|
Items in cart [1] |
Questions? Call 888-624-8373 |
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 123
Adverse Health Effects of Arsenic and Asbestos 123
tors that may account for the observed differences in clinical studies as well
as the effects of diet, race, and climate.
Research should also be designed to evaluate the possible essentiality of
arsenic for humans a requirement that has been demonstrated in four
mammalian species. In the absence of new data, the conclusion reached in
the third volume of Drinking Water and Health remains valid, i.e., "If
0.05 mg/kg of dietary [total] arsenic is also a nutritionally desirable levee
for people, then the adequate human diet should provide a daily intake of
approximately 25 to 50 fig. The current American diet does not meet this
presumed requirement" (National Research Council, 1980~. The unre-
solved status of this issue is further reason for maintaining the current
MCL for arsenic.
ASBESTOS
Asbestos fibers in drinking water and their putative health effects were
reviewed in the first volume of Drinking Water arid Health (National Re-
search Council, 1977, pp. 144-168~. At that time, there were only limited
data from which to evaluate the potential adverse health effects of orally
ingested asbestos. A number of research recommendations suggested in
that volume have been, to some extent, fulfilled. Advances have been made
in the detection, identification, and quantification of asbestos fibers in
drinking water. Several chronic feeding studies completed since that time
have failed to show an effect between the ingestion of various fiber types
and the development of cancer at any site.
There have also been a number of epidemiological studies in which the
exposure to asbestos in drinking water and the incidence of cancer at se-
lected sites have been investigated. This renew is limited to a discussion
and evaluation of those studies and the development of a model to predict
the risks, if any, from such exposure.
BACKGROUND
A marked increase in the incidence rates of lung cancer and both pleural
and peritoneal mesothelioma has been absented in workers exposed to as-
bestos through inhalation (International Agency for Research on Cancer,
1977~. An excess of gastrointestinal tract cancers has also been.found in
these occupationally exposed groups.
The general population may be exposed to asbestos fibers in "air, be`,er-
ages, drinking water, food and pha~..aceutical and dental preparations
and by consumer use of asbestos containing products" (International
OCR for page 124
124 DRINKING WATER AND HEALTH
Agency for Research on Cancer, 1977~. The possible effects of such expo-
sure from drinking water became a matter of some public concern when
the Duluth, Minnesota, water supply obtained from Lake Superior was
found to be heavily contaminated with asbestos fibers.
The potential hazard from asbestos in drinking water was considered by
the American Water Works Association Research Foundation (1974~.
However, the report of this group refers only to asbestos fibers released
into the water from asbestos-cement pipe. It concluded, "Calculations
comparing the probable ingestion exposure in occupational groups to that
likely to occur as a result of ingestion of potable water from asbestos-
cement pipe systems suggest that the probability of risk to health from the
use of such systems is small approaching zero." This conclusion has been
questioned (McCabe and Millette, 1979; U.S. Environmental Protection
Agency, 1979~.
EXPOSURE TO ASBESTOS FROM DRINKING WATER
The results of an extensive EPA survey of asbestos concentrations in drink-
ing water have recently been published by Millette et al. (1979~. A sum-
mary of the data is given in Table III-1. More than 20% of the cities sur-
veyed had water containing more than 1 million fibers per liter, as
measured by transmission electron microscopy (TEM), and almost 11% of
them had water containing more than 10 million fibers per liter. (This EPA
survey was not a representative sampling of U.S. water supplies; therefore.
caution should be exercised when drawing conclusions from the table.)
The asbestos in these supplies was derived from a variety of sources: min-
TABLE III-1 Asbestos Concentrations in Drinking Water from 365
Cities in 43 States, Puerto Rico, and the District of Columbia, as
Measured by Transmission Electron Microscopya
Asbestos Concentration, Number of Percentage of
106 fibers/liter Cities Samples
Below detectable limits 110 30.1
Not statistically significant 90 24.7
Less than 1 90 24.7
From 1 to 10 34 9.3
Greater than 10 41 11.2
TOTAL 365 100.0
. . .
aData from Millette At al., 1979.
OCR for page 125
Adverse Health Effects ot Arsenic and Asbestos 125
ing process discharge into Lake Superior, natural erosion of serpentine
rock in the Bay Area of California and in Seattle, asbestos-cement roofs,
and asbestos-cement pipe. Some of the water supplies surveyed are now
being filtered to reduce the asbestos fiber levels.
Asbestos concentrations in water have also been reported as mass per
liter, but this measurement is not considered useful in the evaluation of
possible carcinogenic effects (see Harington, 1981~. Experiments in which
gelatin pellets of asbestos were implanted into the pleura of rats indicate
that the physical dimensions of the fibers, but not their type, are important
and that the long, thin fibers are the most effective inducers of mesothe-
liomas (Stanton et al., 1981~. The relevance of this work to the carcinoge-
nicity of asbestos in humans has been reviewed by Selikoff and Lee (1978)
and more recently in great detail by Harington (1981~. Harington con-
cluded that the "Stanton hypothesis," emphasizing the fiber dimensions
as important determinants of carcinogenicity, appeared to hold true for
the very limited, relevant data on humans. However, the critical fiber di-
mensions in humans, at least for mesothelioma, are probably much
smaller than those suggested in the rat experiments, the most carcinogenic
fibers apparently being those with diameters less than 0.05 Am and lengths
greater than 3 Am (Harington, 1981~.
The size distribution of asbestos fibers in water varies by source (Millette
et al., 1980~. The smallest fibers are found in water contaminated by the
natural erosion of serpentine rock. These fibers have an average width of
approximately 0.04 Am and an average length of 1 ~`m, compared to an
average width of 0.1 Am and an average length of 1 Am for fibers from the
Whitekloff asbestos mine, which was associated with a high incidence of
mesothelioma (Harington, 1981~. Thus, the average aspect ratio (length:
width) for the waterborne fibers from natural sources is approximately 3
times greater than that for the airborne fibers found in the Whitekloff
mine. Approximately loo of the waterborne fibers are longer than 3 ~m.
Water contaminated by asbestos from asbestos-cement pipe contained fi-
bers with an average diameter of 0.044 Am and an average length of 4.3
~m, which gives an average aspect ratio of 121 (Millette et al., 1980~. Ap-
proximately 30~o of these fibers were longer than 3 ~m. The committee
concluded that there appears to be no reason to consider fibers from either
source as free from risk in comparison to fibers found in occupational set-
tings.
Harington (1981) very tentatively concluded that fiber dimensions "may
in time be applicable to regulatory practice," essentially agreeing with Se-
likoff and Lee (1978) that "there seems to be little basis . . . at present . . .
for seeking to base selective control measures on such hypotheses," i.e.,
fiber dimensions. The committee concurs with these authors.
OCR for page 126
126 DRINKING WATER AND HEALTH
ESTIMATING THE CANCER RISK FROM SWALLOWED
ASBESTOS FOLLOWING OCCUPATIONAL EXPOSURE
There have been a number of epidemiological studies of gastrointestinal
(GI) cancer (in this report this term refers only to cancer of the esophagus,
stomach, small intestine, colon, and rectum) associated with occupational
exposure to asbestos. The reported relative risks (RR's) of observed to ex-
pected cases in these studies are not very large (between 1.5 and 3), and a
number of other studies have failed to detect any risk of GI cancer (Advi-
sory Committee on Asbestos, 1979~. Thus, one must consider the possibil-
ity that the studies with positive results might have been affected by unre-
corded biases. However, careful review of these studies by many authors
and by a Working Group of the International Agency for Research on Can-
cer concluded that the association was one of cause and effect (Advisory
Committee on Asbestos, 1979; International Agency for Research on Can-
cer, 1977; Miller, 19783. The committee concurs with these reviews.
Assuming that the exposed and unexposed workers have the same gen-
eral risk factors for GI cancers and that their observed GI cancer rates are
re and ru, respectively, then the GI cancer burden from the exposure can be
expressed either as a ratio of rates, or relative risk, i.e., RR = rears, or a
difference of rates, i.e., DR = re—rU. For general risk assessment pur-
poses, these can be expressed on a per unit exposure basis by dividing RR
or DR by the exposure dose.
Both RR and DR are valid measures of the risk to the occupational
group, but they implicitly make very different assumptions about the risks
to individuals with different risk factors. The relative risk (or multiplica-
tive) index (RR) implicitly assumes that the risk of GI cancer is increased
in proportion to the individual's underlying risk. The difference of risk (or
additive) index (DR) implicitly assumes that the amount of increased risk
of GI cancer is independent of the individual's underlying GI cancer risk.
None of the occupational studies of exposure to asbestos and Gl cancer
provided data that would enable the committee to distinguish between
these possible models (i.e., the multiplicative or additive models, or some-
thing intermediate). In fact, because of the limited data and lack of any
known strong risk factors for GI cancer, except for increasing age, it is
difficult to know how the studies could shed light on this issue. However,
selection of a model is of critical importance in risk assessment and cannot
be avoided. In the absence of evidence to the contrary, this committee has
usually selected the additive model. For asbestos, however, some informa-
tion suggests that the multiplicative model is to be preferred. The data
relating lung cancer risk to joint exposures to asbestos and cigarettes are
inadequately described by the additive model, whereas the multiplicative
OCR for page 127
Adverse Health Effects of Arsenic and Asbestos 127
model is generally regarded as providing a reasonable description of the
data (see Saracci, 19771. The results of Seidman et al. (1979), who studied
lung cancer in persons exposed for a limited time in an amosite asbestos
factory in New Jersey, indicate that the excess lung cancer increased with
age at exposure. Therefore, the additive model is clearly inadequate, but
the data are again reasonably compatible with a multiplicative model. Un-
fortunately, the observations by Seidman et al. (1979) may have been se-
verely confounded by possible age-related differences in cigarette-smoking
habits, but a very similar pattern of risk was also observed for other "as-
bestos disease," i.e., asbestosis and other noninfectious pulmonary dis-
eases, mesotheliomas, and cancers of the esophagus, stomach, colorec-
tum, larynx, buccal cavity, pharynx, and kidney. The relative risks
associated with exposure to asbestos are similar for all subsites in the GI
tract (Miller, 1978; National Research Council, 1977~. Although the above
information may not constitute proof for the correctness of the multiplica-
tive model, there is clearly no basis for favoring an alternative model for
risk assessment.
Studies of Asbestos Workers
Table III-2 shows the results of five cohort studies of GI cancers in asbestos
workers. The reports of Newhouse and Berry (1979) and Henderson and
Enterline (1979) only give data for GI and certain other sites combined (see
footnote to Table III-2. The committee adjusted the RR's to refer only to
GI cancer by assuming that all excess deaths occurred in this grouping.
To make the observed RR's of use in estimating the possible conse-
quences of ingesting asbestos fibers in drinking water, it is essential to have
some measure of the amount of asbestos swallowed by these workers. Such
estimates can only be approximate because so few measurements of air-
borne asbestos have been made in the workplace (and then, often years
after actual exposure) and because those measurements then had to be
applied to broad categories of employees (Peso, 1979~. Table III-3 shows
the estimates derived either from the published studies or from personal
communication with the authors.
CONVERTING RISE TO ASBESTOS WORKERS TO RISK FROM
SWALLOWED ASBE STO S
In converting the observed risk of GI cancers in asbestos workers to risk of
Gl cancer from ingested asbestos fibers, a number of steps should be
clearly distinguished.
OCR for page 128
128 DRINKING WATER AND HEALTH
TABLE III-2 Results of Five Cohort Studies of Gastrointestinal Tract
Cancer in Asbestos Workers
Deaths
Exposed
Group
Cancer
Site Observed
Codesa (O)
Expected Ratio,
(E) O/E RRh
Reference
U.S. and
Canadian
insulation
workers
N.Y. and N.~. 150-154 43
insulation
workers
U.S. factory 150- 154 32
workers
London 150-158C 40
factory workers
U.S. factory 150- 159c 55
workers
150-154 94 59.4 1.58 1.58
15.1 2.85 2.8S
Selikoff et al.,
1979
Selikoff et al..
1979
21.5 1.49 1.49 Seidman et al.,
1979; U.S.
EPA, 1979
Newhouse and
Berry, 1979
Henderson and
Enterline,
1979
34.0 1.18 1.32
39.9 1.38 1.55
150 = esophagus
151 = stomach
152 = small intestine
153 = large intestine
154 = rectum
155 = liver
156 = gallbladder
157 = pancreas
158 = retropentoneum and peritoneum
IS9 = gastrointestinal ban, not otherwise specified
Standardized mortality ratio = 100 X RR (relative risk) for cancer site codes lSO-154. (See text.)
E:xcluding mesotheliomas.
Step 1: Measurement of Dose and Adoption of Risk Model
Since the committee is interested in effects at low doses, a model relating
relative risk to dose must be used. The standard linear dose-effect model
may be written:
RR= 1 +a X dose,
(1)
where a is a constant to be estimated. For asbestos workers, dose is a func-
tion of both intensity and duration of exposure. Intensity of exposure is
measured in terms of the number of fibers that can be seen with the light
microscope (LM) per milliliter of air. Dose is the simple product of inten-
OCR for page 129
Adverse Health Effects of Arsenic and Asbestos 129
TABLE III-3 Average Asbestos Exposure Estimated for the Workers
in Studies Shown in Table III-2 and Calculation of Increment in
Relative Risk Based on Doses Measured with the Light Microscope
Exposure
Cumulative
intensity Dose Relative Risk
(LM (LM per Unit
fibers/ml Duration fibers/ml Cumulative
RR air) (years) air)(years) Dose ° Reference
1.58 15 34 510 0.00114 Selikoffetal., 1979
2.85 15 40 600 0.00308 Selikoffetal., 1979
1.4 40 1.9 76 0.00645 Seidman e' al.,
1979; U.S. EPA,
1979
1.32 10-30 — 170b 0.00188 Newhouse and Berry,
1979
1.55 — — 49Bc 0.~110 Henderson and
Enterline,
1979
Increase in RR for each year of exposure to one fiber identified by the light microscope (parameter o in
equations 1 and 2).
Calculated from Newhouse and Berry (1979) using the method from U.S. Environmental Protection
Agency, 1979.
'Calculated from Table 2 of Henderson and Enterline (1979), assuming 1 mppef (million particles per
cubic foot) = 2 fibersJml.
sity of exposure and duration of such exposure, usually measured in years
(Y), and expressed as numbers of LM fibers/ml air times Y. It is not self-
endent that equal doses measured in this way must have equal effects
(e.g., cigarette smoking measured as pack-years does not have a fixed ef-
fect on lung cancer incidence but is greater at a low intensity for a long
time); however, authors of all studies on asbestos-induced cancer concur
that measurement of dose on this scale fits the data reasonably well and no
alternative model has been seriously proposed. In particular' there is evi-
dence that even the briefest exposures (less than 6 months) have very long-
term effects on lung cancer rates with no diminution of the associated RR
with the passage of time (up to 35 years) after exposure (Seidman et al.,
1979~. A linear dose-effect relationship for lung cancer and exposure to
asbestos is most clearly shown by results reported by Henderson and En-
terline (1979~. For GI cancers, too few data have been published to estab-
lish or refute linearity, even at high doses. Peto (1979) has discussed these
issues at some length.
OCR for page 130
i30 DRINKING WATER AND HEALTH
Assuming the dose-effect model given by equation (1), one may estimate
the value of parameter a as follows:
a = (RR—1/dose.
(2)
The calculated values of a from the different studies are given in the Table
III-3.
In the dose-effect model given by equations (1) and (2), dose is used to
mean the cumulative dose calculated up to the age at which the cancer is
diagnosed. This may be considered as the correct dose for a no-latent-
period model, and as having general applicability if the dose was received
over a brief period some time before the associated cancer was recorded.
This would result in an overestimation of the dose to workers still occupa-
tionally exposed to asbestos when their cancer rates were being observed.
This overestimation will generally be small, and as long as we adopt the
same convention when calculating the risk from asbestos in the ~ ater sup-
ply, the error will have a very small effect on the computation of risk.
Step 2: Conversion of Dose of Asbestos Inhaled to Dose of Asbestos
Swallowed
Since the excess GI cancers in the workers are assumed to be caused by the
asbestos fibers that these workers swallowed rather than simply inhaled,
the dose calculated in Step 1 must be converted to fibers swallowed. The
committee estimated that
breathing 1 LM fiber/ml for 1 year
= S88 X 106 LM fibers swallowed, (3)
where S88 X 106 is the product of 106 (ml in m3), times 8 (m3 of air
breathed per day at work), times 5 (days worked per week), times 49 (num-
ber of weeks worked per year), times 0.3 (proportion of inhaled fibers that
are subsequently swallowed). Only this last factor of 0.3 needs discussion.
Studies of short-term (30-minute) inhalation exposures of rats to Union
Internationale Contre le Cancer (UICC) standard reference samples of as-
bestos indicated that an average of 40~o of the various types and sizes of
inhaled material is deposited somewhere in the respiratory tract (Morgan
et al., 1975~. Although there is a lack of relevant data on the deposition of
asbestos fibers in humans, Dement (1979) used a mathematical model of
fiber behavior to calculate that roughly 28~o of inhaled chrysotile fibers is
deposited in humans and that approximately twice as much amosite is de-
OCR for page 131
Adverse Health Effects of Arsenic and Asbestos 131
posited. Morgan et al. (1975) did not find such a large difference in their
study on rats, although they did find that the deposition of one form of
chrysotile was 25~o less than that of amosite. Evans and his colleagues
(1973) reported that approximately 65'7o of the material deposited in rats is
cleared through the GI tract within a month. Their results suggest that this
clearance process continues until almost all the deposited material is
cleared. Relevant data on clearance in humans are again lacking, but the
work of Cohen et al. (1979) on the clearance of ferrosoferric oxide (Pe3O4)
dust from experimentally exposed men suggests that lung clearance in hu-
mans is similar to that in rats and that most of the deposited dust will be
cleared in humans. For rats, then, approximately 40~0 of inhaled asbestos
will enter the Gl tract and a somewhat lower figure of 30% appears to be a
reasonable estimate for humans. Note: no allow ance has been made for the
possibility that asbestos is ingested directly. Neglect of this could result in
an overestimation of the effect of a unit dose of swallowed asbestos.
Step 3: Conversion of Number of Fibers Seen by Light Microscope to
Number of Fibers Seen by Transmission Electron Microscope
Asbestos contamination of drinking water is measured in terms of number
of fibers seen with the transmission electron microscope (TEM). To con-
vert from light microscope (LM) measurements to TEM measurements,
the committee has used the following equation:
1 LM fiber = SO TEM fibers.
(.4)
This equation is based on the report of Lynch et al. (1970), who found
that a conversion factor of 50 is roughly appropriate for asbestos exposure
from textile manufacturing, friction work (i.e., mixing, grinding, cutting,
and drilling), and pipe manufacturing. Conversion factors larger than this
have been reported in the literature, e.g., McCabe and Millette (1979)
used 100 and the U.S. Environmental Protection Agency (1979) used 200
in the criteria document, but these estimates do not appear to be based on
the industrial exposure data considered in this chapter.
The relative risk equation (1), which applies to measurements made by
the LM, may thus be expressed as follows for doses swallowed, as mea-
sured by the TEM:
RR = 1 + [a/~588 X 106 X 50~] X dose (in TEM fibers swallowed)
= 1 + (a/0.0294) x dose (in TEM fibers swallowed/10 (S)
= 1 ~ b X dose (in TEM fibers swallowed/10~2~.
OCR for page 132
132 DRINKING WATER AND HEALTH
Estimated values of b derived from the studies listed in Table III-3 are
given in Table III-4. These values vary from 0.039 to 0.22; a "best" value
(obtained by weighting the individual estimates of b inversely proportional
to their estimated variance) is approximately 0.05.
Thus, the RR for GI cancer for a person who has swallowed h X 10~2
TEM fibers can be estimated as follows:
RR= 1 +0.05 Xh.
PREDICTING THE RESULTS OF EPIDEMIOLOGICAL STUDIES
OF GI CANCER RISK FROM ASBESTOS-CONTAhIINATED
DRINKING WATER
(6)
Equation (6) makes direct estimates of relative risks observable in epide-
miological studies correlating exposure to asbestos in drinking water and
GI cancer mortality rates. It is also reasonable to assume that the relative
risks derived from this equation will be approximately correct if applied to
studies of cancer incidence.
A man who has been drinking water containing d X 106 TEM fibers/
liter for n years has consumed h X 10~2 TEM fibers, where
h=n X 365.25 X 2 Xd X10-6
.
His relative risk of GI cancer in this nth year of exposure is:
RR = 1 + O.OS X h
= 1 +3.6525 X 10-5 Xn xd.
(6)
(7)
For example, if n = 20 years and d = 15 X 106 TEM fibers/liter, then the
associated relative risk is:
RR = 1 + 3.6525 X 10-5 X 20 X 15
= 1.011.
Equation (7) is directly applicable to epidemiological studies in which
contamination of drinking water took place for only a limited time, e.g., in
Duluth. To make equation (7) applicable to epidemiological studies in
areas where the contamination has been present for a very long time, the
different durations of exposure of different age segments of the population
first need to be evaluated. The associated relative risks must then be calcu-
lated from equation (7), and then some "average" determined. The exact
OCR for page 133
Adverse Health Effects of Arsenic and Asbestos 133
TABLE III-4 Estimates of Increment in Relative Risk of
Gastrointestinal Cancers per Unit Exposure
a, Increment b, increment
per LM Fibers/ml per 10~2 TEM
Air Times Years Fibers Swallowed Reference
0.00114 0.0388 Selikoffet all, 1979
0.00308 0.1048 Selikoffetal., 1979
0.00645 0.2194 Seidman et al., 1979;
U.S. Environmental
Protection Agency, 1979
0.00188 0.0640 Newhouse and Berry, 1979
0.00110 0.0374 Henderson and Enterline,
1979
fonn of this average would depend on the type of statistic used to describe
the overall relative risk of GI cancer in the exposed community.
CONVERTING RR S TO LIFETIME GI CANCER RISKS
As discussed above, expressing the risk from any agent in terms of relative
risks implies that the effect of the agent is to multiply whatever "normal"
or background cancer rate exists. Table III-5 shows the results of calcula-
tions based on equation (6), indicating the consequences for U.S. white
males who drink 2 liters of water containing d X 106 TEM fibers/liter
daily throughout life. To understand this table, consider a man aged 57.5
years. He has consumed water containing 57.5 X 365.25 X 2 X d X 106
TEM fibers, or 42.0 X 109 TEM fibers up to this point in his life, and his
relative risk of GI cancer is currently:
RR = 1 + 0.05 X 0.0420 X d
= 1 ~ 0.0021 X d.
The additional RR is thus 0.0021d, and the additional probability of GI
cancer occurrence in this, his 58th year is:
0.0021d X 131.1/105 = 0.2754d/105,
where 131.1/lOs is the 1970 GI cancer incidence rate in the United States
for white males in the 55- to 59-year age group (Cutler and Young, 1975).
Summing these additional probabilities of GI cancer deaths up to age 70
gives us a lifetime risk of 9.1060d/lOs.
OCR for page 138
138 DRINKING WATER AND HEALTH
asbestos concentrations of 206 X 106 TEM fibers/liter were found in grab
samples of tap water obtained from this source.
Polissar and his colleagues compared the cancer incidence data for 1974
to 1977 and the cancer mortality data for 1955 to 1975 from the census
tracts with a Sultan River water source to such data from census tracts in
the Puget Sound area with average asbestos concentrations in tap water of
7 X 106 TEM fibers/liter. Table III-7 shows their results for GI cancers.
No effect of asbestos concentration is evident. However, a proportional in-
cidence analysis of GI cancer in which long-term ~ >30 years) residents
were compared with short-term (<30 years) residents of Everett the
main city served by Sultan River water—did show proportional incidence
ratios of 1.49 and 1.39 for males and females, respectively.
It is evident from the "Average Annual Population at Risk" rows in Ta-
bles 3 and 4 in Polissar et al. (1982) that there ulnas tremendous population
growth in the census tracts served by the Sultan River between 1955 and
1977. The average annual total population from 1955 to 1975 was 45,272,
and from 1974 to 1977 it was lS6,099, i.e., the 1975 population appears to
have been roughly 3.5 times the size of the 1965 population. In these Sul-
tan River water tracts, the average duration of residence for persons in the
cancer age range may therefore be as short as 5 years. Using this figure,
equation (1) predicts a relative risk of 1.04.
If we assume that the cases with long-term (more than 30 years) resi-
dence in Everett had been there for 60 years and that cases with short-term
residence had been there for 5 years, then equation (7) predicts a relative
risk of 1.41, which is in close agreement with the observed proportional
incidence ratios.
TABLE III-7 Odds Ratios for Gastrointestinal Cancer incidence
(1974-1977) and Mortality (1955-1975) for Males and Females in Census
Tracts with a Sultan River Water Sourcea
Type of Odds
Sex Data Ratiob
Male Incidence 1.10
Mortality 0.79
Female Incidence 0.97
Mortality 0.86
aFrom Polissar e' al., 1982.
bCalculated from Tables 3, 4, and 7 in Polissar e' al. . 1982; odds ratio for census tracts with a Sultan River
water source versus census tracts in the Seattle-Everett-Tacoma metropolitan areas, which have the Cedar
River, Tolt River, Green River, or Lakewood Wells as their water sources.
OCR for page 139
Adverse Health Effects of Arsenic and Asbestos 139
Within the inherent limitations of the study design, therefore, the
results of this study are compatible with the predictions of equation (7~.
EPA CRITERIA DOCUMENT METHODOLOGY
In the Ambient Water Quality Criteria Document, the U.S. Environmen-
tal Protection Agency (1979) considers much the same data discussed in
this chapter, but uses a very different method of calculating the "addi-
tional lifetime cancer risk of 1 in 100,000." The agency's method of calcu-
lation can be analyzed by dividing it into a number of distinct steps:
Step 1: Cor~versior' of LM Fibers/ml Air to TEM Fibers Swallowed
(This is the equivalent of Steps 3 and 4 in the committee's method.) A man
breathing 1 LM fibers/ml air at work is considered by the EPA to swallow
an average of
200 X 8 X 106 X s/7 = 1,142.9 X 106 TEM fibers/day,
where 1 LM fiber = 200 TEM fibers, 8 = m3 of air breathed per day at
work, 106 = ml in m3, and s/7 = proportion of working days in a week.
(The committee used a conversion factor of SO for LM to TEM, a factor of
49/s2 to represent working weeks in a year, and 0.3 as the proportion of
inhaled asbestos fibers that are subsequently swallowed. As a result, the
risks calculated by the committee are only 7.07~0 of those calculated by the
EPA.)
Step 2: Conversion of LM Fibers/ml Air to TEM Fibers/Liter in
Drinking Water for a 70- Year Exposure
1 LM fiber/ml air at work for 1 year
1,142.9 X 106 TEM fibers swallowed daily for 1 year
(1~142.9 X 106~/2 TEM fibers/liter drinking water for 1 year
571.4 X 106 TEM fibers/liter for 1 year
(571.4 X 1061/70 TEM fibers/liter for a lifetime of 70 years
8.163 X 106 TEM fibers/liter lifetime exposure.
(The committee's method agrees with the principles of this calculation. )
For example, the 510 LM fibers/ml air-year result of Selikoff and his col-
leagues (1979) given in the first row of their Table 3 becomes the equivalent
OCR for page 140
140 DRINKING WATER AND HEALTH
of 510 X 8.163 X 106 = 4,163.6 X 106 TEM fibers/liter in drinking water
for 70 years.
Step 3: Conversion of TEM Fibers/Liter for 70 years to Additional
Lifetime Cancer Risk per 100, 000
The EPA assumes that the number of TEM fibers/liter over 70 years calcu-
lated in Steps 1 and 2 will have the same effect as the original occupational
exposure on GI cancer and peritoneal mesothelioma. In the criteria docu-
ment, this effect (X) is considered to be:
X = (number of excess deaths from GI cancer or peritoneal me-
sothelioma)/(total expected number of deaths from all
causes),
where the expected numbers are calculated in the absence of asbestos ex-
posure. The EPA then equates X with an additional lifetime cancer risk of
X. The required TEM fibers/liter for an additional lifetime cancer risk of 1
in 100,000 is then obtained by simple proportion, i.e.:
X/(calculated TEM fibers/liter over 70 years)
=~1 X 10-51/(required TEM fibers/liter over 70 years).
Thus,
required TEM fibers/liter
= (calculated TEM fibers/liter over 70 years/105 X X).
For example, the 510 LM fibers/ml air-year result of Selikoff et al. (1979)
given in the first row of their Table 3 was associated with 148.9 excess
deaths from GI cancer (39.9) or peritoneal mesothelioma (1091. As given in
Table 30 of the criteria document, the total number of expected deaths
from all causes was 1,660.96,
X= 148.9/1,660.96
= 0.0896
and
required TEM fibers/liter over 70 years
= (4,163.6 X 106~/~105 X 0.0896)
= 0.536 X 106 TEM fibers/liter.
OCR for page 141
Adverse Health Effects of Arsenic and Asbestos 141
(Note: if the conversion factors in Step 1 were replaced by the committee's
estimates, then the last figure would need to be multiplied by 7.07%, giv-
ing required water quality of 0.536 X 106 X 7.07~0 = 0.038 X 106 TEM
fibers/liter.)
The critical assumption in Step 3 is equating excess deaths as a propor-
tion of all deaths with additional lifetime cancer risk. This assumption is
demonstrably false. Suppose that all causes of death except GI cancer and
peritoneal mesothelioma were somehow eliminated from the population.
By definition, this should make no difference to the calculation of lifetime
risk from GI cancer or peritoneal mesothelioma from asbestos exposure,
since these sites were not altered. But X will be drastically changed. In the
study by Selikoff et al. (1979y, the total expected number of deaths from all
causes is reduced from 1,660.96 to 59. 1, so that X changes from 0.0896 to
2.519, or a 28.1-fold increase. The required TEM fibers/liter is reduced
from 0.536 X 106 to 0.019 X 106.
ANIMAL EXPERIMENTS
When there are few or no epidemiological data on which to base estimates
of a compound's carcinogenicity in humans, one recourse is the use of data
from experiments in animals. Although this is not the case for asbestos,
confirmatory evidence that ingested asbestos is a GI tract carcinogen in
animals would add some weight to the epidemiological data.
Asbestos has been shown to cause mesotheliomas in rats and hamsters
when implanted into the pleural cavity and when inhaled (Pott, 1980;
Wagner et al., 1980), but results of long-term asbestos ingestion studies in
rats and hamsters have not produced any convincing evidence of an in-
crease in GI tract tumors in either species. Table III-8 gives details of four
feeding studies with substantial numbers of animals that provided little, if
any, evidence of an effect on GI tract tumors.
Equation (6) suggests that the relative risk of GI cancers from asbestos
exposure of humans may be written as:
RR = 1 + 0.05 X h,
where h X 10~2 is the number of TEM fibers swallowed. If we assume that
the daily dose to an animal over a lifetime had an effect equivalent to that
in humans exposed for 70 years, then the "Maximum Daily Dose" column
of Table III-8 may be multiplied by 1.28 X 10-9 = 0.05 X 70 X 36S.25 X
10-~2 to give the expected excess relative risk (RR—1) as shown in the
table. On this basis only the experiments with chrysotile should have defin-
itely produced a positive result, i.e., an increase in GI tumors.
OCR for page 142
142
~n
~_
~:
._
s"
x
oo
._
o
-
C~
~D
£
o~
o
_,
~:
C)
~:
0
-
~
m
-5:
E~
C)
~:
C~
CY:
~ Ct: ~
~5
C~
C.,~ · ~
~ ~C
X C) ~:
LL Ct _
-
_
~ V1 Ct
,t, ~ E ~
3 ._ ~ ~
~ ~ ~ LU
Lc a~
E 4,
3 C~ ~D
K ._
~ a ~
o ~o
00 —
-
— _
C
_ _
t3 ~
~, E ~
Ct
=: o
_ t_ ._
E o ~
. VO ~ Z
~n 0 0
o ~ o
— o
~ o~
.
E E
{C C~
~o oc
o o
C~
:^ :^
oo oc
o o
o o
~ C~
._ ._
X X
~ ~o
Ct
s:
o
._
-
e~
z
r~
r~ _
. .
o o ~ o
oo
~4
o o o o
X X X X
r~ ~D
oo
4)
_ _
.— o o ._
~ V,
o ~ ~ o
E ~ ~ E
C~
o~
o o o o
_ _ _ _
X X
X X
u) 0 o~ 0
I
_ r~
-
3
oo
~:
._
43V ~ _ _ _
~ ._ ~ 4.) ~ ~
00 ~ ~-
O ·— ._
_
X ~
a,
~ CO CS:
OCR for page 143
Adverse Health Effects of Arsenic and Asbestos 143
The usual method of converting experimental results in animals to hu-
mans is based on mass of the test compound per kilogram of body weight
or per square meter of body surface area. If we use the body weight correc-
tion, then the equivalent asbestos doses to humans are, of course, greatly
increased (see column "Equivalent Daily Dose for Humans"), and the ex-
pected excess relative risks are more than 100 for all the experiments.
The reasons for the negative results in animals are unknown. It may be
that the methods used to extrapolate data from humans to animals are
totally wrong for physical carcinogens such as asbestos.
SUMMARY AND CONCLUSIONS
An excess of GI tract cancers has been observed in some, but not all, occu-
pational groups exposed to asbestos. The reasons for the inconsistent
results have not been established, but careful reviews of all the studies indi-
cate that the results of the positive studies are real and that the demon-
strated association of asbestos exposure and GI tract cancer is most likely
one of cause and effect.
For asbestos this committee considers a multiplicative model for carcin-
ogenic risk assessment to be the most defensible. This model assumes that
the risk of GI tract cancer in persons exposed to asbestos at some given
level will be some fixed multiple of their normal or background GI tract
cancer rate. Many data on the relationship between asbestos exposure and
cancer suggest that this multiplying factor, or relative risk (RR), may be
accurately represented as a linear function of asbestos dose where asbestos
dose is measured in terms of cumulative fiber exposure.
Data from a number of occupational studies suggest that the RR for GI
tract cancer is 1 + 0.05 X dose X 10-~2, where dose is the number of
fibers swallowed, as counted by TEM. Using this equation, and assuming
a daily consumption of 2 liters of water, the committee calculated that
drinking water containing 0.11 X 106 TEM fibers/liter may lead to one
additional GI tract cancer per 100,000 men exposed over a lifetime of 70
years. For women, drinking water containing 0.17 X 106 TEM fibers/liter
may lead to one additional GI tract cancer per 100,000 exposed over a
lifetime of 70 years.
Peritoneal mesotheliomas are associated with occupational exposure to
asbestos, and indeed asbestos may be virtually the sole cause of this tumor.
If peritoneal mesotheliomas are caused by asbestos fibers migrating from
the gastrointestinal tract, which must be considered a definite possibility,
then risk of such tumors needs to be included in evaluating the total cancer
risk from swallowed asbestos. In the five studies shown in Table III-2, ap-
proximately 165 peritoneal mesotheliomas were observed; this compares to
OCR for page 144
144 DRINKING WATER AND HEALTH
the excess of approximately 94 other cancers shown in the table. If these
tumors are simply included in the "Observed" column of Table III-2 and
the calculations redone, then the risk estimates will be approximately 2.75
times as great. A multiplicative model for risk estimation is, however, not
applicable in the absence of an underlying risk, so that for peritoneal me-
sotheliomas an additive model is required. The data from the studies of
asbestos workers does not allow one to construct such a model. For the
present, therefore, the estimates of risk given above should simply be con-
sidered as possible underestimates of the true risk.
The designs of the epidemiological studies of cancer rates in populations
exposed to asbestos in drinking water all have major deficiencies, but the
above estimates of risk are quite compatible with the results of these stud-
ies. Adequate animal feeding studies conducted to date have failed to con-
firm a cancer risk from ingested asbestos. Further work aimed at under-
standing the inconsistent results of GI cancer excess in occupationally
exposed groups is clearly warranted.
REFERENCES
Arsenic
Arguello, R.A., E.E. Tello, and D.D. Cenget. 1939. Cancer and regional endemic chronic
arsenicalism. Br. J. Dermatol. 51:548. abstr.
Borgono, J.M., P. Vicent, H. Venturino, and A. lnfante. 1977. Arsenic in the drinking water
of the city of Antofagasta: Epidemiological and clinical study before and after the installa-
tion of a treatment plant. Environ. Health Perspect. 19:103-105.
Geyer, L. 1898. Ueber die chronischen Hautveranderungen beim Arsenicismus und Betrach-
tungen ueber die Massenerkrankungen in Reichenstein in Schlesien. Arch. Dermatol.
Syphilol. 43:221-280.
Harrington, J.M., J.P. Middaugh, D.L. Morse, and J. Houseworth. 1978. A survey of a pop-
ulation exposed to high concentrations of arsenic in well water in Fairbanks, Alaska. Am.
J. Epidemiol. 108:377-385.
International Agency for Research on Cancer. 1980. Some Metals and Metallic Compounds,
Vol. 23. lARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to
Humans. Lyon: International Agency for Research on Cancer. 438 pp.
Lu, F.-J. 1978. Study on fluorescent compounds in drinking water of endemic areas of Black-
foot disease and reinvestigation of the causes of Blackfoot disease. K'o Hsueh Fa Chan
Yueh K'an 6:388~03.
Morton, W., G. Starr, D. Pohl, J. Stoner, S. Wagner, and P. Weswig. 1976. Skin cancer and
water arsenic in Lane County, O=gon. Cancer 37:2523-2532.
National Research Council. 1977a. Drinking Water and Health. Report of the Safe Drinking
Water Committee, Board on Toxicology and Environmental Health Hazards, Assembly of
Life Sciences. Washington, D.C.: National Academy of Sciences. 939 pp.
National Research Council. 1977b. Arsenic. Medical and Biological Effects of Environmen-
tal Pollutants. Committee on Medical and Biological Effects of Environmental Pollutants.
OCR for page 145
Adverse Health Effects of Arsenic and Asbestos 145
Division of Medical Sciences, Assembly of Life Sciences, National Research Council, Na.
tional Academy of Sciences. Washington, D.C. 332 pp.
National Research Council. 1980. Drinking Water and Health, Vol. 3. Report of the Safe
Drinking Water Committee, Board on Toxicology and Environmental Health Hazards,
Assembly of Life Sciences. Washington, D.C.: National Academy of Sciences. 415 pp.
Schroeder, H.A., and J.J. Balassa. 1966. Abnormal trace metals in man: Arsenic. J. Chron.
Dis. 19:85-106.
Southwick, J.W., A.E. Western, M.M. Beck, T. Whitley, and R. Isaacs. 1981. Community
Health Associated with Arsenic in Drinking Water in Millard County, Utah. Final Report.
EPA 600/1-81-064. Cincinnati, Ohio: U.S. Environmental Protection Agency. Health Ef-
fects Research Laboratory. 74 pp.
Tseng, W.-P. 1977. Effects and dose-response relationships of skin cancer and blackfoot
disease with arsenic. Environ. Health Perspect. 19:109-119.
Tseng, W.-P., H.M. Chu. S.W. How, J.M. Fong. C.S. Lin, and S. Yeh. 1968. Prevalence of
skin cancer in an endemic area of chronic arsenicism in Taiwan. J. Natl. Cancer Inst.
40:453-463.
ITS. Environmental Protection Agency. 1976. Pp. 14-17 in Quality Criteria for Water.
Washington, D.C.: U.S. Environmental Protection Agency.
U.S. Environmental Protection Agency. 1977. National Interim Primary Drinking Water
Regulations. EPA-570/9-76-003. Washington, D.C.: U.S. Environmental Protection
Agency, Office of Water Supply. 159 pp.
U.S. Environmental Protection Agency. 1979. Arsenic. Ambient Water Quality Criteria\for
Arsenic. Washington, D.C.: Criteria and Standards Division, Office of Water Planning
and Standards, U.S. Environmental Protection Agency. 102 pp. Available from National
Technical Info'...ation Service, Springfield, Va., as Publ. No. PB-292-420.
U.S. Environmental Protection Agency. 1981. The Carcinogen Assessment Group's Final
Risk Assessment on Arsenic. EPA 600/6-81-002. Washington, D.C.: Environmental Pro-
tection Agency. 34 pp. Available from National Technical Information Service,
Springfield, Va., as Publ. No. PB-81-206013.
Zaldivar, R. 1974. Arsenic contamination of drinking water and foodstuffs causing endemic
chronic poisoning. Beitr. Pathol. Bd. 151:384-400.
Asbestos
Advisory Committee on Asbestos. 1979. Asbestos, Vols. 1 and 2. London: Her Majesty's Sta-
tione~ Office.
American Water Works Association Research Foundation. 1974. A study of the problem of
asbestos in water. J. Am. Water Works Assoc. 66:1-22.
Cohen, D., S.F. Arai, and J.D. Brain. 1979. Smoking impairs long-tenn dust clearance from
the lung. Science 204:514-517.
Cook, P.M., G.E. Glass, and J.H. Tucker. 1974. Asbestifonn amphibole minerals: Detec-
tion and measurement of high concentrations in municipal water supplies. Science
18S:853-855.
Cutler, S.J., and Y.L. Young, eds. 1975. Third National Cancer Survey: Incidence Data.
National Cancer institute Monograph 41. DHEW Publ. No. (NIH) 7S-781. Bethesda,
Md.: U.S. Department of Health, Education, and Welfare, Public Health Service, Na-
tional Institutes of Health, National Cancer Institute. 4S4 pp.
Dement, J.M. 1979. Estimates of Pulmonary and Gastrointestinal Deposition for Occupa-
OCR for page 146
146 DRINKING WATER AND HEAt TH
tional Fiber Exposures. Publ. No. 79-135. Washington, D.C.: U.S. Department of Health,
Education, and Welfare, National Institute for Occupational Safety and Health.
Donham, K.J., J.W. Berg, L.A. Will, and J.R. Leininger. 1980. The effects of long-term
ingestion of asbestos on the colon of F344 rats. Caneer 45:1073-1084.
Evans, ].C., R.J. Evans, A. Holmes, R.F. Hounam, D.M. Jones, A. Morgan, and M. Walsh.
1973. Studies on the deposition of inhaled fibrous material in the respiratory tract of the rat
and its subsequent clearance using radioactive tracer techniques. 1. U1CC Croeidolite as-
bestos. Environ. Res. 6:180-201.
Harington, J.S. 1981. Fiber eareinogenesis: Epidemiologic observations and the Stanton hy-
pothesis. 1. Natl. Caneer Inst. 67:977-989.
Harrington, J.M., G.F. Craun, and J.W. Meigs. 1978. An investigation of the use of asbestos
cement pipe for public water supply and the incidence of gastrointestinal cancer in Con-
neetieut. Am. J. Epidemiol. 107:96-103.
Henderson, V.L., and P.E. Enterline. 1979. Asbestos exposure: Factors associated with ex-
eess cancer and respiratory disease mortality. Ann. N.Y. Aead. Sei. 330:117-126.
international Ageney for Researeh on Caneer. 1977. MARC Monographs on the Carcinogenic
Risk of Chemicals to Man: Asbestos, Vol. 14. Lyon: international Ageney for Researeh on
Caneer. 106 pp.
Kanarek. M.S., P.Kl. Conforti. L.A. Jackson, R.C. Cooper, and J.C. Murehio. 1980. Asbes-
tos in drinking water and cancer incidence in the San Francisco Bay area. Am. J. Epide-
miol. 112:54-72.
Levy, B.S., E. Sigurdson, J. Mandel, E. Laudon, and J. Pearson. 1976. Investigating possible
effects of asbestos in city water: Surveillance of gastrointestinal cancer incidence in Du-
luth, Minnesota. Am. J. Epidemiol. 103:362-368.
Lynch, J.R., H.E. Ayer, and D.L. Johnson. 1970. The interrelationships of selected asbestos
exposure indices. Am. Ind. Hyg. Assoe. ]. 31:598-604.
Masson, T.J., F.W. McKay, and R.W. Miller. 1974. Asbestos-like fibers in Duluth water
supply: Relation to cancer mortality. 1. Am. Med. Assoe. 228:1019-1020.
McCabe, L.J., and I.R. Millette. 1979. Health effects and prevalence of asbestos fibers in
drinking water. Preprint, Proceedings of the American Water Works Association Annual
Conference, June 24-29, San Franeiseo, Calif. 12 pp.
Miller, A.B. 1978. Asbestos fibre dust and gastro-intestinal malignancies. Review of litera-
ture with regard to a eause/effeet relationship. ]. Chron. Dis. 31:23-33.
Millette, J.R., P.J. Clark, and M.F. Pansing. 1979. Exposure to Asbestos from Drinking
Water in the United States. EPA-600/1-79 028. Washington, D.C.: U.S. Environmental
Protection Ageney. 87 pp. Available from National Technical Information Service,
Springfield, Va., as Publ. No. PB-300-444.
Millette, J.R., P.J. Clark, M.F. Pansing, and J.D. Twyman. 1980. Concentration and size of
asbestos in water supplies. Environ. Health Perspeet. 34:13-25.
Morgan, A., J.C. Evans, RJ. Evans, R.F. Hounam, A. Holmes, and S.G. Doyle. 1975. Stud-
ies on the deposition of inhaled fibrous material in the respiratory tract of the rat and its
subsequent clearance using radioactive tracer techniques. II. Deposition of the UICC stan-
dard reference samples of asbestos. Environ. Res. 10:196-207.
National Researeh Council. 1977. Drinking Water and Health. Report of the Safe Drinking
Water Committee, Advisory Center on Toxicology, Assembly of Life Seienees. Washing-
ton, D.C.: National Academy of Seienees. 939 pp.
National Toxicology Program. 1981a. Draft Careinogenesis Bioassay of Chrysotile Asbestos
in Syrian Golden Hamsters. NTP-81-51. Bethesda, Md.: U.S. Public Health Service. (un-
published) 134] PP.
National Toxicology Program. 1981b. Draft Carcinogenesis Bioassay of Amosite Asbestos in
OCR for page 147
Adverse Health Effects of Arsenic and Asbestos 147
Syrian Golden Hamsters. NTP-82-86. Bethesda, Md.: U.S. Public Health Service. (un-
published) 130] pp.
Newhouse, M.L., and G. Berry. 1979. Patterns of mortality in asbestos factory workers in
London. Ann. N.Y. Acad. Sci. 330:53-60.
Peto, J. 1979. Dose-response relationships for asbestos-related disease: Implications for hy-
giene standards. Part II. Mortality. Ann. N.Y. Acad. Sci. 330:195-203.
Polissar, L., R.K. Severson, E.S. Boatman, and D.B. Thomas. 1982. Cancer incidence in
relation to asbestos in drinking water in the Puget Sound region. Am. J. Epidemiol.
1 16:314-328.
Pott, F. 1980. Animal experiments on biological effects of mineral fibres. Pp. 261-272 in J.C.
Wagner, ed. Biological Effects of Mineral Fibres, Vol. 1. IARC Scientific Publications No.
30. Lyon: International Agency for Research on Cancer.
Saracci, R. 1977. Asbestos and lung cancer: An analysis of the epidemiologica] evidence on
the asbestos-smoking interaction. lot. J. Cancer 20:323-331.
Seidman, H., I.J. Selikoff, and E.C. Hammond. 1979. Short-term asbestos work exposure
and long-term observation. Ann. N.Y. Acad. Sci. 330:61-89.
Selikoff, I.J., and D.H.K. Lee. 1978. Asbestos and Disease. Environmental Sciences. New
York: Academic Press. 549 pp.
Selikoff, I.J., E.C. Hammond, and H. Seidman. 1979. Mortality experience of insulation
workers in the United States and Canada. 1943-1976. Ann. N.Y. Acad. Sci. 330:91-116.
Smith, W.E., D.D. Hubert, H.J. Sobel, E.T. Peters, and T.E. Doerfler. 1980. Health of
e~cperimental animals drinking water with and without amosite asbestos and other mineral
particles. J. Environ. Pathol. Toxicol. 3:277-300.
Stanton, M.~., M. Layard, A. Tegeris, E. Miller, M. May, E. Morgan, and A. Smith. 1981.
Relation of particle dimension to carcinogenicity in amphibole asbestoses and other fibrous
minerals. J. Natl. Cancer lust. 67:965-975.
U.S. En~nronmental Protection Agency. 1979. Asbestos. Ambient Water Quality Criteria.
Washington, D.C.: Criteria and Standards Division, Office of Water Planning and Stan-
dards, I].S. Environmental Protection Agency. 1149] pp. Available from the National
Technical Information Service, Springfield, Va., as Publ. No. PB-297-917.
Wagner, J.C., G. Berry, ].W. Skidmore, and F.D. Pooley. 1980. The comparative effects of
three chrysotiles by injection and inhalation in rats. Pp. 363-372 in J.C. Wagner, ed. Bio-
logical Effects of Mineral Fibres, Vol. 1. lARC Scientific Publications No. 30. Lyon: Inter-
national Agency for Research on Cancer.
OCR for page 148
Representative terms from entire chapter:
relative risk
