-------
2. INTRODUCTION
The purpose of this health effects review is to evaluate the information
that has been developed since 1972 on human disease from asbestos exposure.
The review examines the substantial amount of new health research that has
been reported in the last decade to help evaluate the current standard pro-
mulgated by the U.S. Environmental Protection Agency (EPA) for asbestos emis-
sions. Thus, emphasis will be placed on the literature published after 1972
and on those papers that provide information on the risk from low-level expo-
sures such as those encountered in the non-occupational environment. Specifi-
cally, this report will address the following issues:
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 estimates 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 the different asbestos minerals or of fibers of different
dimensionality?
4. What additional or confirmatory information relating to human carcino-
genicity is provided by animal studies?
5. What are the non-occupational concentrations of asbestos to which popula-
tions are exposed?
6. Is there a basis for estimating numerical risks of asbestos disease that
results from environmental exposure.
Two documents provide a good review of the status of knowledge of the
health effects of asbestos in the early 1970s. One source is the criteria
document for occupational exposure to asbestos produced by the National Insti-
-------
tute of Occupational Safety and Health as part of the Occupational Safety and
Health Administration's consideration of an asbestos standard in early 1972
(NIOSH, 1972). The second document is the proceedings of a conference spon-
sored 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 at that time (IARC, 1973). This latter
document included a report by an Advisory Committee on Asbestos Cancers
appointed by the IARC to review evidence relating exposures to asbestos dust
to cancers.
2.1 SUMMARY OF ASBESTOS HEALTH EFFECTS THROUGH 1972
This summary relies heavily on review articles that are in the pro-
ceedings of the October 1972 IARC meeting and in the report of the IARC Advi-
sory Committee published therein (IARC, 1973) for the summary of health
effects knowledge in 1973.
2.1.1 Occupational Exposure
Diseases considered to be associated with asbestos exposure in 1972
included asbestosis, mesothelioma, bronchogenic carcinoma, and cancers of the
GI tract, including the esophagus, stomach, colon, and rectum. Lung cancer
was associated with exposure to all principal commercial varieties of asbestos
fiber: amosite, anthophyl1ite, crocidolite, and chrysotile. Excess risks of
bronchogenic carcinoma were documented in mining and milling, manufacturing,
and end product use (application of insulation materials). 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 anthophyl1ite 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-yr), 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). Again, exposures
were expressed in mppcf-yr and information on conversion of mppcf to fibers
per milliter was available only for textile production. Despite the paucity
of data, the report of the Advisory Committee on Asbestos Cancers to the IARC
(1973) stated, "The evidence ... suggests that an excess lung carcinoma risk
is not detectable when the occupational exposure has been low. These low
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 manifest by X-ray abnor-
malities of the pleura or parenchymal tissue, had more extensive documentation
than the risk of the asbestos-related malignancies. In part, this documenta-
tion was the result of 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, 1947). Such asbestosis had been documented in a
wide variety of work circumstances and associated with all commercial types of
asbestos fiber. Among some exposed groups, 50 to 80% 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). Company data
supplied to the British Occupational Hygiene Society (BOHS, 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% for accumulated fiber exposure of 100 fibers-yr/ml (e.g., 2
fibers/milliliter (f/ml ) for 50 years). However, shortly after the establish-
ment of the British Standard, additional data from the same factory population
suggested a much greater prevalance of X-ray abnormalities than was believed
-------
CC
o
V)
LU
^w
ait!
C/J <
21
i!
§<
£>
LU
O
cc
20
- _
10
i—i—r
• BASAL RALES
O X-RAY ABNORMALITIES
0 —
O
I
0 100 200 300 400 500
CUMULATIVE EXPOSURE, years x fibres/cm3
Figure 2-1. Dose-response relationship
for prevalance of basal rales in a chryso-
tile asbestos factory.
Source: Berry (1973); x-ray data added
from BOHS (1968).
-------
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 stand-
ard classification of X-rays (ILO, 1971) and the longer time from onset of
employment. Of the 290 employees, only 13 had been employed for 30 or more
years; 172 had less than 20 years of employment. The progression of asbes-
tosis 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.
2.1.2 Environmental and Indirect Occupational Exposure Circumstances.
Four research groups had shown that asbestos disease risk could exist in
circumstances other than direct occupational circumstances. In 1960, 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 occupational exposure. However, all but one of the remainder re-
sulted 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 residen-
tial background of 76 individuals deceased of mesothelioma in the London
hospital. Forty-five of the decedents had been employed in an asbestos indus-
try; 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 Main (1973) further defined the residential risk by documenting
environmental asbestos exposure near a factory in 38 cases in Hamburg. The
final study, which is particularly important because of the size of the popu-
lation implied to be at risk, is that of Harries (1968), who pointed to a risk
of asbestos disease from indirect occupational exposure in the shipbuilding
industry. He described the 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, their work took place where
other employees were 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 micrograms per liter (fjg/1) 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 "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 asbes-
tos to the general public," and "Such evidence as there is does not indicate
any risk" from asbestos fibers in water, beverages, food, or 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 the quantifying of asbestos
in the environment. In 1971, a standardized methodology was established for
the identification of pneumoconiosis: the ILO U/C Classification of Pneumo-
conioses (ILO, 1971). This methodology provided a uniform criteria for asses-
sing the prevalence of asbestos-related X-ray abnormalities. Further, signi-
ficant advances were achieved in the quantification of asbestos aerosols. In
the late 1960s, the membrane filter technique was developed for the measure-
ment of asbestos fibers in workplace aerosols. While this procedure had some
limitations, it established a standardized method, using simple instrumenta-
tion, that was far superior to any that existed previously. This method
subsequently allowed epidemiological studies that based exposure estimates on
a standardized criterion. Additionally, experimental techniques in the quan-
tification of asbestos at concentrations of tenths of ng/m of air and tenths
of ug/1 of water were developed. These techniques extended the sensitivity of
exposure estimates approximately three orders below those of occupational
aerosols and allowed assessment of general population exposures. Finally,
techniques for the analysis of asbestos in lung and other body tissues were
developed. Both digestion techniques, use of electron microscopy to analyze
fibers contained in the digest, and to analyze thin sections of lung tissue
showed that asbestos fibers were commonly present in the lung tissue of gene-
ral population residents, as well as in that of individuals exposed in occu-
pational circumstances.
-------
2.1.4 Animal 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, but
was not described in the open literature in animals until 1967 (Gross et al.,
1967). Mesothelioma, reported in an asbestos 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 carcino-
genic materials produced an enhanced risk of asbestos cancer. Asbestos ex-
posure combined with exposure to benz(a)pyrene was demonstrably more toxic
than exposure to either agent alone. Additionally, organic and metal com-
pounds associated with asbestos fibers were ruled out as an important factor
in the carcinogenicity of fibers. Lastly, animal experimentation involving
the application of fibers onto the pleura of animals indicated that the im-
portant factor in the carcinogenicity was the dimensionality of the fibers
rather than their chemical properties (Stanton, 1973). The greatest carcino-
genicity was related to fibers that were less than 2.5 (jm in diameter and
longer than 10 urn.
2.2 CURRENT ASBESTOS STANDARDS
The current Occupational Safety and Health Administration (OSHA) stan-
dards for an 8-hour time-weighted average (TWA) occupational exposure to
asbestos is 2 fibers longer than 5 jjm 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 min. (29 CFR 1910.001). 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 1 f/ml is now the accepted level for chrysotile. This standard
resulted from recommendations made in 1979 by the Advisory Committee (1979a),
which also recommended a TWA of 0.5 f/ml for amosite and 0.2 f/ml for crocido-
lite. From 1969 to 1983, 2 f/ml (TWA) was the standard for chrysotile (BOSH,
1968). This earlier British standard served as a guide for the OSHA standard
(NIOSH, 1972).
The British standard was developed specifically to prevent asbestosis
among working populations; data that would allow a determination of a standard
-------
for cancer (BOHS, 1968) were felt to be lacking. Unfortunately, among occupa-
tional groups, cancer is the primary cause of excess death among workers (see
Chapter 3). Three-fourths or more of asbestos-related deaths are from malig-
nancy. This fact has led OSHA to propose a lower TWA standard to 0.5 f/ml
(500,000 f/m3) in October, 1975 (29 CFR 1910.001). The National Institute for
Occupational Safety and Health anticipated hearings on a new standard and
proposed a value of 0.1 f/ml (NIOSH, 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 to define a threshold for asbestos carcinogenesis. The
OSHA proposal has been withdrawn, and a new proposal is anticipated. NIOSH
has reaffirmed its position on an 0.1 f/ml standard (1980).
The existing Federal national emission standards for asbestos are pub-
lished in Part 61, Title 40, Code of Federal Regulations. In summary, these
apply to milling, manufacturing, and fabrication sources and to demolition,
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 ng/m ).
10
-------
3. HUMAN HEALTH EFFECTS ASSOCIATED WITH OCCUPATIONAL EXPOSURE TO ASBESTOS
3.1 INTRODUCTION
This review of human health effects associated with occupational exposure
to asbestos is concerned with those studies that aid in the development of an
exposure-response relationship for lung cancer and mesothelioma. While lung
cancer and mesothelioma are the most dominant asbestos-related malignancies,
the strength of the evidence and the relative excess of cancers at extra-
thoracic sites are discussed. Models for assessment of the risk of lung
cancer and mesothelioma are reviewed. Unit exposure risks are estimated from
11 studies that provide information on exposure-response relationships. These
estimates illustrate that considerable variation exists in the unit exposure
risks found for mesothelioma and lung cancer in the different studies. The
possible sources of these different unit risks are also considered. An impor-
tant question is whether the variation is the result of methodological uncer-
tainties (i.e., on the estimates of exposure or of the magnitude of disease)
or whether differences are real and must be reconciled on the basis of the
character of the exposure in terms of fiber size and chemistry.
3.2 MORTALITY ASSOCIATED WITH ASBESTOS EXPOSURE
The study of U.S. and Canadian insulators by Selikoff et al. (1979)
contains the largest excess of asbestos-related deaths among any group of
asbestos workers studied. Thus, this study best demonstrates the full spec-
trum of disease from asbestos exposure. The mortality experience of 17,800
asbestos insulation workers was studied prospectively from January 1, 1967,
through December 31, 1976. These workers were exposed primarily to chrysotile
prior to 1940, to chrysotile and amosite from 1940 through 1965, and primarily
to chrysotile thereafter. No crocidolite is known to have been used in the
U.S. insulation material (Selikoff et al. , 1970). The workers' main activity
was applying new insulation; removal of old materials constituted less than 5%
of their work.
In this group, 2,271 deaths occurred, and their analysis provides impor-
tant insights into the nature of asbestos disease. Table 3-1 lists the expec-
ted and observed deaths by cause and includes data on tumors found less fre-
quently. Lung tumors were common and accounted for approximately 21% of the
deaths; 8% were from mesothelioma of the pleura or peritoneum, and 7% resulted
11
-------
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
Underlying cause of death
Total deaths, all causes
Total cancer, all sites
Cancer of lung
Pleural mesothel ioma
Peritoneal mesothel ioma
Mesothel ioma, n.o.s.
Cancer of esophagus
Cancer of stomach
Cancer of colon-rectum
Cancer of larynx
Cancer of pharynx, buccal
Cancer of kidney
All other cancer
Noninfectious pulmonary
diseases, total
Asbestosis
All other causes
Expected
1658.9
319.7
105.6
_b
_b
_b
7.1
14.2
38.1
4.7
10.1
8.1
131.8
59.0
_b
1280.2
Number
of Deaths
Observed
BE
2271
995
486
63
112
0
18
22
59
11
21
19
184
212
168
1064
DC
2271
922
429
25
24
55
18
18
58
9
16
18
252
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
1.40
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
1.91
3.19
_b
0.91
BE = best evidence. Number of deaths categorizes after review of best
available information (autopsy, surgical, clinical).
DC = Number of deaths as recorded from death certificate information only.
Expected deaths are based upon white male age-specific U.S. death rates of
the U.S. National Center for Health Statistics, 1967-1976. (NCHS, Annually:
1967-1977)
Rates and thus ratios are not available, but these have been rare causes of
death in the general population.
Source: Selikoff et al. (1979)
12
-------
from asbestosis. Considering all cancers, 675 excess malignancies occurred,
constituting 30% of all deaths. In addition to the usual asbestos malignan-
cies, lung cancer, mesothelioma, and gastrointestinal cancer, the incidences
of cancers of the larynx, pharynx and buccal cavity, and kidney were signifi-
cantly elevated. Other tumors were also increased, but not to a statistically
significant degree for individual sites. However, as a group, these other
cancers were significantly in excess, 184 observed deaths (using best avail-
able evidence for classification) versus 131.8 expected deaths (p<0.001).
However, some of this observed excess may be the result of misclassification
of asbestos-related lung cancer or peritoneal mesothelioma. Rather than 184
deaths, certificate of death classification attributed 252 deaths to cancer at
these other sites. Pancreatic, liver, and unspecified abdominal cancers were
commonly misclassified. Pancreatic and abdominal cancers were often found to
be peritoneal mesotheliomas, and several liver cancers were the result of a
primary malignancy in the lung. Because all cases could not be reviewed, some
additional misclassification may still exist. However, the magnitude would
not be great compared to the remaining excess of 52 cases. The excess at
extrathoracic sites may reflect mortality from the dissemination of asbestos
fibers to various organs (Langer, 1974). Alternatively, this trend could
represent a systemic effect of asbestos, perhaps on the immune system, that
leads to a general increased risk of cancer (Goldsmith, 1982).
3.2.1 Accuracy of Cause of Death Ascertainment
Table 3-1 lists the observed deaths according to the cause recorded on
the death certificate (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, information re-
corded on death certificates is usually used because such information, without
verification, serves as the basis for "expected rates." However, mesothelioma
and asbestosis are virtually unseen in the general population; therefore,
their misdiagnosis (which is common) is of little importance. In contrast,
their misdiagnosis among asbestos workers can cause serious distortions in
assessing mortality. Not only are asbestos-related causes understated, but
others, such as pancreatic cancer, might wrongly appear to be significantly
elevated (Selikoff and Seidman, 1981). While substantial differences exist in
the DC and BE characterization of deaths from mesothelioma and asbestosis, the
13
-------
BE and DC deaths from cancer of all sites and lung cancer agree reasonably
wel 1.
Mesothelioma is best described by an absolute risk model and lung cancer
by a relative risk model. Thus, risks for mesothelioma will be expressed in
absolute rates (e.g., deaths/ 1,000 person-years), and the best medical evi-
dence will be used, when available, to establish the number of cases. Risks
for lung cancer will be quantified by the ratio of observed to expected
deaths. In this document, it is assumed that misclassification of lung cancer
deaths would occur as frequently in asbestos workers as in the general popu-
lation (in terms of the percentage of lung cancer cases). Therefore, the
certificate of death cause will be used for establishing the relative risks of
lung cancer in asbestos-exposed groups.
3.3 LINEARITY OF EXPOSURE-RESPONSE RELATIONSHIPS
Some limited direct evidence for linearity of response with asbestos
exposure is available from three studies that compared lung cancer mortality
to the cumulative total dust exposure in asbestos workplaces (Henderson and
Enterline, 1979; Liddell et al. , 1977; Dement et al. , 1982). Figure 3-1 shows
the exposure-response data in these studies in which the ratio of observed
to expected lung cancer mortality is plotted against the measured cumulative
dust exposure. While different exposure -response relationships appear to
exist for the three circumstances, each demonstrates a linear relationship
over the entire range of observation. The differences in the slopes of the
three relationships may relate to differences in the quantity of other dust
present, the fiber size distribution, the age of the population under obser-
vation, and the representativeness of the dust sampling program. These
factors will be discussed later when the exposure-response relationships of
all available studies are compared (Section 3.7). Further, when exposure-
response relationships are analyzed according to duration and intensity of
exposure (McDonald et al., 1980), the results are far less dramatic than those
shown in Figure 3-1. However, this may be the result of small numbers; only
46 excess lung cancer deaths are reported in all exposure categories.
Fewer data are available on the exposure-response relationship for meso-
thelioma. Table 3-2 lists the mesothelioma mortality from three studies
(Seidman et al., 1979; Hobbs et al., 1980; Jones et al., 1980) in terms of
cases per 1,000 person-years of observation beginning 10 years after first
14
-------
"2 12
+••
u
o>
Q.
X
d>
<5
V)
o
cc
UJ
o
z
o
C3
Z
D
O
CC
LU
Q
LL
O
X
CO
cc
3
LLI
oc
10
• Dement et al. (1982)
SI Henderson and Enterline (1979)
A Liddel! et a!. (1977)
I I I
0 1000 2000 3000
ESTIMATED DOSE OF ASBESTOS, MPPCF
Figure 3-1. Exposure-response relationship for lung cancer
observed in three studies. The cumulative exposures are
measured in terms of millions of particles per cubic foot-
years (MPPCF). Note that the exposure values for the
circles are to be multiplied by 1/10.
15
-------
TABLE 3-2. THE RISK OF DEATH FROM MESOTHELIOMA ACCORDING TO THE TIME
OF ASBESTOS EXPOSURE IN THREE STUDIES
Study
Hobbs et al .
Seidman et a'
Jones et al .
Exposure
period,
months
(1980)
< 3
3 - 11
12+
1. (1979)
2.2
7.1
15.4
57
(1980)
< 5
5 - 10
10 - 20
20 - 30
30+
Number
of
deaths
0
10
16
0
3
4
7
0
3
4
4
5
Estimated
person-years
10+ years
from first
exposure
21,213
19,548
14,833
6,640
2,000
2,290
2,480
Deaths/
1000 Percent
Person- Number of
years exposed deaths
0
0.5
1.1
0
1.5
1.7
2.8
314 0
116 2.6
145 2.8
101 4.0
51 9.8
exposure. While few deaths are available for analysis, the data for exposure
periods longer than 3 to 5 months are consistent with a linear relationship.
No deaths from mesothelioma were observed in any of the lowest exposure cate-
gories, whereas 1 to 2 would have been expected in each study on the basis of
a linear dose-response relationship. Similarly, data of Newhouse and Berry
(1979) (Table 3-3) show an increasing risk of mesothelioma with increasing
duration and intensity of exposure. However, a quantitative relationship
cannot be determined.
This document uses a linear exposure-response relationship for estimating
unit exposure risks and for calculating risks at cumulative exposures 10 to
100 times less than those of the occupational circumstances of past years.
This relationship is plausible and no evidence contradicts it, although the
strength of the evidence supporting it is limited. The method has three
16
-------
TABLE 3-3. INCREASING RISK OF MESOTHELIOMA WITH INCREASING DURATION
AND INTENSITY OF EXPOSURE
Males
Females
Duration of
exposure
< 2 yrs
> 2 yrs
< 2 yrs
> 2 yrs
Intensity of
Low-moderate
33
93
{48}
exposure
Severe
104
243
136
360
Source: Newhouse and Berry (1979).
distinct advantages: 1) point estimates of exposure-response can be made
without knowledge of individual exposures, i.e., the excess mortality of an
entire group can be related to the average exposure of the group, 2) extrapo-
lation (or interpolation) to various exposure circumstances can be made easily,
and 3) this procedure is probably conservative from the point of view of human
health. Linearity of exposure-response applies only for similar times of
exposure and observation, among similarly aged individuals, with similar
personal habits.
3.4 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 descriptions of mortality in
terms of observed and expected deaths. Virtually every study of asbestos
workers is described in these terms. Early suggestive evidence supporting
this model 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 have been discussed by Enterline (1976), Peto (1977), and
Nicholson (1982) and have been utilized in projections of lung cancer from
past asbestos exposure by Nicholson et al. (1982). Information on lung cancer
risk from exposures at different ages is available from two studies (Selikoff
et al., 1979; Seidman et al. , 1979). The analyses of these data provide
substantial support for the use of such a formulation for lung cancer.
17
-------
Information from the insulation workers study on the time course of
asbestos cancer risk is given in Figure 3-2, which shows the relative risk of
death from lung cancer (the ratio of observed deaths to expected deaths)
according to age for individuals first employed between ages 15 and 24 and for
those employed between ages 25 and 34. The two curves in Figure 3-2 rise with
the same slope and are separated by the 10 years of difference in age at first
exposure. This result 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
3-2 suggests that the data be combined and plotted according to time from the
onset of exposure. The result is shown in Figure 3-3. The data of Figure 3-3
are plotted according to years from the onset of exposure. However, because
of the great stability of union insulation work, the curve also reflects
effects according to the duration of exposure up to at least 25 years from the
onset of exposure. A linear increase with time from the onset of exposure is
seen for about 35 years (to about the time when many insulators would have
terminated employment), after which the relative risk falls substantially
rather than remain constant as would be expected from the linear increase with
continued exposure. The decrease is partially the result of the earlier
deaths of smokers from the group under study because of their higher mortality
from lung cancer and cardiovascular disease. However, the decrease is not
solely the result of the deaths of smokers; a similar rise and fall occurs
among those individuals who were smokers at the start of the study compared to
smokers in the general population. Part of the decrease may relate to the
elimination of asbestos, particularly chrysotile, from the lung; from select-
ion processes, such as differing exposure patterns (e.g., the survivors may
have avoided intense exposures); from cohort effects; or from differing indi-
vidual biological susceptibilities. While the exact reason for the effect is
not understood, it is a general phenomenon seen in other mortality studies of
asbestos workers (Nicholson et al., 1979; 1983).
The early portions of the curves of Figures 3-2 and 3-3 have three impor-
tant features. First, after a short delay, the curves show a linear increase
in the relative risk of asbestos lung cancer according to time from onset of
exposure. Second, Figure 3-3 shows that this increased relative risk is
18
-------
(0
a
a>
•o
16
Q.
X
0>
2
I 5
(0
®
•o
•o
o>
0) 4
(0 H
xi
o
C/3
OC
111
OC
AGE AT ONSET
« 15-24 YEARS
O 25-34 YEARS
30
40
50
60
AGE
70
80
90
Figure 3-2. 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).
-------
in
.c
*^
CD
0)
T3
m
o.
x
0)
(0
ID
•o
T3
0)
0)
tf)
oc
LU
§
UJ
cc
• ALL WORKERS
O WORKERS WHO SMOKE CIGARETTES
10
20
30
40
50
60
YEARS FROM ONSET OF EXPOSURE
Figure 3-3. 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).
20
-------
proportional to the time worked, and, thus, to the cumulative asbestos ex-
posure. However, the linear rise can occur only if the increased relative
risk that is created by a given cumulative exposure of asbestos continues to
multiply the underlying risk for several decades 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 about
5 years from the onset of exposure. This result shows that the increased
relative risk appropriate to a given exposure is achieved soon after the
exposure takes place. However, if there is a low underlying risk at the time
of the asbestos exposure (e.g., for individuals aged 20 to 30), most of the
cancers that will arise from any increased risk attributable to asbestos will
not occur for many years or even decades until the underlying risk becomes
substantially greater
The data of Seidman et al. (1979) also show that exposure to asbestos
multiplies the pre-existing risk of lung cancer and that the multiplied risk
becomes manifest in a relatively short time. Figure 3-4 depicts the time
course of lung cancer mortality beginning 5 years after the onset of exposure
of a group exposed for short periods of time. The average duration of expo-
sure was 1.46 years; 77% of the population was employed for less than 2 years.
Thus, exposure had largely ceased prior to the beginning of the follow-up
period. Figure 3-4 indicates that a rise to a significantly elevated relative
risk occurred within 10 years, and that the increased relative risk remained
constant throughout the observation period of the study Furthermore, the
relative risk from a specific exposure was independent of the age at which the
exposure began, whereas the excess risk would have increased considerably with
the age of exposure. Seidman et al. (1979) studied the relative risk of death
from lung cancer for individuals exposed for less" than and greater than
9 months. Table 3-4 lists their data according to the individual's age at the
time of entrance into a 10-year observation period. Within a given age cate-
gory, the relative risk was similar during different decades from onset of
exposure, as indicated in Figure 3-4 with the overall data. However, the
relative risk also was independent of the age decade at entry into a 10-year
observation period (see lines labeled "All" in each exposure category). There
was some reduction in the oldest groups. This decrease may be attributed to
the same effects manifest at older ages in insulators and to relatively fewer
cigarette smokers who might be present in the older groups because of selec-
tive mortality.
21
-------
10
.c
*-
(0
0)
•o
•a
a>
+*
o
a)
a
x
a>
tO
0)
•o
TJ
0)
5
in
.a
o
CO
QC
I 2
O
0 10 20 30 40
YEARS FROM ONSET OF EXPOSURE
Figure 3-4. The relative risk of death
from lung cancer among amosite
factory workers according to time
from onset of exposure.
Source: Seidman et al. (1979).
22
-------
TABLE 3-4. RELATIVE RISK OF LUNG CANCER DURING 10-YEAR INTERVALS
AT DIFFERENT TIMES FROM ONSET OF EXPOSURE
Years from
onset of
exposure
Lower exposure
5
15
25
Age
30-39
(<9 months)
0.00 [0.35]a
6.85 (1)
at start of period, years
40-49
3.75 (2)b
4.27 (3)
2.73 (2)
50-59
0.00 [3.04]
2.91 (4)
4.03 (6)
All 3.71 (1) 3.52 (7) 2.58 (10)
Higher exposure (>9 months)
5
15
25
All
0.00 [0.66]
19.07 (2)
11.12 (2)
11.94 (4)
11.45 (5)
13.13 (6)
12.32 (16)
9.93 (8)
5.62 (5)
7.41 (8)
7.48 (21)
a[] = No cases seen. Number of cases expected on the basis of the average
relative risk in the overall exposure category.
() = Number of cases.
Source: Seidman et al. (1979).
In terms of carcinogenic mechanisms, asbestos appears to act like a lung
cancer promoting agent. However, because of the continued residence 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 the fibers
remain continuously available in the lung to act after other necessary carcin-
ogenic processes occur.
A feature of Figure 3-3 that is important in the assessment of asbestos
carcinogenic risk is the decrease in relative risk after 40 years from the
onset of exposure or 60 years of age. As mentioned previously, this decrease
is not completely understood but it has generality. A virtually identical
time course of lung cancer risk occurs in asbestos factory employees (Nichol-
son et al., 1983) and in Canadian chrysotile miners and millers (Nicholson et
al., 1979). Because of the significant decrease at long times from the onset
of exposure and older ages, observations on retiree populations can seriously
23
-------
understate the actual risk of asbestos-related death during earlier years. To
the extent that time periods between 25 and 40 years from the onset of expo-
sure are omitted from observation, a study will underestimate the full impact
of asbectos exposure on death.
To appreciate the effect of the observed lung cancer time-dependence upon
the results of an epidemiological study, the excess risk of lung cancer was
calculated for different observation periods for a hypothetical group that was
exposed for 25 years beginning at age 20. The time course of the risk was set
proportional to that of Figure 3-3, and 1977 general population rates were
used (NCHS, Annually: 1967-1977). Table 3-5 lists the percent excess lung
cancer mortality observed for three follow-up periods, 10 years, 20 years, and
lifetime, beginning at different ages. The table indicates that the percent
excess risk from start of exposure at age 20 to the complete death of all
cohort members is 55% of the maximum that would be achieved 32.5 years after
onset of exposure. The percent excess risk rises up to age 50 because the
follow-up period starts later, reflecting the increased relative risk con-
comitant with increased exposure. For observations starting after age 50, it
falls substantially, such that follow-up begun at age 65 observes only 38% of
the full risk.
To the extent that a group under observation has an age distribution that
is similar to the number alive in each quinquennium in a lifetime follow-up,
an observation for any period of time would reflect the same mortality ratio
TABLE 3-5. 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
02
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
The maximum expressed risk is that manifest 7.5 years after the conclusion
of the 25-year exposure.
24
-------
as an observation from the onset of exposure to the death of the entire co-
hort. To some extent, this situation applies to insulation workers, although
they have fewer older individuals than would occur had their mortality been
governed by general population data. (Their higher risk leads to an earlier
death and there is some loss due to lapse in cohort membership.) Since very
old groups are underrepresented, the excess relative risk of 3.60 (4.60-1)
(BE) documented by Selikoff et al. (1979) overestimates the age 20 to 85+
risk, calculated in this document as 2.53: [excess relative risk at 32.5 years
(5.6-1) (Figure 3.3) X reduction for lifetime exposure (0.55) (Table 3-5)].
The data in Table 3-4 came from observations of long-term exposures to
high concentrations of asbestos (>10 f/ml), where preferential death of sus-
ceptible individuals occurred. Thus, appropriate comparisons between heavily
exposed groups could be made on the basis of lifetime risk (i.e., 55% of the
maximum). 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 susceptibles. A lesser effect would also be
expected from short-term exposures (to less than extreme concentrations). For
such lower exposures, a relative risk unaffected by selection effects probably
would best represent the exposure circumstances. Such risks (at high expo-
sure) are seen in the rising slope of Figure 3-3 and the relative risk in
Figure 3-4. Other studies will likely be affected by selection effects to
some extent.
The above discussion supports a general model for lung cancer in which
the excess risk that occurs t years from the onset of exposure is proportional
to the cumulative exposure to asbestos at time t-10 years times the age and
calendar year risk of lung cancer in the absence of exposure. The incidence
of lung cancer can be formally expressed by
IL(a,y.t,d,f) = IE(a,y) [1 + K^f-d(t-lO)] (3-1)
Here, I.(a,y,t,d,f) is the lung cancer mortality observed or projected in
a population of age a, observed in calendar period y, at t years from the
onset of an asbestos exposure of duration d, and intensity f. I^(a,y) is the
age and calendar year lung cancer mortality expected in the absence of expo-
sure. If smoking data are available, IL and I£ can be smoking-specific
incidences. In this case, f is the intensity of asbestos exposure in fibers
25
-------
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, which is a measure of the
carcinogenic potency of the asbestos exposure. Alternatively, (f-d) is the
cumulative exposure (to 10 years prior to observation) and KL is the slope of
the exposure-response relationship. A delay in manifestation of risk is based
on the data of Seidman et al. (1979) and Selikoff et al. (1979). Equation 3-1
illustrates that the relative risk of lung cancer, IL/IE> 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
dimensionality is an important variable in fiber carcinogenicity. Thus, K^
would be expected to depend, to some extent, on fiber type and dimension. In
practice, however, uncertainties in establishing quantitative dose-response
relations, through the application of Equation 3-1 to observed data, may
preclude the determination of K. by fiber type.
3.5 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, which was first identified by Selikoff et al. (1968). Recent data
on U.S. insulation workers confirm and greatly extend the initial findings
(Hammond et al., 1979a). In this extensive study, 12,051 asbestos workers, 20
or more years from the onset of their exposure, were followed from 1967
through 1976. At the outset, 6,841 workers volunteered a history of cigarette
smoking while 1,379 said they had not smoked cigarettes. By January 1, 1977,
299 deaths had occurred among the cigarette smokers, and 8 deaths occurred
among workers who had not smoked cigarettes.
This experience was compared on an age- and calendar-year-specific basis
with that of comparable workers who had the same smoking habits and were a
part of the American Cancer Society's prospective Cancer Prevention Study
(Hammond, 1966). A total of 73,763 white males who had only a high school
education and were exposed to dusts, fumes, gases, or chemicals during non-
farming work were selected for the control group. The age standardized rates
per 100,000 person-years for each group are shown in Table 3-6. The results
show that both the smoking and non-smoking lung cancer risk is multiplied five
times by the insulator's asbestos exposure. However, the risk is low for
26
-------
TABLE 3-6. 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
smoking?
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).
non-smokers; therefore, multiplying it five times does not result in many
cases, although any excess is undesirable. On the other hand, smoking by
itself causes a major increase and when that high risk is multiplied five
times, an immense increase is found. Corroborative data on the multiplicative
smoking-asbestos interaction are seen in studies by Berry et al. (1979) and
McDonald et al. (1980).
The study by Hammond et al. (1979a) carried the asbestos-smoking interaction
a step further, to show increased risk of death of asbestosis. As noted pre-
viously, insulation work carried a risk of fatal progressive pulmonary fibrosis,
and some workers who never smoked cigarettes died of asbestosis. Nevertheless,
asbestosis mortality for workers who smoked 20 or more cigarettes a day was
2.8 times higher than that for workers who never smoked regularly. Cigarette
smoking, with resulting bronchitis and emphysema, adds an undesirable and som-
etimes unsupportable burden to the asbestos-induced pneumoconiosis. Inter-
active effects between cigarette smoking and the prevalence of X-ray abnorma-
lities have been reported previously (Weiss, 1971). However, no relationship
between cigarette smoking and the risk of death from mesothelioma or gastro-
intestinal cancer was found in the Hammond et al. study (Seidman, quoted in
Frank, 1979).
27
-------
3.6 METHODOLOGICAL LIMITATIONS IN ESTABLISHING DOSE-RESPONSE RELATIONSHIPS
Establishing dose-response relationships for human exposure to asbestos
is associated with substantial difficulties. Perhaps the most important
consideration is that current health effects are the result of exposures to
dust in previous decades when few and imperfect measurements of fiber concen-
trations were made. Current estimates of those concentrations can be inaccu-
rate because individual exposures were highly variable. Further, while
disease response now can be established through epidemiological studies, these
also can be misleading because of methodological limitations. Despite this
possible inaccuracy, useful estimates of risk can be made to provide an
approximate measure of asbestos disease potential in environmental circum-
stances. Limitations of existing data can be taken into account and their
recognition can stimulate appropriate research to fill identified gaps.
One limitation on the accuracy of exposure-response data for asbestos
disease is the lack of information concerning past fiber exposures of those
populations whose mortality or morbidity have been evaluated. Relatively few
measurements were made in facilities that used asbestos fibers before 1965.
Further, those measurements quantified all dust (both fibers and particles)
present in the workplace air Current techniques, which use membrane filters
and phase contrast microscopy for the enumeration of fibers longer than 5 urn,
have been utilized in Great Britain and the United States only since 1964
(Ayer et al., 1965; Holmes, 1965) and have been standardized in the United
States only since 1972 (NIOSH, 1972; 1979) and even later in Great Britain
(Advisory Committee on Asbestos, 1979a, b).
Modern counting techniques may be used 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 and such retrospective estimates are necessarily uncertain. Alterna-
tively, fiber counting techniques and the particle counting instrumentation of
earlier years can be used together to s-imul taneously evaluate a variety of
asbestos-containing aerosols. The comparative readings serve as a "calibra-
tion" of the historic instrument in terms of fiber concentrations. Unfortu-
nately., the calibration depends on the type and size distribution of the
asbestos used in the process under evaluation and the quantity of other dust
present in the aerosol. Thus, no universal conversion has been found between
earlier dust measurements and current fiber counts.
28
-------
In the United States and Canada, those few data that were obtained on
asbestos workers' exposures before 1965 were based primarily upon total dust
concentrations that were measured using a midget impinger. Fibers were in-
efficiently counted with this instrument because bright field microscopy was
used. Attempts to compare fiber concentrations with midget impinger particle
counts generally showed poor correlations (Ayer et al., 1965; Gibbs and
LaChance, 1974). Figure 3-5 provides an illustration of these correlations.
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 does not allow accurate evaluation of fiber concentrations and the
variability in the correlation between fiber measurements and thermal precipi-
tator data is reported to be large (Advisory Committee, 1979b) but no specific
data were given.
Even with the advances in fiber counting techniques, significant errors
may be introduced into attempts to formulate general fiber exposure-response
relationships. The current convention of counting only fibers longer than
5 urn was chosen solely for the convenience of optical microscopic evaluation
(surveillance agencies are generally limited to such instrumentation). This
method does not necessarily correspond to any sharp demarcation of effect for
asbestosis, lung cancer, or mesothelioma. While it is readily understood that
counting only fibers longer than 5 urn enumerates but a fraction of the total
number of fibers present, there is incomplete awareness that the fraction
counted is highly variable. The results depend upon the fiber type, the pro-
cess or products used, and the past history of the asbestos material (e.g.,
old vs. new insulation material), and other factors. For example, the frac-
tion of chrysotile fibers longer than 5 urn in an aerosol can vary by a factor
of 10 (from as little as 0.5% of the total number to more than 5%). When
amosite aerosols are counted, the fraction longer than 5 urn may be 30%, ex-
tending the variability of the fraction counted to two orders of magnitude
(Nicholson et al., 1972; Nicholson, 1976a; Winer and Cossete, 1979).
Even if consideration is restricted to fibers longer than 5 jjm, many
fibers are missed by optical microscopy. Using electron microscopy, Rendall
and Skikne (1980) have measured the percentage of fibers with a diameter less
than or greater than 0.4 p.m (the limit of resolution of an optical microscope)
in various asbestos dust samples. In general, they found that more than 50%
of the 5 urn or longer fibers were less than 0.4 (jm in diameter and, thus,
29
-------
140
120
2 100
D
O
O
cc 80
cc
QQ
S
ILJ
60
40
20
o <1 FIBER PER FIELD
• 3*1 FIBER PER FIELD
10 15 20 25
MIDGET IMPIIMGER COUNT, MPPCF
30
Figure 3-5. A plot of membrane filter and midget impinger counts; MPPCF
represents millions of particles per cubic foot.
Source: Gibbs and LaChance (1974).
30
-------
would not be visible using a standard phase contrast optical microscope.
Moreover, the diameter distribution also varied with activity and fiber type.
As a result, the fraction of fibers that was longer than 5 urn and visible by
light microscopy varied from about 22% in chrysotile and crocidolite mining
and amosite/chrysotile insulation manufacture to 53% in amosite mining.
Intermediate values of 40% were measured in chrysotile brake lining manu-
facture and 33% in amosite mill operations. Thus, even perfect measurement of
workplace air, with accurate enumeration of fibers according to currently
accepted methods, would be expected to lead to different exposure-response
relationships for any specific asbestos disease when different work environ-
ments are studied. Conversely, risks estimated for a given exposure circum-
stance must have a large range of uncertainty to allow for the variability
resulting from fiber size effects.
Those uncertainties that exist in the physical determinations of past
fiber concentrations and the difficulty in evaluating the most important
exposure parameter in current measurements are exacerbated by the sampling
limitations in determining individual or even average exposures of working
populations; only a few workers at a worksite are monitored and then only
occasionally. Variability in work practices, ventilation controls, use of
protective equipment, personal habits, and sampling circumstances add consid-
erable uncertainty to available information on exposure.
Variability in exposure-response relationships obtained in different
studies can also be attributed to statistical variability associated with
small numbers and to methodological difficulties in the estimation of disease.
Studies can be significantly biased by inclusion of recently employed workers
in study cohorts, use of short follow-up periods, and improper treatment of
the various time factors that are important in defining asbestos cancer.
Inadequacies of tracing can lead to significantly inaccurate estimates of
disease. Generally, from 10 to 30% of an observation cohort will be deceased
(sometimes even less). If 10% of the group is untraced and most are deceased,
very large errors in the determination of mortality could result, even if no
person-years are attributed to the lost-to-follow-up group. Finally, the
choice of comparison mortality rates can introduce significant errors. Local
rates are generally 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
31
-------
not available, and existing general population rates may overstate the expec-
ted total mortality because of a "healthy worker effect" (Fox and Collier,
1976). Proper consideration of smoking habits is important in the determina-
tion of lung cancer risks. Unfortunately, full information on the smoking
patterns of all individuals in a cohort often is not available.
Thus in summary, calculations of unit risk values for asbestos must be
viewed with caution as they are uncertain and aspects of them are necessarily
based on estimates that are subjective to some extent because of the following
limitations in data: 1) statistical uncertainties and systematic biases in
epidemiological studies, 2) conversions of particle counts to fiber exposures
are uncertain, and 3) very importantly, the nonrepresentative nature of the
exposure estimates.
3.7 QUANTITATIVE DOSE-RESPONSE RELATIONSHIPS FOR LUNG CANCER
In theory, exposure-response relationships can best be determined from
studies in which individal 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 is deter-
mined from effects observed in all exposure categories. Consistencies in the
observed exposure-response relationships strengthen the risk estimates made
from such studies. In practice, however, the estimation of individual expo-
sures involves considerable uncertainties. An exposure estimate for each
worker must use historic data on particle counts and recent measurements of
the ratio of fiber to particle concentrations. Unfortunately, complete job
histories are not always available for each worker; often only employment
departments are known. Second, relatively few dust counts were made before
1965 and exposure data may not exist for many plant jobs. Third, few fiber-
particle comparison counts are made and these often demonstrate great varia-
bility (Figure 3-5). Finally, worker mobility may significantly alter his or
her exposure from that estimated at a work station. Systematic and random
biases that may occur from any of these uncertainties can lead to significant
alteration of a measured exposure-response relationship, even in studies that
demonstrate a near perfect linear relationship.
In some studies, individual exposures are not determined for each cohort
member, but only for cancer cases of interest and a selected number of con-
trols. Odds ratios are then calculated according to exposure, but they are
limited by the uncertainties of small numbers and confounding effects in
addition to all of the uncertainties discussed above.
32
-------
Finally, two studies will be considered for which information is avail-
able only for the group as a whole. Both studies used recent determinations
of fiber concentrations in work activities believed to represent those of
previous years. While the studies are not affected by the uncertainties of
fiber-particle conversions, they are uncertain because members of each group
were exposed to highly variable asbestos concentrations. In one case (insula-
tors), each worker experienced variable exposure, and in the other case (an
amosite insulation factory), different workers experienced different expo-
sures; however, a plant average exposure was estimated. Both estimates could
be in error to the extent that all jobs were not properly weighted in the
sampling program.
In the following analysis of 11 studies, all available exposure-response
information will be used. When such data are inconsistent or possible biases
are perceived, alternative analyses will also be undertaken (weighted regres-
sion analysis or use of averaged risk-exposure data). A value for K. will be
calculated for each study using either the slope of the observed dose response
data, the odds ratios of case control analyses, or the ratio of excess lung
cancer risk to average exposure. The calculations will generally use Equation
3-1, I. = Ir(l + K. • f • d). Rearranging one obtains
KL = [(IL - IE)/IE]/f • d 3-2a
or
• d 3-2b
The K. values obtained are listed in Table 3-7 and displayed in Figure 3-6.
The 95% confidence limits, calculated from the variance of the observed number
of lung cancer cases are also shown in Table 3-7 For example, consider the
study of Peto (1980). In a cohort exposed after 1950, 11 lung cancers were
observed and 3.35 were expected in the group followed 15 years after first
employment. From Equation 3-2a, KL = (11 3.35)73.35/250 f-yr/ml =
7.65/3.35/250 = 0.0091 f fiber-years/ml is abbreviated (f-yr/ml)]. [As dis-
cussed later, an appropriate exposure for the cohort is 250 f-yr/ml.] The
95% confidence limits on a Poisson variant of 11 are 5.4 and 19.7. Thus, the
range of KL will be from KL = 0.0024. (5.4 - 3. 35)/3. 35/250 to KL = 0.019,
(19.7 3. 35)/3. 35/250. The same procedure will also be used in estimating
33
-------
TABLE 3-7. COMPUTATIONAL DATA ON THE STATISTICAL VARIABILITY ASSOCIATED WITH K,
Selikoff et al .
(1979)
Seidman et al .
(1979)
Henderson and
Enterline (1979)
Weil! et al. (1979)
Finkelstein (1983)
Peto (1980)
(>1950)
Peto (1980)
(<1950)
Dement et al .
(1983a, b)
Berry and
Newhouse (1983)
Li dell et al .
(1977)
Nicholson et al.
(1979)
Rubino et al .
(1979)
KL
0.0091
0.068
0.0044
0.0051
0.067
0.0091
0.0009
0.042
0.0006
0.0006
0.0030
0.0055
Deaths
Expected Observed
105.6 429
18.5 83
23.3 62
(32.2)a 51
2.0 17
3.35 11
16.83 26
9.8 33
Case-control
184 230
7.5 20
Case-control
Excess
324.4
64.5
39.7
(17.8)a
15.0
7.65
9.17
23.2
calculations
46
13.9
calculations
Range of
observed deaths Range of K.
388.4 - 469.6 0.0079 - 0.010
65.1 - 100.9 0.0049 - 0.0087
46.6 - 77.4 0.0026 - 0.0060
37.0 - 65.0 0.0014 - 0.0094
9.9 - 27.2 0.035 - 0.110
5.4 - 19.7 0.0024 - 0.019
17.0 - 38.0 0.00002 - 0.0021
22.7 - 46.3 0.023 - 0.066
200.3 - 259.7 0.0002 - 0.001
12.2 - 30.8 0.0010 - 0.0083
Adjusted for low trace.
-------
0.2000
0.1000
0.0500
0.0200
0.0100
0.0050
0.0020
0.0010
0.0005
0.0002
(3-1
C
UJ
R
o>
§
to
C
eo
E
•Q
E —J
en
01
S3
®
03
S
0>
C
o
_«
o
en
o
[o
e
a.
tn
o
+j
a.
a
®
<-"
®
o>
•o —
Figure 3-6. The values for KL< the fractional increase in lung
cancer per f-yr/ml exposure in 11 asbestos exposed cohorts. The
shaded bar represents the 95% confidence limits on KL
associated with the statistical variability of number of cases
observed. The open bar represents adjustments associated with
possible biases. The fine represents estimated uncertainties
associated with exposure estimates.
35
-------
the variability in studies that provide exposure-response data by cumulative
exposure category. While the variation in K, could be calculated from expec-
ted variances of the individual exposure categories, the above procedure will
yield very similar results. In addition to statistical variations, possible
systematic biases considered in the analysis of each study will be displayed
in Figure 3-6. Finally, the effect of an additive ± two-fold range of un-
certainty in cumulative exposure will be indicated in Figure 3-6. This two-
fold range is a subjective choice, but it is felt to be a realistic estimate
of the uncertainty of all the sampling problems mentioned previously.
3.7.1 Insulation Application; United States (Chrysotile and Amosite)
The previously discussed mortality study of Selikoff et al. (1979) can be
combined with information on asbestos exposure to provide an exposure-risk
estimate. The data on insulators' exposure have been reviewed by Nicholson
(1976a) and are summarized in Table 3-8. Using the standard membrane filter
TABLE 3-8. SUMMARY OF AVERAGE ASBESTOS AIR CONCENTRATION DURING INSULATION
WORK3 Selikoff et al. (1979)
Average fiber concentration, f/ml
Light and heavy
Research group construction Marine work
Nicholson (1975) 6.3
Balzer and Cooper (1968)
Cooper and Balzer (1968) 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 microscopy
Murphy et al. (1971) 8.0
Fleisher et al. (1946) 30-40
Estimates of past exposure based on current membrane-filter data
Nicholson (1976a) 10-15
Average concentrations of fibers longer than 5 p,m evaluated by membrane
filter techniques and phase-contrast microscopy.
Source: Nicholson (1976a).
36
-------
technique of the U.S. Public Health Service for counting asbestos fibers
(NIOSH, 1979), three different laboratories in the United States have 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 used the
same technique in the Devonport Naval Dockyard in Great Britain, and obtained
8.9 f/ml for the average of long-term samples of asbestos concentrations
measured during the application of insulation materials aboard ship (Harries,
1971). The research that led to these data indicated that peak exposures
could be extremely high. For example, it was not uncommon for 2- to 5-minute
concentrations of asbestos to exceed 100 f/ml during the mixing of cement.
However, this mixing would be done perhaps once an hour. Thus, exposures mea-
sured during that hour, including the mixing, would seldom average more than
10 f/ml. Similar experiences were subsequently reported by Cooper and Miedema
(1973), who stated that "peak concentrations may be high for brief periods,
while time-weighted averages are often deceptively low."
Direct information on asbestos fiber concentration, measured by the
currently prescribed analysis procedures, is available only after 1966.
Insulation materials have changed from earlier years. Fibrous glass has found
extensive use, and cork is used rarely. Moreover, the asbestos composition of
insulation products has changed. Pipe covering and insulation block may have
had twice the asbestos content in past years as in the period from 1968 to
1970. However, during this period work practices were virtually identical to
those of previous years, and during the period of these measurements, few
controls of consequence were used. Thus, dust concentrations measured under
these conditions have relevance for the estimate of levels of past years.
Considering the possible doubling of asbestos content of insulation materials,
the data from the studies listed in Table 3-6 would suggest that the insula-
tors' average exposures in the United States during past years could have
ranged from 10 to 15 f/ml for commercial and industrial construction. In
marine construction, it may have been between 15 and 20 f/ml A value of 15
f/ml is used in this document as an overall average. However, because of the
great variability in work activities of this group, the range of uncertainty
in the exposure is estimated to be 10 to 45 f/ml; this range is indicated in
Figure 3-6.
This information and the data in Figure 3-3 allow the calculation of a
lung cancer risk per cumulative unit asbestos exposure (in f-yr/ml) from the
37
-------
linearly rising portion of the curve, the slope of which is 0.16 relative risk
units per year or 0.0107 per f-yr/ml (divided by an average exposure intensity
of 15 f/ml). However, the data of Figure 3-3 used BE in establishing lung
cancer mortality. Adjusting to DC diagnosis reduces the value of KL from
0.0107 to 0.0091; multiply by (3.06/3.60), the ratio of the DC to BE relative
excess risk in Table 3-1.
3.7.2 Insulation Manufacturing; Paterson, NJ (Amosite), Seidman et al. (1979)
The study by Seidman et al. (1979) also can be used for quantitative risk
estimates. While no data exist for air concentrations at the time the Pater-
son factory operated, information, in terms of fiber counts, exists for air
concentrations in two other plants that manufactured the same products with
the same fiber and machinery. One of these plants, located in Tyler, Texas,
opened in 1954 and operated until 1971 and the other plant, located in Port
Allegany, Pennsylvania, opened in 1964 and closed in 1972. Efforts to control
dust were limited in all three facilities. One plant 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 were pub-
lished in the Asbestos Criteria Document of the National Institute for
Occupational Safety and Health (NIOSH, 1972). The arithmetic averages of
these exposure measurements for Tyler (Plant X) and Port Allegany (Plant Y),
obtained using current fiber counting techniques, were 39.1 and 28.9 f/ml,
respectively, with an overall average of 34.9 f/ml. These two recently
operating plants had very similar average exposures; therefore, the Paterson
plant exposures probably did not differ significantly.
The mortality data presented by Seidman et al. (1979) are in a different
format from that usually encountered in epidemiological studies. Seidman
et al. compared the cumulative mortality, by cause, of a cohort of 820 asbes-
tos-exposed workers with a similarly aged hypothetical control population
followed over the same calendar years. Thus, the number of expected deaths in
a time period is based on the number of individuals expected to be alive at
the start of the period, rather than on the number alive in the exposed popu-
lation at the start of the period. Because the mortality of the cohort is
considerably above that expected, the number assumed alive at the start of
later observation periods is much greater than the actual number. Table 3-9
lists the exposure groups of Seidman et al (1979), the average work period of
38
-------
TABLE 3-9. OBSERVED AND EXPECTED CUMULATIVE PROBABILITY OF DEATH FROM LUNG CANCER 5 THROUGH 35
ELAPSED YEARS SINCE THE ONSET OF WORK IN AN AMOSITE ASBESTOS FACTORY,
1941-1951, BY LENGTH OF TIME WORKED
Length of
time worked
>1 mo.
1-2 mo.
2-3 mo.
4-6 mo.
6-12 mo.
1-2 yr.
2+ yr.
All times
Number
of men at
5-year point
61
90
82
148
125
125
188
820
Average
exposure
time, years
0.04
0.09
0.17
0.29
0.59
1.28
4.77
1.46
Estimated
average dose
f-yr/ml
1.4
3.2
5.9
10.2
20.6
44.8
166.9
51. 1C
Expected
percentage
of deaths
2.95
2.70
2.79
2.47
2.15
2.02
2.34
2.40
Observed
percentage
of deaths
6.07
7.34
7.42
5.90
10.21
12.41
18.51
10.71
(DC)
(3)b
(5)
(6)
(8)
(12)
(15)
(34)
(83)
Ratio
2.06
2.72
2.66
2.38
4.74
6.14
7.91
4.46
Adjusted to a person-years-at-risk basis.
() = number of lung cancer deaths.
Person-weighted average.
Source: Seidman et al. (1979).
-------
each group, the estimated cumulative exposure using 35 f/ml as the average
intensity of exposure for the group, the observed cumulative percentages of
deaths (DC), and the expected cumulative percentages of death, adjusted to a
person-years-at-risk basis.
A group average cumulative exposure of 51 f-yr/ml is calculated from the
work duration of all cohort members. This average gives a value of 0.068 for
K, [(10.71 observed/2.40 expected -1)/51 f-yr/ml] (using Equation 3-2b and
data from Table 3-7). The high Standard Mortality Ratios (SMR's) at low
durations of exposure suggest that general population rates may be inappro-
priately low for the study group, because all of the short-term exposure
categories are proportionately higher than expected (by extrapolating from the
longer exposure period data). The underestimate of expected rates may be a
factor of 2; this would correspondingly lower K. in Figure 3-6.
3.7.3 Asbestos Products Manufacturing; United States (Chrysotile and
Crocidol ite), Henderson and Enterline (1979)
The data of Henderson and Enterline (1979) (Figure 3-1) can be used to
establish fiber dose-response data even though their data were presented in
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. Data do exist on the relationship between fiber and
total dust concentrations in textile operations and asbestos cement production.
Dement et al. (1982) found a conversion of 3 f/ml/mppcf was appropriate to
most textile operations, although Ayer (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. The lower value prob-
ably would be most applicable to the Henderson and Enterline circumstance
because of the extensive use of cement and other mineral particles in asbestos
products manufacturing. The least squares regression line through the points
in Figure 3-1 is SMR = 100 + 0.66 x mppcf. Using a value of 1.5 f/ml/mppcf to
represent the conversion relationship, the estimate of K, is 0.0044
(0.66/100/1.5). (Dividing by 100 to convert an increase in SMR to an increase
in relative risk. )
As described previously, observing a cohort beginning at age 65 seriously
understates the full impact of asbestos exposure. Most of the workers whose
mortality experience was graphed in Figure 3-1 began employment before age 25.
40
-------
It was estimated that a study of a retiree cohort could understate mortality
by as much as 60% relative to the maximum observable risk (Table 3-3). A
possible 2.5-fold increase in the value of K, is indicated in Figure 3-6.
3-7-4 Asbestos Cement Products; United States (Chrysotile and Crocidolite),
Wei 11 et al. (1979), Hughes and Wei 11 (1980)
A study of an asbestos cement production facility provides exposure-
response information (Wei 11 et al,, 1979; Hughes and Wei 11, 1980). However,
the data quality is limited because of uncertainties in the mortality data.
While the experience of 5,645 individuals was reported, only 1,791 had been
employed for longer than 2 years. Thus, exposures were limited for most
cohort members. More significantly, tracing was accomplished through informa-
tion supplied on vital status by the Social Security Administration. This
method allowed the vital status of only 75% of the group to be determined.
Those individuals untraced were considered alive in the analyses. This assump-
tion can lead to serious misestimates of mortality because before 1970, many
deaths, particularly of blacks, were not reported to the Social Security
Administration. The percentage of unreported deaths of both sexes ranged from
nearly 80% in 1950 to 15% in 1967 (Aziz and Buckler, 1980). Thus, many cohort
members who were considered alive could be deceased. This inaccuracy is
likely to be the source of the extraordinarily low overall reported mortality
of the cohort, with deficits of about 40% commonly seen in several exposure
categories. (The overall SMR is 68.)
Two methods can be used to adjust an incomplete trace. In one method,
the overall SMR for lung cancer, 104, is divided by the SMR for non-asbestos
related causes to give a corrected relative risk for lung cancer. This method
yields a value for K. of 0.0060, using a value of 64 mppcf for the group
exposure and a fiber-particle conversion factor of 1.4 (Harnmad et al., 1979)
[(104/68) - l]/64/1.4 (Cf. equation 3-2b). Alternatively, a regression of SMR
on dose yields SMR = 77 + 0.46 x mppcf. The low value of SMR at zero exposure
probably is the result of missing deaths. If the percent missing is similar
in each category, then KL = 0.0043, (0.46/100/1.4/0.77), where the 3 divisions
account for conversion of SMR to relative risk, mppcf to f/ml, and to a SMR of
100 at zero dose. The average of these values, 0.0052, will be used for the
point estimate of K. . The assumption that there is an equal percentage of
missing deaths in each category is uncertain. There are more untraced deaths
41
-------
in the lowest category (J. Hughes, personal communication). However a greater
percentage of those untraced in the most exposed group may be deceased
(because of longer exposure and greater age). If all of the untraced deaths
are assumed to be in the three lowest exposure categories and the regression
line for SMR is forced through the origin, its slope is 0.040; (mppcf); K. is
0.0029. This downward adjustment is indicated in Figure 3-6.
3.7.5 Asbestos Cement Products; Ontario, Canada (Chrysotile and Crocidolite),
Finkelstein (1983)
A recent study by Finkelstein (1983) 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. The cohort consisted of all individuals who had 9 or more
years of employment beginning before 1960. Their mortality experience was
followed through October 1980. (An expanded cohort of 751 workers who had 1
or more years of employment has also been reported by Finkel stein (1982b), but
is not yet published. This cohort yields virtually identical unit risk values.)
Impinger particle counts of varying degrees of comprehensiveness were avail-
able from various sources (government, insurance company, employer) from 1949
until the 1970's. After 1973, membrane fiber counts were taken. Individual
exposure estimates were constructed, based on recent fiber concentrations at a
particular job, and modified for earlier years by changes in dustiness of that
job, as determined by the impinger particle counts. For example, exposure
estimates for the years 1948 to 1954 for willow operators, forming machine
operators, and lathe operators were 40 f/ml, 16 f/ml, and 8 f/ml, respectively.
The average cumulative 18-year exposure for the production group in the
asbestos cement work was 112.5 f-yr/ml. Seventeen lung cancer deaths were
observed versus 2.0 expected deaths from Ontario rates for an SMR of 850 or a
relative risk of 8.5. Three deaths versus 2.3 expected occurred in an unex-
posed group. This result yields a value of KL = 0.067 [(8.5-1)7112.5]. Data
also are presented on the lung cancer SMRs for separate cumulative exposure
categories, but they are so variable because of the few deaths in each ex-
posure category that no exposure-response relationship can be obtained. The
first two exposure categories show risk increasing steeply with exposure, but
the last falls significantly, although an extreme mesothelioma and GI cancer
risk occurs in the category.
42
-------
The reasons for the very significant difference in risk seen in two
plants (of the same company) producing the same product are unknown. The
point estimate of risk from Finkelstein (1983) (K = 0.067) is 13 times that
of Well! et al. (1979) (KL = 0.0052) even after an attempt to correct for the
incomplete trace of the latter study. The exposure estimates of Finkelstein
would appear reasonable. In a study of asbestosis in the Ontario plant
(Finkelstein, 1982a), data comparable to that of Berry et al. (1979) were
obtained. Finkelstein observed prevalence rates of asbestosis of 4% at 50 to
99 f-yr/ml and 6% at 100 to 149 f-yr/ml versus Berry et al.'s 2.5% and 8.5%,
respectively. Henderson and Enter!ine (1979) observed SMR's of 231 and 522,
respectively, among retirees of cement sheet and shingle work and cement pipe
work. These values are more consistent with the higher risk of Finkelstein
than the lower one of Wei 11 et al.
3.7.6 Textile Products Manufacturing; Rochdale, England (Chrysotile), Peto
(1980)
The mortality experience from an oft-studied British textile plant (BOHS,
1968; Berry et al. , 1979; Knox et al. , 1968; Peto, 1980) is difficult to
interpret. First, dust concentrations have changed fairly dramatically over
the past 5 decades of plant operations. Subsequent estimates of those con-
centrations have changed also. No measurements of dust concentrations were
made before 1951. Between 1951 and 1964 thermal precipitators were used to
evaluate total dust levels, and thereafter, filter techniques similar, but not
identical to those in the United States, were used. Average fiber concen-
trations have been published for earlier years, based on a comparison of fiber
counting with thermal precipitator techniques (Berry, 1973). Unfortunately,
no published data exist on the variability of the correlation between these
two techniques, although they are stated to correlate "relatively poorly"
(Advisory Comm., 1979b). Earlier published estimates have been stated to be
inaccurate; Berry et al. (1979) reported that a re-evaluation of the work his-
tories 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%. Recently, coincident with the finding of consider-
able asbestos-related disease among recent (post-1951) employees and the
British Government's review of its asbestos standard, the hygiene officers of
the plant have re-evaluated previously reported exposure data. Data now
43
-------
suggest 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). Unfortunately, the data on which such revised estimates
were made were not provided in the text of the British Advisory Committee
Reports when the Advisory Committee accepted them (Advisory Comm. 1979a). The
comparative fiber concentration estimates are provided by Peto (1980) and
listed in Table 3-10. However, no background data are available.
TABLE 3-10. PREVIOUS AND REVISED ESTIMATES OF MEAN DUST LEVELS IN FIBERS/ML
(WEIGHTED BY THE NUMBER OF WORKERS AT EACH LEVEL IN SELECTED YEARS)
1936 1941 1946 1951 1956 1961 1966 1971 1974
Previous estimates
corresponding to 13.3 14.5 13.2 10.8 5.3 5.2 5.4 3.4
early fiber counts
(Peto et al., 1977)
Revised estimates 32.4 23.9 12.2 12.7 4.7 1.1
corresponding to No measurements
modern counting prior to 1951
of static samples3
These estimates are based on preliminary data on 126 workers, first employed
between 1951 and 1955, and should be regarded as provisional.
Source: Peto (1980)
Evaluation of the new estimates is further clouded by questions con-
cerning the appropriateness of multiplying static sampler concentrations by
two. This approach is directly contradicted by published factory data
(Table 3-11) on the comparison of static and personal sampling results by job
(Smither and Lewinsohn, 1973). Dr. Lewinsohn (personal communication) con-
firmed these results. He stated that the static sampler concentrations were
generally higher than those of the personal samplers of men workers at the
monitored job. The company placed the static samplers to best reflect the
breathing zone dust concentrations of operators while they tended machines.
Dr. Lewinsohn stated that if the machines were running smoothly, the worker
would often leave the site (e.g., to talk with fellow workers, go to the rest
room) and experience a lower dust concentration. The difference between
static and personal sampling data was greatest in the dustier jobs (compare
44
-------
TABLE 3-11. DUST LEVELS: ROCHDALE ASBESTOS TEXTILE FACTORY, 1971
Department
Fiberisl ng
Cardi ng
Spinning
Weaving
Plaiting
Process
Bag slitting
Mechanical bagging
Fine cards
Medium cards
Coarse cards
Electrical sliver cards
Fine spinning
Roving frames
Intermediate frames
Beami ng
Pirn weaving
Cloth weaving
Listing weaving
Medium plaiting
Static
3
4
3.5
4.5
8
1.5
2.5
6
5.5
0.5
1.5
2
0.5
4
Sampler
Personal
1
1
2
3.5
6
1
3
3
3
0.5
1
1
0.5
2
Source: Smither and Lewinsohn, 1973.
weaving vs. carding) because workers tended to leave a dusty area more fre-
quently 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). Thus, the fiber estimates published
by Peto (1980) reflect what is believed to be an improper adjustment and the
range of uncertainty in K will reflect this.
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. The published fiber concentrations (Peto, 1980) suggest that
the pre-1951 group was exposed to about 30 to 40 f/ml prior to 1965 and that
the post-1950 group was exposed to about 15 to 20 f/ml. In the pre-1951
group, 26 lung cancers occurred vs. 16.85 expected; in the post-1950 group
eleven occurred vs. 3.35 expected. It is anomalous that proportionally more
incidents of disease were seen in the latter group. An analysis by Peto
(1980) suggests that the cumulative exposure of the post-1950 group is 250
(200 to 300) f-yr/ml. This dose and mortality data 15 years after the onset
of exposure yields a value of KL = 0.0091, [(11-3.35)73.35/250] (using
Equation 3-2a). The corresponding estimate for the pre-1951 group, using
45
-------
600 f-yr/ml for the cumulative exposure, is 0.0009. The values for the older
group suffer from uncertainties in exposure estimates and those of the younger
group suffer from few deaths in the cohort. Both sets of data are negatively
influenced by the relatively short time since first exposure for many of the
cohort members. As indicated above, uncertainties in exposure estimates could
raise these estimates by a factor of 3.
The differences between the two subcohorts employed in this facility are
difficult to reconcile. The data are severely limited by the relatively small
size of the cohort and the few deaths available for analysis. Nevertheless,
the nearly 10-fold difference in the estimated risk of death from lung cancer
suggests the possible existence of some unidentified bias in the pre-1951
group. The finding of only a 50% increase in lung cancer in exposure circum-
stances where 5.3% of deaths were from asbestosis is certainly unusual, as is
the finding that virtually as many deaths occurred from mesothelioma as lung
cancer
3.7.7 Textile Products Manufacturing; United States (Chrysotile), Dement
et al. (1982, 1983a, 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. In this study, data from impinger measurements of total dust, in
terms of mppcf were available, characterizing dust concentrations since 1930.
Further, 1,106 paired and concurrent impinger-membrane filter measurements
allowed conversion of earlier dust measurements to fiber concentrations.
These conversions showed that 3 f/ml were equivalent to 1 mppcf for all oper-
ations except fiber preparation. (The 95% confidence interval was 2 to 3.5
f/ml/mppcf. ) A value of 8 f/ml/mppcf characterized fiber preparation work
(95% confidence interval: 5 to 9). After 1940, average fiber concentrations
in most operations were estimated to range from 5 to 10 f/ml with the ex-
ception of fiber preparation and waste recovery, where mean concentrations
were from 10 to 80 f/ml. A weighted regression line through all data plotted
according to cumulative fiber exposure yields SMR = 150 + 4.20 x f-yr/ml for a
KL of 0.042 (4.20/100).
Dement etal. (1982) used U.S. rates for calculating expected deaths.
County rates were 75% higher. Dement et al.'s arguments for the use of
national rates are persuasive. (Local rates were probably influenced by
46
-------
nearby shipyard employment and the smoking habits of the study population
reflected those of the U.S. general population.) However, a value of KL
reduced by 33% will be indicated in Figure 3-7 This value will bring the SMR
at zero exposure to 100 and allow for some consideration of unusually high
local rates.
3.7.8 Friction Products Manufacturing; Great Britain (Chrysotile and
Crocidolite), Berry and Newhouse (1983)
Newhouse and Berry (1983) analyzed the mortality of a large workforce
employed to manufacture friction products. All individuals employed after
1940 were included in the study and the mortality experience through 1979 was
determined. Exposure estimates were made by reconstructing work and ventila-
tion conditions of earlier years. Fiber measurements from these reconstructed
conditions suggested that exposures before 1931 exceeded 20 f/ml but those
afterwards seldom exceeded 5 f/ml. From 1970, exposures were less than 1
f/ml. These relatively low intensities of exposure kept the average cumula-
tive exposure for the group to less than 50 f-yr/ml.
The overall mortality of all study participants, 10 years and more after
the onset of exposure, was no greater than expected for all causes. The
number of deaths from cancer of the lung and pleura was slightly elevated in
men (151 observed vs. 139.5 expected) but the excess was largely accounted for
by eight mesothelioma deaths. No unusual mortality was found in study parti-
cipants employed 10 or more years. Using a case-control analysis according to
cumulative exposure, Newhouse and Berry estimated that the lung cancer in-
creased risk was 0.06% per f-yr/ml (K, = 0.0006) with an upper 90% confidence
limit of 0.8% per f-yr/ml.
3.7.9 Mining and Milling; Quebec, Canada (Chrysotile), Liddell et al. (1977),
McDonald et al. (1980)
The results reported by Liddell et al. (1977) on mortality with respect
to total dust exposure in Canadian mines and mills can be converted to rela-
tionships expressed in terms of fiber exposures. Using a slope of 0.0019
mppcf-yr as indicated in Figure 3-1, and a value of 3 f/ml/mppcf for the
particle fiber conversion factor, KL = 0.00063. The factor of 3 f/ml/mppcf is
the midpoint of the range of 1 to 5 f/ml/mppcf suggested by McDonald et al.
(1980) as applicable to most jobs in mining and milling.
47
-------
These studies of the Canadian miners are highly anomalous and indicate a
lung cancer risk lower than virtually any other study of asbestos workers.
First, the overall risk of lung cancer mortality in all miners is 1.25 times
that expected for the general population. Yet in studies of the mortality of
male residents of Thetford, in the midst of the Canadian asbestos mining area
(Toft et al., 1981; Wigle, 1977), an excess risk of 1.84 is seen in lung
cancer and 2.30 in cancer of the stomach. No corresponding increases were
seen in female cancer rates, therefore, Toft et al. (1981) and Wigle (1977)
attributed the excesses to occupational exposure in the mines. Siemiatycki
(1982) recently showed data from Asbestos and Thetford Mines, Quebec, which
indicated an SMR for lung cancer in males of 148 compared to Quebec rates
[which may be high by a factor of 1.5 compared to local rates (McDonald et
al., 1971). Second, internal inconsistencies exist in the McDonald et al.
(1980) analysis of the combined effect of asbestos exposure and cigarette
smoking. In the lower cumulative asbestos exposure category, the relative
risk of death of smokers compared to that of non-smokers is 11.8, as expected.
However, in the medium and high cumulative asbestos exposure categories, the
relative mortality risks of smokers to non-smokers are 6.6 and 3.6, respec-
tively. This result suggests the possibility of some misclassification of
asbestos exposure or of smoking. A final uncertainty of the studies is the
large percentage (10%) of untraced cohort members. The bias introduced by
such a large proportion of individuals is unknown. The studies do not indi-
cate how the untraced individuals were treated.
3.7.10 Mining and Milling; Thetford Mines, Canada (Chrysotile), Nicholson
et al. (1976b, 1979)
Higher risks were obtained by Nicholson et al. (1976b, 1979) from the
mortality experience of a smaller group of miners and millers employed 20 or
more years at Thetford Mines, Quebec. The 1979 publication indicates that 178
deaths occurred among 544 men who were employed during 1961 in one of four
mining companies. In the ensuing 16 years of follow-up, 26 deaths resulted
from asbestosis, 28 (25 on DC) resulted from lung cancer (11.1 expected), and
1 resulted from mesothelioma.
In this study, fiber measurements were made during 1974 in five mines and
mills, and data on particle counts were supplied by the Canadian Government.
From these data, exposure estimates were made for each of the 544 individuals
48
-------
according to their job history. Fiber exposures for earlier years were esti-
mated by adjusting current measurements by changes in particle counts observed
since 1950.
The mortality experience of the whole group has been reported by two
exposure categories (Nicholson, 1976b). The first exposure category corres-
ponded to a 20-year cumulative dust exposure of 560 f-yr/ml. The lung cancer
SMR in this group was 1.55 (7 observed, 4.52 expected). In the second cate-
gory, with a cumulative exposure of 1,760 f-yr/ml, the SMR was 4.33 (13 ob-
served, 3.00 expected). The ratio of the difference in excess risk to the
difference in cumulative exposure suggests that K, = 0.0023, (3.33 - 0.55)/
(1760 - 560). However, Quebec rates were used to estimate expected deaths,
and these may overestimate mortality. McDonald et al. (1971) stated that the
local rates of five contiguous counties are two-thirds those of the Province.
Thus, KL should be increased by a factor of 1.5 to 0.0034, or 0.0030 on the
basis of DC lung cancer diagnosis. Such an adjustment also makes a straight
line through the two SMR's that pass close to the value of 100. The effect of
not adjusting K. is indicated in Figure 3-6.
3.7.11 Mining and Milling; Italy (Chrysoti1e), Rubino et al. (1979)
A final study of chrysotile mining and milling is 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. However, the excess mortality was largely confined to non-
malignant respiratory disease, cardiovascular diseases, and accidents. The
overall SMR for all malignant neoplasms was 106, with only cancer of the
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 the relative risk for a
high exposure group (376 f-yr/ml) was 2.54 times that of a low exposure group
(75 f-yr/ml) [«L = 0.0051, (2.54-1)7(376-75)]. A case-control analysis of
the lung cancer according to cumulative dust exposure showed a relative risk
of 2.89.
49
-------
Thus, K. lies between 0.005 and 0.006 from the analyses according to dust
exposure. However, the relatively low overall risk for lung cancer in the
entire group (11 cases observed and 10.4 expected) suggests that the excess
risk coi'ld be zero.
3.7.12 Summary Dose-Response Relationships for Lung Cancer
The results of all the determinations of K. , the fractional increases in
lung cancer risk per f-yr/ml exposure are displayed in Figure 3-6, 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 3-7. The range of individual values of K. is large, and many of the
differences may be the result of statistical variation associated with small
numbers. Several studies have 95% statistical confidence limits exceeding an
order of magnitude. While the study of insulators could have the widest
uncertainty in exposure estimates, its low statistical variance gives it
considerable strength. Considering the statistical variability and other
uncertainties in the data, the agreement is fairly good. The ranges of all
but one estimate of K. lie between 0.005 and 0.03. The only estimate of K,
that lies outside this range is that made from the study of Liddell et al.
(1977). An average for K., weighted by the reciprocol of the variance of the
value of each study (with a lower cutoff at 0.0001), is 0.0095. No evidence
in this analysis suggests that a special carcinogenic potency is ascribable to
an individual type of fiber. Some of the highest and lowest values for K. are
obtained from pure chrysotile exposures. Exposures involving a mixture of
fibers, including amosite and crocidolite, also span a large range of values
for K^. Wide differences occur in the results of separate epidemiological
studies of nearly identical work conditions. This difference suggests a
midpoint estimate for KL of about 0.01, but with an uncertainty of about
three-fold.
3.8 TIME AND AGE DEPENDENCE OF MESOTHELIOMA
In contrast to lung cancer, for which a relative risk model accurately
explains the data, mesothelioma is best described by an absolute risk model,
in which the incidence of death is independent of the age at first exposure
and increases according to a power of time from the onset of exposure. The
rationale for such a model describing human carcinogenesis has been discussed
by several authors (e.g., Armitage and Doll, 1960; Pike, 1966; Cook et al.,
50
-------
1969). Such a model was utilized by Newhouse and Berry (1976) to predict
mesothelioma mortality among a cohort of factory workers in England. Specifi-
cally, they matched the incidence of mesothelioma to the relationship IM -
k M
c(t - d) , where IM is the mesothelioma incidence at a time t from onset of
exposure, d is a delay in the expression of the risk, and k is an empirically
derived constant. Additionally, the incidence of asbestos-induced mesotheli-
oma in rats (Berry and Wagner, 1969) followed this time course. In the case
of the analysis of Newhouse and Berry, the data suggested that the value of k
was between 1.4 and 2 and d was between 9 and 11 years. However, the rela-
tively small number of cases available for analysis led to a large uncertainty
in the values estimated for either k or d. Peto et al (1982) have recently
analyzed mesothelioma incidence in five groups of asbestos-exposed workers.
In one analyzed study, by Selikoff et al. (1979), the number of cases of meso-
thelioma was sufficiently large that the age dependence of the mesothelioma
risk could be investigated. Peto et al. (1982) showed that the absolute inci-
dence of mesothelioma was independent of the age at first exposure and that a
3 2
function, 1^ = ct , accurately represented the data for individuals between
20 and 45 years from the onset of exposure. However, observed incidence rates
for earlier times were less than those projected, and the authors suggested
2
that an expression proportional to (t - 10) better fit the data up to 45
years from the onset of exposure. The analysis of Peto et al. (1982) was
confined primarily to individuals who were first employed between 1922 and
1946; the fit to the mortality of the entire group (including those exposed
before and after that span) suggests a value of k greater than 3.2.
Figure 3-7 shows the risk of death from mesothelioma according to age for
individuals exposed first between ages 15 and 24 and between ages 25 and 34.
Although these data are somewhat uncertain because of small numbers, they
roughly parallel one another by 10 years as did the increased relative risk
for lung cancers. 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 mesothelioma among the
insulation workers is plotted according to time from the onset of exposure on
the right side of Figure 3-7 The risk increases at about 45 or 50 years from
the onset of exposure and then appears to decrease. Whether the decrease is
real or simply the result of misdiagnosis of the disease in individuals age 70
and older or the result of statistical fluctuations associated with small
numbers is not certain.
51
-------
DC
<
HI
>•
Z
o
AGE25yr.
I I
J L
10
20
40 60 80
20
40 60
AGE, years
YEARS FROM ONSET
OF EXPOSURE
Figure 3-7. 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).
-------
Mesothelioma risk from a short-term exposure can be considered to in-
!<•
crease at c(t - 10) , where k is between 2 and 4 and c is proportional to the
short-term cumulative exposure. Using a value of k = 3 (which best fits the
data for insulators) leads to the following relations for varying times of
exposure.
IM(t,d,f) = KM f[(t-10)3 - (t-10-d)3] t > 10+d (3-3a)
= KM f(t-10)3 10+d > t > 10 (3-3b)
=0 10 > t (3-3c)
IM is the mesothelioma mortality at t years from the onset of exposure to
asbestos for duration d at a concentration f KM represents the carcinogenic
potency and may depend on fiber type and dimensionality. IM depends only upon
exposure variables and not upon age or calendar year period.
Mesothelioma incidence is better represented by a model with a delay
if
period versus one that rises as t First, the delay period model fits the
full time course of insulator data better. Second, after 45 years from onset,
this model rises less rapidly than a function with no delay. The evidence
from two studies (Selikoff et al., 1979 - See Figure 3-7; Nicholson et a!.,
1983) shows that mesothelioma risk after 45 years from the onset of exposure
ceases to rise and perhaps falls. Thus, a function with a 10-year delay is
less likely to overstate the lifetime risk of mesothelioma in individuals who
were exposed early in life.
3.9 QUANTITATIVE DOSE-RESPONSE RELATIONSHIPS FOR MESOTHELIOMA
Four of the above studies provide information on the incidence of meso-
thelioma (pleural and peritoneal combined) according to time from the onset of
exposure and data that would allow estimates to be made of the duration and
intensity of asbestos exposure. Thus, values for K^, the potency factor for
mesothelioma risk in Equations 3~3a to 3-3c, can be estimated. Other studies
have reported cases of mesothelioma, but incidence data are lacking. In some
of these other studies, the incidence data are not provided. In others, data
were not given because very few mesothelioma deaths were seen. Thus, some
studies with missing data could be those in which a lower value of KM is
obtained and values of KM were estimated from a biased sample of those studies
in which K. was estimated. A measure of the bias can be estimated by compari-
53
-------
son of the values of KM and K. obtained in each analysis. The estimate of K^
for each of the four studies was made by calculating a relative mesothelioma
incidence using Equation 3-3 and data on duration and intensity of asbestos
exposure. The relative incidence curves were then superimposed on the ob-
served incidence data in each study and a value for K^ established. These
fits are depicted on Figures 3-8 and 3-9. The four studies are described
below and summary data are listed in Table 3-12.
3.9.1 Insulation Application; Selikoff et al. (1979), Peto et al. (1982)
A follow-up through 1979 of the cohort of insulators provides data on the
incidence of mesothel ioma with time from the onset of exposure (Peto et al.,
1982). Their time-weighted average exposure was estimated to be 15 f/ml
(Nicholson, 1976a). Using these data and 25 years for their average duration
_Q
of exposure, a value of KM = 1.5 x 10 is estimated.
3.9.2 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 35 f/ml for the average
_Q
exposure yields an estimate for KM of 5.7 x 10
3.9.3 Textile Products Manufacturing; Peto (1980), Peto et al. (1982)
A value of 30 f/ml is suggested by the data presented by Peto (1980).
However, this value is uncertain because discrepancies exist in the relative
exposures measured using personal samplers and static samplers (see above). If
the exposure measured by personal samplers are less than those from static
samplers, as suggested by the data of Smither and Lewinsohn (1973), the
average exposure could be about 15 f/ml. Using 30 f/ml and an employment
-ft
period of 25 years, a value of KM = 0.7 x 10 is estimated.
3.9.4 Asbestos Cement Products; Ontario, Canada, Finkelstein (1983)
The cumulative exposure of the cohort over 18 years was 112 f/yr. Only
men with 9 or more years of employment were included in the cohort. When data
on the exact duration and intensity of exposure are unavailable, a value of 12
years for duration of exposure and 9 f/ml for the intensity of exposure were
used. These figures yield a value of K= 1.2 x 10 .
54
-------
100
50
CO
cc
<
in
>
z
o
CO
cc
LLJ
Q.
CO
I
<
LLJ
Q
20
10
SELIKOFF ET AL (1979)
INSULATORS
SEIDMAN ET AL (1979)
AMOSITE FACTORY WORKERS
20 40 60 20
YEARS FROM ONSET OF EXPOSURE
40
60
Figure 3-8. The match of curves calculated using Equation 3-3 to data on
the incidence of mesothelioma in two studies. The fit is achieved for
"8
~8
KM = 1.5 x 10"° for insulators data and KM = 5.7 x 10"° for the
amosite workers data.
Source: Peto et al. (1982); Selikoff et al. (1979); Seidman et al. (1979).
55
-------
100
50
Cfl
cc
<
111
>
Z
O
CA
DC
LU
Q.
CO
I
LU
O
20
10
i—r~n—r
PETO (1980)
TEXTILE WORKERS
FINKELSTESN (1983)
CEMENT WORKERS
20 40 60 20
YEARS FROM ONSET OF EXPOSURE
40
60
Figure 3-9. The match of curves calculated using Equation 3-3 to data on
the incidence of mesothelioma in two studies. The fit is achieved for
KM = 0.7 x 10'8 for the textile workers data and KM = 1.2 x 10'7 for the
cement workers data.
Source: Peto (1980); Finkelstein (1983).
56
-------
TABLE 3-12. SUMMARY OF THE DATA ON KM> THE MEASURE OF MESOTHELIOMA RISK PER
FIBER EXPOSURE IN FOUR STUDIES OF ASBESTOS WORKERS
Study
Average
employment
duration
Average
exposure,
f/ml K,
M
Insulators
(Selikoff et al. , 1979;
Peto et al., 1982)
Textile workers
(Peto, 1980;
Peto et al., 1982)
Amosite factory workers
(Seidman et al., 1979)
Cement factory workers
(Finkelstein, 1983)
25
15
1.5 x 10
-8
1.6 x 10
-6
25
1.5
12
30
35
0.7 x 10
5.7 x 10
0.8 x 10
-6
0.8 x 10
-6
9 1.2 x 10"7 1.7 x 10 5
3.9.5 Summary of Quantitative Dose-Response Relationships for Mesothelioma
These data for these four studies are plotted in Table 3-12 and show re-
markable consistency between the ratio of KM/K, . The four studies suggest
— f> ML _q
that a ratio of KM/K, of 10 is appropriate and that a range of 3 x 10 to 3
_o " L
x 10 for KM would appear to represent most exposure situations, but several
studies suggest values outside this range.
3.10 ASBESTOS CANCERS AT EXTRATHORACIC SITES
The consistency of an increased cancer risk at extrathoracic sites and
its magnitude, either in absolute (observed-expected deaths) or relative
(observed/expected deaths) terms, is less for cancer at other sites than for
lung cancer. Nevertheless, many studies document significant cancer risks at
various GI sites. Cancer of the kidney has also been found to be signifi-
cantly elevated in two large studies (Selikoff et al. , 1979; Puntoni et al.,
1979). Among female workers, ovarian cancer has been found in excess (New-
house et al., 1972). While no other specific sites have been shown to be
elevated at the 0.05 level of significance, the category of all cancers other
than lung, GI tract or mesothelial is significantly elevated (e.g., Selikoff
et al., 1979)
57
-------
Table 3-13 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. This choice eliminated many small studies, which
have statistically uncertain data, from consideration, as well as several
large studies that demonstrated a low risk of lung cancer, either because of
exposure or follow-up circumstances. Because the excess risk of GI cancer is
less than that of the lung, significantly elevated risks are unlikely to be
seen in studies that demonstrate little lung cancer risk. Negative data in
such studies do not carry great significance. Data in Table 3-13, show that
all but one of the listed studies has an excess GI cancer risk, albeit in
three studies, the risk is small. However, five of the 13 studies demonstrate
the risk at a 0.05 level of significance. Figure 3-10 displays the relation-
ship between the relative risk of lung cancer and relative risk of GI cancer
in the 12 studies with excess GI cancer risk. A consistent relationship
exists between a greater GI cancer risk and an increased lung cancer risk. The
GI tract obviously is exposed to fibers because the majority of inhaled fibers
are brought up from the respiratory tract and swallowed (Morgan et al., 1975).
Additionally, some fibers may become entrapped within the gut wall (Storeygard
and Brown, 1977). Nevertheless, the magnitude of the excess fibers at GI
sites is much less than that for the lung. In recent studies, the GI excess
is about 10 to 15% of the lung excess.
Table 3-13 also lists the observed and expected mortality for cancers
other than mesothelioma and the GI or respiratory tract. The elevation is not
as consistent as that for GI cancer. Only three studies have elevated risks
that are significant at the 0.05 level and deficits are observed in four. 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 cancer
was not properly identified. In the study of insulators, Selikoff et al.
(1979) found that 26 of 49 pancreatic cancers were misclassified; most of the
misclassified were peritoneal mesotheliomas. As with GI cancer, the excess
at other sites is much less than the excess for lung cancer and generally less
than that for GI cancer.
58
-------
TABLE 3-13. OBSERVED AND EXPECTED DEATHS FOR VARIOUS CAUSES IN SELECTED MORTALITY STUDIES
en
Respiratory cancer
ICD 162-164
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Henderson and Enter! ine (1979)
McDonald et al . (1980)
Newhouse and Berry (1979) (male)
Newhouse and Berry (1979) (female)
Selikoff et al. (1979) (NY-NJ)
Selikoff et al. (1979) (U.S.)
Nicholson et al. (1979)
Peto (1977)
Mancuso and El-attar (1967)
Puntoni et al. (1979)
Seidman et al. (1979)
Dement et al. (1983b)
Jones et al. (1980)
0
63
230
103
27
93
429
28
51
30
123
83
33
12
E
23.3
184.0
43.2
3.2
13.3
105.6
11.0
23.8
9.8
54.9
21.9
9.8
3.8
0-E
39.7
46.0
59.8
23.8
79.7
381.4
17.0
17.2
20.2
68.1
61.1
23.2
8.2
0
55
276
40
20
43
122
10
16
15
94
28
10
10
Digestive cancer
ICD 150-159
E
39.9
272.4
34.0
10.2
15.0
84.1
9.5
15.7
7.1
76.6
22.7
8.1
20.3
0-E
15.1
3.6
6.0
9.8
28.0
37.9
0.5
0.3
7.9
17.4
5.3
1.9
(10.3)
(0-E)
WTJ
r
0.380
0.078
0.100
0.412
0.351
0.099
0.029
0.019
0.527
0.255
0.087
0.082
def.
ICD
0
55
237
38
33
28
184
10
18
20
88
39
11
35
Other cancers
except 150-49, 162-4, meso
E
45.6
217.4
27.4
20.4
24.5
131.8
16.1
24.8
6.8
81.3
35.9
14.1
39.5
0-E
9.4
19.6
10.6
12.6
3.5
52.2
(6.1)
(6.8)
13.2
6.7
3.1
(3.1)
(4.5)
rFi7°
0.237
0.426
0.177
0.529
0.044
0.137
def.
def.
0.653
0.098
0.037
def.
def.
0 = observed deaths.
E = expected deaths.
D = digestive cancer
R = respiratory cancer
o = other cancer
def. = no ratio when deficient in 0-E
-------
3.0
OBSERVED/EXPECTED DEATHS
FROM GASTROINTESTINAL CANCER
Figure 3-10. The ratio of observed to expected
mortality from lung cancer versus the ratio of
observed to expected mortality from
gastrointestinal cancer. See Table 3-13 for study
reference number 1-12. The point of Jones et. al.
(1980) with an SMR of 0.49 for digestive cancer is
not plotted.
60
-------
3.11 ASBESTOSIS
Asbestosis, the long-term disease entity resulting from the inhalation of
asbestos fibers, is a chronic, progressive pneumoconiosis. The disease is
characterized by fibrosis of the lung parenchyma, usually radiologically
evident only after 10 years from first exposure, although changes can occur
earlier following more severe exposures. Shortness of breath is the primary
symptom; coughing is a less common symptom; and signs such as rales, finger
clubbing, and, in later stages of the disease, weight loss appear in a propor-
tion of cases. The disease was first reported 8 decades ago (Murray, 1907)
and has occurred frequently among workers occupationally exposed to asbestos
fibers in ensuing years. Characteristic 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, parietal pleura may be involved.
Currently, 50 to 80% of individuals in some occupational groups with
exposures beginning more than 20 years earlier have been found to have ab-
normal X-rays. These individuals include asbestos insulation workers
(Selikoff et al., 1965), miners and millers (Nicholson, 1976b), and asbestos
factory employees (Lewinsohn, 1972). In many circumstances, the disease
progresses following cessation of exposure. In a group of workers employed in
an asbestos factory for various periods of time between 1941 and 1954, X-ray
changes were observed years following exposure in individuals having exposures
as short as 1 week (Personal communication, I.J. Selikoff).
In addition to disease and disablement during life, asbestosis accounts
for a large proportion of deaths among workers. The first reports of the
disease (Auribault, 1906; Murray, 1907) described complete eradication of
working groups. Much improvement in dust control has taken place in the
industry since the turn of the century, but even recently, those exposed in
extremely dusty environments, such as textile mills, may have up to 40% of
their deaths attributed to asbestos (Nicholson, 1976a). Groups experiencing
less severe exposures for 20 or more years, such as occurs in mining and
milling (Nicholson, 1976b) or insulation work (Selikoff et al , 1979) may have
from 5 to 20% of their deaths attributed to pneumoconiosis. All varieties of
asbestos appear equally capable of producing asbestosis (Irwig et al., 1979).
In groups exposed at lower concentrations, such as the families of workers,
there is less incapacitation, and death from asbestosis has not been reported.
61
-------
3.12 MANIFESTATIONS OF OTHER OCCUPATIONAL EXPOSURE TO ASBESTOS
In the past decade, considerable evidence has been developed on the
prevalence of asbestos disease in workers who were exposed to a variety of
work activities. Shipyard trades (other than insulation work), were shown to
have particularly significant exposure. By 1975, Harries (1976) had identi-
fied 55 mesothelial cases 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 3 years.
However, no exposure data are available. X-ray abnormalities among non-
insulator shipyard employees also are common. Among long-term (mostly 30+
years) shipyard workers, 86% were found to have X-ray abnormalities character-
istic of asbestos exposure (Selikoff et al. , 1981). Maintenance personnel
have also been shown to be at risk from asbestos disease. Lilis et al. (1979)
reported the finding of X-ray abnormalities among 55% of X-rays of 20+ year
chemical plant workers.
These findings are important because they point to future sources of
asbestos emission to the environment. The removal of asbestos from friable
products, including insulation material, and the installation of engineering
controls in factories have significantly reduced the exposure and emissions
from primary manufacturing or primary using sources. However, over one mil-
lion tons of asbestos is contained in friable materials in ships, buildings,
power plants, chemical plants, refineries, and other locations of high temper-
ature equipment (Nicholson, 1976a). The maintenance, repair, and removal of
this material will account for the principal exposures to workers and emis-
sions into the environment (both in and out of buildings) in the future.
3.13 DEPOSITION AND CLEARANCE
Some limited 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 analyzed
cases were selected because of death from mesothelioma, often coupled with an
investigation of a specific work group (Wagner et al. , 1982; Berry and New-
house, 1983). However, the cases have not been correlated with known cumula-
tive exposures. Generally, amphibole burdens of individuals who were heavily
7 fi
exposed range from 10 to 19 f/g dry weight; general population controls (in
62
-------
Great Britain) are usually less than 106 f/g 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 to provide a basis for establishing a model
for the deposition and clearance of fibers in humans. Both short- and long-
term clearance mechanisms are expected to 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 continu-
ous exposure) diminished by a clearance that is proportional to (by factor p1)
the number of fibers present.
~ = A - pN (3-4)
For the number of fibers present after a constant exposure of duration t,,
Equation 3-4 yields,
N = -
and at a time t? after cessation of a constant exposure of duration t,,
N = |(l-e'ptl)e"pt* (3-6)
Such a model would be applicable at times t and t?, which are long compared
to any short-term clearance mechanisms. This model is clearly very simplistic
in that it considers only one characteristic time for long-term removal pro-
cesses. Nevertheless, the model illustrates the difficulty of applying even
the simplest model. In order to systematize lung burdens, information on the
duration and intensity of the exposure and the time from last exposure is re-
quired 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.
63
-------
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 workers were exposed between 10 and 16 years. While data on the indi-
vidual exposure times were unknown, fewer fibers were found in the lung wash-
ings of those workers who were removed from exposure for the longest period.
The data are consistent with a decrease of 50% in the number of washable
fibers at 5 to 7 years after cessation of exposure. However, washable fibers
may not be proportional to the residual lung burden or to the number of fibers
trapped within lung tissue. The lung washings were largely amphibole; no
corresponding data are available for chrysotile fibers.
Data on the fiber dimensionality 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.
3.13.1 Models of Deposition and Clearance
The Task Group on Lung Dynamics of the International Commission on Radio-
logical Protection has proposed a model for the deposition and retention of
particles (See Brain and Valberg, 1974). The results of this model are shown
in Figure 3-11, which depicts the percentages of particles of different sizes
deposited in the various compartments of the respiratory tract. Alveolar
deposition is dominant for particles with a mass median diameter of less than
0.1 pm. As the particle size increases, deposition in this area decreases,
falling to 25% at 1 pm and to 0 at 10 urn or above. Nasal and pharyngeal sur-
face deposition becomes important above 1 urn and rises rapidly to be the domi-
nant deposition site for particles 10 pm in diameter or greater. The above
model was developed for spherical particles. Timbrel! (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
finding has been corroborated by calculations of Harris and Fraser (1976).
Thus, few fibers with diameters as large as 2 pm are likely to deposit in the
alveolar spaces, although finer fibers, even as long as 200 (jm, may do so.
64
-------
o
LLJ
55
o
a.
iu
o
BRONCHO ALVEOLAR
NASO PHARYNGEAL
TRACHEO-
BRONCHIAL
10 —
0.01
0.05 0.1 0.5 1.0 5 10
MASS MEDIAN DIAMETER, M
Figure 3-11. 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).
-------
3.14 EFFECTS OF INTERMITTENT EXPOSURE VERSUS CONTINUOUS EXPOSURE
Two distinct kinds of exposure occurred to workers in the studies review-
ed above. On the one hand, workers in some production operations (e.g.,
textiles) would be exposed to a relatively constant concentration of asbestos
fiber throughout their work day. On the other hand, insulators, maintenance
mechanics, and some production workers were exposed to extremely variable
concentrations of asbestos, with most of their cumulative exposure being the
result of intense exposures of short duration. 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 re-
ceived 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
for only short time periods. Because of the time dependence of mesothelioma
risk, 100 f/ml for 1 year is not equivalent to 2 f/ml for 50 years.) Short,
intense exposures could have an effect different from longer, lower exposure
if clearance mechanisms are altered by very high concentrations of inspired
asbestos. There are no data that directly address this question. However,
indirect information suggests that the magnitude of the effect is less than
the variability between studies with continuous exposure. First, Henderson
and Enterline (1979) found that the excess lung cancer risk for plant wide
maintenance mechanics was only slightly higher (21%) than that for production
workers, on a unit exposure basis. The risk of pneumoconiosis was much less
per unit cumulative exposure among maintenance workers. Second, the simila-
rity of unit exposure risks of insulators compared to that for groups with
continuous exposure would suggest that the character of their exposure is not
important. However, both comparisons depend upon the exposure estimates of
these groups. Clearly, average exposures are difficult to estimate from epi-
sodic exposures and the above numerical similarities may be fortuitous. The
unusually low pneumoconiosis risk among the mechanics in the Henderson and
Enterline study may be the results of exposure misestimates.
3.15 RELATIVE CARCINOGENICITY OF DIFFERENT ASBESTOS VARIETIES
Information on the effect of specific asbestos varieties in different
exposure circumstances is limited. 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). Both
66
-------
Great Britain and Sweden have imposed far more rigid standards for crocidolite
than for other varieties of asbestos. In contrast, the United States has no
special standard for any specific asbestos mineral.
A special role has been attributed to crocidolite by some investigators,
perhaps because the first environmental mesotheliomas were found in an area
where crocidolite exposure was likely (Wagner et al., 1960). Subsequently, in
Great Britain, where crocidolite was often used, many individuals who devel-
oped mesotheliomas were found to have had opportunities, for exposure to this
fiber, although such association was not unique. In fact, equal opportunity
for exposure to chrysotile was demonstrated (Greenberg and Lloyd-Davies,
1974). While crocidolite is a factor in an increased risk of death from
mesothelioma in some circumstances, in others this cannot be demonstrated.
Considerable data indicate that significant risks of mesothelioma exist in
particular circumstances from exposure to other asbestos varieties.
Enterline and Henderson (1973) and Weil! et al. (1979) suggested that
workers who were exposed to chrysotile and crocidolite may have had a greater
lung cancer risk than those exposed to only chrysotile. These suggestions
were based on air concentrations of total particles in the respective work
environments, and they included much other dust, such as cement. A signifi-
cantly added crocidolite exposure could have been present in the combined
exposure work circumstances without significantly affecting the total particle
count.
The manufacture of amosite insulation has been shown to be associated
with a high risk of mesothelioma (Table 3-12), while amosite mining has demon-
strated little excess risk of death from mesothelioma (Webster, 1970). Simi-
larly, data on chrysotile use is ambiguous. Exposures in the British factory
studied by Peto (1980), which predominently used chrysotile, carried a high
risk of mesothelioma, but recently questions were raised over the use of some
crocidolite in the facility. No data are available on the relative amounts
used of each fiber. Over 4% (4.3%) of the deaths were caused by mesothelioma
in a long-term follow-up of a facility that used 5000 to 6000 tons of chry-
sotile, 50 tons of amosite, and 4 tons of crocidolite annually (except for 3
years when 375 tons of amosite were used) (Robinson et al., 1979). In con-
trast, only one mesothelioma occurred in 175 deaths in the factory studied by
Dement et al. (1982).
67
-------
Much of these differences in risk may be accounted for by the differences
in fiber size distributions in the three work environments rather than fiber
type. The greatest percentage of longer and thicker fibers would occur in the
work environment of miners and millers. When asbestos is used in manufac-
turing processes, it is broken apart as it is incorporated in finished pro-
ducts. During application or removal of insulation products, asbestos is
further manipulated and the fibers become reduced in length and diameter. As
these smaller fibers can readily be carried to the periphery of the lung,
penetrate the visceral pleura, and lodge in the visceral or parietal pleura,
they may be important to the etiology of mesothelioma. Bignon, Sebastien, and
their colleagues (1978) have reported data from a study of lungs and pleura of
shipyard workers. Larger fibers, often amphibole, were usually found in lung
tissue. In the pleura, the fibers were generally chrysotile, but finer and
smaller. The early association of mesothelioma with crocidolite occurred
because, even in mining, crocidolite is readily broken apart and its extensive
use in Great Britain in extremely dusty circumstances (e.g., spray insulation)
created high exposures for many individuals with a concomitant high risk of
death from mesothelioma. On the other hand, the mining and milling of chryso-
tile involved exposure to long and curly fibers, which are easily counted but
not easily inspired.
In Turkey, recent exposure to the fibrous zeolite mineral erionite has
been associated with mesothelioma. Results reported by Baris et al. (1979)
demonstrate an extraordinary risk. Annual incidence rates for mesothelioma of
nearly 1% exist. In 1974, 11 of 18 deaths in Karain, Turkey were from meso-
thel ioma. Seventy-five percent of the fiber diameters are reported to be less
than 0.25 urn. The lengths were highly variable, but most fibers were shorter
than 5 urn. Asbestos minerals in identified geological deposits are not re-
ported to occur in the area.
3.16 SUMMARY
Data are available that allow a unit risk to be made for lung cancer and
allow such risks to be made for mesothel ioma. The values for K, , the frac-
tional risk per f-yr/ml, vary widely among the studies, largely because of
the statistical variability associated with smaller numbers, but also because
of uncertainties associated with methodology and exposure estimates. Never-
theless, even with this variability, a ten-fold range of KL from 0.003 to 0.03
overlaps the ranges of K, observed in all studies but one.
68
-------
Data on K the potency coefficient for mesothelioma risk, suggests a range
between 3 x lo"9 and 3 x 10"8. However, the data available to establish KM
are much more limited than that for K, Differences in asbestos type cannot
explain the variation seen in K, and KM in different studies. However, lower
risk values found in chrysotile mining suggest that fiber dimensionality may
be important.
Thus in summary, calculations of unit risk values for asbestos must be
viewed with caution as they are uncertain and aspects of them are necessarily
based on estimates that are subjective to some extent because of the following
limitations in the data: 1) statistical uncertainties and systematic biases
in epidemiological studies, 2) conversions of particle counts to fiber ex-
posures are uncertain, and 3) very importantly, the nonrepresentative nature
of the exposure estimates.
69
-------
4. ANIMAL STUDIES
4.1 INTRODUCTION
Most animal studies of asbestos health effects have been used to confirm
previously established human data rather than to predict human disease. This
situation has occurred in part because asbestos usage predated the use of
animal studies for ascertainment of risk; in part because some animal models
used were relatively resistant to the human diseases of concern; and finally
because the principal carcinogenic risk from asbestos, lung cancer, is the
result of a multifactorial interaction between other agents, principally
cigarette smoking, and asbestos exposure and is difficult to elicit in a
single exposure circumstance. All of the asbestos-related malignancies were
first identified in humans. Nevertheless, the experimental studies have
confirmed the identification of disease and provided important information not
available from human studies on the deposition, clearance, and retention of
fibers, as well as on cellular changes at short times after exposure. Unfor-
tunately, 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 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 in the human studies concerning
the relative carcinogenicity of different asbestos varieties still remains.
4.2 FIBER DEPOSITION AND CLEARANCE
The deposition and clearance of fibers from the respiratory tract of rats
has been studied directly by Morgan and his colleagues (Morgan et al. , 1975;
Evans et al., 1973) using radioactive asbestos samples. Following 30-minute
inhalation exposures in a nose breathing apparatus, the deposition and clear-
ance from the respiratory tract were followed. At the conclusion of the
inhalation, the distribution of fibers in various organ systems was deter-
mined. Thirty-one to 68% of the inspired fibrous material was deposited in
the respiratory tract. The distributions of that deposited material are shown
in Table 4-1. Rapid clearance, primarily from the upper respiratory tract
(airways above the trachea), occurred within 30 minutes; up to two-thirds of
the fibers were swallowed and found in the GI tract.
70
-------
TABLE 4-1. DISTRIBUTION OF FIBER AT THE TERMINATION OF 30-MINUTE EXPOSURES
(PERCENT OF TOTAL DEPOSITED)
Fiber
Chrysotile A
Chrysotile B
Amosite
Croc idol ite
Anthophyllite
Fluoramphibole
Nasal
passages Esophagus
9
8
6
8
7
3
± 3
± 2
± 1
± 3
± 2
± 2
2 ±
2 ±
2 ±
2 ±
2 ±
1 +
1
1
1
1
1
1
Gastro-
intestinal
tract
51 ±
54 ±
57 ±
51 ±
61 ±
67 ±
9
5
4
9
8
5
Lower
respiratory
tract
38
36
35
39
30
29
+
+
+
+
+
+
8
4
5
5
8
4
Percent,
deposited
31 ±
43 ±
42 ±
41 ±
64 +
68 ±
6
14
14
11
24
17
Mean and standard deviation
Percent of total inspired.
Source: Morgan et al., 1975
Clearance from the lower respiratory tract (trachea to alveoli) proceeds
slower; two distinct components were observed. The first component, believed
to be caused by macrophage movement, leads to the elimination of a consider-
able portion of the material deposited in the lower respiratory tract with a
half life of 6 to 10 hours. The slower phase that follows has a half life of
60 to 80 days and involves the clearance from alveolar spaces. Data for a
synthetic f 1 uoramphibole (Figure 4-1) show one short and two long-term compo-
nents for the clearance of fibers. Other data on the lung content of animals
sacrificed at various times after exposure show only a single long-term clear-
ance 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
deposited in the bronchiolar-alveolar spaces. These data are shown for dif-
ferent fibers in Figure 4-2, along with estimates of the percentage of mate-
rial deposited in the upper respiratory tract. The relative similarity of the
71
-------
100
03
o
03
T3
03
•*-*
U
o
U
DC
o
2
I-
2
D
O
O
LL
O
I-
2
LLJ
O
CC
10
= 86 5e-° 693t/0
3e
-°
2e
-° 693t/118
20 40 60 80 100
TIME AFTER ADMINISTRATION, days
120
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 js a percentage of the counting
rate measured immediately after exposure.
Source: Morgan et al. (1977).
72
-------
O)
1 20
•o
a>
a
§ 15
c
0)
o
O
a.
LU
O
ce
Key
Glass fibre # 108
UICC Anthophyllite
Fluoramphibole
UICC Chrysotile A
UICC Chrysotile B
UICC Amosite
UICC Crocidolite
Kuruman Crocidolite
Malipsdrift Crocidolite
Cerium Oxide
•
A.
A
•
X
0
1 2 3
ACTIVITY MEDIAN AERODYNAMIC DIAMETER, Mm
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).
73
-------
percentage deposited in the lower bronchioles or alveoli for different fiber
diameters is a reflection of two competing processes. At lower fiber diame-
ters, fibers can be inspired and then expired without impaction in the lower
respiratory tract. As the fiber diameter increases, impaction in the upper
respiratory tract becomes important; this leads to a lower percentage being
carried to the alveolar spaces.
Morgan et al. (1978) have 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
after 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
was relatively uniform. However, at later times, there was a movement of
fibers to the periphery of the lung where they accumulated in subpleural foci
consisting of alveoli filled with fiber-contained cells.
Other data on the deposition and retention of inhaled asbestos have been
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, Evans, and
Holmes (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% of the
total crocidolite or amosite inhaled is 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, excess cancer of the stomach, colon, and
rectum, and possibly cancers at other non-respiratory sites, such as the
kidney, could result from the migration of such fibers to and across the
74
-------
AFTER REMOVAL
FROM EXPOSURE
TIME (MONTHS)
10000
20000
30000
CUMULATIVE DOSE, mg/m'/hr.
Figure 4-3. Mean weight of dust in lungs of rats in
relation to dose and time.
Source: Wagner et al. (1974).
-------
gastrointestinal mucosa. Additionally, fibers may reach organs in 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) described early (14-day) local inflammatory
lesions found in the terminal bronchioles of rats following inhalation of
asbestos fibers. These lesions consisted of multinucleated giant cells,
lymphocytes, and fibroblasts. Progressive fibrosis followed within a few
weeks of the first exposure to dust. Davis et al. (1978) described similar
early lesions that were found in rats and consisted of a proliferation of
macrophages and cell debris in the terminal bronchioles and alveolae.
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 esophagus, stomach, or cecum
tissue, but structural changes in the ileum were seen, particularly of the
villi. Considerable cellular debris was detected by light microscopy in the
ileum, colon, and rectum tissue. The electron microscopic data confirmed the
light microscopy data and indicated that the observed changes were consistent
with a mineral-induced cytotoxicity.
A single oral administration of from 5 to 100 mg/kg of chrysotile to rats
has produced a subsequent increase in thymidine in the stomach, duodenum, and
jejunum (Amacher et al. , 1975). This result suggests that an immediate re-
sponse of cellular proliferation and DNA synthesis may be stimulated by chryso-
tile ingestion.
4.4 MUTAGENICITY
Asbestos has not been shown to be mutagenic in Escherichia coli or
Salmonella typhimurium tester strains (Chamberlain and Tarmy, 1977). Newman
et al. (1980) reported that asbestos has no mutagenic ability in Syrian ham-
ster embryo cells, but may increase cell permeability and allow other mutagens
into the cell. However, Sincock (1977) used several chrysotile, amosite, and
crocidolite samples to show that an increased frequency of polyploids and
cells with fragments resulted from passive inclusion of asbestos in the cul-
ture media of Chinese hamster ovary (CHO)-Kl cells. Similarly, Lavappa et al.
(1975) showed that chrysotile induced a significant and exposure-related
76
-------
increase in chromosome aberrations in cultured Syrian hamster embryo cells.
Amosite, chrysotile, and crocidolite have been found to be weakly mutagenic in
Chinese hamster lung cells in the 6-thioguanine-resistance assay (Huang,
1979). Finally, Livingston et al. (1980) have shown that exposure to croci-
dolite and amosite can increase the sister chromatid exchange rate in Chinese
hamster ovarian fibroblasts.
4.5 INHALATION STUDIES
The first unequivocal data that showed a relationship between asbestos
inhalation and lung malignancy in laboratory animals was that of Gross et al.
(1967), who observed carcinomas in rats exposed to a mean concentration of 86
mg/m chrysotile for 30 hr/wk 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 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 in 31 rats sacri-
ficed after 2 years following exposure to about 48 mg/m of crocidolite. No
malignant tumors were reported in rabbits, guinea pigs, or hamsters or in
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% 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 crocidolite inhalation. Again, significant experi-
mental details were lacking; information on survival times and times of sacri-
fice would have been useful. Available details of the exposures and results
are given in Table 4-3. While the relative carcinogen!city of the fiber types
was similar, the fibrogenic potential of chrysotile, which had been substanti-
ally reduced in length and possibly altered (Langer et al., 1978) by milling,
was much less than that of the amphiboles. These results were also discussed
in a later paper by Reeves (1976).
The most important series of animal inhalation studies is that of Wagner
et al. (1974, 19775). Wagner exposed 849 Wistar SPF rats to the five UICC
(Union Intranationale Contra le Cancer) asbestos samples at concentrations
3
from 10.1 to 14.7 mg/m for times ranging from 1 day to 24 months. These
77
-------
TABLE 4-2. SUMMARY OF EXPERIMENTS ON THE EFFECTS OF INHALATION OF ASBESTOS
Study Animal species Material administered
Gross et al. (1967) 132 male white rats Ball- and hammer-milled
Canadian chrysotile
with/without 0.05 ml
intratracheal 5 per-
cent NaOH
Dosage
42-146 mg/ml
(mean concentra-
tration, 86 mg/
m3) for 30 hrs/
week
Animals Examined Findings
for tumors (malignant tumors)
72 17 adenocarcinomas
4 squamous-cell sarcomas
7 f ibrosarcomas
1 mesothel ioma
verage survival
time
not available
Reeves et al. (1971)
55 male white rats
206 rats
106 rabbits
139 guinea pigs
214 hamsters
Controls with/without
5 percent NaOH
control
Ball-milled chrysotile, 48±2 mg/m3 for
amosite, and crocidolite 16 hours/week up
to 2 years
Reeves et al. (1974) 219 rats
216 gerbils
100 mice
72 rabbits
108 guinea pigs
Wagner et al. (1974) 13 groups of approxi-
mately 50 and 15 of
about 25 Wistar SPF
rats
Wagner et al. (1977a) CO Wistar male and
female rats
Ball- and hammer-
mil led chrysotile,
amosite and crocidolite
Amosite, anthophyl1ite,
crocidolite, Canadian
chrysotile, Rhodesian
chrysotile (UICC sam-
ples)
Superfine chrysotile
48±2 mg/m3 for
16 hours/week
up to 2 years
10.1 to 14.7
mg/m3 for 1 day
to 24 months,
35 hours/week
10.8 mg/m3 37.5
hours/week for
3, 6, or 12 months
CO Wistar male and
female rats
Nonfibrous cosmetic talc
Davis et al. (1978)
46 groups of approxi- UICC samples of amosite, 2 mg/m3 and
mately Han SPF rats chrysotile, and 10 mg/m3 35
and 20 Han SFP rats crocidolite hours/week
for 224 days
39
not available
120 rats
116 gerbils
10 mice
30 rabbits
43 guinea pigs
849
208
2 squamous-cell carcino-
mas in 31 animals from
crocidolite exposure
not available
no information
periodic sacri-
fices were made
10 malignant tumors in no information
rats, 2 in mice (Table 4-3) periodic sacri-
fices were made
(See Tables 4-4 and 4-5) 669 to 857 days
20 Han SPF rats
control
control
20
All asbestos varieties
produced mesothelioma and
lung cancer, some from ex-
posure as short as 1 day
1 adenocarcinoma of the
lung in 24 animals ex-
posed for 12 months
none
7 adenocarcinomas
3 squamous-cell
sarcomas, 1 pleural
mesothelioma, 1
peritoneal mesothelioma
none
versus 754 to 803
for controls.
Survival times
not significantly
affected by expo-
sure.
not available
sacrificed at 29
months
-------
TABLE 4-3. EXPERIMENTAL INHALATION CARCINOGENESIS
Exposure3
Mass Fiber,
Fiber mg/m3 f/ml
Chrysotile 47.9 54
Amosite 48.6 864
Crocidolite 50.2 1,105
Control s
Animals
exami ned
43
46
46
5
Rats
Mai ignant tumors
1 lung papillary carcinoma
1 lung squamous-cell carcinoma
1 pleural mesothel ioma
2 pleural mesothel iomas
3 squamous-cell carcinomas
1 adenocarcinoma
1 papillary carcinoma - all of
the lung
None
Animals
examined
19
17
18
6
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.
morphology (3:1 aspect ratio) by light microscopy.
Source: Reeves et al. (1974).
of the airborne mass was of fibrous
-------
concentrations are typically 10 times those measured in dusty asbestos work-
places during earlier decades. For all the exposure times, 50 adenocarci-
nomas, 40 squamous-cel1 carcinomas, and 11 mesotheliomas were produced. All
varieties of asbestos produced mesothelioma and lung malignancies, in some
cases from exposures as short as 1 day. Data from these experiments are
presented in Tables 4-4 and 4-5. These tumors follow a reasonably good linear
relationship for exposure times of 3 months or greater. However, the inci-
dence 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 was also documented. There were 17
lung tumors, 6 in rats with no evidence of asbestosis and 11 in rats with
minimal or slight asbestosis. Cancers at extrapulmonary sites were listed.
Seven malignancies of ovaries and eight malignancies of male genitourinary
organs were observed in the exposed groups of approximately 350 male and
female rats. No malignancies were observed in control groups of 60 males and
females. The incidence of malignancy at other sites varied little from that
of the controls. However, the authors note that if controls from other ex-
periments in which ovarian and genitourinary tumors were present are included,
the comparative incidence in the exposed groups in the first study lacks
statistical significance. However, no data were provided on the variation of
tumor incidence at extrapulmonary sites with asbestos dosage.
Wagner et al. (1977a) also compared the effects of inhalation of a super-
fine chrysotile to those of inhalation of a pure nonfibrous talc. One adeno-
3
carcinoma was found in 24 rats exposed to 10.8 mg/m of chrysotile for 37.5
hr/wk for 12 months.
In a study similar to Wagner's, Davis et al. (1978) exposed rats to 2.0
3
or 10.0 mg/m of chrysotile, crocidolite, and amosite (equivalent to 430 to
1950 f/ml). Adeno- and squamous cell carcinomas were observed in chrysotile
exposures, but not in crocidolite or amosite exposures (Table 4-6). One
pleural mesothelioma was observed with crocidolite exposure, and extrapulmo-
nary neoplasms included a peritoneal mesothelioma. A relatively large number
of peritoneal connective tissue malgnancies also were observed; these included
a leimyofibroma on the wall of the small intestine. The meaning of these
tumors is unclear.
80
-------
TABLE 4-4. NUMBER OF RATS WITH LUNG TUMORS OR MESOTHELIOMAS AFTER EXPOSURE
TO VARIOUS FORMS OF ASBESTOS THROUGH INHALATION
Form of Asbestos
Amosite
Anthophyllite
C roc i do lite
Chrysotile
(Canadian)
Chrysotile
(Rhodesian)
None
Number of
animals
146
145
141
137
144
126
Adenocarcinomas
5
8
7
11
19
0
Squamous-cel 1
carcinomas
6
8
9
6
11
0
Mesothelioma
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
Length of
exposure
None
1 day
3 months
6 months
12 months
24 months
Number
of Animals
Tested
126
219
180
90
129
95
Number of Animals
with lung
carcinomas
0
3a
8
7
35
37
Number of Animals
With pleural
mesotheliomas
0
2b
1
0
6
2
Percent
of animals
with tumors
0.0
2.3
5.0
7.8
31.8
41.0
aTwo rats exposed to Chrysotile and one to crocidolite.
One rat exposed to amosite and one to crocidolite.
Source: Wagner et al. (1974).
81
-------
TABLE 4-6. EXPERIMENTAL INHALATION CARCINOGENESIS IN RATS
Exposure
Mass Fiber,
mg/m3 f>5|j/ml
Chrysotile 10 1,950
Chrysotile 2 390
Amosite 10 550
Crocidolite 10 860
Crocidolite 5 430
Control
Number of
animals
examined Malignant tumors
40
42
43
40
43
20
6 adenocarcinomas
2 squamous-cell carci
1 squamous-cell carci
1 peritoneal mesothel
None
None
nomas
noma
ioma
1 pleural mesothel ioma
None
Source: Davis et al. (1978)
Inhalation exposures result in concomitant GI exposures from the asbestos
that is swallowed after clearance from the bronchial tree. While all inhala-
tion 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-
cluded a search for tumors at extrathoracic sites. A limited number of these
tumors were found, but no association can 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
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 would result in
mesothelioma was forthcoming in 1970 when Donna (1970) produced mesotheliomas
82
-------
in Sprague-Dawley rats treated with a single dose of 67 mg of chrysotile,
amosite, or crocidolite. Reeves et al (1971) produced mesothelial tumors in
rats (1 of 3 with crocidolite 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% of animals into which material had
been surgically implanted onto the pleural surface. The authors concluded
that the carcinogenicity of asbestos and other fibers is strongly related to
their physical size; fibers that have a diameter of less than 3 pm would
be carcinogenic and those that have a larger diameter would not be carcino-
genic. Further, samples treated by grinding in a ball mill to produce shorter
length fibers were less likely to produce tumors. While the authors attri-
buted the reduced carcinogenicity to a shorter fiber length, the question has
been raised as to the effect of the destruction of crystal 1inity and perhaps
other changes in the fibers occasioned by the extensive ball milling (Langer
et al., 1978).
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 urn
in length, although a good correlation oi carcinogenicity is also obtained for
fibers < 1.5 urn in diameter and >_ 4 urn in length. The logit distribution of
tumor incidence against the log of the number of particles < 0.25 urn x > 8 urn
is shown in Figure 4-4. The regression equation for the dotted line is:
ln[p/(l-p)] = -.".62 + 0.93 log x, (4-1)
where p is the tumor probability and x the number of particles < 0.25 urn x > 8 urn,
A reasonable relationship exists for the equation results and available data, but
substantial discrepancies occur, suggesting the possibility that other relation-
ships may better fit the data. Bertrand and Pezerat (1980) have suggested that
carcinogenicity may correlate as well with the ratio of length to width (aspect
ratio).
83
-------
TABLE 4-7. SUMMARY OF 72 EXPERIMENTS WITH DIFFERENT FIBROUS MATERIALS
cc
Experiment
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Compound
Titanate 1
Titanate 2
Si Icarbide
Dawson 5
Tremolite 1
Tremolite 2
Dawson 1
Croc id 1
Crocid 2
Crocid 3
Amosite
Crocid 4
Glass 1
Crocid 5
Glass 2
Glass 3
Glass 4
Alumin 1
Glass 5
Dawson 7
Dawson 4
Dawson 3
Glass 6
Crocid 6
Crocid 7
Crocid 8
Alumin 2
Alumin 3
Crocid 9
Wollaston 1
Alumin 4
Crocid 10
Alumin 5
Glass 20
Glass 7
Wollaston 3
Actual
tumor
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
tumor
probabi 1 ity
± SO
95±4.7
100
100
100
100
100
95±4.8
94+6.0
93±6.5
93+6.9
93+7.1
86±9.0
85±13.2
78+10.8
77+16.6
74±8.5
71±9.1
70±10.2
69±9.6
68±9.8
66±12.2
66113.4
64117.7
63113.9
56111.7
53112.9
44111.7
41110.5
33+9.8
31112.5
28+12.0
37113.5
22+9.8
22+10.0
21+8.7
19+10.5
Common log
f ibers/ug
<0.25 |jm x
>8 urn
4.94
4.70
5.15
4.94
3.14
2.84
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
Halloy 1
Halloy 2
Glass 8
Crocid 11
Glass 19
Glass 9
Alumin 6
Dawson 6
Dawson 2
Wol laston 2
Crocid 12
Attapul 2
Glass 10
Glass 11
Titanate 3
Attapul 1
Talc 1
Glass 12
Glass 13
Glass 14
Glass 15
Alumin 7
Glass 16
Talc 3
Talc 2
Talc 4
Alumin 8
Glass 21
Glass 22
Glass 17
Glass 18
Crocid 13
Wollaston 4
Talc 5
Talc 6
Talc 7
Actual
tumor
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
1/28
2/47
1/45
0/28
0/115
0/29
0/24
0/30
0/30
0/29
Percent
tumor
probabi 1 ity
+ SD
20+9.0
23+9.3
19+10.3
1918.5
15±9.0
14±9.4
13+8.8
13±6.9
12±7.9
12±8.0
10±7.0
11+7.5
8+5.6
8+5.5
8+8.0
8+5.3
7+6.9
7+5.4
615.7
615.5
615.9
5+5.1
5+4.4
414.3
4+3.8
5+4.9
3+3.4
614.4
2+2.3
0
0
0
0
0
0
0
Common log
f ibers/ug
<0.25 UN x
>8 urn
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
SD = Standard deviation.
Source: Stanton et al. (1981)
-------
cc
o
I
D
1-
LL
0
>
u
ffi
ffi
0
cc
1 .VJ
0.9
0.8
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
ADDW
.AGGPT .
LTGGGG
LTTGG
CWTTGG
! I I
= crocidolite
= glass
= dawsonite
= aluminum oxide
= silicon carbide
= attapuigite
= titanate
= talc
= tremolite
= wollastonite
= halloysite
= amosite
4
/
S
S
s'
^
. ^ G
-" L
G
I I I
I M I M I I I P D S I
D P' C ^ — — '
^^»
^^
S —
C - fi
G S
G L, G —
^' D
/ DG D
s C
s —
r
c s'
/
/ L
I —
^ C
_
L
G L __
^
C —
I I T | | | | |
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
LOG NUMBER PARTICLES MEASURING < 0.25 nmx> ^m PER MICROGRAM
Figure 4.4 Regression curve relating probability of tumor to logarithm of
number of particles per ^g with diameter < 0.25 /jm and length >8
6.0
Source: Stanton et al. (1981).
-------
Another comprehensive set of experiments was conducted by Wagner (Wagner
et al., 1973; 1977b). Wagner also produced mesothelioma from intrapleural
administration of asbestos to CD Wistar rats and demonstrated a strong dose-
response relationship. Tables 4-8 and 4-9 list the results of these experi-
ments.
Pylev and Shabad (1973) and Shabad et al. (1974) reported mesotheliomas
in 18 of 48 rats and in 31 of 67 rats injected with three doses of 20 mg of
Russian chrysotile. Other experiments by Smith and Hubert (1974) produced
mesothel iomas in hamsters injected with 10 to 25 mg of chrysotile, 10 mg of
amosite or anthophyl1ite, and 1 to 10 mg of crocidolite.
Various suggestions have been made that natural oils and waxes contamina-
ting asbestos fibers might be related to their carcinogenicity (Harington,
1962; Harington and Roe, 1965; Commins and Gibbs, 1969). However, this theory
was not borne out in the previously mentioned experiments by Wagner et al.
(1973) or Stanton and Wrench (1972).
4.7 INTRATRACHEAL INJECTION
Intratracheal injection has been used to study the combined effect of the
administration of chrysotile with benzo(a)pyrene in rats or hamsters. In rats
given three doses of 2 mg of chrysotile (Shabad et al., 1974) or in hamsters
given 12 mg of chrysotile (Smith et al., 1970), no lung tumors were observed.
However, the coadministration of benzo(a)pyrene resulted in lung tumors, and
this suggests a cocarcinogenic or synergistic effect.
4.8 INTRAPERITONEAL ADMINISTRATION
Intraperitoneal injections of 20 mg of crocidolite or chrysotile produced
three peritoneal mesotheliomas in 13 Charles River CD rats. Twenty mg of
amosite produced no tumors in a group of 11 rats (Maltoni and Annoscia, 1974).
Malton and Annoscia also injected 25 mg of crocidolite into 50 male and 50
female 17-week-old Sprague-Dawley rats and observed 31 mesothelial tumors in
males and 34 in females.
In an extensive series of experiments, Pott and Friedrichs (1972) and
Pott et al. (1976) produced peritoneal mesothel iomas in mice and rats that
were injected with various commercial varieties of asbestos and other fibrous
material. These results are shown in Table 4-10. Using experiments with
intrapleural administration, the malignant response was altered by ball-
86
-------
TABLE 4-8. PERCENTAGE OF RATS DEVELOPING MESOTHELIOMAS AFTER INTRAPLEURAL
ADMINISTRATION OF VARIOUS MATERIALS
Material
Percent of Rats
with Mesotheliomas
SFA chrysotile (superfine Canadian sample)
UICC crocidolite
LJICC amosite
UICC anthophyllite
UICC chrysotile (Canadian)
UICC chrysotile (Rhodesian)
Fine glass fiber (code 100), median diameter,
0.12 pro
Ceramic fiber, diameter, 0.5-1 pma
Glass powder
Coarse glass fiber (code 110), median diameter,
1.8 jm
66
61
36
34
30
19
12
10
3
0
From Wagner et al. (1973).
Source: Wagner (1977b)
TABLE 4-9. DOSE-RESPONSE DATA FOLLOWING INTRAPLEURAL ADMINISTRATION
OF ASBESTOS TO RATS
Material
SFA chrysotile
Croc idol ite
Dose
mg
0.5
1
2
4
8
0.5
1
2
4
8
Number of
rats with
mesothel ioma
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)
87
-------
TABLE 4-10. TUMORS IN ABDOMEN AND/OR THORAX AFTER INTRAPERITONEAL INJECTION OF GLASS FIBERS, CROCIDOLITE, OR CORUNDUM IN RATS
Dust
Glass fibers
MN 104
Glass fibers
MN 104
Glass fibers
MN 104
Crocidol ite
Corundum
UICC Rhodesian
00 chrysotile
cc
UICC Rhodesian
chrysotile
UICC Rhodesian
chrysotile
UICC Rhodesian
chrysotile
UICC Rhodesian
chrysotile
UICC Rhodesian
milled
Palygoescite
Form
f
f
f
f
g
f
f
f
f
f
f
f
Intraperi toneal
dose
2
10
2 x 25
2
2 x 25
2
6.25
25
4 x 25
3 x 25
s. c.
4 x 25
3 x 25
Effective
number of
dissected
rats
73
77
77
39
37
37
35
31
33
33
37
34
Number of
days before
first tumor
421
210
194
452
545
431
343
276
323
449
400
257
Average
survival time
of rats with
tumors, days
after injection
703
632
367
761
799
651
501
419
361
449
509
348
Rats
with
tumors ,
percent
27.4
53.2
71.4
38.5
8.1
16.2
77.1
00.6
54.5
3.0
32.4
76.5
1
17
36
47
12
1
4
24
21
16
_
9
24
Tumor/type
23 456
3 - - 1 1
4 - 13-
62 -
3 - - 2 1
222
2 - - 1 -
3 -
21 1 - -
2 - -
1 -
s.c.
3 - -
2 -
-------
TABLE 4-10. (continued)
cc
Oust
Glass fibers
s + s 106
Glass fibers
S + S 106
Glass fibers
S + S 106
Gypsum
Henalite
Actinolite
Biotite
Haematite
(precipitation)
Haematite
(mineral)
Pectolite
Sanidine
Talc
NaCl (control)
Form
f
f
f
f
f
9
g
g
g
g
g
g
-
Intraperitoneal
dose
2
10
4 x 25
4 x 25
4 x 25
4 x 25
4 x 25
4 x 25
4 x 25
4 x 25
4 x 25
4 x 25
4 x 2 ml
Effective
number of
dissected
rats
34
36
32
35
34
39
37
34
38
40
39
36
72
Number of
days before
first tumor
692
350
197
579
249
-
-
-
-
569
579
587
-
Average
survival time
of rats with
tumors, days
after injection
692
530
325
583
315
-
-
-
-
569
579
587
-
Rats
with u
tumors, Tumor/type
percent 123 456
2.9 1 -
11.1 2 2 - 1 -
71.9 20 3 - ---
5.7 - - 1 11-
73.5 17 8 - - - -
-
-
-
-
2.5 - 111
2.6 - 1 - - - -
2.8 1 - - -
-
f = fibrous; g = granular.
Tumor Types are: 1 Mesothel ioma; 2 Spindle cell sarcoma; 3 Polym-cell sarcoma; 4 Carcinoma; 5 Reticulum
6 Benign -- not evaluated in tumor rates.
cell sarcoma;
Sources: Pott and F'riedrichs (1972); Pott et al. (1976).
-------
milling fibers for 4 hours. The rate of tumor production was reduced from 55
to 32% and the time from the onset of exposure to the first tumor was length-
ened from 323 to 400 days following administration of four doses of 25 mg of
UICC Rhodesian chrysotile. In the case of the ball-milled fibers, 99% of the
fibers were reported to be smaller than 3 urn, 93% were less than 1 urn, and 60%
were less than 0.3 urn.
Pott (1980) has proposed a model for the relative carcinogenicity of
mineral fibers according to their dimensionality using the results of injec-
tion and implantation data. Figure 4-5 shows the schematic features of this
model. The greatest carcinogenicity is attributed to fiber lengths between 5
and 40 urn with diameters between 0.05 and 1 urn.
A strong conclusion that can be drawn from the above experimental data is
that long (4 urn) and fine diameter (<1 pm) 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 reduced carcinogenicity for shorter,
ball-milled fibers is less clear because the relative contributions of shorter
fiber length and the significant alteration of the crystal structure by input
of physical energy are not yet defined. However, the extrapolation of data
developed on size-dependent effects, from intrapleural or intraperitoneal
administration to inhalation (where movement of the fibers in airways and
subsequently through body tissues is strongly size-dependent), presents signi-
ficant difficulties. Moreover, the number of shorter (<5 urn) fibers in an
exposure circumstance may be 100 times greater than the number of longer
fibers; therefore, their carcinogenicity must be 100 times less before their
contribution can be neglected.
4.9 TERATOGENICITY
There is no evidence that asbestos is teratogenic. Schneider and Maurer
(1977) fed pregnant CD-I mice doses of 4 to 400 mg/kg body weight (1.43 to
143) for days 1 to 15 of gestation. They also administered 1, 10, or 100 ug
of asbestos to day 4 blastocysts, which were transferred to pseudopregnant
mice. No positive effects were noted in either experiment.
4.10 SUMMARY
The animal data on the carcinogenicity of asbestos fibers confirm and
extend epidemiological human data. Mesothelioma and lung cancer have been
90
-------
.031\
0.031
100 -i
80-
60 -
40 -
20 -
cc
O
o
o
z
111
o
o
z
o
o
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).
-------
produced by all the principal commercial asbestos varieties, chrysotile,
amosite, crocidolite and anthophyl1ite, even by exposures as short as 1 day.
The deposition and clearance of fibers from the lung suggest that most inhaled
fibers O99%) are eventually cleared from the lung by ciliary or phagocytic
action. Chrysotile appears to be more readily removed, and dissolution of the
fibers occurs in addition to other clearance processes. Implantation and
injection studies suggest that the carcinogenicity of durable mineral fibers
is related to their dimensionality and not to their chemical composition.
Long (>4 urn) and thin (<1 urn) fibers are most carcinogenic when they are in
place at a potential tumor site. However, deposition, clearance, and migra-
tion of fibers is also size dependent, and the importance of all size-
dependent effects in the carcinogenicity of inhaled fibers is not fully estab-
lished.
92
-------
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 ug/m3 of particulate
matter in which the researcher is attempting to quantify asbestos concentra-
tions from about 0.1 ng/m to perhaps 1000 ng/m . Thus, asbestos may con-
stitute only 0.0001 to 1% of tht- particulate matter in a given air sample.
Moreover, the asbestos found in the ambient air had a size distribution in
which the vast majority of the fibers were too short or thin to be seen in an
optical microscope. In many cases, these fibers and fibrils will be agglomer-
ated with a variety of other materials present in the air samples.
The only effective method of analysis has used the electron microscope to
enumerate and size all asbestos fibers (Nicholson and Pundsack, 1973; Samudra
et al., 1978). Samples from such analysis were collected on Millipore fil-
ters, usually with a nominal pore size of 0.8 pm and in some cases, backed by
a nylon mesh. To prepare a sample for analysis, a portion of the filter was
ashed in a low temperature oxygen furnace, which removed the membrane filter
material and all organic material collected in the sample. The residue was
recovered in a liquid phase, dispersed by ultrasonification, and filtered on a
Nuclepore filter. The refiltered material was coated by carbon to entrap the
collected particles. A segment of the coated filter was then mounted on an
electron microscope grid, which was placed on a filter paper saturated with
chloroform, the vapors of which serve to dissolve the filter material Ear-
lier 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 that film was mounted on an electron microscope grid for scanning.
Chrysotile asbestos was identified on the basis of its morphology in the
electron microscope and amphiboles were identified by their selected area
electron diffraction patterns, supplemented by energy dispersive X-ray analy-
sis. Because of the dispersal of the fibers and their disruption by ultrason-
ification, no information was obtained on the size distribution of the origi-
nal aerosol. Air concentrations were recorded only in terms of the total mass
of asbestos present in a given air volume, usually in nanograms per cubic
meter. (See Section 5-9 for data on the interconvertibil ity of optical fiber
93
-------
counts and electron microscopic mass determinations.) Environmental measure-
ments can also be made by using Nuclepore filters and eliminating the ashing
and refiltration steps mentioned above. However, great care must be taken to
assure that fibers are not lost from the filter prior to processing.
An analysis of 25 samples collected in buildings with asbestos surfacing
material, some of which showed evidence of contamination, demonstrated the
inadequacy of phase contrast optical microscopic techniques for the quantifi-
cation of asbestos (Nicholson et al., 1975). Figure 5-1 shows the correlation
of optical fiber counts determined using NIOSH prescribed techniques (1972)
and asbestos mass measurements obtained on the same sample. In determining
the fiber concentrations, 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 because the
study was designed to evaluate phase contrast microscopy. 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.
5.2 GENERAL ENVIRONMENT
Asbestos of the chrysotile variety has been found to be a ubiquitous
contaminant of ambient air A study of 187 quarterly samples collected in 48
U.S. cities from 1969 to 1970 showed chrysotile asbestos to be present in
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 (EPA,
1974). Each value represents the chrysotile concentration in a composite of
from five to seven 24-hour samples and, thus, averages over possible peak
concentrations, which could occur periodically or randomly. Of the three sam-
ples greater than 20 ng/m analyzed by Mount Sinai, one sample was in a city
that had a major shipyard and another was in a city that had four brake manu-
facturing facilities. Thus, these samples may have included a contribution
from a specific source in 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 of Paris (Sebastien et al., 1980). These values were
obtained during 1974 and 1975 and were generally lower than those measured in
94
-------
'--D
30
-------
TABLE 5-1. THE 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
63.5
87.7
94.2
98.5
99.0
100.0
Microscopic Analysis
Battell e .
Memorial Institute
Number Percentage
of of
samples samples
27 21.3
60 47.2
102 80.1
124 97.6
125 98.5
127 100.0
127 100.0
Paris, France0
Percentage
of
samples
70
85
98
100
Sources: aNicholson (1971); bEPA (1974); cSebastien et al. (1980).
the United States, perhaps reflecting a diminished use of asbestos in con-
struction compared to that of the United States during 1969-1970.
In a study of the ambient air of New York City, in which samples were
taken only during daytime working hours, higher values than those mentioned
above were obtained (Nicholson et al., 1971). These 6-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 night time
periods, for example. Table 5-2 records the chrysotile content of 22 samples
collected in the five boroughs of New York and their overall cumulative dis-
tribution. The samples analyzed in all the studies discussed above were taken
during a period when fireproofing of high rise buildings by spraying asbestos-
containing materials was permitted. The practice was especially common in New
York City. While no sampling station was known to be located adjacent to an
active construction site, unusually high levels could nevertheless have resul-
ted from the procedure. Other sources that may have contributed to these air
concentrations include automobile braking, other construction activities,
consumer use of asbestos products, and maintenance or repair of asbestos-con-
taining materials (e.g., thermal insulation).
96
-------
TABLE 5-2. DISTRIBUTION OF 4- TO 8-HOUR DAYTIME CHRYSOTILE ASBESTOS
CONCENTRATIONS IN THE AMBIENT AIR OF NEW YORK CITY 1969-1970
Asbestos
(ng/m3)
concentration
less than
1
2
5
10
20
50
100
Cumulative number
of samples
0
1
4
8
16
21
22
Cumulative percentage
of samples
0.0
4.5
18.1
36.4
72.7
95.4
100.0
Distribution by borough
Asbestos air level L ncj/m3
Sampling locations
Manhattan
Brooklyn
Bronx
Queens
Staten Island
Number of samples
7
3
4
4
4
Range
8-65
6-39
2-25
3-18
5-14
Average
30
19
12
9
8
Source: Nicholson et al. (1971).
5.3 CHRYSOTILE ASBESTOS CONCENTRATIONS ABOUT 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 conduc-
ted in lower Manhattan in 1969 about sites where extensive spraying of asbes-
tos-containing fireproofing material was taking place. Eight sampling sites
were established about the World Trade Center construction site during the
period when asbestos material was sprayed on the steelwork of the first tower.
Table 5-3 shows the results of building-top air samples located at sites
within one-half mile of the Trade Center site and demonstrates that spray
fireproofing did contribute significantly to asbestos air pollution (Nicholson
et al., 1971; Nicholson and Pundsack, 1973). In some instances, chrysotile
asbestos levels approximately 100 times the concentrations typically found in
the ambient air were observed.
97
-------
TABLE 5-3. DISTRIBUTION OF 6- TO 8-HOUR CHRYSOTILE ASBESTOS
CONCENTRATIONS WITHIN ONE-HALF MIi_E OF THE SPRAYING OF ASBESTOS MATERIALS ON
BUILDING STEELWORK 1969-1970
Asbestos
(ng/m3
concentration
) less than
5
10
20
50
100
200
500
Cumulative number
of samples
0
3
8
14
16
16
17
Cumulative percentage
of samples
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/m^
Average
60
25
18
Source: Nicholson et al (1971).
5.4 ASBESTOS CONCENTRATIONS IN tJUILDINGS IN THE UNITED STATES AND FRANCE
During 1974, 116 samples of indoor and outdoor air were collected in 19
buildings in five U.S. cities to assess whether contamination of the building
air resulted from the presence of asbestos-containing surfacing material in
rooms or return air plenums (Nicholson et al., 1975). The asbestos material
in the buildings was of two main types: 1) a cementitious or pi aster-like
material that had been sprayed as a slurry onto steelwork or building surfaces
and 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 the air sampling in these build-
ings (Table 5-4) provide evidence that the air of buildings with fibrous
asbestos-containing materials may often be contaminated.
98
-------
TABLE 5-4. THE CUMULATIVE DISTRIBUTION OF 8- TO 16-HOUR CHRYSOTILE ASBESTOS
CONCENTRATIONS IN BUILDING WITH ASBESTOS-CONTAINING SURFACING MATERIAL
IN ROOMS OF AIR PLENUMS
Asbestos
concentration
ng/m3 less than
1
2
5
10
20
50
100
200
500
1000
Arithmetic average
concentration
Friable
Number of
samp! es
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
Cementiti
Number of
samples
3
6
10
17
26
27
27
28
ous spray
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
samples
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).
-------
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 of which had at least one measurement higher
than 7 ng/m , the upper limit of the outdoor asbestos concentrations measured
by these workers. The distribution of the 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 of 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, in two
highly contaminated areas, significant reductions were measured (500 to 750
3 3
ng/m decreased to less than 1 ng/m ). The importance of proper removal
techniques and cleanup cannot be overemphasized.
Additionally, Sebastien et al. (1982), measured concentrations of indoor
3
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
A recent concern was the discovery of extensive asbestos use in public
school buildings (Nicholson, 1978b). Asbestos surfaces were found in more
than 10% of pupil use areas in schools of New Jersey, with two-thirds of these
surfaces having some evidence of damage. Because these values appear to be
typical of conditions in many other states, it has been estimated that from 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 contami-
nation for this use of asbestos, 10 schools were sampled in the urban centers
of New York and New Jersey and suburban areas of Massachusetts and New Jersey.
Schools were selected for sampling because of visible damage, in some cases
extensive, and thus are not typical of all schools.
Table 5-6 lists the distribution of chrysotile concentrations found in
samples taken over 4 to 8 hours in these 10 schools. Chrysotile asbestos
concentrations ranged from 9 ng/m to 1950 ng/m3, with an average of 217
3 3
ng/m . Outside air samples at three of the schools varied from 3 ng/m , with
3 3
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 this area. The highest value (1950 ng/m ) was in a
100
-------
TABLE 5-5. THE CUMULATIVE DISTRIBUTION OF 5-DAY ASBESTOS CONCENTRATIONS
IN PARIS BUILDINGS WITH ASBESTOS-CONTAINING SURFACING MATERIALS
Asbestos concentration
(ng/m3) less than
Building samples
Number Percentage
Outdoor control samples
Number Percentage
Chrysotile
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
95.6
96.3
97.8
100.0
25 ng/m3
14
16
17
19
73.7
84.2
89.5
100.0
1 ng/m^
Amphiboles"
1
2
5
10
20
50
100
200
500
Arithmetic average
concentration
112
115
122
125
129
131
132
133
135
83.0
2
,4
85.
90.
92.6
95.6
97.0
97.8
98.5
100.0
10 ng/m3
19
100.0
0.1 ng/m;
aNo value reported for 104 building samples. Some materials would have con-
tained no amphibole asbestos.
Source: Sebastien et al. (1980).
101
-------
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, 1978b
sample that followed routine sweeping of a hallway in a school with water
damage to the asbestos surface. However, no visible asbestos was seen on the
hallway floor. Because the schools were selected on the basis of visible
damage, these results cannot be considered typical of all schools with asbes-
tos surfaces. However, the results illustrate the extensive contamination
that can occur.
A recent study suggests that the above New Jersey samples in schools may
not be atypical (Constant, Jr et al., 1983). Concentrations identical to
those indicated above were found in the analysis of samples collected during a
5-day period in 25 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. An arithmetic mean concentra-
tion of 237 ng/m was measured in 54 samples collected in rooms or areas that
had asbestos surfacing material In contrast, a concentration of 8 ng/m was
measured in 31 samples of outdoor air taken at the same time. Of special
concern are 31 samples that were collected in the schools that used asbestos,
but in areas where asbestos was not used. These data showed an average con-
3
centration of 54 ng/m , indicating tl
The data are summarized in Table 5-7.
3
centration of 54 ng/m , indicating the dispersal of asbestos from the source.
Finally, Sawyer (1977; 1979) has reviewed a variety of data on air con-
centrations, measured by optical microscopy, that have been observed in
102
-------
TABLE 5-7. CUMULATIVE DISTRIBUTION OF 5-DAY CHRYSOTILE ASBESTOS CONCENTRATIONS IN
25 SCHOOLS WITH ASBESTOS SURFACING MATERIALS, 1980-1981
o
OJ
Asbestos
Rooms wi
concentration Number of
ng/m3 less than
1
2
5
10
20
50
100
200
500
1000
Arithmetic average
concentration
1
2
10
20
50
100
200
500
Arithmetic average
concentration
samples
4
6
7
10
16
25
33
43
48
54
44
45
48
50
52
52
53
54
th asbestos
Percentage
of samples
7.4
11.1
13.0
18.5
29.6
46.3
61.1
79.6
88.9
100.0
231 ng/m3
81.5
83.3
88.9
92.6
96.3
96.3
98.1
100.0
6. 1 ng/m3
Rooms without asbestos
Number of
samples
Chrysoti
6
7
10
12
13
17
27
29
31
Amp hi bo 1
21
22
26
27
27
29
31
Percentage
of samples
le
19.3
22.6
38.7
41.9
54.8
87.1
93.5
96.8
100.0
54 ng/m3
es
67 7
71.0
83.9
87.1
87 1
93.5
100.0
8.7 ng/m3
Outdoor
Number of
sampl es
18
21
26
28
29
30
31
26 '
29
30
30
31
control s
Percentage
of samples
58.1
67 7
83.9
90.3
93.5
96.8
100.0
8 ng/m3
83.9
93.5
96.8
96.8
100.0
0.7 ng/m3
Source: Constant, Jr et al. (1983).
-------
circumstances where asbestos materials in schools and other buildings are
disturbed by routine or abnormal activity. These results are shown in Table
5-8, demonstrate that a wide variety of activities can lead to high asbestos
concentrations during disturbance of asbestos surfacing material. Maintenance
and renovation work, particularly if performed improperly, can lead to sub-
stantially elevated asbestos levels.
TABLE 5-8. AIRBORNE ASBESTOS IN BUILDINGS
Friable asbestos material
Classification
Quiet, non-
specific,
routine
Maintenance
Custodial
Renovation
Vandal ism
Main mode of
contamination
Fallout
Reentrainment
Contact
Mixed: contact
reentrainment
Mixed: contact
reentrainment
Contact
Mean
count of
Activity fibers per
description cm3 n
None
Dormitory
University, schools
offices
Re lamping
pi umbing
cable movement
Cleani ng
dry sweeping
dry dusting
by stander
heavy dusting
Ceiling repair
track light
hanging light
partition
pipe lagging
Ceiling damage
0.
0.
0.
0.
1.
1.
0.
15.
1.
4.
0.
2.
17.
7.
I.
3.
4.
12.
0
1
1
2
4
2
9
5
6
0
3
8
7
7
1
1
1
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.
0.
6.
0.
1.
0.
1.
8.
2.
0.
1.
1.
8.
0
o-o.
1
1-0.
1
1-2.
2-3.
7
7
3
3
6
2
9
8
1
8-5.
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 ly-exposed to the fiber directs attention to air concentrations in the
homes of such workers. Thirteen samples have been collected in the homes of
asbestos mine and mill employees and analyzed for chryostile (Nicholson et
al. , 1980). The workers were employed at mine operations in California and
Newfoundland and at the time of sampling (1973 and 1976), they did not have
104
-------
access to shower facilities nor did they commonly change clothes before going
home. Table 5-9 lists the concentrations range of the home samples. Three
samples taken in homes of non-miners in Newfoundland yielded concentrations of
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 as asbestos cancers have been documented in
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).
family contacts of workers, concentrations such as those described in this
document should be viewed with particular concern.
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/m3 (two U.S. samples that may have been affected by specific sources were
not included). Short-term daytime samples are generally higher; this reflects
the possible contributions of traffic, construction, and other human activi-
ties. Of buildings with asbestos-surfacing materials, average concentrations
100 times those of the ambient air are seen in some schools. Concentrations
of 5 to 30 times background are seen in some other building circumstances.
5.8 OTHER EMISSION SOURCES
The weathering of asbestos cement wall and roofing materials has been
shown to be a source of asbestos air pollution in the analysis of air samples
105
-------
TABLE 5-10. SUMMARY Oir ENVIRONMENTAL ASBESTOS SAMPLING
Sample set
Col lection
period
Number
of samples
Mean
concentration, ng/m"
Quarterly composites of 5 to 7 1969-70
24-hour U.S. samples (Nicholson,
1971; Nicholson and Pundsack, 1973)
5 day samples of Paris, France 1974-75
(Sebastien et al., 1980)
6- to 8-hour samples of New York 1969
City (Nicholson et al., 1971)
5 day, 7 hour control samples 1980-81
for U.S. school study (Constant,
Jr. et al , 1982)
16-hour samples of five U.S. 1974
cites (EPA, 1974)
New Jersey schools with damaged 1977
asbestos surfacing materials
in pupil use areas (Nicholson,
1978b)
U.S. school rooms/areas with 1980-81
asbestos surfacing material
(Constant, Jr et al., 1983)
U.S. school room/areas in 1980-81
building with asbestos
surfacing material
(Constant, Jr et al , 1983)
Buildings with asbestos 1976-77
materials in Paris, France
(Sebastien et al. , 1980)
U.S. buildings with friable 1974
asbestos in plenus or as
surfacing material (Nicholson
et al., 1975, 1976)
U.S. buildings with cementi- 1974
tious asbestos material in
plenum or as surfacing material
(Nicholson et al , 1975, 1976)
187
161
22
31
34
27
54
31
135
54
28
C = chrysotile.
A = amphibole.
3.3 C
0.96 C
16 C
9 (8C,lAb)
13 C
217 C
237 (231C,6A)
63 (54C.9A)
35 (25C,10A)
48 C
15 C
106
-------
taken in buildings constructed of such material (Nicholson, 1978a). Seven
samples taken in a school after a heavy rainfall showed asbestos concentra-
T ~
tions from 20 to 4500 ng/m (arithmetic mean = 780 ng/m -- all but two sam-
ples 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 in a concurrent study of houses con-
structed of asbestos cement materials. Roof water runoff from the homes
landed on the ground and was not reentrained, while that of the schools fell
to a smooth walkway, which allowed easy reentrainment, when dry. Contamina-
tion 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 by automobiles and other
vehicles. Measurements of brake and clutch emissions revealed 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 in occupational environments. In these
studies, concentrations of fibers that are longer than 5 pm were determined
using optical microscopy or were estimated from optical microscopic measure-
ments of total particulate matter. On the other hand, all current measure-
ments of low-level environmental pollution utilize electron microscopic tech-
niques, which determine the total mass of asbestos present in a given volume
of air. 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.
Some data relate optical fiber counts (longer than 5 urn) to the total
mass of asbestos as determined by electron microscopic techniques or other
weight determinations. These relationships (Table 5-11) provide crude esti-
mates of a conversion factor relating fiber concentrations fibers per milli-
liter to airborne asbestos mass micrograms per cubic meter. The proposed
standards for asbestos in Great Britain set by the British Occupational
107
-------
TABLE 5-11. MEASURED RELATIONSHIPS BETWEEN OPTICAL FIBER COUNTS
AND MASS AIRBORNE CHRYSOTILE
Sampling situation
r ., a
Fiber
counts
f/ml
Mass
concentration
jjg/m3
Conversion factors
[jg/m3 or fjg
fTmT 10"*? 103 f/mg
Textile factory
British Occupational
Hygiene Society
(1968) (weight vs.
fiber count)
2
Air chamber monitoring
Davis, et al. (1978) 1950
Monitoring brake
repair work
Rohl et al. (1976)
Electron Microscopy
(E.M. mass vs. 0.1 to 4.7
fiber count) (7 samples)
Textile mill
Lynch et al. (1970)
Friction products manufacturing
Lynch et al. (1970)
Pipe manufacturing
Lynch et al. (1970)
120
10,000
60
16
200
0.1 to 6.6
0.7 to 24
mean = 6
150C
170
45
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.
Conversion factor may be high because of overestimate of asbestos mass on the
basis of total magnesium.
Hygiene Society (BOHS) stated that a "respirable" mass of 0.12 mg of asbestos
per cubic meter was equivalent to 2 f/ml (BOHS, 1968). The standard did not
state how this relationship was determined. However, if the relationship was
obtained from magnesium determinations in an aerosol, the weight determination
would likely be high because of the presence of other nonfibrous, magnesium-
containing compounds in the aerosol. Such was the case in 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
108
-------
because of possible losses in the determination of mass by electron micro-
scopy. 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 is 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
3
range of conversion factors, 30 (jg/m /f/ml, will be used. The geometric
standard deviation of this value is 4, and this uncertainty severely limits
any extrapolation in which it is used. In the case of amosite, the data of
Davis et al. (1978) suggest that a conversion factor of 18 is appropriate.
However, these data yielded lower chrysotile values than all other chrysotile
estimates; therefore, they may also be low for amosite.
5.10 SUMMARY
Measurements using electron microscopic techniques have established the
presence of asbestos in the urban ambient air, usually at concentrations less
o 33
than 10 ng/m . Concentrations of 100 ng/m to 1000 ng/m have been measured
near specific asbestos emission sources, in schools where asbestos-containing
materials are used for sound control, and in office buildings where similar
materials are used for fire control. Most ambient measurements were taken
over ten years ago. More current data would be informative.
109
-------
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 range of values determined for K. and KM in Chapter 3 will be used to
calculate a range of risks from daytime exposure to 0.01 f/ml. This concen-
tration corresponds to about 300 ng/m , a concentration previously found in
several environmental exposure circumstances.
Tables 6-1, 6-2, and 6-3 list a range of calculated lifetime risks of
mesothelioma and lung cancer for a 40 hr/wk exposure to 0.01 f/ml for various
time periods. The risks from longer or shorter exposures/week can be esti-
mated by directly scaling the data in the tables. Values of K, = 0.3 to 3 x
_ 9 _ n
10 and of K,= 0.3 to 3.0 x 10 were used in these calculations. U.S. 1977
mortality rates (NCHS, Annually: 1967-1977) were utilized as the basic data
for the calculation. The tables utili/ed both smoking specific (Tables 6-1
and 6-2) and general population (Table 6-3) rates. We will assume that
current U.S. male mortality rates reflect the experience of 67% smokers (many,
however, are now exsmokers) and current female rates reflect the experience of
33% smokers. Using these percentages and the data of Hammond (1966) on the
mortality ratio of smokers to nonsmokers, smoking-specific total mortality
rates were calculated. Current lung canc.er rates for males will be multiplied
by 1.5 to represent the rates for smoking males. This factor comes from the
fact that current male rates largely result from the 67% of men who are smokers
or exsmokers. Correspondengly, current female lung cancer rates will be
multiplied by 3 to to reflect the fact that approximately 33% of women are
current or exsmokers. This factor for women may, in fact, 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 were taken from
110
-------
TABLE 6-1. THE RANGE OF LIFETIME RISKS PER 100,000 FEMALES OF DEATH FROM
MESOTHELIOMA AND LUNG CANCER FROM AN ASBESTOS EXPOSURE OF 0.01 F/ML FOR
40 HR/WK ACCORDING TO AGE AT FIRST EXPOSURE, DURATION OF EXPOSURE, AND SMOKING
Age at onset
of exposure
1
5
Years
10
of exposure
20
Li fetime
Mesothel ioma in Female Smokers
0
10
20
30
50
0
10
20
30
50
1.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0
6 -
4 -
o _
04 -
2
2
2
2
1
9.9
6.4
3.8
2.0
0.4
2.0
2.0
2.0
2.0
1.4
4.6
2.9
1.7
0.9
0.1
Lung
1.0
1.0 -
1.0 -
1.0
0.6
45.7
28.8
16.8
8.8
1.4
Cancer
9.6
9.6
9.6
9.5
6.3
Mesothel ioma
0
10
20
30
50
1.
0.
0.
0.
0.
1 -
7
4
2
,04
10.6
6.8
4.1
2.2
0.4
4.8
3.1
1.8
1.0
0.2
48.7
31.0
18.3
9.7
1.6
Lung Cancer i
0
10
20
30
50
0.
0.
0.
0.
0.
02
02
0?
0?
02
0.2
0.2
0.2
0.2
0.2
0.09
0.09
0.09
0.09
0.08
- 0.9
- 0.9
- 0.9
- 0.9
- 0.8
8.2
5. 1
2.9
1.5
0.2
in Femal
1.9
1.9
1.9
1.9
1. 1
in Female
8.8
5.8
3.2
1.6
0.2
n Female
0.2 -
0.2
0.2
0.2
0.2
82.2
51.0
29.1
14.7
2.1
e Smokers
19.1
19.1
19.1
18.5
11.1
13.3
8.0
4.4
2.1 -
0.3
3.8
3.8
3.8
3.4 -
1.6 -
133.0
80.0
43.8
21.0
2.5
38.1
38.1
37.5
34.2
16.2
18.0 -
10.2 -
5.2
2.3
0.3
10.7 -
8.8 -
7 0
5.1
1.7
180.0
102.0
52.0
23.4
2.5
107 1
88.2
69.2
50.7
17 4
Nonsmokers
87 7
58.0
31.7
16.4
2.4
14.2
8.7
4.8 -
2.4 -
0.3
142.4
86.6
48.0
23.5
2.9
19.4
11.1
5.7
2.6
0.3
194.4
111.3
57.6
26.3
2.9
Nonsmokers
1.9
1.9
1.9
1.9
1.5
0.4
0.4 -
0.4
0.4
0.3
3.7
3.8
3.7
3.6
2.5
1.2
1.0
0.8 -
0.6
0.3
11.7
9.9
8. 1
6.2
2.8
111
-------
TABLE 6-3. THE RANGE OF LIFETIME RISKS PER 100,000 PERSONS OF DEATH FROM
MESOTHELIOMA AND LUNG CANCER FROM AN ASBESTOS EXPOSURE OF 0.01 F/ML FOR
40 HR/WK ACCORDING TO AGE AND DURATION OF EXPOSURE. U.S. GENERAL POPULATION
DEATH RATES WERE USED AND SMOKING HABITS WERE NOT CONSIDERED
Age at onset Years of exposure
of exposure 1 5 10 20 Lifetime
Mesothelioma in Females
0
10
20
30
50
1.0
0.7
0.4
0.2
0.04
10.4
6.7
4.0
2.2
- 0.4
4.8
3.0
1.8
1.0
0.2
47.9
30.4
17 9
9.5
1.5
8.7
5.4
3.1
1.6
0.2
- 86.3
53.9
31.1
- 16.0
2.3
14.0
8.5
4.7
2.3 -
0.3
140.0
84.8
46.9
22.8
2.8
19.
10.
5.
2.
0.
,7 -
.9 -
6 -
.6 -
,3 -
196.6
108.9
56.3
25.5
2.8
Lung Cancer in Females
0
10
20
30
50
0.
0.
0.
0.
0.
.07
.07
.07
.07
.05
0.7
0.7
0.7
0.7
0.5
0.3 -
0.3 -
0.3 -
0.3
0.2
3.3
3.3
3.3
3.3
2.2
0.7
0.7
0.7
0.6
0.4
6.6
6.6
6.6
6.4
3.9
1.3 -
1.3 -
1.3 -
1.2 -
0.6 -
13.2
13.3
13.0
11.9
5.8
3.8 -
3.1 -
2.5 -
1.8 -
0.6 -
37.5
31.0
24.5
17.9
6.3
Mesothelioma in Males
0
10
20
30
50
0.
0.
0.
0.
0.
.8
.5
3
2 -
02 -
8.0
5.0
2.9
1.5
0.2
3.6
2.2
1.3
0.6
0.1
36.4
22.3
12.5
6.3
0.8
6.5
4.2
2.2
1.0
0.1
65.1
41.6
21.5
10.4
1.3
10.4 -
6.1 -
3.2 -
1.5 -
0.1 -
104.1
60.5
31.8
14.6
1.4
13.8 -
7.5 -
3.7 -
1.6 -
0.1 -
137.7
76.3
36.9
15.9
1.5
Lung Cancer in Males
0
10
20
30
50
0.2 -
0.2 -
0.2
0.2
0.2
2.1
2.1
2.2
2.2
1.8
1.1
1.1
1.1
1.1
0.8 -
10.6
10.6
10.7
10.7
8.2
2. 1
2.1
2.1
2.1 -
1.5
21.2
21.3
21.4
21.3
14.5
4.2 -
4.3 -
4.2 -
4.0 -
2.1 -
42.3
42.5
42.4
40.4
20.8
12.
10.
8.
6.
2.
2 -
1 -
1 -
1 -
2 -
121.8
101.4
80.7
60.6
21.6
112
-------
TABLE 6-2. THE RANGE OF LIFETIME RISKS PER 100,000 MALES OF DEATH FROM
MESOTHELIOMA AND LUNG CANCER FROM AN ASBESTOS EXPOSURE OF 0.01 F/ML FOR
40 HR/WK ACCORDING TO AGE AT FIRST EXPOSURE, DURATION OF EXPOSURE, AND SMOKING
Age at onset
of exposure
Years of exposure
1
5
Mesothel ioma
0
10
20
30
50
0
10
20
30
50
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
8 -
5 -
3 -
1 -
02
3
3
3
3
3 -
7.6
4.7
2.6
1.4
0.2
3.0
3.0
3.0
3.0
2.6
3.5
2.1
1.2
0.6
0.08
Lung
1.5
1.5
1.5
1.5 -
1.2 -
34.5
21.0
11.7
5.8
0.8
Cancer
14.9
15.0
15.2
15.2
11.6
Mesothel ioma
0
10
20
30
50
0
10
20
30
50
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
9
fi
3
?
03
n?
02
0?
02
02
8.9
5.6
3.2
1.7
0.3
- 2.1
- 2.1
-2.1
2.2
2.0
4.1
2.5
1.5
0.8
0.1
Lung
0.1
0.1
0.1 -
0.1 -
0.1 -
40.7
25.2
14.6
7.5
1.1
Cancer
1.1
1.1
1.1
1.1
0.9
in Male
6.2
3.7
2.0
1.0
0.1
in Mali?
3.0
3.0
3.0
3.0
2.0
in Male
7.3
4.5
2.5
1.3
0.2
in Male
0.2
0.2
0.2
0.2
0.2
10
Smokers
61.1
36.8
20.0
- 9.6
1.1
Smokers
29.9
30.0
30.2
30.0
20.3
9.8
5.6
2.9
1.3 -
0.1
6.0
6.0
5.0
5.7 -
2.9
20
98.2
55.6
29.4
13.2
1.3
59.6
59.9
59.6
56.6
28.8
Lifetime
12.9
7.0 -
3.4
1.4 -
0.1 -
17.0 -
14.1 -
11.3 -
8.4 -
3.0 -
129.3
70.2
34.2
14.4
1.3
170.1
141.3
112.5
84.0
30.0
Nonsmokers
- 73.1
44.7
- 25.1
- 12.5
- 1.6
Nonsmokers
-2.1
2.1
2.1
2.1
- 1.6
11.8
7.0
3.7
1.8
0.2
0.4
0.4
0.4
0.4
0.3
117.5
69.5
37.4
17.6
1.9
4.2
4.2
4.2
4.1
2.8
15.7 -
8.8 -
4.4 -
1.9 -
0.2 -
1.3 -
1.1 -
0.9 -
0.7
0.3 -
157.2
87 6
44.1
19.2
1.9
13.2
11.1
9. 0
6.9
3. 0
113
-------
Garfinkel (1981). The results show the importance of the time course of
mesothelioma. 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 un-
certain because of the variability of the data from which values of K, were
calculated and from uncertainties in extrapolating from risks estimated at
high occupational exposures to concentrations more than 100 times lower.
Thus, actual risks in a given environmental exposure could be outside the
listed ranges.
6.2 OBSERVED ENVIRONMENTAL ASBESTOS DISEASE
Asbestos-related disease in persons who had not been directly exposed at
the workplace has been known 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 Newhouse and Thomson (1965),
who showed that mesothelioma could occur among individuals whose potential
asbestos exposure consisted of having resided near an asbestos factory or in
the household of an asbestos Worker. Twenty of 76 cases from the files of the
London Hospital were the result of such exposures.
Of considerable importance are the forthcoming data on the prevalence of
X-ray abnormalities and the incidence of mesothelioma in family contacts of
the amosite factory employees in Paterson, New Jersey. Anderson and Selikoff
(1979) have shown that 35% of 685 family contacts of former asbestos factory
workers had abnormalities that were characteristic of asbestos exposure, when
they were x-rayed 30 or so years after their first household contact. The
data are shown in Tables 6-4 and 6-5, which compares 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 con-
cern was 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
was also significant.
114
-------
TABLE 6-4. PREVALENCE OF RADIOGRAPHIC ABNORMALITIES ASSOCIATED WITH ASBESTOS
EXPOSURE AMONG HOUSEHOLD MEMBERS OF AMOSITE ASBESTOS WORKERS
Exposure group
New Jersey urban residents**
Entered household after active
worker employment ceasedt
Household resident during active
worker employment!
Household resident and personal
occupational asbestos exposure
Total
exami ned
326
40
One or more radiographi
abnormalities present*
r-15 ( 5%>n
6 (15%^
685 ^240 (35%)
51 23 (45%)
X^ = 7.1 p <.
X2 = 114 p <.
c
01
001
*ILO U/C Pneumoconiosis 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. A MATCHED COMPARISON GROUP: CHEST X-RAY ABNORMALITIES AMONG
685 HOUSEHOLD CONTACTS OF AMOSITE ASBESTOS WORKERS AND 326 INDIVIDUAL
RESIDENTS IN URBAN NEW JERSEY
Group
Pleural Pleural Pleural Irregular*
Total thickening calcification plaques opacities
examined present present present present
Household contacts
of asbestos
workers 685
Urban New Jersey
residents 326
*ILO U/C Pneumoconioses Classification irregular opacities 1/0 or greater.
Source: Anderson and Selikoff (1979).
146 (18.8%) 66 (8.5%) 61 (7.9%) 114 (16.6%)
4 ( 1.2%) 0 (8.5%) 2 (0.6%) 11 ( 3.4%)
115
-------
TABLE 6-6. MESOTHELIOMA FOLLOWING ONSET OF FACTORY ASBESTOS
EXPOSURE, 1941-1945*
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 all 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)
Through 1977, four deaths from mesothelioma occurred among the family
contacts of these same factory workers. Table 6-6 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).
The study compared the mortality of a group of 1,779 residents within 0.5 mile
of the Paterson amosite asbestos plant with 3,771 controls in a different, but
economically similar section of town. No differences in the relative mortal-
ity experiences were seen, except for one mesothelioma in the neighborhood
group. This one case was in an electrician and occupational exposure may have
contributed to the disease.
116
-------
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) and adjusted to a continuous rather than day-
time exposure. If the air concentration in both circumstances 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 2,000, 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, would
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.
6.4 LIMITATIONS TO EXTRAPOLATIONS AND ESTIMATIONS
These calculations of unit risk values for asbestos must be viewed with
caution as they are uncertain and aspects of them are necessarily based on
estimates that are subjective to some extent because of the following limi-
tations in data: 1) one is extrapolating from high occupational levels to
much lower ambient levels, 2) the mass to fiber conversion is uncertain, 3)
various confounding aspects of the medical data, and 4) very importantly the
nonrepresentative nature of the exposure estimates.
117
-------
REFERENCES
Advisory Committee on Asbestos. (1979a) Vol I: Final report of the Advisory
Committee. Health and Safety Commission. Her Majesty's Stationery Office,
London.
Advisory Committee on Asbestos. (1979b) Vol. II: Papers prepared for the
Advisory Committee. Health and Safety Commission. Her Majesty's Stat-
ionery Office, London.
Amacher, O.E.; Alarif, A.; Epstein, S.S. (1975) The dose-dependent effects
of ingested chrysotile on DMA synthesis in the gastrointestinal tract,
liver, and pancreas of the rat. Environ. Res. 10: 208-216.
Anderson, H.A. (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. Grune and
Stratton. New York pp. 253-273.
Armitage, P ; Doll, R. (1960) Stochastic models for carcinogenesis. In: Pro-
ceedings of the fourth Berkeley symposium on mathematical statistics and
probability. Vol 4. Berkeley, CA: Univ. of Calif. Press; pp. 19-38.
Auribault, M. (1906) Bull del'Inspect, du Travail p. 126.
Ayer, H.E.; Lynch, J.R.; Fauney, 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. Proc. Am. Statistical Assoc. Meeting. Houston, Tex. Aug 11-14,
1980.
Balzer, J.L.; Cooper, W.C. (1968) The working environment of insulating workers
American Industrial Hygiene Association J. 29; 222-221.
Ban's, Y.I.; Artvinli, M. ; Sahin, A.A. (1979) Environmental mesothelioma in
Turkey. Ann. N. Y Acad. Sci. 330: 423-432.
Berry, G. (1973) Hygiene standards-theory and application. In: Bogovski,
P.; Timbrel!, V; Gil son, J.C., Wagner, J.C. (eds.). Biological Effects of
Asbestos. I.A.R.C. Scientific Pub. No. 8, Lyon, France, pp. 145-149.
Berry. G.; Newhouse, M.L. (1983) Mortality of workers manufacturing friction
materials using asbestos. Br J. Ind. Med. 36: 98-112.
Berry, G. ; Newhouse, M; Turok, M. (1972) Combined effect of asbestos expo-
sure and smoking on mortality from lung cancer in factory workers.
Lancet 2: 476-479.
118
-------
Berry, G. ; Wagner, J.C. (1969) The application of a mathematical model
describing the times of occurrence of mesotheliomas in rats following in-
oculation with asbestos. Br J. Cancer 23: 582-586.
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 dimen-
sional characteristics. In: Wagner, J.C. (ed.) Biological Effects of
Mineral Fibres. I.A.R.C. Scientific Publication No. 30, Lyon, France,
pp. 901-911.
Bignon, J. ; Sebastien, P ; Gaudichet, A. (1978) Measurement of asbestos re-
tention in the human respiratory system related to health effects. In:
Gravett, D.D. et al. , eds. Workshop on asbestos: definitions and meas-
urements methods. NBS Special Publication 506. Washington, DC, pp.
95-115.
Blot, W.J.; Harrington, J.M.; Toledo, A.; Hoover, R. ; Heath, C.W.; Fraumeni,
Jr., J.F. (1978) Lung cancer after employment in shipyards during World
War II. N. Engl J. Med. 299: 620-624.
Bohlig, H. ; Hain, E. (1973) Cancer in relation to environmental exposure.
In: Bogovski, P ; Timbrel!, V; Gilson, J.C.; Wagner, J.C. (eds.) Bio-
logical Effects of Asbestos. I.A.R.C. Scientific Publication No. 8,
Lyon, France, pp. 217-221.
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.
C.F.R. (1975) 29: §1910.1001; Occupational Safety and Health Standards
Asbestos.
C.F.R. (1982) 40: §61.20; Subpart B--National Emission Standard for Asbestos.
Chamberlain, M. ; Tarmy, E.M. (1977) Asbestos and glass fibres in bacterial
mutation tests. Mutat. Res. 43: 159-164.
Commins, B.T. ; Gibbs, G.W. (1969) Contaminating organic material in asbestos.
Br. J. Cancer 23: 358-362.
Constant, Jr., P.C.; Bergman, F.J.; Atkinson, G.R. (1983) Airborne asbestos
levels in schools. Final Report, E.P.A. Contract 68-01-5915. Midwest
Research Institute.
Cook, P.J.; Doll, R. ; Fellingham, S.F. (1969) A mathematical model for the H
distribution of cancer in man. Int. J. Cancer 4; 93-112.
Cooper, W.C.; Balzer, J.L. (1968) Evaluation and control of asbestos exposures
in the insulation trade. In: Holstein and Anspach (eds.), International
Konferiaz Uber die Biologischen Vierkunger des Asbestos, Deutsche Zent.
Arbeitsmed. Berlin, Germany
119
-------
Cooper, W. C.; Miedema, J. (1973) Asbestosis in the manufacture of insula-
ting materials. In: Bogovski, P.; Timbrel!, V ; Gilson, J.C.; Wagner,
J.C. (eds.) Biological Effects of Asbestos. I.A.R.C. Scientific Publi-
cation No. 8, Lyon, France, pp. 175-178.
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.
Dement, J.M.; Harris, R.L. ; Symons, M.J.; Shy, C. (1982) Estimates of dose-
response for respiratory cancer among chrysotile asbestos textile workers.
In: Walton, W.H. ed. Inhaled Particles V. Oxford: Pergamon Press.
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: 399-420.
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: 421-434.
Donna, A. (1970) Tumori sperimentali da amianto di crisotilo, crocidolite e
amosite in ratto Sprague-Dawley. Med. Lav. 61: 1.
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.
Environmental Protection Agency. (1973) National emissions standards for ha-
zardous pollutants, asbestos, beryllium and mercury. 38 FR 8820.
Environmental Protection Agency. (1974) A preliminary report on asbestos in
the Duluth, Minnesota area. Office of Technical Analysis.
Evans, J.C.; Evans, R.J.; Holmes, A.; Houram, 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.
Ferris, Jr , B.C.; Ranadive, M.V.; Peters, J.M.; Murphy, R.L.H.; Burgess,
W.A.; Pendergrass, H.P (1971) Prevalence of Chronic Respiratory Disease
Arch. Env. Health 23; 220-225.
Finkelstein, M.M. (1982a) Asbestosis in long-term employees of an Ontario
asbestos-cement factory Am. Rev. Respir. Dis. 125: 496-501.
Finkelstein, M.M. (1982b) Mortality in asbestos-cement workers. Presented
at: 2nd Int. Symp. on Epidemiology in Occupational Health; August;
Montreal.
120
-------
Finkelstein, M.M. (1983) Mortality among long-term employees of an Ontario
asbestos-cement factory. Br J. Ind. Med. 40: 138-144.
Fleisher, W.E.; Viles, Jr , F J. ; Gade, R.L.; Drinker, P. (1946) A health
survey of pipe covering operations in construction of naval vessels. J.
Ind. Hygiene and Toxicology 28; 9-16.
Fox, A.JV 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 inter-
action. 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.
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
chrysotile industry. Arch. Environ. Health 28: 69-71.
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.
Greenberg, M. ; Lloyd-Davies, T.A. (1974) Mesothelioma Register 1967-68. Br.
J. Ind. Med. 31: 91-104.
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.
Hammad, Y.Y.; Diem, J. ; Weill, W. (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: Epidemiological Study of Cancer and Other Chronic Diseases
N.C.I. Monograph 19. Washington, DC: U.S. Govt. Printing Office, pp.
127-104.
Hammond, E.G.; 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.; Selikoff, I.J.; Seidman, H. (1979b) Mortality Experience of
Residents in the Neighborhood of an Asbestos Factory, Ann. N.Y Acad.
Sci. 330: 417-422.
121
-------
Harington, J.S. (1962) Occurrence of oils containing 3,4-benzpyrene and re-
lated substances in asbestos. Nature (London) 193: 43-45.
Harington, J.S.; Roe, F.J.C. (1965) 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. Occu'p. 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.
Harris, Jr , R.L.; Eraser, D.A (1976) A model for deposition of fibers in the
human respiratory system. Am. Ind. Hyg. Assoc. J. 37: 73-89.
Henderson, V.I.; Enterline, P.E. (1979) Asbestos exposure: factors associa-
ted 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 can-
cer in men engaged in mining and milling crocidolite in Western Australia.
In: Wagner, J.C. (ed. ) Biological Effects of Mineral Fibres. I.A.R.C.
Scientific Publication No. 30. Lyon, France, pp. 615-625.
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, O.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 muta-
genic in Chinese hamster lung cells. Mutat. Res. 68: 263-274.
Hughes, J.; Wei 11, H. (1980) Lung cancer risk associated with manufacture
of asbestos-cement products. In: Wagner, J.C. (ed.) Biological Effects
of Mineral Fibres. I.A.R.C. Scientific Publication No. 30, Lyon, France,
pp. 627-634.
International Agency for Research in Cancer. (1973) In: Bogovski, P.;
Timbrel!, V.; Gilson, J.C.; Wagner, J.C. (eds.) Biological Effects of
Asbestos. I.A.R.C. Scientific Pub. No. 8, Lyon, France, 346 pp.
International Labour Office. (Revised 1971) International Classification of
Radiographs of Pneumoconioses. Occupational Safety and Health Series No.
22. Geneva, Switzerland.
Irwig, L.M.; DuToit, R.S.J.; SIuis-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: 34-52.
122
-------
Jacko, M.G.; DuCharme, R.T ; Somers, J.T. (1973) How much asbestos do vehi-
cles emit? SAE J. Automot. Eng. 81: 38-40.
Jacobs, k. ; Humphys, K.S.; Dodgson, K.S.; Richards, R.J. (1978) Light and
electron microscope studies of the rat digestive tract following pro-
longed 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.; Wignell,
B.K. ; Madeley, R.J.; Aggarwal, A. (1980) The consequences of exposure
to asbestos dust in a wartime gas-mask factory. In: Wagner, J.C. (ed.)
Biological Effects of Mineral Fibres. I.A.R.C. Scientific Publication
No. 30, Lyon, France, pp. 637-653.
Knox, J.F.; Holmes, S. ; Doll, R.; Hill, I.D. (1968) Mortality from lung can-
cer and other causes among workers in an asbestos textile factory. Br.
J. Ind. Med. 25: 293-303.
Langer, A.M. (1974) Inorganic particles in human tissues and their associa-
tion with neoplastic disease. EHP 9: 229-233.
Langer, A.M.; Wolff, M.S.; Rohl , A.M.; Selikoff, I.J. (1978) Variation of
some physical, chemical, and biological properties of chrysotile asbestos
subjected to prolonged milling. Toxicol. Environ. Health 4: 173-176.
Lavappa, K.S.; Fu, M.M.; Epstein, S.S. (1975) Cytogenetic studies on chryso-
tile asbestos. Environ. Res. 10: 165-173.
Lewinsohn, H.C. (1972) The medical surveillance of asbestos workers. R. Soc.
Health J. 92: 69-77
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 a-
nalysis: appraised by application to asbestos mining. J. Roy. Statis.
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, J.R.; Ayer, H.E. (1968) Measurement of dust exposures in the asbestos
textile industry. Am. Ind. Hyg. Assoc. J. 271: 431-437.
Lynch, J.R.; Ayer, H.E.; Johnson, D.L. (1970) The interrelationships of se-
lected asbestos exposure indices. Am. Ind. Hyg. Assoc. J. 31: 598-604.
Lynch, K.M.; Smith, W.A. (1935) Pulmonary Asbestosis III: carcinoma of lung
in asbesto-si1icosis. Am. J. Cancer 24: 56-64.
123
-------
Maltoni, C.; Annoscia, C. (1974) Mesotheliomas in rats following the intra-
peritoneal injection of crocidolite. In: Davis, W. and Maltoni, C. eds. ,
Advances in Tumour Prevention, Detection, and Characterization. Vol. 1:
Characterization of Human Tumours. Amsterdam, Excerpta Medica, p. 115.
Mancuso, J.F.; El-attar, A.A. (1967) Mortality patterns in a cohort of asbes-
tos workers. J. Occupational Medicine vol. 9, pp. 147-162.
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, J.C.; McDonald, A.D.; Gibbs, G.W.; Siemiatycki, J. ; Rossiter, C.E.
(1971) Mortality in the chrysotile asbestos mines and mills of Quebec.
Arch. Environ. Health 22: 677-686.
Merewether, E.R.A. (1947) Annual Report to the Chief Inspector of Factories.
London, H.M. Stat. Office, pp. 66-81.
Morgan A.; Evans, J.C.; Evans, R.J.; Hounam, 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.; Evans, R.J.; Hounam, F.R.; Holmes, A.; Doyle, S.G.
(1979) Fiber dimensions: their significance in the deposition and clear-
ance of inhaled fiber dust. In: Lemen, R. ; Dement, J.R. eds. Dusts and
Disease. Park Forest, IL: Pathotox Publishers; pp. 131-144.
Morgan, A.; Evans, J.C.; Holmes, A.
haled fibrous minerals in the
techniques. In: Walton, W.H.
Pergamon Press, pp. 259-273.
(1977) Deposition and clearance of in-
rat. Studies using radioactive tracer
ed. Inhaled Particles IV. London:
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.
Murphy, R.L.H.; Ferris, Jr., B.G.; Burgess, W.A., Worcester, J.; Gaenser, E.A.
(1971). Effects of low concentrations of asbestos. New England J. of
Med. 285; 1271-1278.
Murray. H.M. (1907) Report of the Departmental Committee on Compensation for
Industrial Disease. London, H.M. Stationary Office, p. 127.
National Center for Health Statistics (Annually, 1967-1977) Vital statistics
of the United States Vol. II Mortality. DHEW-DHHS publication
Hyattsville, MD.
National Institute of Occupational Safety and Health. (1972)
recommended standard: occupational exposure to asbestos.
Washington, DC, U.S. Government Printing Office.
Criteria for a
HMS 72-10267.
124
-------
National Institute for Occupational Safety and Health (1976) Revised recommended
Asbestos Standard DHEW Publication 79-164, Washington, D.C., U.S. Govern-
ment Printing Office.
National Institute for Occupational Safety and Health. (1979) USPHS/NIOSH
membrane filter method for evaluating airborne asbestos fibers. DHEW
Publication 79-127 Washington, DC, U. S. Government Printing Office.
National Institute for Occupational Safety and Health. (1980) NIOSH/OSHA
Asbestos Work Group. Workplace exposures to asbestos: review and recom-
mendations. Dept. of Health and Human Services Publication 81-103
(NIOSH), Washington, DC: U.S. Government Printing Office.
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 disease among long-term asbestos
workers in the United Kingdom. Ann. N. Y Acad. Sci. 330: 53-60.
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.
Newhouse, M.L.; Thompson, H. (1965) Mesothelioma of the pleura and peri-
toneum following exposure to asbestos in the London area. Br. J. Ind.
Med. 22: 261.
Newman, H.A., 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. EHP., Environ. Health
Perspect. 34: 103-111.
Nicholson, W.J. (1971) Measurement of asbestos in ambient air. Final Report,
Contract CPA 70-92, National Air Pollution Control Administration.
Nicholson, W.J. (1975) Insulation Hygiene. Progress Reports Vol. 3 No. 1, Mt.
Sinai School of Medicine, NY, NY
Nicholson, W.J. (1976a) Asbestos - the TLV approach. Ann. N. Y. Acad. Sci.
271: 152-169.
Nicholson, W.J. (1976b) Submission to the Comite d'etude sur la salubrite
dans 1'industrie de 1'amiante (Beaudry Commission), Annexe, pp. 151-160.
Nicholson, W.J. (1978a) Chrysotile asbestos in air samples collected in
Puerto Rico. Report to C.P.S.C., Contract 77128000.
Nicholson, W.J. (19785) Control of sprayed asbestos surfaces in school
buildings: a feasibility study. Final report to the Nat. Inst. of
Environ. Health Sci., Contract l-ES-2113. See also: Nicholson, W.J.;
Swoszowski, Jr., E.J.; Rohl , A.M.; 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.
125
-------
Nicholson, W.J. (1981) Criteria Document for Swedens Occupational Standards:
Asbestos and Inorganic Fiber; Arbete Och Ha'lsa Vol. 17, 103 pp.
Nicholson, W.J. (1982a) The doso and time dependence of occupational cancer.
In: Prevention of Occupational Cancer: International Symposium, Interna-
tional Labor Office Occupational Safety and Health Series, No. 46; Geneva,
Switzerland, pp. 44-67
Nicholson, W.J.; Holaday, D.A.; Heimann, H. (1972) Proceedings of the Inter-
national Symposium on Safety and Health in Shipbuilding and Shiprepairing,
Helsinki International Labour Organization Occupational Safety and
Health Series, Geneva, Switzerland: Press 27: 27~47.
Nicholson, W.J.; Pundsack, F L. (1973) Asbestos in the environment. In:
Bogovski, P.; Timbrel!, V.; Gilson, J.C.; Wagner, J.C. (eds.) Biological
Effects of Asbestos. I.A.R.C. Scientific Publication No. 8, Lyon, France,
pp. 126-130.
Nicholson, W.J.; Rohl, A.N.; Ferrand, E.F (1971) Asbestos air pollution in
New York City. In: England, H.M. ; Beery, W.T. eds. Proceedings of the
Second Clean Air Congress. New York, NY: Academic Press, pp. 136-139.
Nicholson, W.J.; Rohl, A.N.; Weisman, I. (1975) Asbestos contamination of
the air in public buildings. Final Report, Contract 68-02-1346, E.P.A.
See also: Nicholson, W.J.; Rohl, A.N.; Weisman, I. (1976) Asbestos
contamination of building air supply systems. Proceedings of the Inter-
national Conf on Environ. Sensing and Assessment. 1st Electrical and
Electronic Engineers Vol. II. Paper 29-6.
Nicholson, W.J.; Perkel, G.; Selikoff, I.J. (1982b) Occupational exposure to
asbestos: population at risk and projected mortality - 1980-2030. Am. J.
Ind. Med. 3: 259-312.
Nicholson, W.J.; Rohl, A.N.; Weisman, I.; Selikoff, I.J.
mental asbestos concentrations in the United States.
(ed.) Biological Effects of Mineral Fibres. I.A.R.C.
cations, No. 30, Lyon, France, pp 823-827.
(1980) Environ-
In: Wagner, J.C.
Scientific Publi-
Nicholson, W.J.; Selikoff, I.J.; Seidman,
ience of asbestos factory workers:
asbestos exposure. Environ. Res. In
H.; Hammond, E.C. Mortality exper-
effect of differing intensities of
Press (1983).
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.
Peto, J. (1977) The establishment of Industrial Hygiene Standards: An example.
In: Whittemore, A. (ed. ) Health Quantitative Methods, Society for
Industrial and Applied Mathematics. Phi!a., PA pp. 104-114.
Peto, J. (1978)
484-489.
The hygiene standard for chrysotile asbestos. Lancet 1
126
-------
Peto, J. (1980) Lung cancer in relation to measured dust levels in an
asbestos factory. In: Wagner, J.C. (ed. ) Biological Effects of Mineral
Fibres. I.A.R.C. Scientific Publication No. 30, Lyon, France, pp.
829-936.
Peto, J.; Seidman, H. ; Selikoff, I.J. (1982) Mesothelioma incidence in
asbestos workers: implications for models of carcinogenesis and risk
assessment. Br. J. Cancer 45: 124-135.
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. (ed.) Biological Effects of Mineral Fibres. I.A.R.C.
Scientific Publication No. 30, Lyon, France, pp. 261-272.
Pott, F.; Friedrichs, K.H. (1972) Tumoren der ratte nach i.p.-injektion fas-
formiger Staube. Naturwissenschaften 59:318. Abs. 2321.
Pott, F.; Friedricks, K.H.; Huth, F. (1976) Ergebnisse aus tierversuchen zur
kanzerogenen wirkung faserforminger staube und ihre deutung im hinblick
auf die tumorentseehung beim menschen. Zentralbl. Bakteriol. Parasiterkd.
Infektionskr. Hyg. Abt. 1: Orig. Reihe B 162: 467.
Puntoni, R. ; Vercelli, F ; Merlo, F.; Valerio, F. ; Santi, L. (1979) Morta-
lity 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.; Timbrel!, V.; Gilson, J.D.;
Wagner, J.C. (eds.), Biological Effects of Asbestos. I.A.R.C. Sci. Publ.
No. 8, Lyon, France, p. 99.
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.C. (1974) Inhalation carcinogenesis from
various forms of asbestos. Environ. Res. 8: 178-202.
Reeves, A.L.; Puro, H.E.; Smith, R.G.; Vorwald, A.J. (1971) Experimental
asbestos carcinogenesis. Environ. Res. 4: 496-511.
Rendall, R.E.G.; Skikne, M.I. (1980) Submicroscopic fibres in industrial
atmospheres. In: Wagner, J.C. (ed.) Biological Effects of Mineral Fibres.
I.A.R.C. Scientific Pub. No. 30, Lyon, France, pp. 837-843.
Robinson, C. ; Lemen, R. ; Wagoner, J.K. (1979) Mortality patterns, 1940-75,
among workers employed in an asbestos textile friction and packing pro-
ducts manufacturing facility. In: Lemen, R.; Dement, J.R. eds. Dusts and
Disease. Park Forest, IL: Pathotox Publishers; pp. 131-144.
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.
127
-------
Rom, W.N.; Livingston, G.K.; Casey, K.R.; Wood, S.D.; Egger, M.J.; Chiu, G.L. ;
Jerominski, L. (1983) Sister chromatic exchange frequency in asbestos
workers. J. Natl. Cancer Inst. 70: 45-48.
Rubino, G.F.; Piolatto, G. ; Newhouse, M.L.; Scansetti, G. ; Aresini, G.A.;
Murray, 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.; Stockham, J.D. (Revised: 1978) Electron micro-
scopic measurement of airborne asbestos concentrations - a provisional
methodology manual. EPA-600/2-77-178.
Sawyer, R.N. (1977) Asbestos exposure in a Yale building: analysis and reso-
lution. Environ. Res. 13: 146-168.
Sawyer, R.N. (1979) Indoor air pollution: application of hazard criteria.
Ann. N. Y. Acad. Sci 330: 579-586.
Schneider, V.; Maurer, R.R. (1977) Asbestos and embryonic development.
Teratology 15: 273-280.
Sebastien, P ; Bignon. J. ; and Martin M. , (1982) Indoor airborne abestos
pollution: from the ceiling and the floor. Science 216: 1410-1413.
Sebastien, P ; Bi11 ion-Gal land, M.A.; Dufour, G. ; Bignon, J. (1980) Measure-
ment of asbestos air pollution inside buildings sprayed with asbestos.
EPA 560/13-80-026.
Sebastien, P ; Janson, X.; Bonnand, G. et al. (1979) Translocation of
asbestos fibers through respiratory tract and gastrointestinal tract
according to fiber type and size. In: Lemen, R. ; Dement, J.M. eds.
Dusts and Disease. Park Forest, IL: Pathotox Publishers; pp. 65-85.
Selikoff, I.J.; Churg, J. ; Hammond, E.C. (1964) Asbestos exposure and neo-
plasia. J. Am. Med. Assoc. 188: 22-26.
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: 106-112.
Selikoff, I.J.; Hammond, E.C.; Churg, J. (1970) Mortality experience of
asbestos insulation workers, 1943-68. In: Shapiro, M.S. ed. Proceedings
International Conference Pneumoconiosis, Johannesburg. Capetown, Oxford
Univ. Press; pp. 97-103.
Selikoff, I.J.; Hammond, E.C.; Seidman, H. (1979) Mortality experience of
insulation workers in the United States and Canada. Ann. N. Y. Acad.
Sci. 330: 91-116.
Selikoff, I.J.; Nicholson, W.J.; Lilis, R. (1981) Radiological evidence of
asbestos disease among ship repair workers. Am. J. Ind. Med. 1: 9-22.
128
-------
Selikoff, I.J.; Seidman, H. (1981). Cancer of the pancreas among asbestos
insulation workers. Cancer 47: 1469-1473.
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.
Shabad, L.M.; Pylev, L.M.; Krivosheeva, L.V.; Kulagina, T.F.; Nemenko, B.A.
(1974) Experimental studies on asbestos carcinogenicity. JNCI. , J.
Nat! Cancer Inst. 52: 1175-1187
Siemiatycki, J. (1982) Health effects on the general population (Mortality
in the general population in asbestos mining areas). Presented at: World
Symposium on Asbestos; May; Montreal, Quebec.
Sincock, A.M. (1977) I_n Vitro Chromosomal effects of asbestos and other
materials. In: Origins of Human Cancer. Cold Spring Harbor, NY, Press,
1976.
Smith, W.E.; Miller, L. ; Elsasser, R.E.; Hubert, D.D. (1965) Tests for car-
cinogenicity in animals. Ann. N. Y. Acad. Sci. 132: 456-488.
Smith, W.E.; Hubert, D.D. (1974) The intrapleural route as a means for esti-
mating carcinogenicity. In: Karbe, E. ; Park, J.R. eds. , Experimental
Lung Cancer. Berlin: Springer-Verlag, p. 92.
Smith, W.E.; Miller, L.; Churg, J (1970) An experimental model for study of
cocarcinogenesis in the respiratory tract. In: Nettesheim, P. ed. Mor-
phology of Experimental Respiratory Carcinogenesis. U.S. Atomic Energy
Comm. Oak Ridge, Tennessee, pp. 299-316.
Smither, W.J.; Lewinsohn, H.C. (1973) Asbestos in textile manufacturing.
In: Bogovski , P ; Timbrel 1, V.; Gilson, J.C.; Wagner, J.C. (eds.), Bio-
logical Effects of Asbestos. I.A.R.C. Scient. Publ. No. 8, Lyon, France,
pp. 169-174.
Spurny, K.R.; Stbber, W. ; Weiss, G.; Opieta, H. (1980) Some special problems
concerning asbestos fiber pollution in ambient air. Atmos. Pollut. 8:
315-322.
Stanton, M.F. (1973) Some etiological considerations of fibre carcinogenesis.
In: Bogovski, P ; Timbrel!, V.; Gilson, J.C.; Wagner, J.C. (eds.), Bio-
logical Effects of Asbestos. I.A.R.C. Sci. Publ. No. 8, Lyon, France, p.
289.
Stanton, M.F.; Layard, M. ; Teyeris, A.; Miller, E.; May, M.; Kent, E. (1977)
Carcinogenicity of fibrous glass: pleural response in the rat in relation
to fiber dimensions. JNCI., 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. JNCI., J. Natl. Cancer Inst. 64:
965-975.
129
-------
Stanton, M.F ; Wrench, C. (1972) Mechanisms of mesothelioma induction with
asbestos and fibrous glass. JNCI., J. Natl. Cancer Inst. 48: 797-821.
Steel, J. (1979) Asbestos control limits. In: Vol II, Papers prepared for
the Advisory Committee on Asbestos. London, Her Majesty's Stationery
Office, pp. 85-88.
Storeygard, A.R.; Brown, A. L. (1977) Penetration of the small intestinal
mucosa by asbestos fibers. Mayo Clin. Proc. 52: 809-812.
Timbrel!, V. (1965) The inhalation of fibrous dusts. Ann. N. Y. Acad. Sci.
132: 255-273.
Toft, P.; Wigle, D.; Meranger, J.C.; Mao, Y. (1981) Asbestos and drinking
water in Canada. Science Total Environ. 18: 77-89.
Wagner, J.C. (1977b) Studies of the carcinogenic effect of fibre glass of
different diameters following intrapleural inoculation in experimental
animals. In: Natl Inst. Occup. Safety and Health Symp. Occupational
Exposure to Fibrous Glass. Univ. of Maryland, 1977.
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.
Wagner, J.C.; Berry, G.; Cook, T.J.; Hill, R.J.; Pooley, F.D.; Skidmore, J.W.
(1977a) Animal experiments with talc. In: Walton, W.C. ed. Inhaled
Particles and Vapors, IV. New York: Pergamon Press; pp. 647-654.
Wagner, J.C.; Sleggs, C.A.; Marchand, P. (1960) Diffuse pleural mesothelioma
and asbestos exposure in the north western Cape Province. Br. J. Ind.
Med. 17: 260-271.
Wagner, J.C.; Berry, G. ; Skidmore, J.W. ; Timbrel 1, V. (1974) The effects of
the inhalation of asbestos in rats. Br. J. Cancer 29: 252-269.
Wagner, J.C.; Berry, G. ; Timbrel 1, V. (1973) Mesotheliomata in rats after
inoculation with asbestos and other materials. Br. J. Cancer 28:
173-185.
Webster, I. (1970) Asbestos exposure in South Africa. In: Shapiro, H.A. ed.
Proceedings of International Conference Pneumoconiosis, Johannesburg,
Capetown, Oxford Univ. Press; pp. 209-212.
Weill, H. ; Hughes, J. ; Waggenspick, C. (1979) Influence of dose and fiber
type on respiratory malignancy in asbestos cement manufacturing. Am.
Rev. Respir Dis. 120: 345-354.
Weiss, A. (1953) Pleurakrebs bei 1ungensabestos, rn vivo morphologisch gesi-
chert. Medizinische 3: 93-94.
Weiss, W. (1971) Cigarette smoking, asbestos and pulmonary fibrosis. Am.
Rev. Respir. Dis. 104: 223-227.
130
-------
Wigle, D.T. (1977) Cancer mortality in relation to asbestos in municipal
water supplies. Arch. Environ. Health 32: 185-189.
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.
131
-------