.6
25.1
13.3
2.1
4.6
4.6
4.6
4.6
3.1
51.0
31.2
17.5
8.8
1.1
120.8
75.5
43.5
22.4
3.2
9.2
9.2
9.2
9.0
5.5
91.1
58.2
30.1
14.6
1.8
196.0
118.7
65.7
31.9
3.9
IB. 5
IB. 6
IB. 2
16.7
8.1
145.7
84.7
44.5
20.4
2.D
275.2
152,5
7B.8
35.7
3.9
52.5
43.4
34.3
25.1
8.8
192.8
106.8
51.7
22.3
2.1
Cancer in Males
1.7
1.4
1.1
0.8
0.3
2.9
2.9
3.1
3.1
2. 5
14.8
14.9
15.0
14.9
11.5
29.7
29. B.
30.0
29. S
20.3
59,2
59.5
59.4
56.6
29.1
170.5
142.0
113.0
84.8
30.2
The 95% eonfidence limit on the risk values for lung cancer for an unstudied exposure cir-
cumstance is a factor of 10. The 953! confidence limil. on the risk values for lung cancer on
the average determined from 11 unit exposure risk studies is a factor of 2.5. The 95% con-
fidence limit on the risk values for mesothelioma for an unstudied exposure circumstance is
a factor of 20. The 95% confidence limit on the risk values for mesothelioma for a studied
circumstance can be reasonably averaged as a factor of 5. The values for continuous expo-
sure were derived by multiplying 40 hr/wk risks, obtained from occupational exposures, by
4.2 (the ratio of hours in a week to 40 hours.)
165
image:
6-2) and general population (Table 6-3) rates. We are assuming that the
current U.S. male mortality rates reflect the experience of 67 percent smokers
(many, however, are now ex-smokers) and that current female mortality rates
reflect the experience of 33 percent smokers. Using these percentages and the
data of Hammond (1966) on the mortality ratio of smokers to nonsmokers, smoking-
specific total mortality rates are calculated. Current lung cancer mortality
rates for males are multiplied by 1.5 to represent the rates for smoking
males. The multiplication factor comes from the fact that the current male
rates result from a population where 67 percent of men are smokers or ex-
smokers. Correspondingly, current female lung cancer mortality rates are
multiplied by 3 to reflect the fact that approximately 33 percent of women are
current or ex-smokers. This factor for women may be low, because the current
rapid increase in female rates may not yet fully reflect the full impact of
women's smoking; however, they should not exceed the male smoker's rates.
Nonsmoking lung cancer rates for both males and females are taken from Garfinkel
(1981).
The results show the importance of the time course of mesothel ioma.
Children exposed at younger ages are especially susceptible because of their
long life expectancy. The time of exposure plays little role in the lifetime
excess risk of lung cancer; any exposure before the age of 45 or 50 contributes
equally to the lifetime risk. The risk estimates are uncertain because of the
variability of the data from which values of K, are calculated and from uncer-
tainties in extrapolating from risks estimated at high occupational exposures
to concentrations 1/100 and less. Thus, actual risks in a given environmental
exposure could be outside the listed ranges.
The risks in tables 6-1, 6-2, and 6-3 would appear to be the best esti-
mates for exposure to fibers released from the variety of asbestos products
used in the United States, including products containing small amounts of
crocidolite and substantial quantities of amosite. As noted in the tables,
the 95 percent confidence limits on the risk estimate for an unstudied exposure
circumstance are a factor of times 1/10 and times 10. As indicated in section
3.17, exposures to crocidolite appear to carry a proportionately greater
mesothelioma risk. Thus tables 6-1, 6-2, and 6-3 will likely underestimate
(by perhaps a factor of 4) the mesothelioma risk to aerosols containing predomi-
nantly crocidolite asbestos. Conversely, in some pure chrysotile exposure
circumstances (such as in mining and milling), the risk will be overestimated.
166
image:
6.1.1 Alternative Analyses
As discussed previously, the data strongly support a relative risk model
for lung cancer and a linear dose-response relationship. No data indicate the
existence of a threshold, although one cannot be ruled out.
If a threshold does exist, there would be a corresponding reduction in
the calculated lung cancer risk. There is no evidence of a quadratic term in
the dose-response relationship nor is it indicated by existing models for
asbestos lung cancer. If, however, a small quadratic term is present, there
would be some reduction in the calculated risk.
Alternative models do exist for mesothelioma. There are uncertainties in
the power of time at which mesothelioma risk increases. The uncertainty,
however, has relatively little effect on calculated lifetime risk values,
because a fit must be made to existing occupational risk over a time span of
four or five decades, leaving only two or three decades of life for manifesta-
tion of different power function effects. A lower power requires a much
greater multiplying coefficient. Table 6-4 shows the effect on the calculated
lifetime risk of three different time functions that are matched to best fit
the time course of risk among insulation workers. Table 6-4 shows that the
extremes of effect differ by less than a factor of two. As was shown in
Table 3-4, there is very little empirical evidence for quadratic or higher
terms in the mesothelioma dose-response relationship, although they are compat-
ible with existing cancer models. If higher than linear terms were present,
they would reduce the calculated risks by less than a factor of two.
TABLE 6-4. COMPARISON OF THE EFFECT OF DIFFERENT MODELS FOR THE
TIME COURSE OF MESOTHELIOMA RISK FOR A FIVE-YEAR EXPOSURE TO 0.01 F/ML
Age at onset
of exposure
0
10
20
30
50
Eq. 3-6
51.0
31.2
17.5
8.8
1.1
Calculated deaths/100,000 males
t5
76.0
38.0
17.5
7.0
1.0
ts-z
46.0
27. 2
15.0
7.0
1.0
167
image:
6.2 OBSERVED ENVIRONMENTAL ASBESTOS DISEASE
Asbestos-related disease in persons who have not been directly exposed at
the workplace has been reported since 1960. In that year, Wagner et al.
(1960) published a review of 47 cases of mesothelioma found in the Northwest
Cape Province of South Africa in the previous 5 years. Approximately half of
the cases described were in individuals who, decades before, had lived or
worked near an area of asbestos mining. The hazard from environmental asbestos
exposure was further documented in the findings of Newnouse and Thomson (1965),
showing that mesothelioma could occur among individuals whose potential asbes-
tos exposure consisted of having resided near an asbestos factory or in the
household of an asbestos worker; 20 of 76 cases from the files of the London
Hospital were the resul.t of such exposures.
Of considerable importance are data on the prevalence of X-ray abnormali-
ties and the incidence of mesothelioma in family contacts of amosite factory
employees in Paterson, New Jersey. Anderson and Selikoff (1979) showed that
35 percent of 685 family contacts of former asbestos factory workers had
abnormalities characteristic of asbestos exposure when they were X-rayed 30 or
so years after their first household contact. The data, shown in Tables 6-5
and 6-6, compare the household group with 326 New Jersey urban residents. The
overall difference in the percentage of abnormalities between the two groups
is highly significant. Of special concern is the finding that the difference
in the prevalence of abnormalities in a group of children born into a worker's
household after his employment ceased is also significant.
Four mesothelioma cases also occurred among the family contacts of these
same factory workers (Anderson et al., 1976). Table 6-7 lists the cases by
time from onset of exposure, along with the number of deaths from other causes
in the same time period (1961-1977; one death occurred subsequent to 1977).
One percent of the deaths after 20 years from first exposure were from mesothe-
lioma; however, further observations will be necessary to fully establish the
incidence of this neoplasm among family contacts. An additional contribution
of asbestos-related lung cancer could also exist, but studies in this regard
have not yet been completed.
A second population-based mortality study of mesothelioma and other
cancer risks in environmental circumstances is that of Hammond et al. (1979b).
This study compared the mortality of a group of 1779 residents within 0.5 mile
of the Paterson amosite asbestos plant with 3771 controls in a different, but
168
image:
TABLE 6-5. PREVALENCE OF RADIOGRAPHIC ABNORMALITIES ASSOCIATED WITH ASBESTOS
EXPOSURE AMONG HOUSEHOLD MEMBERS OF AMOSITE ASBESTOS WORKERS
Exposure group
New Jersey urban res 1 dents**
Entered household after active
worker employment ceasedt
Household resident during active
worker employmentt
Household resident and personal
occupational asbestos exposure
Total
examined
326
40
685
51
One or more radlographic
abnormalities present*
15 ( 5%)
6 (15%) x2 = 7.1 p <.01
240 (35%) x2 = H4 p <.001
23 (45%)
*ILO U/C Pneumoconlosls Classification categories; Irregular opacities 1/0
or greater; pleural thickening; pleural calcification; pleural plaques.
**No known direct occupational or household exposure to asbestos.
tNo known direct occupational exposure to asbestos.
Source: Anderson and Selikoff (1979).
TABLE 6-5. CHEST X-RAY ABNORMALITIES AMONG 685 HOUSEHOLD CONTACTS OF
AMOSITE ASBESTOS WORKERS AND 326 INDIVIDUAL RESIDENTS IN
URBAN NEW JERSEY, A MATCHED COMPARISON GROUP
Pleural Pleural Pleural Irregular*
Total thickening calcification plaques opacities
Group examined present present present present
Household contacts
of asbestos
workers 685
Urban New Jersey
residents 326
146 (18.8%) 66 (8.5%) 61 (7.9%) 114 (16.6%)
4 ( 1.2%) 0 (0.0%) 2 (0.6%) 11 ( 3.4%)
*ILO U/C Pneumoconlosls Classification irregular opacities 1/0 or greater.
Source: Anderson and Selikoff (1979).
169
image:
TABLE 6-7. MESOTHELIOMA FOLLOWING ONSET OF FACTORY ASBESTOS
EXPOSURE, 1941-1945a
Years from onset
Factory workers (933)
Total
deaths Mesothelioma
Household contacts (2205)
Total
deaths Mesothelioma
<20 years
20-24 years
25-29 years
30-34 years
35+ years
Total >20 years
Total al 1 years
270
102
113
84
5
304
574
0
2
5
7
0
14
14
280
93
111
124
56
384
664
0
0
0
3
1
4
4
Data of Selikoff and Anderson,
Source: Nicholson (1981).
economically similar section of town. No differences in the relative mortal-
ity experiences are seen, except for one mesothelioma in the neighborhood
group. This one case was an electrician; thus, occupational exposure may have
contributed to the disease.
6.3 COMPARISON OF OBSERVED MORTALITY WITH EXTRAPOLATED DATA
The mortality data in these two population-based studies can be compared
with estimates from the data that led to Table 6-3 but calculated for 35
years, rather than a lifetime. If the air concentration in both circumstances
3
was 200 ng/m , approximately 2 mesothelioma deaths/100,000 would be expected
in 35 years of observation. In both cases, the exposed population was about
2000; so, the expected number of mesotheliomas would be 0.04 (range: 0.004 to
0.4). The higher numbers observed, particularly in the household group,
suggest that higher exposures (e.g., from shaking dusty overalls) may have
occurred in workers' homes or that the extrapolations based on occupational
data may understate risks.
170
image:
6.4 COMPARISON OF ESTIMATED MESOTHELIOMAS WITH SEER DATA
The risk estimates of Table 6-1 through 6-3 can also be used to compare
estimated mesothelioma risk with that observed in the National Cancer Institute's
Surveillance, Epidemiology and End Results (SEER) Cancer Registry Program,
Between 1973 and 1978, 170 cases of mesothelioma were identified among females
in the SEER program which is based on 10% of the U.S. population (Connelly,
1980), Thus, about 280 cases occur annually in the U.S. among females. Using
Equations 3-6d and the current female population of the U.S., it is estimated
that 32 cases would occur annually from a continuous lifetime exposure to
0.0001 f/tnl (about 3 ng/m ). However, such a concentration, which was measured
in urban areas during 1970-71 would be influenced by the substantial use of
asbestos building products. The "background" concentrations during 1910-1940
would likely be less. Nicholson (1983) has estimated that about 20 mesothelio-
mas would occur among men and women if an average concentration of 2 ng/m
existed from 1930.
6.5 LIMITATIONS TO EXTRAPOLATIONS AND ESTIMATIONS
The above calculations of unit risk values for asbestos must be viewed
with caution because they are uncertain and are necessarily based on estimates
that are subjective, to some extent, because of the following limitations in
data: (1) extrapolation from high occupational levels to much lower ambient
levels, (2) mass-to-fiber conversion is uncertain, (3) various confounding
aspects of the medical data and, very importantly (4) the nonrepresentative
nature of the exposure estimates. The ranges of uncertainty estimated may in
fact be greater than those stated here, but insufficient information exists by
which to make more precise or definite estimates of uncertainty.
171
image:
7. OTHER REVIEWS OF ASBESTOS HEALTH EFFECTS
7.1 INTRODUCTION
Recently several government agencies in different countries reviewed
asbestos health effects. The most important of the reviews outside the United
States are those of the Advisory Committee on Asbestos (1979a,b) (ACA) of the
British Health and Safety Commission and the report of the Ontario Royal
Commission (ORC) (1984). Updates on the British report have been published by
Acheson and Gardner (1983), and most recently by Doll and Peto (1985). Each
of these major reports was the result of lengthy testimony by many scientists
and deliberation by a selected committee over a long period of time. In the
United States, the National Academy of Sciences (NAS) has reviewed the non-
occupational health risk of asbestiform fibers (National Academy of Sciences,
1984) and a Chronic Hazard Advisory Panel convened by the U.S. Consumer Product
Safety Commission (1983) reported on the hazards of asbestos. There are large
areas of agreement and some of disagreement between these other reviews and
those of this document with regard to the spectrum of asbestos-related disease,
the models describing asbestos-related lung cancer and mesothelioma, unit
exposure risks in occupational circumstances, possible differences in carcino-
genic potency of different asbestos minerals, and risk estimates at low,
non-occupational exposures. These are discussed below.
7.2 THE SPECTRUM OF ASBESTOS-RELATED MORTALITY AND FIBER TYPE EFFECTS
There was unanimity that all commercial varieties of asbestos, including
chrysotile, crocidolite, amosite, and anthophyllite, produced lung cancer in
humans. The Ontario Royal Commission (1984) noted the considerable difference
in lung cancer risk in different chrysotile-using processes. The reports
implicated chrysotile, crocidolite and amosite in increased risks of mesothe-
lioma. However, they disagreed on the importance of the role of each fiber
type. The various British and Canadian reports view chrysotile as being a
substantially less potent mesothelial carcinogen than amosite and amosite to
be somewhat less potent than crocidolite. In the view of Acheson and Gardner
(1983) "exposure to chrysotile alone so far has rarely been shown to cause
mesothelioma." The British and Canadian views are based on the high frequency
of mesothelioma deaths associated with crocidolite and amosite exposures, even
172
image:
though, in some circumstances, the amphibole usage may have been vary small
relative to chrysotile. The CPSC report viewed chrysotile as being important
in the production of pleural mesothelioma but not for peritoneal tumors. This
view is based on similar ratios of pleural mesothelioma to excess lung cancer
found among chrysotile-exposed workers compared to mixed or amphibole-exposed
workers. The NAS believed that information was insufficient to establish a
differential risk based on chemistry. It stated, "many of the apparent differ-
ences (in carcinogenic potency) may be explained by the differences in physical
properties and concentrations used by the various industries."
All reports noted that the strength of the evidence associating asbestos
exposure with cancers other than mesothelioma o!A of. the lung is less. (Gastro-
intestinal and laryngeal cancers were attributed to asbestos exposure by the
Ontario Royal Commission.(1984) and by the Advisory Committee on Asbestos
(1979a,b), although Acheson and Gardner felt in 1983 that the evidence linking
asbestos and GI cancer was "less convincing than in 1979." Doll and Peto
(1985), in their review, conclude that there are no grounds for believing that
gastrointestinal cancers in general are peculiarly likely to be caused by
asbestos exposure. They further state that: (1) for laryngeal cancer, on the
^.iier hand, <.,,> _..,ce is quite strong; (2) they reserve judgment about the
possibility that asbestos causes cancer of the esophagus; and (3) they also
note what evidence would be needed to weaken their view regarding possible
gastrointestinal tract cancer linkage to asbestos exposure. Both the U.S.
Consumer Product Safety Commission Panel (1983) and National Academy of
Sciences (1984) noted the increased risk of GI cancers in several cohorts, but
each declined to take a firm position on causality. The CPSC Report specifi-
cally noted a disagreement on the issue among panelists.
7,3 MODELS FOR LUNG CANCER AND MESOTHELIOMA
All reports adopted models for lung cancer and mesothelioma similar to
those of this report, a relative risk model for lung cancer and an absolute
risk model for mesothelioma, in which the risk increased as a power function
of time from exposure. All noted the limitations on the data establishing a
dose-response relationship, but all felt a linear model was most appropriate,
particularly for regulatory purposes. None suggested there was any evidence
173
image:
of a threshold for asbestos cancer (although the data were insufficient to
exclude one).
7.4 EXTRAPOLATIONS TO LOW EXPOSURE CIRCUMSTANCES
All of the major reviews by government agencies mentioned above undertook
quantitative risk assessments for non-occupational or low exposures to asbestos.
Because of agreement on the models for lung cancer and mesothelioma, very
similar unit risks were estimated. Differences were largely the result of the
choice of studies considered and were relatively small. All of the groups
recognized the limitations in the data on which extrapolations were based, the
dependence of the extrapolation on a linear dose-response relationship, the
uncertainties of estimation of asbestos exposure in past years, and the diffi-
culties of converting between different methods of measurement. Two groups
(National Academy of Sciences, 1984; U.S. Consumer Product Safety Commission,
1983), estimated risks at lower exposures using average unit exposure risks as
was done in this document; the other two (Ontario Royal Commission, 1984;
Advisory Committee on Asbestos, 1979a,b) used risk estimates from data in
different occupational studies and a range of the results was presented.
Various estimates of the uncertainty of these risks were provided; most were
of an ad hoc nature. A comparison of these different risk estimates is shown
in Table 7-1. There is reasonable agreement between the estimates when consid-
eration is taken of the different exposure circumstances. The NAS value for
mesothelioma risk appears to be low relative to their lung cancer risk (the
lifetime exposure risk barely exceeds that for lung cancer in a non-smoker).
This may be the result of separately choosing b and k in the risk relationship
k
= bt , rather than determining b after selecting a value for k.
When making the extrapolation from the work place exposure to the ambient
exposure, one must be aware that the physical structure and other properties
of asbestos may make the exposure risks substantially different.
174
image:
TABLE 7-1. THE RISKS OF DEATH/100,000 INDIVIDUALS FROM MESOTHELIOMA AND
LUNG CANCER FROM A LIFETIME ASBESTOS EXPOSURE TO 0.01 f/ml
Population
Lung cancer
Mesothelioma
Female smokers
Female nonsmokers
Male smokers
Male nonsmokers
Males exposed 40
years from age 20
from Table 6-3
Female smokers
Female nonsmokers
Male smokers
Male nonsmokers
This Document
150.0 (15 - 1500)
16.4 (1.64 - 164)
238.0 (23.8 - 2380)
18.5 (1.85 - 185)
88.5 (8.9 - 885)
252.0 (12.6 •
272.0 (13.6 -
181,0 (9,1 -
220.0 (11.0 •
46.5 (2.3 -
5040)
5440)
3620)
4400)
920)
National Academy of Science (1984)
57.5 (0 -
7.5 (0 -
160.0 (0 -
15.0 (0 -
275)
32.5)
725)
55)
22.
22.
22.
22.5
(0
(0
(0
(0
875)
875)
875)
875)
U.S. Consumer Product Safety Commission (1983)
Female smokers
Female nonsmokers
Male smokers
Male nonsmokers
95.2 (30.1 - 301.2)
15.7 (5.0 - 496)
155.0 (49.0 - 490.1)
17.5 (5.54 - 55.4)
246.0 (78.0 - 779.9)
266.6 (84.3 - 842.9)
174.2 (55.1 - 551.0)
215.3 (68.1 - 680.8)
Ontario Royal Commission (1984)
A hypothetical workforce
of 385 male smokers,
385 male nonsmokers,
115 female smokers, and
115 female nonsmokers
0.4 - 76
1.4 - 187.5
Males and females
Males
Advisory Committee on Asbestos (1979a,b)
8.6 - 286
Doll and Peto (1985)c
25.2 5.6
Exposure of 25 years from age twenty-two.
years exposure.
Exposure of 35 years from age 20.
175
image:
7.5 RELATIVE CARCINOGENICITY OF DIFFERENT FIBER TYPES
As briefly mentioned above, some differences exist among the major reports
by different national organizations on the relative carcinogenicity of different
asbestos fiber types. The view of the British in the Report of the Advisory
Committee on Asbestos (1979a,b) and of Acheson and Gardner (1983), who wrote
the background health effects paper and a 1983 update, is that crocidolite is
a very potent mesothellal carcinogen, amoslte is less so, and chrysotile
rarely produces such a tumor. Their view 1s based on data similar to that of
Table 3-35 and on the finding that in surveys of Individuals with mesothelloma,
particularly in Great Britain, an exposure to crocidolite or amphiboles can
usually be documented either in a history or in analysis of lung tissue for
asbestos fibers (a history of exposure to chrysotile is equally common). It
is not certain how much weight one should place upon this latter evidence. In
Great Britain, as in the United States, occupational exposure to asbestos
largely involves exposure to mixtures of fibers. Thus, an association between
amphibole exposure and mesothelloma would be expected. It is found that
amphibole asbestos varieties are retained 1n the lung for decades after exposure,
whereas chrysotile undergoes removal processes of various types. Thus, with
even brief or low intensity amphibole exposures, fibers are commonly found in
lung tissue analysis.
The Ontario Royal Commission (1984) also noted that there is a convincing
case against amphiboles in relation to the incidence of mesothelioma and that,
while chrysotile is capable of causing mesothelioma in humans, the incidence
among chrysotile-exposed cohorts has been relatively low. For this, they cite
the example of the Charleston, South Carolina textile plant with an extraordi-
narily high incidence of lung cancer, but only one mesothelioma.
Doll and Peto (1985) state that, in their opinion, the epidemiological
data show that chrysotile can cause both mesothelioma and lung cancer but that
peritoneal mesothelioma is rarely caused by chrysotile exposure and that
crocidolite and amoslte are more dangerous then chrysotile when used in the
same way. Doll and Peto (1985) particularly noted the much greater mesothe-
lioma risk in the experience of gas mask manufacturing workers who used crocido-
lite compared to those who used chrysotile (Acheson et al., 1982). However,
no exposure data were available.
The view of the National Academy of Sciences (1984) report was that the
epidemiological literature on the relative ability of different fiber types to
176
image:
cause disease does not present a clear picture. The observed variation in
risk may be due to different effects caused by different fiber types or dimen-
sions used in processes in which other contaminants are present. They state
that the magnitude of the difference in reported risks is not likely to be
explained by fiber or process differences alone.
7.6 NON-MALIGNANT EFFECTS
All reviews of asbestos did not consider a non-malignant disease to be of
importance at the exposures found in environmental circumstances. For example,
the Ontario Royal Commission (1984) concluded that "at low levels of occupational
exposure to asbestos the fibrotic process in the lungs, if indeed it can be
initiated, will not likely progress to the point of clinical manifestation or
even the mildest discomfort. On the basis of the available data our best
judgement as to the lifetime occupational exposure to asbestos at which the
fibrotic process cannot advance to the point of clinical manifestation of
asbestosis is in the range of 25 f-y/m£ and below."
177
image:
8. REFERENCES
Acheson, E. 0.; Gardner, M. J. (1983) Asbestos: the control limit for asbestos.
London, United Kingdom: Her Majesty's Stationery Office,
Acheson, E. D. ; Gardner, M. J.; Pippard, E. C.; Grime, L, P. (1982) Mortality
of two groups of women who manufactured gas masks from chrysotile and
crocidolite asbestos: a 40-year follow-up, Br. J. Ind. Med. 39: 344-348.
Acheson, E. D. ; Gardner, M. J.; Winter, P. D.; Bennett, C. (1984) Cancer in a
factory using amosite asbestos. Int. J, Epidemiol. 13: 3-10.
Advisory Committee on Asbestos. (1979a) Asbestos volume 1: final report of the
Advisory Committee. London, United Kingdom: Health and Safety Commission.
Advisory Committee on Asbestos. (19795) Asbestos volume 2: papers prepared for
the Advisory Committee. London, United Kingdom: Health and Safety Commis-
sion.
Albin, M.; Jakobsson, K.; Englander, V.; Ranstan, J.; Welinder, H. ; Westrup,
C.; Holier, T. (1984) Mortality and cancer morbidity in a cohort of
asbestos cement workers. In: Vlth international pneumoconiosis conference
1983 = VI. Internationale pneumoconiosis - konferenz 1983; v. 2; September
1983; Bochum, West Germany. Bremerhaven, West Germany: Wirtschaftsverlag
NW, Verlag fur neue Wissenschaft GmbH; pp. 825-829.
Amacher, D. E.; Alarif, A.; Epstein, S.S. (1975) The dose-dependent effects of
ingested chrysotile on DNA synthesis in the gastrointestinal tract,
liver, and pancreas of the rat. Environ. Res. 10: 208-216.
Anderson, H. A.; Lilis, R. ; Daum, S, M. ; Fischbein, A. S. ; Selikoff, I. J.
(1976) Household-contact asbestos neoplastic risk. Ann. N. Y. Acad. Sci.
271: 311-323,
Anderson, H. A.; Selikoff, I. J. (1979) Asbestos-associated radiographic
changes among household contacts of amosite asbestos workers. In: Preger,
L., ed. Induced disease: drugs, irradiation, occupation. New York, NY:
Grune and Stratton; pp. 253-273.
Armitage, P.; Doll, R. (1954) The age distribution of cancer and a multi-stage
theory of carcinogenesis. Br, J. Cancer 8: 1-12.
Armitage, P.; Doll, R. (1957) A two-stage theory of carcinogenesis in relation
to the age distribution of human cancer. Br. J. Cancer 11: 161-169.
Armitage, P.; Doll, R. (1961) Stochastic models for carcinogenesis. In: Pro-
ceedings of the fourth Berkeley symposium on mathematical statistics and
probability: v. 4. Berkeley, CA: University of California Press; pp. 19-38.
Auribault, M. (1906) Note sur 1'hygiene de la securite des ouvriers dan la
filatures et tissage d'amiante [Note on the health and safety of workers
in spinning and weaving of asbestos]. Bull. Insp. Trav. 14: 120-132.
178
image:
Ayer, H. E.; Lynch, J. R.; Fanney, J. H. (1965) A comparison of impinger and
membrane filter techniques for evaluating air- samples in asbestos plants.
Ann. N. Y. Acad. Sci. 132: 274-287.
Aziz, F.; Buckler, W. (1980) Mortality and the continuous work history sample.
In; 1980 proceedings of the section on survey research methods: papers
presented at the annual meeting of the American Statistical Association;
August; Houston, TX. Washington, DC: American Statistical Association;
pp. 461-466.
Baris, Y. I,; Artvinli, M.; Sahin, A. A. (1979) Environmental mesothelioma in
Turkey. Ann. N. Y. Acad. Sci. 330: 423-432.
Beebe, G. W.; Kato, H; Land, C. E. (1978) Studies of the mortality of A-bomb
survivors. 6. Mortality and radiation dose, 1950-1974. Radiat. Res. 75:
138-201.
Berenblum, L; Shubik, P. (1949) An experimental study of the initiating stage
of carcinogenesis, and a re-examination of the somatic cell mutation
theory of cancer. Br. J. Cancer 3: 109-118.
Berry, G. (1973) Hygiene standards—theory and application. In: Bogovski, P.;
Gil son, J. C.; Timbrel 1, V.; Wagrfer, J. C.., eds. Biological effects of
asbestos: proceedings of a working conference; October 1972; Lyon, France.
Lyon, France: World Health Organization, International Agency for Research
on Cancer; pp. 145-149. (IARC scientific publication no. 8).
Berry, G. ; Newhouse, M. L. (1983) Mortality of workers manufacturing friction
materials using asbestos. Br. J. Ind. Med. 40: 1-7.
Berry, G.; Wagner, J. C. (1969) The application of a mathematical model de-
scribing the times of occurrence of mesotheliomas in rats following
inoculation with asbestos. Br. J. Cancer 23: 582-586.
Berry, G. ; Newhouse, M. L. ; Turok, M. (1972) Combined effect of asbestos
exposure and smoking on mortality from lung cancer in factory workers.
Lancet 2: 476-479.
Berry G.; Gilson, J. C.; Holmes, S.; Lewinsohn, H. C. ; Roach, S. A. (1979)
Asbestosis: a study of dose-response relationships in an asbestos textile
factory. Br. J. Ind. Med. 36: 98-112.
Bertrand, R. ; Pezerat, H. (1980) Fibrous glass: carcinogenicity and dimensional
characteristics. In: Wagner, J. C. ; Davis, W., eds. Biological effects
of mineral fibres = Effets biologiques des fibres minerales: v. 2, pro-
ceedings of a symposium; September 1979; Lyon, France. Lyon, France:
World Health Organization, International Agency for Research on Cancer;
pp. 901-911. (IARC scientific publication no. 30; INSERM symposia series:
v. 92).
179
image:
Bignon, J.; Sebastien, P.; Gaudichet, A. (1978) Measurement of asbestos reten-
tion in the human respiratory system related to health effects. In:
Gravett, C. C.; La Fleur, P. D.; Heinrich, K. F. J., eds. Proceedings of
workshop on asbestos: definitions and measurement methods; July 1977;
Gaithersburg, MD. Washington, DC: National Bureau of Standards; pp.
95-119; NBS special publication 506. Available from: NTIS, Springfield,
VA; PB-289703/4.
Blot, W. J. ; Harrington, J. M. ; Toledo, A.; Hoover, R. ; Heath,- C. W., Jr.;
Fraumeni, J. F. , Jr. (1978) Lung cancer after employment in shipyards
during World War II. N. Engl. J. Med. 299: 620-624.
Blot, W. J, : Stone, B. J. ; Fraumeni,
mortality in U.S. counties with
Environ. Res. 18: 281-290.
J. F., Jr.; Morris, L. E. (1979) Cancer
shipyard industries during World War II.
Bogovski, P.; Gilson, J. C.; Timbrel 1, V.; Wagner, J. C., eds. (1973) Biolog-
ical effects of asbestos: proceedings of a working conference; October
1972; Lyon, France. Lyon, France: World Health Organization, International
Agency for Research on Cancer. (IARC scientific publication no. 8).
Bohlig, H.; Main, E. (1973) Cancer in relation to environmental exposure. In:
Bogovski, P.; Gilson, J. C.; Timbrel 1, V.; Wagner, J. C., eds. Biological
effects of asbestos: proceedings of a working conference; October 1972;
Lyon, France. Lyon, France: World Health Organization, International
Agency for Research on Cancer; pp. 217-221. (IARC scientific publication
no. 8).
Brain, J. D.; Valberg, P. A. (1974) Models of lung retention based on ICRP
task group report. Arch. Environ. Health 28: 1-11.
British Occupational Hygiene Society. (1968) Hygiene standard for chrysotile
asbestos dust. Ann. Occup. Hyg. 11: 47-69.
British Occupational Hygiene Society, (1983) Report from the Committee on
Asbestos: a study of the health experience on two U.K. asbestos factories.
Ann. Occup. Hyg. 27: 1-55.
Brown, D. P.; Dement, J. M.; Wagoner, J. K. (1979) Mortality patterns among
miners and millers occupationally exposed to asbestiform talc. In: Lemen,
R. ; Dement, J. M. , eds. Dusts and disease: proceedings of the conference
on occupational exposures to fibrous and particulate dust and their
extension into the environment. Forest Park, IL: Pathotox Publishers,
Inc.; pp. 317-324.
Burdett,
for
transmission
G. J.; Rood, A. P. (1983) Membrane-filter, direct-transfer technique
the analysis of asbestos fibers or other inorganic particles by
electron microscopy. Environ. Sci. Technol. 17: 643-648.
Chamberlain, M. ; Brown, R. C. (1978) The cytotoxic effects
other mineral dust in tissue culture cell lines. Br. J.
183-189.
of asbestos and
Exp. Pathol. 59:
180
image:
Chamberlain, M.; Tarmy, E. M. (1977) Asbestos and glass fibres in bacterial
mutation tests. Mutat. Res. 43: 159-164.
Churg, A.; Wiggs, B.; Depaoli, L,; Kampe, B,; Stevens, B, (1984) Lung asbestos
contents in chrysotile workers with mesothelioma. Am. Rev. Respir. Dis.
130: 1042-1045.
Clemmesen, J.; Hjalgrim-Jensen, S. (1981) Cancer incidence among 5,686 asbes-
tos-cement -workers followed from 1943 through 1976. Ecotoxicol. Environ.
Saf. 5: 15-23.
Code of Federal Regulations. (1984a) Occupational safety and health standards:
subpart Z--toxic and hazardous substances, asbestos. C. F. R. 29: §
1910.1001.
Code of Federal Regulations. (1984b) National emission standards for hazardous
air pollutants: subpart M--national emission standards for asbestos.
C. F. R. 40: §§ 61.140-61.156.
Commins, B. T. ; Gibbs, G. W. (1969) Contaminating organic material in asbestos.
Br. 0. Cancer 23: 358-362.
Connelly, R. R. (1980) Mesothelioma incidence in the United States. Presented
at: workshop on opportunities for research and education in asbestos-
related disease; June, Bethesda, MD, Bethesda, MD: U.S. Department of
Health and'Human Services, National Cancer Institute.
Constant, P. C., Jr.; Bergman, F. J.; Atkinson, G. R.; Rose, D. R.; Watts, D.
L.; Logue, E. E.; Hartwell, T. D.; Price, B. P.; Ogden, J. S. (1983)
Airborne asbestos levels in schools, Washington, DC: U.S. Environmental
Protection Agency, Office of Toxic Substances; EPA report no. 560/5-83-003.
Available from: NTIS, Springfield, VA; PB84-129683/REB.
Cook, P.; Doll, R, ; Fellingham, S. A. (1969) A mathematical model for the age
distribution of cancer in man. Int. J. Cancer 4: 93-112.
Cooper, W. C.; Balzer, J. L. (1973) Evaluation and control of asbestos expo-
sures in the insulation trade. In: Holstein, E.; Anspach, M., eds. Inter-
national Konferenz u'ber die biologischen Wirkungen des Asbestos, [Inter-
national conference on the biological effects of asbestos]; April 1968;
Dresden, East Germany. Berlin, East Germany: Deutsche Zentralblatt
Arbeitsmedizin; pp. 151-160.
Cooper, W. C. ; Miedema, J. (1973) Asbestosis in the manufacture of insulating
materials. In: Bogovski, P.; Gilson, J. C.; Timbrell, V.; Wagner, J. C.,
eds. Biological effects of asbestos: proceedings of a working conference;
October 1972; Lyon, France. Lyon, France: World Health Organization,
International Agency for Research on Cancer; pp. 175-178. (IARC scientific
publication no. 8).
Davis, J. M. G. ; Beckett, S. T.; Bolton, R. E.; Collings, P.; Middleton, A. P.
(1978) Mass and number of fibers in the pathogenesis of asbestos-related
lung disease in rats. Br. J. Cancer 37: 673-688.
181
image:
Day, N-. E. ; Brown, C. C. (1980) Multistage models and primary prevention of
cancer, J. Nat!. Cancer Inst. 64; 977-989.
Dement, J. M.; Harris, R, L., Jr.; Symons, M. J.; Shy, C. M. (1982) Estimates
of dose-response for respiratory cancer among chrysotile asbestos textile
workers. In; Inhaled particles V: proceedings of an international sympo-
sium; September 1980; Cardiff, South Glamorgan, Wales. Ann. Occup. Hyg.
26(1-4): 869-887.
Dement, J. M.; Harris, R. L., Jr.; Symons, M. J.; Shy, C. M. (1983a) Exposures
and mortality among chrysotile asbestos workers. Part I: Exposure esti-
mates. Am. J. Ind. Med. 4: 399-419.
Dement, J. M.; Harris, R. L., Jr.; Symons, M. J.; Shy, C. M. (1983b) Exposures
and mortality among chrysotile asbestos workers. Part II: mortality. Am.
J. Ind. Med. 4: 421-433.
Doll, R. (1955) Mortality from lung cancer in asbestos workers. Br. J. Ind.
Med. 12: 81-86.
Doll, R.; Peto, R. (1978) Cigarette smoking and bronchial carcinoma: dose and
time relationships among regular smokers and lifelong non-smokers. J.
Epidemiol. Comrnun. Health 32: 303-313.
Doll, R.; Peto, J. (1985) Asbestos: effects on health of exposure to asbestos.
London, United Kingdom; Health and Safety Commission,
Donna, A. (1970) Tumori sperimentali da amianto di crisotilo, crocidolite e
amosite in ratto Sprague-Dawley [Experimental tumours from asbestos
(chrysotile, crocidolite and amosite) in the Sprague-Dawley rat]. Med.
Lav. 61: 1-32.
Elmes, P. C.; Simpson, M. J. C. (1977) Insulation workers in Belfast. A further
study of mortality due to asbestos exposure (1940-75). Br. J. Ind. Med.
34: 174-180.
Enterline, P. E. (1976) Estimating health risks in studies of the health ef-
fects of asbestos. Am. Rev. Respir. Dis. 113: 175-180.
Enterline, P. E.; Henderson, V. (1973) Type of asbestos and respiratory cancer
in the asbestos industry. Arch. Environ. Health 27: 312-317.
Evans, J. C. ; Evans, R. J. ; Holmes, A.; Hounam, R. F.; Jones, D. M. ; Morgan,
A.; Walsh, M. (1973) Studies on the deposition of inhaled fibrous material
in the respiratory tract of the rat and its subsequent clearance using
radioactive tracer techniques, I. UICC crocidolite asbestos. Environ.
Res. 6: 180-201.
Federal Register. (1975) Occupational exposure to asbestos: notice of pro-
posed rulemaking. F. R. (October 9) 40: 47652-47665.
Federal Register. (1984) Occupational exposure to asbestos. F. R. (April 10)
49: 14116-14145.
182
image:
Ferris, B. G., Jr.; Ranadive, M. V.; Peters, J, M.; Murphy, R, L. H.; Burgess,
W. A.; Pendergrass, H. P. (1971) Prevalence of chronic respiratory disease:
asbestosis in ship repair workers. Arch. Environ. Health 23: 220-225.
Finkelstein, M. M. (1982) Asbestosis in long-term employees of an Ontario
asbestos-cement factory. Am. Rev. Respir. Dis. 125: 496-501.
Finkelstein, M, M. (1983) Mortality among long-term employees of an Ontario
asbestos-cement factory. Br. J. Ind. Med. 40: 138-144.
Fisher, L. C. (1958) Multiple-mutation theory of carcinogenesis. Nature (London)
181: 651-652.
Fisher, .J, C.; Holloman, J. H, (1951) A hypothesis for the origin of cancer
foci. Cancer (Philadelphia) 4: 916-918.
Fleischer, W. E.; Viles, F. J., Jr.; Gade, R. L.; Drinker, P. (1946) A health
survey of pipe-covering operations in constructing naval vessels. J. Ind.
Hyg. Toxicol. 28: 9-16.
Fox, A. J. ; Collier, P. F. (1976) Low mortality rates in industrial cohort
studies due to selection for work and survival in the industry. Br. J.
Prev. Soc. Med. 30: 225-230.
Frank, A. L. (1979) Public health significance of smoking-asbestos interaction.
Ann. N. Y. Acad. Sci. 330: 791-794.
Garfinkel, L. (1981) Time trends in lung cancer mortality among nonsmokers
and a note on passive smoking. J. Natl. Cancer Inst. 66: 1061-1066.
Gehring, P. J.; Watanabe, P. G. ; Park, C. N. (1978) Resolution of dose-response
toxicity data for chemicals requiring metabolic activation: example—vinyl
chloride. Toxicol, Appl. Pharmacol. 44: 581-591.
Gibbs, G. W. ; Hwang, C. Y. (1975) Physical parameters of airborne asbestos
fibres in various work environments—preliminary findings. Am. Ind. Hyg.
Assoc. J, 36: 459-466.
Gibbs, G. W.; LaChance, M. (1974) Dust-fiber relationships in the Quebec chry-
sotile industry. Arch. Environ. Health 28: 69-71.
Gibbs, G. W.; Rowlands, N.; Brulotte, R. (1980) A pilot study on the measure-
ment of airborne asbestos fibre concentrations in ambient air. Presented
at: 73rd annual meeting of the Air Pollution Control Association; June;
Montreal, Quebec, Canada. Pittsburgh, PA: Air Pollution Control Associa-
tion; paper no. 80-27.3.
Gloyne, S. R, (1936) A case of oat cell carcinoma of the lung occurring in
asbestosis. Tubercle 18: 100-101.
Goldsmith, J. R. (1982) Asbestos as a systemic carcinogen: the evidence from
eleven cohorts. Am. J. Ind. Med. 3: 341-348.
183
image:
Graham, S. ; Blanchet, M.; Rohrer, T. (1977) Cancer in asbestos-mining and
other areas of Quebec. J. Natl. Cancer Inst. 59: 1139-1145.
Gross, P.; deTreville, R. T. P.; Tolker, E. B.; Kaschak, M.; Babyak, M. A,
(1967) Experimental asbestosis: the development of lung cancer in rats
with pulmonary deposits of chrysotile asbestos dust. Arch. Environ.
Health 15: 343-355.
Guess, H. A.; Crump, K. S. (1976) Low-dose extrapolation of data from animal
carcinogenicity experiments—analysis of a new statistical technique.
Math. Biosci. 32: 15-36.
Guess, H, A.; Crump, K. S. (1978) Best-estimate low-dose extrapolation of car-
cinogenicity data. Environ. Health. Perspect, 22: 149-152.
Guess, H. A.; Crump, K. S,; Peto, R. (1977) Uncertainty estimates for low-dose-
rate extrapolations of animal carcinogenicity data. Cancer Res. 37:
3475-3483.
Hammad, Y. Y. ; Diem, J. ; Weill, H. (1979) Evaluation of dust exposure in
asbestos cement manufacturing operations. Am. Ind. Hyg. Assoc. J. 40:
490-495.
Hammond, E. C. (1966) Smoking in relation to death rates of one million men
and women. In: Haenszel, W. , ed. Epidemiological approaches to the study
of cancer and other chronic diseases. Washington, DC: U.S. Department of
Health, Education, and Welfare, National Cancer Institute; pp. 127-204.
(National Cancer Institute monograph 19).
Hammond, E. C. ; Selikoff, I. J. ; Seidman, H. (1979a) Asbestos exposure, ciga-
rette smoking and death rates. Ann. N. Y. Acad. Sci. 330: 473-490,
Hammond, E. C. ; Garfinkel, L. ; Selikoff, I. J. ; Nicholson, W. J. (1979b)
Mortality experience of residents in the neighborhood of an asbestos
factory. Ann. N. Y. Acad. Sci. 330: 417-422.
Harington, J. S. (1962) Occurrence of oils containing 3:4-benzpyrene and
related substances in asbestos. Nature (London) 193: 43-45.
Harington, J. S. ; Roe, F. J. C. (1969) Studies of carcinogenesis of asbestos
fibers and their natural oils. Ann. N. Y. Acad. Sci. 132: 439-450.
Harries, P. G. (1968) Asbestos hazards in naval dockyards. Ann. Occup. Hyg.
11: 135-145.
Harries, P. G. (1971) A comparison of mass and fibre concentrations of asbestos
dust in shipyard insulation processes. Ann. Occup. Hyg. 14: 235-240.
Harries, P. G. (1976) Experience with asbestos disease and its control in
Great Britain's naval dockyards. Environ. Res. 11: 261-267.
184
image:
Harris, J. E. (1979) Appendix: cigarette smoking in the United States, 1950-
1978. In: Smoking and health: a report of the Surgeon General.
Washington, D.C.: U.S. Department of Health, Education, and Welfare,
Public Health Service; pp. A-l to A-29;; DHEW publication no. (PHS)
79-50066. Available from: GPO, Washington, DC; S/N 017-000-00218-0,
Harris, R. L. , Jr.; Fraser, D. A. (1976) A model for deposition of fibers in
the human respiratory system. Am. Ind. Hyg. Assoc. J. 37: 73-89.
Hart, R. W.; Fertel, R.; Newman, H. A. I.; Daniel, F. B.; Blakeslee, J. R.
(1979) Effects of selected asbestos fibers on cellular and molecular
parameters. Cincinnati, OH: U.S. Environmental Protection Agency, Health
Effects Research Laboratory; EPA report no. EPA-600/1-79-21. Available
from: NTIS, Springfield, VA; PB-299199/0.
Henderson, V. L. ; Enterline, P. E. (1979) Asbestos exposure:' factors associ-
ated with excess cancer and respiratory disease mortality. Ann. N. Y.
Acad. Sci. 330: 117-126.
Hobbs, M. S. T.; Woodward, S. D.; Murphy, B.; Musk, A. W.; Elder, J. E. (1980)
The incidence of pneumoconiosis, mesothelioma and other respiratory
cancer in men engaged in mining and milling crocidolite in Western
Australia. In: Wagner, J.C.; Davis, W., e>ds. Biological effects of mineral
fibres = Effets biologiques des fibres minerales: v. 2, proceedings of a
symposium; September 1979; Lyon France. Lyon, France: World Health
Organization, International Agency for Research on Cancer; pp. 615-625.
(IARC scientific publication no. 30; INSERM symposia series: v. 92).
Hoel, D. G. ; Kaplan, N. L. ; Anderson, M. W. (1983) Implication of nonlinear
kinetics on risk estimation in carcinogenesis. Science (Washington, DC)
219: 1032-1037.
Holmes, S. (1965) Developments in dust sampling and counting techniques in the
asbestos industry. Ann. N. Y. Acad. Sci. 132: 288-297.
Holt, P. F. ; Mills, J. ; Young, D. K. (1964) The early effects of chrysotile
asbestos dust on the rat lung. J. Pathol. Bacteriol. 87: 15-23.
Huang, S. L. (1979) Amosite, chrysotile and crocidolite asbestos are mutagenic
in Chinese hamster lung cells. Mutat. Res. 68: 265-274.
Hughes, J. ; Weill, H. (1980) Lung cancer risk associated with manufacture of
asbestos-cement products. In: Wagner, J. C. ; Davis, W., eds. Biological
effects of mineral fibres = Effets biologiques des fibres minerales: v.
2, proceedings of a symposium; September 1979; Lyon France. Lyon,
France: World Health Organization, International Agency for Research on
Cancer; pp. 627-635. (IARC scientific publication no. 30; INSERM sym-
posia series: v. 92),
International Agency for Research on Cancer. (1982) Chemicals, industrial
processes and industries associated with cancer in humans, IARC mono-
graphs, volumes 1 to 29: report of an IARC ad hoc working group; February;
Lyon, France. Lyon, France: World Health Organization; pp. 13-15, 52-53.
(IARC monographs on the evaluation of the carcinogenic risk of chemicals
to humans: suppl. 4),
185
image:
International Labour Office. (1971) International Classification of radiographs
of pneumoconioses (revised, 1968). Geneva, Switzerland; International
Labour Office. (Occupational safety and health series no. 22),
Irwig, L. M.; duToit, R. S. J.; Sluis-Cremer, G. K.; Solomon, A.; Thomas, R.
G.; Hamel, P. P. H.; Webster, I.; Hastie, T. (1979) Risk of asbestosis in
crocidolite and amosite mines in South Africa. Ann. N. Y. Acad. Sci. 330:
35-52,
Jacko, M. G.; DuCharme, R. T. ; Somers, J. T. (1973) How much asbestos do
vehicles emit? Automot. Eng. 81: 38-40.
Jacobs, R. ; Humphrys, J.; Dodgson, K. S.; Richards, R. J, (1978) Light and
electron microscope studies of the rat digestive tract following
prolonged and short-term ingestion of chrysotile asbestos. Br. J. Exp.
Pathol. 59: 443-453.
Jones, J. S. P.; Smith, P. G. ; Pooley, F. D.; Berry, G.; Sawle, G. W.;
Madeley, R. J. ; Wignell, B. K, ; Aggarv/al, A. (1980) The consequences of
exposure to asbestos dust in a wartime gas-mask factory. In: Wagner, J.
C. ; Davis, W., eds. Biological effects of mineral fibres = Effets
biologiques des fibres minerales: v. 2, proceedings of a symposium;
September 1979; Lyon, France. Lyon, France: World Health Organization,
International Agency for Research on Cancer; pp. 637-653. (IARC scientific
publication no. 30; INSERM symposia series: v. 92).
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. Washington, DC: U.S. Department of Health, Education, and
Welfare, National Cancer Institute; pp. 1-125. (National Cancer Institute
monograph 19).
Kleinfeld, M.; Messite, J.; Kooyman, 0. (1967) Mortality experience in a group
of asbestos workers. Arch. Environ. Health. 15: 177-180.
Kleinfeld, M.; Messite, J.; Zaki, M. H. (1974) Mortality experience among talc
workers: a follow-up study. J. Occup. Med. 16: 345-349.
Knox, J. F. ; Holmes, S. ; Doll, R. ; Hill, I. D. (1968) Mortality from lung
cancer and other causes among workers in an asbestos textile factory. Br.
J. Ind. Med. 25: 293-303.
Kolonel, L. N.; Hirohata, T.; Chappell, B. V.; Viola, F, V.; Harris, D. E.
(1980) Cancer mortality in a cohort of naval shipyard workers in Hawaii:
early findings. J. Natl. Cancer Inst. 64: 739-743.
Langer, A. M. (1974) Inorganic particles in human tissues and their associa-
tion with neoplastic disease. Environ. Health Perspect. 9: 229-233.
Langer, A. M. ; Wolff, M. S.; Rohl, A. N.; Selikoff, I. J. (1978) Variation of
properties of chrysotile asbestos subjected to milling. J. Toxicol.
Environ. Health 4: 173-188.
186
image:
Lavappa, K. S, ; Fu, M. M. ; Epstein, S. S. (1975) Cytogenetic studies on
chrysotile asbestos. Environ. Res. 10: 165-173.
Lechner, J. F.; Haugen, A.; Trump, B. F.; Tokiwa, T.; Harris, C. C. (1983)
Effects of asbestos and carcinogenic metals on cultured human bronchial
epithelium. In: Harris, C. C. ; Autrup, H. N., eds. Human carcinogenesis.
New York, NY: Academic Press; pp. 561-585.
Lee, P. N.; O'Neill, J. A. (1971) The effect both of time and dose applied on
tumour incidence rate in benzopyrene skin painting experiments. Br. J.
Cancer 25: 759-770.
Leidel, N. A.; Bayer, S. G.; Zumwalde, R. D.; Busch, K. A. (1979) USPHS/NIOSH
membrane filter method for evaluating airborne asbestos fibers.
Cincinnati, OH: U.S. Department of Health, Education, and Welfare, National
Institute of Occupational Safety and Health; DHEW (N10SH) publication no.
79-127. Available from: NTIS, Springfield, VA; PB-297731.
Lewinsohn, H. C. (1972) The medical surveillance of asbestos workers. R. Soc,
Health J. 92: 69-77.
Lewinsohn, H. C. (1983) [Personal communication with Dr. W. J. Nicholson].
Danbury, CT: Union Carbide Corporation.
Liddell, F. D. K.; Hanley, J. A. (1985) Relations between asbestos exposure
and lung cancer SMRs in occupational cohort studies. Br. J. Ind. Med.
42: 389-396.
Liddell, F. D. K.; McDonald, J. C. (1980) Radiological findings'as predictors
of mortality in Quebec asbestos workers. Br. J. Med. 37: 257-267.
Liddell, F. D. K. ; McDonald, J. C. ; Thomas, D. C. (1977) Methods of cohort
analysis: appraisal by application to asbestos mining. J. R. Stat. Soc. A
140; 469-491.
Lilis, R. ; Daum, S.; Anderson, H.; Sirota, M.; Andrews, G. ; Selikoff, I. J.
(1979) Asbestos disease in maintenance workers of the chemical industry.
Ann. N. Y. Acad. Sci. 330: 127-136.
Livingston, G. K. ; Rom, W. N. ; Morris, M. V. (1980) Asbestos^induced sister
chromatid exchanges in cultured Chinese hamster ovarian fibroblast cells.
J. Environ. Pathol. Toxicol. 4: 373-382.
Lynch, K. M. ; Smith, W. A, (1935) Pulmonary asbestosis. Ill: Carcinoma of
lung in asbesto-silicosis. Am. J. Cancer 14: 56-64.
Lynch, J. R.; Ayer, H. E.; Johnson, D. L. (1970) The interrelationships of se-
lected asbestos exposure indices. Am. Ind. Hyg. Assoc. J. 31: 598-504.
187
image:
Maltoni, C.; Annoscia, C. (1974) Mesotheliomas in rats following the intra-
peritoneal injection of crocidolite. In: Davis, W,; Maltoni, C., eds.
Advances in tumour prevention, detection, and characterization: v. 1,
characterization of human tumours: proceedings of fifth international
symposium on the biological characterization of human tumours; April
1973; Bologna, Italy. Amsterdam, The Netherlands: Excsrpta Medica; pp.
115-116.
Mancuso, T. F.; Coulter, E. J. (1963) Methodology in industrial health studies:
the cohort approach, with special reference to an asbestos company. Arch.
Environ. Health 6: 210-226.
Mancuso, T. F.; El-Attar, A. A, (1967) Mortality pattern in a cohort of asbestos
workers, J. Occup. Med. 9: 147-162.
Mantel, N. ; Haenszel, W. (1959) Statistical aspects of the analysis of data
from retrospective studies of disease. J. Natl. Cancer Inst. 22:
719-748.
Mason, T. J. ; McKay, F. W. (1974) U.S. cancer mortality by county: 1950-1969.
Washington, DC: U.S. Department of Health, Education, and Welfare,
National Cancer Institute; DHEW publication no. (NIH) 74-615.
McDonald, J. C.; Liddell, F. D. K. (1979) Mortality in Canadian miners and
millers exposed to chrysotile. Ann. N. Y. Acad. Sci. 330: 1-10.
McDonald, A. D.; McDonald, J. C. (1978) Mesothelioma after crocidolite exposure
during gas mask manufacture. Environ. Res. 17: 340-346.
McDonald, A. D. ; McDonald, J. C, (1980) Malignant mescthelioma in North America.
Cancer (Philadelphia) 46: 1650-1656,
McDonald, J. C.; McDonald, A. D.; Gibbs, G. W.; Siemiatyckit J.; Rossiter, C.
E. (1971) Mortality in the chrysotile asbestos mines and mills of Quebec,
Arch. Environ. Health. 22: 677-686.
McDonald, J. C. ; Liddell, F. D. K.; Gibbs, G. W.; Eyssen, G. E. ; McDonald, A.
D. (1980) Dust exposure and mortality in chrysotile mining, 1910-75. Br.
J. Ind. Med. 37: 11-24.
McDonald, A. D. ; Fry, J. S. ; Wool ley, A. J.; McDonald, J. C. (19B3a) Dust
exposure and mortality in an American chrysotile textile plant. Br. J.
Ind. Med, 40: 361-367.
McDonald, A. D.; Fry, J. S.; Woolley, A. J.; McDonald, J. C. (19B3b) Dust
exposure and mortality in an American factory using chrysotile, amosite
and crocidolite in mainly textile manufacturing. Br. J. Ind. Med. 40:
368-374.
McDonald, A. D. ; Fry, J. S. ; Woolley, A. J. ; McDonald, J. C. (1984) Dust
exposure and mortality in an American chrysotile asbestos friction pro-
ducts plant. Br. J. Ind. Med. 41: 151-157.
188
image:
Merewether, E. R. A. (1949) Annual report of the Chief Inspector of Factories
for the year 1947, London, United Kingdom: Her Majesty's Stationery Office;
pp. 66-81.
Meurman, L. 0.; Kiviluoto, R.; Hakama, M. (1974) Mortality and morbidity among
the working population of anthophyllite asbestos miners in Finland. Br;
J. Ind. Med. 31: 105-112.
Morgan, A. (197.9) Fiber dimensions: their significance in the deposition and
clearance of inhaled fibrous dust. In: Lemen, R,; Dement, J. M., eds.
Dusts and disease: proceedings of the conference on occupational exposures
to fibrous and particulate dust and their extension into the environment.
Forest Park, IL: Pathotox publishers, Inc.; pp. 87-96.
Morgan, A.; Evans, J. C. ; Evans, R. J. ; Hounarn, R. F. ; Holmes, A.; Doyle, S.
G. (1975) Studies on the deposition of inhaled fibrous material in the
respiratory tract of the rat and its subsequent clearance using radio-
active tracer techniques. II. Deposition of the UICC standard reference
samples of asbestos. Environ. Res. 10: 196-207.
Morgan, A.; Evans, J. C.; Holmes, A. (1977) Deposition and clearance of inhaled
fibrous minerals in the rat. Studies using radioactive tracer techniques.
In: Walton, W. H. , ed. Inhaled particles IV (in two parts): part 1,
proceedings of an international symposium; September 1975; Edinburgh,'
United Kingdom. Oxford, United Kingdom: Pergamon Press; pp. 259-273.
Morgan, A.; Talbot, R. J. ; Holmes, A. (1978) Significance of fibre length in
the clearance of asbestos fibres from the lung. Br. J. Ind. Med. 35:
146-153,
Mossman, B. T. ; Eastman, A.; Landesman, J. M.; Bresnick, E. (1983) Effects of
crocidolite and chrysotile asbestos on cellular uptake and metabolism of
benzo(a)pyrene in hamster tracheal epithelial cells. Environ. Health
Perspect. 51: 331-335.
Muller, H. J. (1951) Radiation damage to the genetic material. In: Baitsell,
G. A., ed. Science in progress. New Haven, CT: Yale University Press;
pp. 93-165, 481-493. (Sigma Xi national lectureships: seventh series,
1949 and 1950).
Murphy, R. L. H.; Ferris, B. G., Jr.; Burgess, W. A.; Worcester, J.; Gaensler,
E. A. (1971) Effects of low concentrations of asbestos: clinical, environ-
mental, radiologic and epidemiologic observations in shipyard pipe coverers
and controls. N. Engl. J. Med. 285; 1271-1278.
Murray, H. M. (1907) Report of the departmental committee on compensation for
industrial diseases. London, United Kingdom: His Majesty's Stationery
Office; p. 127.
National Academy of Sciences. (1983) Risk assessment in the federal government:
managing the process. Washington, DC: National Academy Press.
National Academy of Sciences. (1984) Asbestiform fibers: nonoccupational
health risks. Washington, DC: National Academy Press.
189
image:
National Center for Health Statistics, (1977) Vital statistics of the United
States: volume II - mortality. Hyattsville, MD: U.S. Department of Health
and Human Services.
National Institute for Occupational Safety and Health. (1972) Criteria for a
recommended standard: occupational exposure to asbestos. Washington, DC;
U.S. Department of Health, Education, and Welfare; DHEW (NIOSH) publica-
tion no. HMS 7210267. Available from; NTIS, Springfield, VA; PB-209510.
National Institute for Occupational Safety and Health. (1976) Revised recom-
mended asbestos standard. Washington, DC: U.S. Department of Health,
Education, and Welfare; DHEW (NIOSH) publication no. 79-164. Available
from: NTIS, Springfield, VA; PB-273965.
National Institute for Occupational Safety and Health. (1980) Workplace expo-
sures to asbestos: review and recommendations. Washington, DC: U.S.
Department of Health and Human Services; DHHS (NIOSH) publication no.
81-103. Available from: NTIS, Springfield, VA; PB83-176677.
Newhouse, M. L.; Berry, G. (1976) Predictions of mortality from mesothelial
tumours in asbestos factory workers. Br. J. Ind. Med. 33; 147-151.
Newhouse, M. L.; Berry, G. (1979) Patterns of mortality in asbestos workers in
London. Ann. N. Y. Acad. Sci. 330: 53-60.
Newhouse, M. L.; Thompson, H. (1965) Mesothelioma of the pleura and peritoneum
following exposure to asbestos in the London area. Br. J. Ind. Med. 22:
261-269.
Newhouse, M. L. ; Berry, G.; Wagner, J. C.; Turok, M. E. (1972) A study of the
mortality of female asbestos workers. Br. J. Ind. Med. 29: 134-141.
Newman, H. A. I.; Saat, Y. A,; Hart, R. W. (1980)-Putative inhibitory effects
of chrysotile, crocidolite and amosite mineral fibers on the more complex
surface membrane glycolipids and glycoproteins. Environ. Health
Perspect. 34: 103-111.
Nicholson, W. J. (1971) Measurement of asbestos in ambient air. Washington,
DC: National Air Pollution Control Administration; contract no. CPA
70-92.
Nicholson, W. J., ed, (1975) Insulation hygiene progress reports: v. 3, no. 1.
New York, NY: City University of New York, Mount Sinai School of Medicine.
Nicholson, W. J, (1976a) Case study 1: asbestos—the TLV approach. Ann. N. Y.
Acad. Sci. 271: 152-169.
Nicholson, W. J. (1976b) [Tables] In: Annexe: edition revue et corrigee du
rapport preliminaire. Comite d'etude sur la salubrite olans I1 Industrie
de 1'amiante (Beaudry Commission); pp. 151-160.
Nicholson, W. J. (1978) Chrysotile asbestos in air samples collected in Puerto
Rico: report to the Consumer Products Safety Commission. New York, NY:
City University of New York, Mount Sinai School of Medicine; CPS contract
no. 77128000.
190
image:
Nicholson, W. J. (1981) Criteria document for Swedish occupational standards:
asbestos and inorganic fibers. Arb. Halsa (17).
Nicholson, W. J. (1982) The dose and time dependence of occupational cancer.
In: Prevention of occupational cancer—international symposium; April
1981; Helsinki, Finland. Geneva, Switzerland: International Labour
Office; pp. 44-67. (Occupational safety and health series no. 46).
Nicholson, W. J. (1983) Tumour incidence after asbestos exposure in the USA:
cancer risk of the non-occupational population. VDI Ber. (475): 161-177.
Nicholson, W. J. ; Pundsack, F. L, (1973) Asbestos in the environment. In:
Bogovski, P.; Gil son, J. C.; Timbrel 1, V.; Wagner, J. C. , eds. Biological
effects of asbestos: proceedings of a v/orking conference; October 1972;
Lyon, France. Lyon, France: World Health Organization, International
Agency for Research on Cancer; pp. 126-130. (IARC scientific publication
no. 8).
Nicholson, W. J.; Rohl, A. N.; Ferrand, E. F. (1971) Asbestos air pollution in
New York City. In: Englund, H. M. ; Beery, W. T., eds. Proceedings of the
second international clean air congress; December 1970; Washington, DC.
New York, NY: Academic Press; pp. 136-139.
Nicholson, W. J.; Holaday, D. A.; Heimann, H. (1972) Direct and indirect occu-
pational exposure to insulation dusts in United States shipyards. In:
Safety and health in shipbuilding and ship repairing', proceedings of a
symposium; August-September 1971; Helsinki, Finland. Geneva, Switzerland:
International Labour Office; pp. 37-47. (Occupational safety and health
series no. 27).
Nicholson, W. J. ; Rohl, A. N. ; Weisman, I. (1975) Asbestos contamination of
the air in public buildings. Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards; EPA
report no. EPA 450/3-76/004, Available from: NTIS, Springfield, VA;
PB-250980/0.
Nicholson, W. J. ; Rohl, A. N. ; Weisman, I. (1976) Asbestos contamination of
building air supply systems. In: International conference on environmental
sensing and assessment: v. 2; September 1975; Las Vegas, NV. New York,
NY: Institute of Electrical and Electronics Engineers, Inc.; paper no.
29-6.
Nicholson, W. J. ; Rohl, A. N, ; Sawyer, R. N.; Swoszowski, E. J. , Jr.; Todaro,
J. D. (1978) Control of sprayed asbestos surfaces in school buildings: a
feasibility study. New York, NY: Mount Sinai School of Medicine, Environ-
mental Sciences Laboratory; NIEHS contract no. N01-ES-7-2113. See also:
Nicholson, W. J. ; Swoszowski, E. J. , Jr.; Rohl, A. N.; Todaro, J. D. ;
Adams, A. (1979) Asbestos contamination in United States schools from use
of asbestos surfacing materials. Ann. N. Y. Acad. Sci. 330: 587-596.
Nicholson, W. J. ; Selikoff, I. J. ; Seidman, H.; Lilis, R.; Formby, P. (1979)
Long-term mortality experience of chrysotile miners and millers in Thetford
Mines, Quebec. Ann. N. Y. Acad. Sci. 330: 11-21.
191
image:
Nicholson-, W. J. ; Rohl, A. N. ; Weisman, I.; Selikoff, I. J. (1980) Environ-
mental asbestos concentrations in the United States. In: Wagner, J. C. ;
Davis, W., eds. Biological effects of mineral fibres = Effets bibliogiques
des fibres minerales: v. 2, proceedings of a symposium; September 1979;
Lyon, France. Lyon, France: World Health Organization, International
Agency for Research on Cancer; pp. 823-827. (IARC scientific publication
no. 30; INSERM symposia series:'v. 92),
Nicholson, W. J. ; Perkel, G. ; Selikoff, I. J. (1982) Occupational exposure to
asbestos: population at risk and projected mortality—1980-2030. Am. J.
Ind. Med. 3: 259-311.
Nicholson, W. J.; Selikoff, I. J.; Seidman, H.; Hammond, E. C. (1986) Mortality
experience of asbestos factory workers; effect of differing intensities
of asbestos exposure. Environ. Res.: In press.
Nordling, C. 0. (1953) A new theory on the cancer-inducing mechanism. Br. J.
Cancer 1: 68-72.
Occupational Safety and Health Administration. (1983) Quantitative risk assess-
ment for asbestos related cancers. Washington, DC: Directorate of Health
Standards; OSHA contract no. J-9-F-2-0074. Available for inspection at:
U.S. Department of Labor, OSHA Technical Data Center, Francis Perkins
Building; docket no. H033C, exhibit no. 84-392.
Occupational Safety and Health Administration. (1984) [Docket of current rule-
making for revision of the asbestos (dust) standard]. Washington, DC:
U.S. Department of Labor. Available for inspection at: U.S. Department of
Labor, OSHA Technical Data Center, Francis Perkins Building; docket
no. H033C.
Ontario Royal Commission. (1984) Report of the Royal Commission on matters of
health and safety arising from the use of asbestos in Ontario: v. 1-3.
Toronto, ON, Canada: Ontario Ministry of the Attorney General.
Peto, J. (1977) The establishment of industrial hygiene standards: an example.
In: Whittemore, A., ed. Environmental health: quantitative methods, pro-
ceedings of a conference; July 1976; Alta, UT. Philadelphia, PA: Society
for Industrial and Applied Mathematics; pp. 104-114.
Peto, J. (1980) Lung cancer mortality in relation to measured'dust levels in
an asbestos textile factory. In: Wagner, J. C.; Davis, W.,'eds. Biological
effects of mineral fibres = Effets biologiques des fibres minerales:
v. 2, proceedings of a symposium; September 1979; Lyon, France. Lyon,
France: World Health Organization, International Agency for Research on
Cancer; pp. 829-836. (IARC scientific publication no. 30; INSERM symposia
series: v. 92).
Peto, R. ; Roe, F. J. C.; Lee, P. N.; Levy, L.; Clack, J. (1975) Cancer and
ageing in mice and men. Br. J. Cancer 32: 411-426.
Peto, J.; Seidman, H.; Selikoff, I. J. (1982) Mesothelioma mortality in asbes-
tos workers: implications for models of carcinogenesis and risk assessment.
Br. J. Cancer 45: 124-135.
192
image:
Peto, J. ; Doll, R.; Goffe> J. ; Clayton, R. (1985) Mortality of workers in the
Rochdale asbestos textile factory. Ann. Occup. Hyg.: In press.
Pike, M.C. (1966) A method of analysis of a certain class of experiments in
carcinogenesis. Biometrics 22: 142-161.
Pott, F. (1980) Animal experiments on biological effects of mineral fibres.
In: Wagner, J. C.; Davis, W., eds. Biological effects of mineral fibres =
Effets biologiques des fibres minerales: v, 1, proceedings of a symposium;
September 1979; Lyon France. Lyon, France: World Health Organization,
International Agency for Research on Cancer; pp. 251-272. (IARC scientific
publication no. 30; INSERM symposia series: v. 92).
Pott, F.; Friedrichs, K. H. (1972) Tumoren der Ratte nach i.p.-Injektion fas-
erf ormiger Staube [Tumors in the rat developing after intraperitoneal
injection of fibrous dusts], Naturwissenschaften 59; 318.
Pott, F.; Friedrichs, K. H.; Huth, F. (1976) Ergebnisse aus Tierversuchen zur
kanzerogenen Wirkung faserformiger Staube und ihre Deutung im Hinblick
auf die Tumorentstehung beim Menschen [Results from animal experiments on
the carcinogenic effect of fibrous dusts and their significance with
regard to the origin of tumors in the human], Zentralbl. Bakteriol.
Parasitenkd. Infektionskrankh. Hyg. Abt. 1: Orig. Reihe B 162: 467-505.
Puntoni, R. ; Vercelli, M.; Merlo, F.; Valeric, F.; Santi, L. (1979) Mortality
among shipyard workers in Genoa, Italy. Ann. N. Y. Acad. Sci. 330: 353-377.
Pylev, L. N. ; Shabad, L. M. (1973) Some results of experimental studies in
asbestos carcinogenesis. In: Bogovski, P.; Gilson, J. D. ; Timbrell, V.;
Wagner, J. C,, eds. Biological effects of asbestos: proceedings of a
working symposium; October 1972; Lyon France. Lyon, France: World Health
Organization, International Agency for Research on Cancer; pp. 99-105.
(IARC scientific publication no. 8).
Reeves, A. L. (1976) The carcinogenic effect of inhaled asbestos fibers. Ann.
Clin. Lab. Sci. 6: 459-466.
Reeves, A. L. ; Puro, H. E. ; Smith, R. G. ; Vorv/ald, A. J. (1971) Experimental
asbestos carcinogenesis. Environ. Res. 4: 496-511.
Reeves, A. L.; Puro, H. E.; Smith, R. G. (1974) Inhalation carcinogenesis from
various forms of asbestos. Environ. Res. 8: 178-202.
Rendall, R. E. G. ; Skikne, M. I. (1980) Submicroscopic fibres in industrial
atmospheres. In: Wagner, J. C. ; Davis, W. , eds. Biological effects of
mineral fibres = Effets biologiques des fibres minerales: v, 2, proceed-
ings of a symposium; September 1979; Lyon, France. Lyon, France: World
Health Organization, International Agency for Research on Cancer; pp.
837-843. (IARC scientific publication no. 30; INSERM symposia series: v.
92).
193
image:
Robinson, C. ; Lemen, R. ; Wagoner, J. K. (1979) Mortality patterns, 1940-75,
among workers employed in an asbestos textile friction and packing products
manufacturing facility. In: Lemen, R.; Dement, J. R., eds. Dusts and
disease: proceedings of the conference on occupational exposures to
fibrous and particulate dust and their extension into the environment.
Forest Park, IL: Pathotox Publishers, Inc.; pp. 131-143.
Rohl, A. N.; Langer, A. M.; Wolff, M. S. ; Weisman, I. (1976) Asbestos exposure
during brake lining maintenance and repair. Environ. Res. '12: 110-128.
Rossiter, C. E. ; Coles, R. M. (1980) HM Dockyard, Devonport: 1947 mortality
study. In: Wagner, J. C. ; Davis, W., eds. Biological effects of mineral
fibres = Effets biologiques des fibres minerales: v. 2, proceedings of a
symposium; September 1979; Lyon, France. Lyon, France: World Health
Organization, International Agency for Research on Cancer; pp. 713-721.
(IARC scientific publication no. 30; INSERM symposia series: v. 92).
Rubino,
10, G. F. ; Piolatto, G. ; Newhouse, M. L.; Scansetti, G.; Aresini, G. A.;
Hurray, R. (1979) Mortality of chrysotile asbestos workers at the Balangero
Mine, northern Italy. Br. J. Ind. Med. 36: 187-194.
Samudra, A. V. ; Harwood, C. F. (1977) Electron microscopic measurement of
airborne asbestos concentrations: a provisional methodology manual.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Environ-
mental Sciences Research Laboratory; EPA report no. EPA-600/2-77-178-REV.
Available from: NTIS, Springfield, VA; PB-285945.
Sawyer, R. N. (1977) Asbestos exposure in a Yale building: analysis and resolu-
tion. Environ. Res. 13: 146-169.
Sawyer, R. N. (1979) Indoor asbestos pollution: application of hazard criteria.
Ann. N. Y Acad. Sci. 330: 579-586.
Schneider, V.; Maurer, R. R, (1977) Asbestos and embryonic development. Tera-
tology 15: 273-279.
Sebastien, P.; Janson, X.; Bonnand, G.; Riba, G.; Masse, R.; Bignon, J. (1979)
Translocation of asbestos fibers through respiratory tract and gastroin-
testinal tract according to fiber type and size. In: Lemen, R.; Dement,
J. M. eds. Dusts and disease: proceedings of the conference on occupa-
tional exposures to fibrous and particulate dust and their extension into
the environment. Forest Park, IL: Pathotox Publishers, Inc.; pp. 65-85.
Sebastien, P.; Billion-Galland, M. A.; Dufour, G.; Bignon, J. (1980) Measure-
ment of asbestos air pollution inside buildings sprayed with asbestos.
Washington, DC: U.S. Environmental Protection Agency, Survey and Analysis
Division; EPA report no. EPA-560/13-80-026. Available from: NTIS,
Springfield, VA; PB81-147001.
Sebastien, P.; Bignon, J. ; Martin, M. (1982) Indoor airborne asbestos pollu-
tion: from the ceiling and the floor. Science (Washington, DC) 216:
1410-1413.
194
image:
Seidman, H. (1984) Short-term asbestos work exposure and long-term observation.
In: [Docket of current rulemaking for revision of the asbestos (dust)
standard]. Washington, DC: U.S. Department of Labor, Occupational Safety
and Health Administration. Available for inspection at: U.S. Department of
Labor, OSHA Technical Data Center, Francis I'erkins Building; docket no.
H033C, exhibit nos. 261-A and 261-B.
Seidman, H. ; Selikoff, I. J.; Hammond, E. C. (1979) Short-term asbestos work
exposure and long-term observation. Ann. N. Y. Acad. Sci. 330: 61-89.
Selikoff, I. J.; Seidman, H. (1981). Cancer of the pancreas among asbestos
insulation workers. Cancer (Philadelphia) 47(suppl.): 1469-1473.
Selikoff, I. J. ; Churg, J.; Hammond, E. C. (1965) The occurrence of asbestosis
among insulation workers in the United States. Ann, N. Y. Acad. Sci. 132:
139-155.
Selikoff, I. J.; Hammond, E. C.; Churg, J. (1968) Asbestos exposure, smoking,
and neoplasia. J. Am. Med. Assoc. 204: 104-110.
Selikoff, I. J. ; Hammond, E. C. ; Churg, J. (1970) Mortality experience of
asbestos insulation workers, 1943-68. In: Shapiro, H. A., ed. Pneumoconi-
osis: proceedings of the international conference; 1969; Johannesburg,
South Africa. Capetown, South Africa: Oxford University Press; pp. 180-186.
Selikoff, I. J.; Hammond, E. C.; Seidman, H. (1S79) Mortality experience of
insulation workers in the United States and Canada, 1943-1976. Ann. N. Y.
Acad. Sci. 330: 91-116.
Selikoff, I. J. ; Nicholson, W. J. ; Lilis, R. (1980) Radiological evidence of
asbestos disease among ship repair workers. Am. J. Ind. Med. 1: 9-22.
Shabad, L. M.; Pylev, L. N.; Krivosheeva, L. V.; Kulagina, T. F.; Nemenko, B.
A. (1974) Experimental studies on asbestos carcinogenicity. J. Natl.
Cancer Inst. 52: 1175-1187.
Siemiatycki, J. (1982) Health effects on the general population (mortality in
the general population in asbestos mining areas). In: Asbestos, health
and -society, proceedings of the world symposium on asbestos; May; Montreal
Quebec, Canada, Montreal, Quebec, Canada: Canadian Asbestos Information
Centre; pp. 337-348.
Sincock, A. M. (1977) Preliminary studies of the in vitro cellular effects of
asbestos and fine glass dusts. In: Hiatt, H. H.; Watson, J. D.; Winsten,
J. A., eds. Origins of human cancer: book B, mechanisms of carcinogenesis.
Cold Spring Harbor, NY: Cold Spring Harbor Laboratory; pp. 941-954.
(Cold Spring Harbor conferences on cell proliferation: v. 4).
Sincock, A. M.; Delhanty, J. D. A.; Casey, G. (1982) A comparison of the cyto-
genic response to asbestos and glass fibre in Chinese hamster and human
cell lines: demonstration of growth inhibition in primary human fibro-
blasts. Mutat. Res. 101: 257-268.
195
image:
Smith, W. E.; Hubert, D, D. (1974) The intraplaural route as a means for esti-
mating carcinogenicity. In: Karbe, E.; Park, J. F., eds. Experimental
lung cancer; carcinogenesis and bioassay. Berlin, West Germany; Springer-
Verlag; pp. 93-106.
Smith, W. E.; Miller, L.; Elsasser, R. E,; Hubert, D. D. (1965) Tests for
carcinogenicity of asbestos. Ann. N. Y, Acad. Sci. 132: 456-488.
Smith, W. E. ; Miller, L.; Churg, J. (1970) An experimental model for the study
of carcinogenesis in the respiratory tract. In: Nettesheim, P.; Hanna, M.
G., Jr.; Deatherage, J. W,, Jr., eds. Morphology of experimental respira-
tory carcinogenesis: proceedings of a Biology Division, Oak Ridge National
Laboratory, conference; May; Gatlinburg, TN. Oak Ridge, TN: U.S. Atomic
Energy Commission; pp. 299-316. (AEC symposium series 21). Available
from: NTIS, Springfield, VA; CONF-70D501.
Smither, W. J, ; Lewinsohn, H. C. (1973) Asbestosis in textile manufacturing.
In: Bogovski, P.; Gilson, J. C.; Timbrell, V.; Wagner, J. C., eds. Biolog-
ical effects of asbestos: proceedings of a working conference; October
1972; Lyon, France. Lyon, France: World Health Organization, International
Agency for Research in Cancer; pp. 169-174. (IARC scientific publication
no. 8).
Solomons, K. (1984) Malignant mesothelioma - clinical and epidemiological
features. S, Afr. Med. J. 66: 407-414.
Spumy, K. R. ; Stb'ber, W. ; Weiss, G. ; Opiela, H. (1980) Some special problems
concerning asbestos fiber pollution in ambient air. in; ^-ndrie, to, M. ,
ed. Atmospheric pollution 1980: proceedings of the 14th international
colloquium; May; Paris, France, Amsterdam, The Netherlands: Elsevier
Scientific Publishing Company; pp. 315-322. (Studies in environmental
science: 8).
Stanton, M. F. (1973) Some etiological considerations of fibre carcinogenesis.
In: Bogovski, P.; Gilson, J. C. ; Timbrell, V.; Wagner, J. C. eds. Biolog-
ical effects of asbestos: proceedings of a working conference; October
1972; Lyon, France, Lyon, France: World Health Organization, International
Agency for Research in Cancer; pp. 289-294. (IARC scientific publication
no, 8).
Stanton, M. F.; Wrench, C. (1972) Mechanisms of mesothelioma induction with
asbestos and fibrous glass. J. Natl. Cancer Inst. 48: 797-816.
Stanton, M. F. ; Layard, M.; Tegeris, A.; Miller, E.; May, M.; Kent, E. (1977)
Carcinogenicity of fibrous glass: pleural response in the rat in relation
to fiber dimensions. J. Natl. Cancer Inst. 58; 587-603.
Stanton, M. F.; Layard, M.; Tegeris, A.; Miller, E. ; May, M.; Morgan, E.;
Smith, A, (1981) Relation of particle dimension to carcinogenicity in
amphibole asbestos and other fibrous minerals. J. Natl. Cancer Inst.
67; 965-975.
196
image:
Steel, J. (1979) Asbestos control limits. In: Asbestos volume 2: papers prepared
for the advisory committee. London, United Kingdom: Health and Safety
Commission; pp. 85-88,
Storeygard, A, R.; Brown, A. L., Jr. (1977) Penetration of the small intestinal
mucosa by asbestos fibers. Mayo Clin. Proc. 52: 809-812.
Teta, M. J.; Lewinsohn, H. C. ; Meigs, J. W.; Vidone, R. A.; Mowad, L. Z. ;
Flannery, J. T. (1983) Mesothelioma in Connecticut, 1955-1977: occupa-
tional and geographic associations. J. Occup. Med. 25: 749-756.
Timbrel!, V. (1965) The inhalation of fibrous dusts. Ann. N. Y. Acad. Sci.
132: 255-273.
Thomas, H. F. ; Benjamin, I. T.; Elwood, P. C.; Sweetnam, P. M. (1982) Further
follow-up study of workers from an asbestos cement factory. Br. J. Ind.
Med. 39: 273-276.
Toft, P.; Wigle, D. T.; Meranger, J. C.; Mao, Y. (1981) Asbestos and drinking
water in Canada. Sci. Total Environ. 18; 77-89.
U.S. Consumer Product Safety Commission. (1983) Report to the U.S. Consumer
Product Safety Commission by the chronic hazard advisory panel on as-
bestos. Washington, DC: Directorate for Health Sciences.
U.S. Environmental Protection Agency. (1974) A preliminary report on asbestos
in the Duluth, Minnesota, area. Duluth, MN: Office of Enforcement and
General Counsel, Office of Technical Analysis.
Valeric, F.; de Ferrari, M.; Ottaggio, L. ; Repetto, E.; Santi, L. (1980) Cyto-
genic effects of Rodesian chrysotile on human lymphocytes i_n vitro. In:
Wagner, J. C.; Davis, W. , eds. Biological effects of mineral fibres =
Effets biologiques des fibres minerales: v. 1, proceedings of a symposium;
September 1979; Lyon, France. Lyon, France: World Health Organization,
International Agency for Research on Cancer; pp. 485-489. (IARC scientific
publication no. 30; INSERM symposia series: v. 92).
Wagner, J. C.; Berry, G.; Skidmore, J. W. (1976) Studies of the carcinogenic
effect of fibre glass of different diameters following intrapleural
inoculation in experimental animals. In: Occupational exposure to fibrous
glass: proceedings of a symposium; June 1974; College Park, MD, Cin-
cinnati, OH: U.S. Department of Health, Education, and Welfare, National
Institute for Occupational Safety and Health; pp. 193-204; DHEW publica-
tion no. (NIOSH) 76-151. Available from: NTIS, Springfield, VA; PB-258869.
Wagner, J. C.; Sleggs, C. 0
and asbestos exposure
Med. 17: 260-271.
.; Marchand, P. (1960) Diffuse pleural mesothelioma
in the north western Cape Province. Br. J. Ind.
Wagner, J. C. ; Berry, G.; Timbrell, V. (1973) Mesotheliomata in rats after
ino'culation with asbestos and other materials. Br. J. Cancer 28: 173-185.
Wagner, J. C. ; Berry, G.; Skidmore, J. W.; Timbrell, V. (1974) The effects of
the inhalation of asbestos in rats. Br. J. Cancer 29: 252-269.
197
image:
Wagner, J. C. ; Berry, G.; Cook, T. J. ; Hill, R. J.; Pooley, F. D.; Skidmore,
J. W. (1977) Animal experiments with talc. In; Walton, W.C., ed. Inhaled
particles IV (in two parts): part 2, proceedings of an international
symposium; September 1975; Edinburgh, United Kingdom. Oxford, United
Kingdom Pergamon Press, pp. 647-654.
Wagner, J. C.; Berry, G.; Pooley, F. D. (1982) Mesotheliomas and asbestos type
in asbestos textile workers: a study of lung contents, Br, Med. J, 285:
603-606.
Webster, I. (1978) Discussion. In: Glen, W. H, , ed. Proceedings of asbestos
symposium; October 1977; Johannesburg, South Africa. Randburg, South
Africa: National Institute for Metallurgy; p. 79.
Weill, H. (1984) [Testimony of Hans Wei 11, M.D.]- In: [Docket of current rule-
making for revision of the asbestos (dust) standard], Washington, DC:
U.S. Department of Labor, Occupational Safety and Health Administration.
Available for inspection at: U.S. Department of Labor, OSHA Technical
Data Center, Francis Perkins Building; docket no. H033C, exhibit no. 93-2.
Weill, H.; Hughes, J.; Waggenspack, C. (1979) Influence of dose and fiber type
on respiratory malignancy risk in asbestos cement manufacturing. Am. Rev.
Respir. Dis. 120: 345-354.
Weiss, A. (1953) Pleurakrebs bei Lungenasbestose, _in vivo morphologisch gesi-
chert [Cancer of pleura in pulmonary asbestosis determined morphologi-
cally in vivo]. Medizinische (3): 93-94.
Weiss, W. (1971) Cigarette smoking, asbestos and pulmonary fibrosis. Am. Rev.
Respir. Dis. 104: 223-227.
Weiss, W. (1977) Mortality of a cohort exposed to chrysotile asbestos. J. Occup.
Med. 19: 737-740.
Whipple, H. E.; van Reyen, P. E., eds, (1965) Biological effects of asbestos.
Ann. N. Y. Acad. Sci. 132 (art. 1).
Whittemore, A. (1977a) Epidemiological implications of the multistage theory
of carcinogenesis. In: Whittemore, A.; ed. Environmental health: quanti-
tative methods, proceedings of a conference; July 1976; Alta, UT.
Philadelphia, PA: Society for Industrial and Applied Mathematics.; pp.
72-87.
Whittemore, A. S. (1977b) The age distribution of human cancer for carcinogenic
exposures of varying intensity. Am. J. Epidemiol. 106: 418-432.
Wigle, D. T, (1977) Cancer mortality in relation to asbestos in municipal
water supplies. Arch. Environ. Health 32: 185-190.
Wignall, B. K.; Fox, A. J. (1982) Mortality of female gas mask assemblers. Br.
J.' ,Ind. Med. 39: 34-38.
Winer, A. A.; Cossette, M. (1979) The effect of aspect ratio on fiber counts:
a preliminary study. Ann. N. Y. Acad. Sci. 330: 661-672.
GOVERNMENTPRIKTOIGOmCEil966 -6"t6-llG/ <*D611
image:
TECHNICAL REPORT DATA
(Please read Instructions on the reverie before completing)
.REPORT NO.
EPA/60078-84/003F
3. RECIPIENT'S ACCESSION NO.
4, TITLE AND SUBTITLE
Airborne Asbestos Health Assessment Update
i. REPORT DATE
June 1935
6. PERFORMING ORGANIZATION CODE
AUTHOFMS)
See List of Authors, Contributors, and Reviewers
8. PERFORMING ORGANIZATION REPORT NO.
9, PERFORMING ORGANIZATION NAME AND ADDRESS
1O. PROGRAM ELEMENT NO.
11, CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
EnvirTonnrantal Criteria and Assessment Office
Office of Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
13, TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/23
15. SUPPLEMENTARY NOTEU
16. ABSTRACT
Recent data from population studies strengthened the association of asbestos
with disease. Lung cancer and mesothelioma are the most important asbestos-related
causes of death. The data suggest that the excess risk of lung cancer from
asbestos exposure is proportional to cumulative exposure (duration X intensity) and
underlying risk in the absence of exposure. Risk of death from iresothelioma
appears proportional to cumulative exposure in a given period. Aninal studies
confirm the human epiderniological results. All major asbestos varieties produce
lung cancer and mesotheliona, with only limited differences in carcinogenic
potency. One can extrapolate the risks of asbestos cancers from occupational
exposures, although the uncertainty is approximately tenfold or greater.
Calculations of asbestos unit risk values are uncertain and based on subjective
estimates because of the following limitations in data: (1} extrapolation from
high occupational levels to much lower ambient levels; (2) the uncertainty of
mass-to-fiber conversion,- (3) statistical uncertainties; (4) various biases and
confounding aspects of medical data; and very importantly, (5) nonrepresentative
exposure estimates.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFlEFIS'GPEN ENDED TERMS
c. COSATI Field/Croup
18. DISTRIBUTION STATIMENT
Release to Public
IS. SECURITY CLASSJTliis Report]
Unclassified
21, NO OF PAGES
20. SECURITY SLASS (Thispuge)
Unclassified
22. PRICE
EPA Fotm 2220-1 (H«», 4-77) PREVIOUS EDITION is OBSOLETE .
1
image:
