United States
                Environmental Protection
                Agency
Office of Health and
Environmental Assessment
Washington DC 20460
                Research and Development
EPA/600/S8-84/003F Apr. 1991
EPA       Project  Summary
                Airborne Asbestos  Health
                Assessment  Update
                W.J. Nicholson
                  Data developed since the early 1970s,
                from large population studies with long
                follow-up, have added to our  knowl-
                edge of asbestos-related diseases and
                strengthened the evidence for associa-
                tions  between asbestos  and specific
                types of health effects. Lung  cancer
                and mesothelioma are the most Impor-
                tant asbestos-related causes of death
                among  exposed individuals. Cancer
                other than lung has also been associ-
                ated with asbestos exposure. The accu-
                mulated data suggest that the excess
                risk of lung cancer from asbestos expo-
                sure is proportional to the cumulative
                exposure (the duration times the inten-
                sity) and the underlying risk in the ab-
                sence of exposure. The risk of death
                from mesothelioma Is approximately
                proportional to the cumulative  expo-
                sure to asbestos and increases sharply
                with time since onset of exposure.
                  Animal studies confirm the  human
                epidemiologies! results and indicate that
                all major asbestos varieties produce
                lung cancer and  mesothelioma, with
                only limited differences in carcinogenic
                potency. Some measurements demon-
                strate that asbestos exposures exceed-
                ing 100 times  background occur  in
                non-occupational  environments.  Cur-
                rently, the most important of these non-
                occupational exposures is the release
                of fibers from asbestos-containing sur-
                facing building materials or from
                sprayed asbestos fireproofing in high-
                rise buildings.
                  Extrapolations of risks of asbestos
                cancers  from occupational circum-
                stances can be made, although numeri-
 cal estimates  in a specific exposure
 circumstance  have a  large (approxi-
 mately tenfold) uncertainty. Because of
 this uncertainty, calculations of unit risk
 values for asbestos at low concentra-
 tions must be viewed with caution and
 are subject to the following limitations:
 1) variability in the exposure-response
 relationship at high exposures; 2) un-
 certainty in extrapolating to exposures
 1/100 as much; and 3)  uncertainties in
 conversion of optical  fiber counts  to
 electron microscopic fiber counts  or
 mass determinations.
    This Project Summary was devel-
 oped by EPA's Environmental Criteria
 and Assessment Office, Research Tri-
 angle Park, NC, to announce key find-
 Ings of the research project published
 In 1986 that Is fully documented In a
 separate report of the same title (see
 Project Report ordering Information at
 back).

 Introduction
   The  principal objective of this docu-
 ment is to provide the U.S. Environmental
 Protection Agency (EPA) with a sound sci-
 entific basis for review  and revision,  as
 appropriate, of the national emission stan-
 dard for asbestos, 40 CFR 61, subpart B,
 as  required by the 1977 Clean Air Act
 Amendments, Sections 111 and 112. The
 health effects basis for designating asbes-
 tos as a hazardous pollutant and minimiz-
 ing emissions via the original 1973 National
 Emissions Standard for Hazardous Air
 Pollutants was  scrutinized, at that  time,
 during two public hearings and a public
 comment period. Once a pollutant has been
                                                                 Printed on Recycled Paper

-------
designated as a "hazardous" air pollutant,
the burden of proof is placed on proving
that designation wrong.
   The  original health effects  basis for
designating asbestos as a hazardous air
pollutant was qualitative evidence estab-
lishing asbestos-associated  carcinogenic
effects.  However, insufficient bases then
existed by which to define pertinent quanti-
tative dose-response relationships; i.e., unit
risk values could not be credibly estimated.
The main focus of the update document is
to describe asbestos-related health effects
developments since  1972 and to deter-
mine if new data warrant the specification
of unit risk values for asbestos. The docu-
ment forms part of the basis to perform a
risk assessment.
   The  National  Academy  of  Sciences
(NAS) in  1983 suggested a definition of
risk assessment as the use of the factual
data base to define the health  effects of
exposure  of individuals or populations to
hazardous materials, such as  asbestos.
Therefore, the update document is mainly
concerned with the excess risk of cancer
from  inhalation of asbestos fibers, with
emphasis placed  on the  literature  pub-
lished after 1972,  and on those  reports
that provide information on the risk from
low-level  exposures, such as  those en-
countered in the  non-occupational  envi-
ronment.

Occupational Exposure
   The International Agency for Research
on Cancer (IARC) lists asbestos as a group
1 carcinogen, meaning that exposure to
asbestos is carcinogenic to humans. EPA's
proposed  guidelines would categorize as-
bestos as Group A, human carcinogen.
   Diseases considered to be associated
with asbestos exposure include asbestosis,
mesothelioma, bronchogenic carcinoma,
and cancers of the gastrointestinal  (Gl)
tract,  including the esophagus, stomach,
colon, and rectum. Lung cancer is associ-
ated with  exposure to four principal com-
mercial varieties of asbestos fiber: amosite,
anthophyllite, crocidolite, and  chrysolite.
Excess  risks of bronchogenic  carcinoma
are documented  in mining  and  milling,
manufacturing, and end product use (ap-
plication of insulation materials). Mesothe-
lioma is a cause of death among  factory
employees, insulation applicators, and
workmen employed in the mining and mill-
ing  of crocidolite. A much lower  risk of
death from mesothelioma  is  observed
among chrysolite or amosite mine and mill
employees, and no cases  are associated
with anthophyllite exposure. The IARC Ad-
visory Committee suggests that the risk of
death from mesothelioma is greatest with
crocidolite, less with amosite, and still less
with chrysotile. This suggestion was based
on the association of disease with expo-
sures.  No unit exposure risk information
existed prior to 1972.
   Information on exposure-response re-
lationships for lung cancer risk among vari-
ous exposed groups was scanty. Data from
Canadian mine and mill employees clearly
indicated an increasing risk with increasing
exposure, measured in terms of millions of
particles  per cubic  foot-years (mppcf-y),
but data  on the risk at minimal exposure
were  uncertain because the number of
expected deaths calculated using adjacent
county rates suggested that all exposure
categories were at elevated  risk. 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. Exposures were
expressed in mppcf-y, and information on
conversion of mppcf to fibers per milliliter
was available  only for  textile production.
Despite the paucity  of data,  the 1973 re-
port of the Advisory Committee on Asbes-
tos Cancers to  the IARC stated, "The
evidence ... suggests that an excess lung
carcinoma risk is not detectable when the
occupational exposure has been bw. These
low occupational exposures  have almost
certainly  been  much greater than that to
the public from general air pollution." Lim-
ited data existed on the association of Gl
cancer with asbestos exposure, but  the
"excess is relatively small compared with
that for bronchial cancer."
   The prevalence of asbestosis, particu-
larly as manifested by X-ray abnormalities
of the  pleura or parenchyma! tissue, had
been documented more extensively than
the risk of the asbestos-related malignan-
cies. In part, this  documentation  resulted
from knowledge of this  disease extending
back to the turn  of the century, whereas
the malignant  potential of asbestos was
not suggested until  1935 and not widely
appreciated until  the  1940s. Asbestosis
had been documented in a wide variety of
work circumstances and  associated with
all commercial types of asbestos fibers.
Among some heavily exposed groups, 50
to 80 percent of individuals employed for
20  or  more years  were found to have
abnormal X-rays characteristic of asbestos
exposure. A lower percentage of abnormal
X-rays was present in lesser  exposed
groups.
   Company data supplied  to the British
Occupational Hygiene Society (BOHS) on
X-ray and clinical abnormalities among 290
employees of  a large  textile production
facility  in Great Britain were analyzed in
terms of a fiber exposure-response rela-
tionship. The results were utilized in estab-
lishing  the  1969  British  regulation on
asbestos. These data suggested that the
risk of  developing the  earliest signs of
asbestosis was  less than 1  percent for
accumulated fiber exposure of 100 fiber-
years/ml (f-y/ml), e.g., 2 fibers/millili-
ter (f/ml) for 50 years.  However, shortly
after the establishment of the British stan-
dard,  additional data from the same fac-
tory population suggested a much greater
prevalence of X-ray abnormalities than was
believed to exist at the time the British
standard was set in  1972.  These data
resulted from use of the  new International
Labour Office (ILO) standard classification
of X-rays and the longer time from onset of
employment. Of the 290 employees whose
clinical data were reviewed by the BOHS,
only 13 tiad been employed for 30 or more
years; 172 had  less than 20 years of em-
ployment. The progression of asbestosis
depends on both cumulative exposure and
time from exposure; therefore, analysis in
terms of only one variable can be mislead-
ing.

Environmental Exposure
   Several research groups had shown
that asbestos disease risk could develop
from other than direct occupational expo-
sures. Various researchers in 1960 have
shown that a mesothelioma risk in environ-
mental 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 resulted from
exposure occasioned by living or working
in the area of the mining activity. Another
study in 1965 that showed an extra-occu-
pational risk investigated the occupational
and residential background of 76 individu-
als deceased of mesothelioma in a Lon-
don 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. In  1973,
investigators identified environmental as-
bestos exposure in 38 mesothelioma cases
without occupational exposure who resided
near an asbestos factory, further defining
residential risk. A final study, which is par-
ticularly important because of the size of
the population implied to be at risk  of as-
bestos disease from indirect occupational
exposure in the shipbuilding industry, was
conducted in 1968. It described the pres-
ence  of asbestosis in 13 individuals and
mesothelioma in 5 others who were em-

-------
ployed in a shipyard but were not mem-
bers of trades that regularly used asbes-
tos. Rather, they were exposed to the dust
created by other employees placing or re-
moving insulation.
   Evidence of ubiquitous general popula-
tion exposure and environmental contami-
nation from the spraying of asbestos on
the steel-work of  high-rise buildings was
established by 1972. Asbestos was com-
monly found  at concentrations  of nano-
grams per cubic meter (ng/m3) in virtually
all United States cities, and at concentra-
tions of micrograms per liter (u.g/1) in  river
systems of the United States. Concentra-
tions 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 it suggested
that  "the evidence  relates to conditions
many years ago .... There is no evidence
of a risk to the general public at present."

Analytical Methodology
   During the late 1960s and  early 1970s,
significantly improved methods were de-
veloped for assessing asbestos  disease
and quantifying asbestos in the  environ-
ment. In 1971, a standardized methodol-
ogy was established. It provided a uniform
criterion for assessing the prevalence  of
asbestos-related X-ray abnormalities.
   Significant advances  were  also
achieved in the quantification of asbestos
aerosols. In the late 1960s, the membrane
filter technique was developed for the
measurement  of asbestos fibers in work-
place aerosols. While this procedure has
some limitations,  it did establish a stan-
dardized method, using simple instrumen-
tation, that was far superior to any that
existed previously. This method subse-
quently allowed epidemiological studies to
be done that based exposure estimates on
a  standardized criterion.  Experimental
techniques in the quantification of asbes-
tos at concentrations of tenths of ng/m3 of
air and tenths of u.g/1 of water were  also
developed, extending the sensitivity of ex-
posure estimates approximately three or-
ders  of  magnitude below those  of
occupational  aerosols and allowing as-
sessment of general population exposures.
Finally, techniques for the analysis of as-
bestos in lung and other body tissues were
developed. Tissue digestion techniques and
the use of electron microscopy to analyze
fibers contained in the digest and in thin
sections of lung tissue showed that asbes-
tos fibers were commonly  present in the
lung tissue of general population residents
as well as individuals exposed in occupa-
tional circumstances.

Experimental Studies
   Experimental animal studies using as-
bestos fibers confirmed the risks of lung
cancer and  mesothelioma from  amosite,
crockJolite, and chrysotile. In each case,
the establishment of a risk in animals fol-
lowed the association of  the malignancy
with human  exposure. For  example, a
causal relationship between  lung  cancer
and  asbestos exposure in humans was
suggested in 1935 and confirmed in the
late  1940s but was not described in the
open literature in animals until 1967. Me-
sothelioma, reported in an asbestos worker
in 1953, was produced in animal  experi-
mentation in  1965. Other animal  experi-
mentation showed that combinations of
asbestos and other carcinogenic materials
produced  an enhanced risk  of asbestos
cancer. Asbestos exposure combined with
exposure  to benz(a)pyrene was demon-
strably more carcinogenic than exposure
to either agent alone. Additionally, organic
and metal compounds associated with as-
bestos fibers were ruled out as important
factors in the carcinogenicity  of  fibers.
Lastly, in  1973 experimentation  involving
the application of fibers onto the pleura of
animals indicated that the important factor
in the carcinogenicity was the length and
width of the fibers rather than their  chemi-
cal properties. The greatest carcinogenic-
ity was related to fibers that were less than
2.5 u.m in diameter and longer than  10 u.m.

Current Asbestos Standards
   The current Occupational Safety and
Health Administration  (OSHA) standards
for an 8-hour time-weighted average (TWA)
occupational  exposure to asbestos  is 2
fibers longer than 5 urn in length per millili-
ter of air  (2 f/ml or 2,000,000 f/m3). Peak
exposures of up to 10 f/ml are permitted
for no more than 10 min. This standard
has  been in effect since July  1, 1976,
when it replaced an earlier one of 5 f/ml
TWA. In Great Britain, a value of 0.5 f/ml is
now the accepted level for chrysotile. This
standard  has evolved from recommenda-
tions made in 1979 by the Advisory Com-
mittee  on    Asbestos,  which   also
recommended a TWA of 0.5 f/ml for amosite
and 0.2 f/ml  for crocidolite. From 1969 to
1983, 2 f/ml  TWA was the standard for
chrysotile. This  earlier British standard
served as a  guide  for the 1972  OSHA
standard.
   The 1969 British standard was devel-
oped  specifically  to  prevent asbestosis
among working populations; data that would
allow a determination of  a standard for
cancer were felt to  be lacking. Among
occupational groups, cancer is the primary
cause  of excess death among workers.
Three-fourths or more of asbestos-related
deaths are from  malignancy. This fact led
OSHA to propose a  lowered TWA stan-
dard to 0.5 f/ml (500,000 f/m3) in October
1975. The  National Institute for Occupa-
tional Safety and Health (NIOSH) antici-
pated  hearings on a new standard and
proposed a value of 0.1 f/ml in 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 ana-
lytical techniques using optical microscopy,
and that 0.1 f/ml may not necessarily pro-
tect against cancer. The preamble to the
OSHA proposal acknowledges that no in-
formation  exists by  which to  define a
threshold for asbestos carcinogenesis. The
OSHA proposal has been withdrawn, and
a new proposal was submitted on April 10,
1984. In it,  OSHA proposed a TWA stan-
dard of either 0.2  or 0.5 f/ml depending
upon information to be obtained in hear-
ings (held during the summer of 1984).
NIOSH reaffirmed its position on a 0.1 f/ml
TWA standard.
   The existing federal national emission
standards  for asbestos are published  in
Part 61, Title 40, Code of Federal Regula-
tions. In summary, these apply to milling,
manufacturing, and fabrication sources, and
to demolition, renovation, and waste dis-
posal,  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 limita-
tions. However, some local governmental
agencies have numerical standards (e.g.,
New York: 27 ng/m3), but these have little
regulatory relevance.

Other Reviews of Asbestos
Health Effects
   Recently, several  government agen-
cies in  different countries reviewed asbes-
tos health  effects. The  most important of
the reviews outside the United States are
those of the Advisory Committee on As-
bestos of the British Health and  Safety
Commission and the report of the Ontario
Royal Commission. Each of these major
reports was the result of lengthy testimony
by  many scientists and deliberations by
selected committees over a long period of
time. In the United States, the NAS has
reviewed the non-occupational health risk
of asbestiform fibers, and a Chronic Haz-
ard Advisory Panel convened by the U.S.

-------
 Consumer Product Safety Commission re-
 ported on the hazards of asbestos.
    There are large  areas  of  agreement
 and some of disagreement between these
 reviews  and those of  the  full document
 with regard  to the spectrum of asbestos-
 related disease, the models describing as-
 bestos-related   lung   cancer   and
 mesothelioma, unit exposure risks in occu-
 pational  circumstances, possible differ-
 ences in carcinogenic potency of different
 asbestos minerals,  and risk estimates  at
 low, non-occupational exposures.

 Summary
    Lung cancer and mesothelioma are the
 most important asbestos-related causes of
 death among  exposed individuals. Gas-
 trointestinal  cancers are also increased in
 most studies of  occupationally  exposed
 workers. Cancer at other sites (larynx, kid-
 ney,  ovary)  has also been shown to be
 associated with asbestos exposure in some
 studies, but  the degree of excess risk and
 the strength of the association are less for
 these and the gastrointestinal cancers than
 for lung cancer or mesothelioma. The IARC
 lists asbestos as a group  1  carcinogen,
 meaning that exposure to asbestos is car-
 cinogenic to humans.  EPA's proposed
 guidelines categorize asbestos as Group
 A, human carcinogen.
    Data  from a study  of U.S. insulation
 workers allow models to be developed for
 the time and age dependence of lung can-
 cer and mesothelioma risk. Thirteen other
 studies provide exposure-response infor-
 mation. The  accumulated data suggest that
 the excess risk  of death from lung cancer
 from asbestos exposure is proportional  to
 the cumulative exposure (the duration times
 the intensity) and the underlying risk in the
 absence of exposure. The time course  of
 lung cancer  is determined primarily by the
 time course of  the underlying risk. How-
 ever,  the risk of death from mesothelioma
 increases very  rapidly  after the onset  of
 exposure and is independent of age  and
 cigarette smoking. As  with lung cancer,
 the risk appears to be proportional to the
 cumulative exposure to asbestos in a given
 period. The dose and time relationships for
 other asbestos cancers are uncertain.
   Values characterizing lung cancer risk
obtained from different studies vary widely.
 Some of the  variability can be attributed to
 specific processes. Chrysotile mining  and
 milling, and perhaps friction product manu-
facture, appear to have lower unit expo-
sure risks than chrysotile textile production
and other uses of asbestos. Other variability
can be associated with the uncertainties of
small  numbers in epidemiological  studies
and improper estimates of  exposures  in
 earlier years. Some differences between
 studies may be related to differences in
 fiber type, but these are much less than
 those associated with specific processes.
    Four studies provide similar quantita-
 tive data on  the  unit exposure risk for
 mesothelioma and six additional studies
 provide corroborative,  but  less accurate,
 quantitative data.  The same factors that
 affect the lung cancer unit exposure risk
 appear to affect that of mesothelioma, as
 the  ratio of a measure of mesothelioma
 risk to excess lung cancer risk is roughly
 constant across the ten studies. However,
 in other studies the  ratio of  number  of
 mesothelioma deaths to lung cancer deaths
 among groups  exposed to substantial
 quantities of crocidolite is two to four times
 higher than among groups exposed pre-
 dominantly to other fibers. Further, the risk
 of peritoneal mesothelioma appears to be
 less from exposure to chrysotile than  to
 either crocidolite or amosite, but this sug-
 gestion is tempered by uncertainties asso-
 ciated  with  the  greater  possibility  of
 misdiagnosis of the disease.
    Animal studies  confirm the human epi-
 demiological results. All major asbestos
 varieties produce lung  cancer and meso-
 thelioma with only limited  differences  in
 carcinogenic  potency. Implantation  and
 injection studies show that fiber dimen-
 sionality, not chemistry, is the most impor-
 tant factor in fiber-induced carcinogenicity.
 Long (>4 urn) and  thin (<1  urn) fibers are
 the most carcinogenic  at a cancer-induc-
 ible site. However,  the size dependence of
 the deposition and migration of fibers also
 affects their carcinogenic action in humans.
    Measurements  demonstrate that as-
 bestos exposures exceeding 100 times the
 background occur  to individuals in some
 non-occupational settings.  Currently, the
 most important of these non-occupational
 exposures is  from the  release  of fibers
 from asbestos-containing surfacing build-
 ing materials, or from sprayed  asbestos-
 containing fireproof ing in high-rise buildings.
 A high potential exists for future exposure
 from maintenance, repair, and removal of
 these materials.
   Extrapolations of risks of asbestos can-
 cers from occupational circumstances can
 be made, although  numerical estimates in
 a specific exposure circumstance have a
 large (approximately tenfold) uncertainty.
 Because of this uncertainty, calculations of
 unit  risk values  for asbestos at the low
 concentrations measured in the environ-
 ment must be viewed  with caution. The
 best estimate of risk to the United States
 general population for  a  lifetime continu-
 ous exposure to 0.0001 f/ml is 2.8 meso-
thelioma deaths and 0.5 excess lung cancer
deaths per 100,000 females. Correspond-
ing numbers  for males are 1.9 mesothe-
lioma deaths  and 1.7 excess lung cancer
deaths per 100,000 individuals. Excess Gl
cancer  mortality is approximately 10-30
percent that of excess lung cancer mortality.
These risks are subjective, to some extent,
and  are also subject to variability in the
exposure-response relationship at high
exposures, uncertainty  in extrapolating to
exposures  1/100 as much, and uncertain-
ties in conversion of optical fiber counts to
electron microscopic fiber counts or mass
determinations.
   Recently, several government agencies
in different countries reviewed asbestos
health  effects. Areas of  agreement and
disagreement between various reviews are
presented in the full document. A compari-
son of the  different risk estimates is also
provided.
                                                                               &U. S. GOVERNMENT PRINTING OFFICE: 199 1/548-028/20206

-------
 W.J. Nicholson is with Mt. Sinai School of Medicine, New York, NY 10029.
 D. Kotchmar is the EPA Project Officer, (see below).
 The complete report, entitled "Airborne Asbestos Health Assessment Update," (Order
   No. PB86-242864/AS; Cost: $31.00, subject to change) will be available only from:
         National Technical Information Service
         5285 Port Royal Road
         Springfield,  VA 22161
         Telephone:  703-487-4650
 The EPA Project Officer can be contacted at:
         Environmental Criteria and Assessment Office
         U.S. Environmental Protection Agency
         Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati. OH 45268
      BULK RATE
POSTAGE & FEES PAID
 EPA PERMIT NO. G-35
Official Business
Penalty for Private Use $300
EPA/600/S8-84/003F

-------