nr>Acr                    EXTERNAL
                   UKAr                  REVIEW DRAFT
             DO NOT QUOTE OR CITE               MAY 1978
  CARCINOGEN ASSESSMENT GROUP'S
 ASSESSMENT OF CARCINOGENIC RISK
      FROM POPULATION EXPOSURE
   TO CADMIUM  IN THE AMBIENT AIR
                     NOTICE
         This document is a preliminary draft. It has not been
         formally released by EPA and should not at this stage be
         construed to represent Agency policy. It is being
         circulated for comment on its technical accuracy and
         policy implications.
                DO NOT QUOTE OR CITE
  LIBRARY SERVICES RTF NC
                  CENTAL PROTECTION AGENCY
                  Research and Development
TECHNICAL DOCUMENT COLLECTION shington, D.C. 20460

-------
                   The  Carcinogen  Assessment  Group's  DO NOT QUOTE OR Cl'T
           Assessment of  Carcinogenic  Risk  from Population
                Exposure  to  Cadmium  in  the  Ambient Air.
    This  document  is  being  released by EPA for external review

                                            MAY 1 9 19/8
                                   oy  E^^lbert,  M.D,
                                  Chai rman
Elizabeth L.  Anderson,  Ph.D.
Charles Brown, Ph.  D.  (NCI  Liason)
Richard N. Hil 1 ,  Ph.D,  M.D.
Charles B. Hiremath, Ph.D.
Robert E. McGaughy, Ph.D.
Lakshmi C. Mishra,  Ph.D.
Ruth Pertel,  Ph.D.
Wade T. Richardson,J.D.
Dharm Singh,  Ph.  D.
Todd Thorslund,  Sc. D.

STAFF_ASSI^STA]JTS

Irene Schapiro

-------
                          CONTENTS

I.    SUMMARY AND CONCLUSIONS                           1

II.   INTRODUCTION                                      4

III.  METABOLISM                                        4

IV.   MUTAGENICITY AND OTHER RELAVENT TESTS             6

V.    TOXIC EFFECTS                                     9

VI    CARCINOGENICITY              '                     9

     A.  Human Studies                                 9

         1.  OccupationalExposure                     9

           a) Potts Study (1965)       .                9
           b) Kipling and Water-house Study (1967)     12
           c) Lemen,et al. Study (1976)               13
           d) McMichael, et al.  Study (1976)          15
           e) Summary of Occupational Epidemiology    18
              Reports

     B.  Chronic Animal  Studies                        21

         1.  Injection Studies                         21
         2.  Oral Studies                              21

             a) Schroeder, et  al. Prat Study (1965)    21
             b) Levy and Clack Rat Study (1975)        21
             c) Levy, et al. Mouse Study (1975)        25

         3.  Summary of Animal Experiments             26

VII. RISK ASSESSMENT                                  27

-------
                                                       DRAFT
I.   SUMMARY AND CONCLUSIONS                    DO NOT QUOTE OR CITE
     Cadmium compounds have been shown to  produce  adverse
effect in experimental animals,  and studies  in  man  link
cadmium exposures with several  clinical  disease states.
Body concentrations of cadmium  increase  with age,  at  least
up to middle age, due to the poor renal  excretion  of  the
compound; thus, each source of  exposure  to cadmium  adds an
increment to the body burden.
     Injection of cadmium in experimental  animals  results  in
                                       *
the development of malignant tumors (sarcomas)  at  the site
of injection.  In addition, following subcutaneous  injection
of soluble cadmium salts, a significant increase in tumors
at a distant site, namely interstital cell tumors  of  the
testis, was found in treated rats and mice.
     Orally administered cadmium has not produced  a
;ignificant increase  in tumors in experimental  animals, but
the three  available studies have deficiencies which preclude
accepting  them as negative studies.
     Cadmium salts increase the  frequency of point and
chromosomal mutations.  They induce jji vitro mammalian
cellular transformation and enhance transformation of
virus-infected mammalian cells.  The outcome of these tests
is known to be highly correlated with oncogenicity.

-------
                                -2-
     Occupational epidemiology studies at three independent
locations with known exposure to cadmium have reported
associations between cadmium exposure and excess rates of
prostate cancer.   At an alkaline battery plant a
proportionate mortality study reported a high frequency of
prostatic cancer  amongst deaths in workers exposed to
cadmium for at least 10 years.  Another study at the same
plant involving a follow up of the population of workers who
were exposed at least one year to cadmium showed a
statistically significant increase in the incidence of
prostatic cancer.  At a cadmium smelter a retrospective
cohort study of workers exposed to cadmium for at least two
years reported a  significant increase in cancer deaths,
mainly due to respiratory cancer.  A statistically signifi-
cant excess prostate cancer was found only  in workers ex
poseed for more than 20 years.  In a survey of four rubber
producing plants  the authors found that one of the four job
catergories that  was associated with workers dying of
prostrate cancer-involved exposure to cadmium and other
metal oxides.
     When all of  this information is considered together
it is possible that cadmium is a human carcinogen, although
the evidence is not decisive.

-------
                                -3-
     A very crude quantitative assessment of population
risks of prostrate cancer due to cadmium in the ambient  air
can be done on the basis of the prostate cancer data on
alkaline battery and smelter workers together with EPA
estimates of ambient concentrations and number of people
exposed.  The conclusion is that a, lifetime exposure to
atomospheric cadmium from any of the individual sources
results in a lifetime risk of less than 2 x 10'5 and the
total number of prostate cancer deaths,in the United States
population exposed to ambient airborne cadmium is less than
about 5 deaths per year.  Since the daily cadmium retention
from ambient air is only about  1/20 of the retention from
food and about 1/3 of the intake from smoking a half-pack of
cigarettes per day, this report deals with a relatively
small part of the total  problem of cadmium contamination

-------
              ION
     Environmental  exposure to cadmium occurs by several
routes.  It is estimated that the exposure to an individual
in food is about 50 ug per day;  water, 2 ug per day;  and
air, less than 1 ug per day.   The major sources of cadmium
in the ambient air  are non-ferrous metal smelters, iron  and
steel mills and municipal  incinerators.  In addition  sources
of cadmium exposure are wide  ranging.   For example it is
present in cigarettes at relatively  high levels and is used
as a pesticide,  largely on turf  grasses.
III. METABOLISM
     This section is based on literature reviews by Friberg
et al . , (1971) and  the International  Agency for Research  on
Cancer (1976).
     The two primary routes of cadmium absorption in  man  are
the respiratory  and gastrointestinal  tract.  The extent  of
absorption varies with the route of  exposre.
     In acute  and chronic  inhalation  studies  of mice, dogs
and rabbits about 10 to 40% of the inhaled dose was
absorbed.  Salts tested were  cadmium  chloride, oxide,
sulfide and cadmium iron dust.  It is possible that
absorption differs  with the form of  cadmium;  however, these
studies were not adequate  to  determine this.
     There is  markedly less absorption from oral  dosing  of
cadmium salts  than  by inhalation.  The average absorption by
mice, rats and goats was only a  few  percent of the dose
(range: 0.5 to 8%).  Compounds tested were cadmium chloride
and cadmium nitrate.  Dietary factors  have also been  shown
to influence absorption.

-------
                                 -5-
Low,. calcium and low iron diets have increased cadmium absorption in mice
and rats.  Vitamin D has been reported to  increase  intestinal absorption
in chickens.
     Skog and Wahlberg  (1964) studied the  absorption of cadmium chloride
from the guinea pig skin.  Absorption was  determined by "disappearance
measurements" using a scintillator to measure radioactive skin deposits.
Six molar concentrations ranging  from 0.005  to 4.87 were tested.  Dermal
absorption of cadmium was very low.  It  increased with increasing molar
concentrations to a maximum  at 0.235 M when  the mean absorption was 1.8%/
5 hours.  At other concentrations mean absorption values were less than
1.0%/5 hours.
     Cadmium steadily accumulates in the body with  normal exposure/ reaching
a peak level at age 50.  The background  level of a  newborn baby is about lug
total; the body burden  of a  50-year-old  American man is about 30 mg.  Fri-
berg estimates that daily cadmium intake from food  in the United States is
in the  upper portion of the 20 to 50 ug range and  calculates that cadmium reten-
tion would have to be 3 to 8% to  account for the 30 mg accumulation.  This
does not account for intake  from  sources other than food and assures a
linear accumulation with negligible excretion.
     In both animal and human studies, very  small amounts of cadmium were
excreted in the urine.  In humans, urinary levels have been less than 5 ug/
day; generally excretion is  1 to  2 ug/day.  Smckers excrete more cadmium than
nonsmokers; cigarettes  contribute to smokers increase in cadmium body burden
and excretion as compared with nonsmckers.  Excretion also increases in persons
with renal tubular dysfunction, a condition  which can be produced by prolonged

-------
                                 -6-
cadrnium exposure.  Cac3mium has a long biological half-life.  In
humans it is estimated to be 10 to'30 years,- and in other species it
usually has been found to be at least 200 days.
     About one third of the total body burden of cadmium in man is in the
kidneys; together the kidneys and liver store about 50% of the burden.
Renal concentrations of cadmium have generally been 5 to 100% higher in
smokers than nonsmokers.
     Cadmium has also been found in significant amounts in the pancreas.
After a single intravenous injection of 109 Cd-dloride, mice were found
to accumulate cadmium in bone marrow, periostium, tastes and hypophysis.
The latter two sites are of interest in light of the tumors observed in
Cd-injected animals.
     The metabolism of cadmium is closely associated with zinc metabolism.
Added zinc can protect animals from some cadmium-induced toxic effects.
Metallothionein, a low molecular weight cadmium and zinc-binding protein,
is believed to transport cadmium.  Some evidence suggest that low cadmium
exposures are nontoxic because the metal is bound to the protein.  When all
metal.lothionein sites are bound to cadmium, any additional cadmium is free
to produce toxic effects.
IV.  MEJTAGENICITY AND Qi'H£R pPTTVaNT TESTS
     Three secondary as well as several primary sources were consulted in
the preparation of this section.  The secondary sources were Sanchu, 1977;
Pertel, 1977 and FDA, 1977; all primary references are cited in those
three sources.

-------
                                    -7-
Oroanism
Reported
Mutageni-
citv
Garments
Authors
(+)
1.  Point Mutations

    Sacchrcmyces

    Drosophila
    Chinese hamster cells
      in culture

2.  Chromosome Mutations

    a.  Maimalian cells,
          in vitro
Human leukocytes    (+)


Chinese hanster     (+)
  line

Human leukocytes
  & fibroblast line (-)

Mairnials, in vivo

Sheep leukocytes    (+)

Mouse oocytes       (+)

Human leukocytes
    b.
            Sex linked lethals;
            low power of test
                                        Numerical  not struc-
                                        tural aberrations
                                         Itai-itai patients
                                         Itai-itai patients
                                            and Cd workers

                                         Workers exrosed to  Cd
                                               & Pb"

                                         Workers exposed to  Cd,
                                              Pb and Zn
                                                          Takahashi,  1975

                                                          Friberg et  al.,
                                                                   1974
                                8-azaguanine resist-      Costa,  1976
                                  ance
                                                                  Shirashi et al.,
                                                                    1972

                                                                  Rohr & Bauchinger,
                                                                    1976
                                                                  Paton & Allison,
                                                                    1972
                                      Doyle et al.,  1974

                                      Shimada et al.,  1976

                                      Shirashi & Yosida,
                                        1972

                                      Eui et al., 1975
                                      Deknudt & Leonard,
                                        1975

                                      Bauchinger et al.,
                                        1976

-------
                                                       DOtuJTOliC'Ej:


Oraanisra
Dominant
lethal

Reported
Mutsgeni-
citv
(-)

(-)


Ccimtents
Mouse

Mouse


Authors
Epstein et al. ,
1972
Gilliavcd s
  Other Tests
  B.  subtilis
      S.  cerevisiae

      Syrian hanster
       cells
Reported
Effect
                for Cl but N03 salt
                 in "rec" assay

                for mitotic segregation

                for alteration in DMA
                sedimentation in CsCl2
                gradients

                for unscheduled DMA
                synthesis

                for _in vitro trans-
                  formation
                                                                 Leonard, 1975
Nishioka, 1975


Takahashi, 1972

Costa, 1976



Costa, 1976


Costa, 1976
                                       for enhancement of SA7     Costa,  1976
                                       viral transformation,
                                       in vitro
              v, x
-------
                                    -9-
V.  TOXIC EFFECTS
     Chronic occupational exposure to excessive cadmium in the air
produces emphysema after an exposure period of-about 20 years.  Kidney
malfunction also occurs in these workers, manifested by glucosuria,
proteinuria, excess cadmium excretion, and increased incidence cf
renal stones.   In rats, many  studies have shown that cadmium affects
the testis, causing vascular  degeneration and, testicular atrophy.
In addition it causes hypertension and kidney damage in experimental
animals. Itai-itai disease, due to chronic ingestion of cadmium-con-
taminated  rice, was characterized  by multiple bone fractures mainly in
multiparous post mencpausal women  (Friberg,  1971).

VI.  CAEONOGENICLTY
     A. Human Studies
         The  epidemiology studies  reviewed here include  four  studies  of
workers occupationally exposed to  cadmium,  two  of them with overlapping
 samples.  Two additional studies analyze cancer mortality in relation to
 metals in drinking water.
         1.  occupational Exposure
             a.  Potts Study  (1965)
             This article reports the health status of current and past
 employees of a British alkaline battery factory who were exposed to cadmium
 oxide dust for at least 10 years.   The plant had been in operation since
 1920, and in 1946 it was moved to a nev location  in the sane town.  Before
 1950, when an  extensive exhaust ventilation system was installed, cacmiun
 concentrations of 236 mg/m3  and 0.6 to 2.8 mg/m3^ were

-------
                                -10-
in two separate areas of the plant.  Between 1950 and 1956 the levels
were less than 0.5 rag/in^ in most parts of the factory but seme regions
were well above this.  After 1956, when further improvements were made,
most areas had levels below 0.1 rog/ra3, and the operating policy sir.ce
then has been to take special action when the level exceeds 0.5 mg/nr.
     The health survey covered workers employed by the plant since 1920.
In 70 employees working over 10 years, 44% had protein in their urine.
The protein had the same characteristics as that observed in workers at
another factory who were exposed to cadmium funes.  Kidney function tests
were not done routinely but four case reports of workers with proteinuria
revealed no kidney damage.
     Potts reports that "careful search" showed a total of 74 men had been
exposed to cadmium for more than 10 years, and eight of those workers had
died.  He leaves the  ijnpression that he succeeded in determining the vital
status of all 74 men without indicating the source of his mortality figures
or cause-specific death data.  Of  the eight deaths he reports, five were
from cancer and  three of  these were cancer of the prostate.  The data from
Potts paper are  summarized below.

-------
                                 11
   Mortality Data for Cadmium Workers Exposed for More Than Ten Years
Year
of Death
1960
1960
1961
1962
1962
1963
1964
1964
Aae
65
75
65
63
78
53
65
59
Length of
Cadmium
Exoosure
(Years)
31
14
37
34
18
35
38
24
Cause of Death
Auricular Fibrillation
Carcinoma of Prostate
Carcinoma of Prostate
Bronchitis and Atheroma
Bronchitis
Carcinoma of Bronchus
Carcinoma of Prostate
Carcinoma tosis
     All of these people were exposed to the high cadmium concentrations
that existed before 1950; two of the prostate cases were thus exposed
for at least 24 years and the third for at least 4 years.  The exact
time of the start of exposure was not stated' and the cadmium concentra-
tions in the earlier period are known only in 1949, when the decision
was made to improve conditions at the new plant.  No report was made of
workers exposed less than 10 years.
     While Pott's cancer mortality findings are striking, it is impossible
to make a meaningful comparison of rates with the general population since
the age structure of the exposed population is not reported.

-------
                               -12-
     b.  Kipling and Waterhouse Study (1967)

     In a letter to the editor of Lancet, Kipling and Waterhouse

report on a survey of 248 British workers occupationally exposed to

cadmium oxide.  This work was conducted on the same population

as Potts' study, however/ Kipling and Waterhouse examined incidence of

prostate cancer and included workers with at least one year of exposure

to cadmium.

     The rate of cancer for all sites as well as for cancer of the prostate,

bronchus, bladder, and testes was determined, using an analysis based on total

time-to-tumor/ and compared with the expected rate for "such a group of men of

known age".  No other description of the control group was provided.  The results

are shown in the table below.  There were 4 cases of prostate cancer compared to

the expected number of 0.58, a significant increase at the .003 level.  Three of

these cases are the same as Pott's prostate mortality cases, according to a

personal communication frcm Kipling reported in IABC (1976).  There was no report

of the &ge distribution of the cancer cases or the level or duration of exposure

to cadmium.


    Cancer Incidence In 248 Workers Exposed To Cadmium  Oxide
                     For More Than 1 Year
Site of Cancer
All sites —
Bronchus —
Bladder —
Prostate —
Testis —
No. of
Excected
— 13.13
— 4.40
— ' 0.51
— 0.58
— 0.11
cases
Observed
12
5
1
4
0
Probability
of occurrence
.0.660
0.449
0.398
0.003
0.898

-------
                           13-
C.  Lenten et al., Study  (1976)
     These investigators did a retrospective cohort study of
workers at a cadmium smelter at which cadmium metal powder, cadmium
oxide and cadmium sulf ide were present.  The smelter has been producing
cadmium since 1925.  Cadmium concentrations were measured on two occasions.
In 1947 the average concentration of cadmium fumes ranged from 0.04 to
6.59 mg/nr and of cadmium dust was  17.23 mg/m  , with most plant operations
below 1.5 mg/nr.  In 1973,  another  measureirent shewed  average air concentra-
tions ranging below 1 mg/nr on an 8-hour tome-weighted average basis, with
excursions up to 24 mg/nr during  infrequent operations.  The authors reported
that in 1973 a respirator program was  in use at the plant which reduced
exposures by a factor of about 10.  No information was given about how long
before 1973 the respirator  program  had been in operation or whether the
respirators were routinely  used by  workers in all sections of the plant.
In two sections of  the plant concentrations of 75 to 90 ug/nr were measured
in 1973.  The arsenic air concentrations were generally less than accut 1/1CO
of the cadmium levels.   The ratios  of zinc and arsenic to cadmium in the ores
which are processed are  about 1/10  and 1/200,  respectively, and the authors
concluded  that exposure  to metals other than cadmium were  insignificant.
     The study  included  292 white male workers who worked at the plant
between 1940  and 1969 and who were  employed  for  at  least two years.  The
vital  status  of  these workers was determined as  of  1974, and it was found
 that 92 had diedr  180 were still alive and 20  were  not accounted for; the
 latter were assumed to  be alive in the analysis  of  the data.  Person-years
 at risk of dying were coipuued  and observed cancer  rates were ccnpared  to age-
 adjusted rates  for the  total U.S. white male population.

-------
                               -14-
     A significant increase in cancer mortality among the exposed
group was observed; most of this excessive risk was accounted for by
neoplasms of the respiratory system which was significant at the .05
level.  Of the 12 malignant respiratory disease cases, the cell type
of eight was known.  Three of these eight were squamous cell carcinomas,
the type most conroonly associated with cigarette smoking.  No smoking
histories for the groups were included.  An association between respira-
tory cancer and cadmium has not been observed in other studies.  In one
dog study, the cause of death for cadmium-exposed animals was non-necplastic
respiratory problems (personal communication, Dr. Richard Lemen, 8/10/77).
     Because of previous reports of increased prostate cancer in workers
exposed to cadmium, the prostate cancer mortality of the cohort was also
studied.  The results are given in the table below:
Prostate
Case
1
2
3
4
Cancer
Age
71
77
79
64
Deaths Amorc
More Than 2
ExDosure
4
13
18
17
Cadmium Smelter
Years Exposure
Latencv
32
25 *''
31
26
Workers with
Date of Death
2/26/72
3/19/68
12/10/60
4/3/51
     The four deaths from prostate cancer were not significantly higher
than the 1.15 expected to occur in the full cohort.  However, an increased
risk was found (4 cases compared to 0.88 expected) at the .05 level among
workers with a 20 year latency period since initial cadadum exposure.

-------
                                                      DRAFT
d.  McMlchael  et al.  (1976)                    DO NOT QUOT; CR
     These investigators  identified  cohorts  of  active  and
retired workers, aged 40  to  84,  at four  rubber-producing
plants and followed their mortality  experience  for  10  years.
     The mortality experience was  determined from  group life
insurance records  and confirmed  with  internal companny
records.  In the case of  one cohort,  a random sample  of
apparent survivors was checked against Internal  Revenue
                                     t
Service records to verify that they  were still  alive.
     It is important  to note that  rubber workers are
potentially exposed to numerous  organic  and  inorganic
chemicals, some of them known or suspected carcinogens.
While cadmium is present  in  these  factories, mortality  rates
and cause-specific deaths in this  study, which  is  not
controlled for type of exposure, can not be  attributed  to
cadmium alone.
     A total of 18,903 male  workers  from the four  facctories
were included.  About 1%  of  this group — active  workers  who
switched jobs during the  course  of the study--  was  lost to
follow-up.  Death  certificates were  obtained from  98%  of the
remaining cohort members  who had died.
     For the entire group 18,903 men, 98% had worked  at the
plant for at least 10 years.  The  study  did  not include
former workers who had transferred to another place of
employment or had  died as of the date the cohort was
identified.  As a  result, the sample could have selected
for the stronger workers  who survived and could have

-------
                             -16-
excluded some who might have died due to occupational
exposure.  While such a study might show an excess mortality
in  the worker population, the effects of the work
environment could be underestimated or overlooked.
     The mortality data for the workers was compared to
mortality figures for the 1968 U.S. male population.  The
analysis was made using standardized Mortality Ratios (SMR),
controlling for race and using age categories  of 40-64,
65-84, and 40-84.  A SMR of 100 indicates no difference
between the study group and the standard populaton.   An SMR
greater than 100 indicates an excess of deaths in the test
group.  However, since worker populations are  generally
healthier than the general population, in a non-hazardous
work environment the SMR is expected to be slightly  less
than 100.
     The total  number of deaths from all  causes was  5,106,
with 1,014 due to cancer.   The SMR for deaths  from all
causes for the full  cohort was 94, it was relatively higher
in the older group than in the younger workers.  The SMR for
all  cancer sites showed similar trends,  with a value of 100
for all ages and companies combined, 104 for the older  group
and 92 for the younger group.   When specific sites were
examined, SMR's for all companies combined were considerably
higher than 100 for stomach (148), rectum (116),  prostate
(119),  all  leukemias (130), lymphatic leukemia (15), and
lymphosarcoma  and Hodgkin's disease (129).   No striking
excesses were  seen in the  combined-company  SMR's

-------
                             -17-

for cancers of the rectum,  pancreas,  respiratory  system,
bladder ,  and brai n.
     These figures reflect  exposure to a variety  of
substances in the  workplace.   In an attempt  to  correlate
cancer sites with  specific  job exposures,  an exploratory
study was made of  the job classifications  of workers  with
stomach,  prostate  and bladder cancer and lymphosarcoma  and
Hodgkin's disease.  The work  histories of men dying  from
each of these causes was compared to an age-stratified
random sample of 23% of the full cohort (n=1476)  to  see
whether differences existed in the porportion of  men  who  had
worked for at least two years in any job category.
     Four jobs, compounding and mixing, cement mixing,
janitoring»and trucking, were assocciated with  prostate
cancer.  The compounding and mixing association is  of
interest because these workers are exposed to cadmium oxides
and other metal oxides that are used as vulcanization
accelerators.  The exposure levels are not reported.  While
this is of interest as a basis for future investigations,  it
is  clearly insufficient as  a  basis for associating  cadmium
with prostate cancer since  the worker's exposure  to  other
substances and the extent of  cadmium exposure have  not  been
a n ay 1 s e d.

-------
                             -18-
e.  Summary of Occupational Ep_j_denji'o1 ogy Reports
     Four occupational  epidemiology studies have been
reviewed.  In all of them an association between prostate
cancer and cadmium is suggested.  In one study (Kipling &
Waterhouse, 1967) an excess number of  cases of prostate
cancer occurred in exposed workers.  Another study on the
same population shows a high death rate from prostate cancer
(Potts, 1965).  Excess  mortality from  prostate cancer among
workers whose initial cadmium exposure had been at least 20
years before the study  was observea in a third paper (Lemen
et al . , 1976).  Although it was not a  controlled study, a
fourth report (McMichael et al., 1976) suggested a possible
correlation between excess prostate cancer mortality and
possible exposure to cadmium.
     All of these studies involved small numbers which  makes
the estimation of risk  difficult.   However, because of  the
consistency of the findings, together  they raise a serious
question as to a possibility of a  relationship between
cadmium and prostate cancer.

-------
                                   -19-
f.  Non-occupational Epidgniology Stariies
     Correlations between cancer mortality and cadmium in drinking water
have been studied by Berg and Burbank (1972) and by Zyka (1973).  In Berg
and Burbank1s study, the drinking water concentrations of eight metals,
including cadmium, were determined in 15 water basins of the United States.
Death rates for 34 types of cancer were also determined for the regions
corresponding to the sane water basins.  The water basins were then ranked
with respect to each metal and each cancer site, and rank correlation
coefficients were calculated.
     For cadimum 10 positive (statistically significant) correlations were
observed.  No correlation was noted for testicular cancer, which might be
expected from animal experiments.  The authors did not report looking for
prostate cancer correlations.  Positive associations were observed for
intestinal cancer and  "smokers cancer" of the mouth, esophagus, larynx, lung,
and bladder.  The authors were unable to separate the effects of tobacco
from possible effects  of cadmium in the "smokers cancer".  Cigarettes contain
cadmium, and smokers have excess body burdens of cadmium; this could account
for the correlation.   Other studies cited by the authors have found no cor-
relation between cadmium levels and intestinal cancer.
     The authors note  that, depending how cadmium is measured, rank orders
of cadmium levels can  be vastly different from one another.  Berg and Burbaak
report that their results "are based on less than perfect analytical results
and sampling."  In light of this study's limitations, no conclusions can be
drawn about cadmium's  relationship to cancer mortality.  It does suggest areas
for further investigation.

-------
                                -20-
                                                            DRAh
                                                   DOM,"QUOit ORCilL
     Zyka (1973)  chemically analyzed current and past drinking watsr
sources in districts of the city of  Kutna Hora, near Prague, Czecholslovakia
and correlated metal concentrations  with cancer mortality rates.  Considerable
disparity existed in cancer mortality rates among the several districts
of the city, and Zyka reported  that  the source of drinking water was the only
environmental factor that varied.
     He found that the water from  the Havirna mine, the water source for the
area with the highest cancer mortality/ had high levels of Cd, Zn, As, Cu, Al,
Pe and Se.  In contrast, in the districts with the lowest cancer mortality,
the only drinking water contaminants found were NO^ and low levels of Pbf Zn,
As and Eg.
     There was apparently no attempt to control for other factors possibly
associated with the variation in mortality rates  (e.g. age structure of
the population), and the effect of cadmium independent of other trace metals
was not identified.  High concentrations  in drinking water of several of the
trace metals found in the Eavirna water have been associated with cancer
in previous studies.  Cie effect of  this  study is to suggest a possible
relationship between high concentrations  of trace metals in drinking water
and excess cancer mortality without  specifically  implicating cadmium.

-------
                                -21-
B.  Chronic Anrnal Studies
     1.  Injection Studies
     In 14 chronic studies on rats reviewed by a working group for the Inter-
national Agency for Research on Cancer {IMC) (1976), malignant tumors at the
injection site wer  reported in 10.  Tumors followed administration of soluble
cadmium salts  (chloride and sulfate), insoluble salts (sulfide and oxide) and
finely divided cadmium metal.  Rats  receiving 'injections at four sites
(0.04 or 0.08 rag/site) developed local tuners in .areas originating frcra mese-
nchymal mesoderm  (subcutaneous,  intramuscular, subperiostial) but not from
those originating frcin ectodenaal  (intracutaneous), endodermal  (liver, ventral
prostate, salivary gland) or epithelial mescdenn  (kidney)  (Gunn et al., 1967).
     In several studies,  subcutaneous injection of soluble cadmium salts has
resulted  in the development of  interstitial cell  tuners of the  testis in rats
and mice  (IAEC, 1976).   Injection  of cadmium produces degeneration of semini-
ferous  tubules followed by  proliferation  of interstitial  cells  and f ir^aUy
tumor?.  A  sumrary of the findings is given on the next page.
   2. Oral  Studies
      a.   Schroeder et al. (1965) Rat study"
      In a laboratory designed  for  lew  trace-metal aninal  experiments
Long-Evans  rats were fed a  lew-metal diet supplemented with  5 pan of
cadmium,  lead, or chromium  (III) as acetate salts in drinking water.
This treatment was continued  frcm weaning until death  of  the animals.
Initially there were 69 males  and  58 females;  but high early mortality
 (before 3 months) occurred  in treated  animals; and the groups were replen-
 ished before 3 months so that each group had more than 52 animals.  Further

-------
                                   -22-
               TESnOJLAR INTERSTITIAL m*- TUMORS IN ANIMALS

                   GIVEN CADMIUM BY SOBOJTMffiOGS INJECTION
Strain
Cd C12   a/
Dose(inaAa)
                               Tumor Bearira Animals  Observation
Treated
Control  Time{mo.)
References
Mouse

 Charles River  5.5

Rat
                  20/26
                               Gunn et al., 1963C
Wistar
Albino CB
Wistar
Wistar
Fischer 344
5.5
0. Sing/rat
5.5
3.7 to 5.5
5.5
17/25
10/18
10/25
13/13
16/20
0/20^/ 11
0/16^ 20
23
11
0/10^' 12
Gunn et al. , 1963f/
Roe et al. , 1964f/
Knorre, 1971
Lucis et al. ,1972
Reddy et al. , 1973
a/ Single S.C. injection except Roe et al., who used 10 weekly injections
 "  of CdSO^
b/ Difference significant at p<.001.
c/ Reviewed by IARC (1976); original pacers not received in tins to be reviewed
   by CAG.

-------
                                   -23-
nortality occurred because of a pneumonia epidemic during weeks 107 to 117,
reducing the effective size of the groups to 42 males and 44 females in
controls and 50 niales and 46 females .in treated groups. In the cadmium
group there was significantly higher mortality as compared with controls
during the first 33 months of observation.  (Animals survived up to 42
months).
     The major gross causes of death were pneumonia and tumors.  Hypertension
contributed to the progressive mortality of males, especially since treated
animals had enlarged hearts and thickened left ventricles, and the authors
had previously reported hypertension in rats.  Of the animals which were
autopsied, gross tumors were the  cause of death in 13/48 treated males compared
to 9/35 control males.  For females the corresponding observations were 15/36
in treated versus 15/35 in controls.  The differences are not statistically
significant.  No microscopic examination of the tumors was reported.  All but
one animal with tumors died after 86 weeks. Sixteen of 20 cadmium-fed rats
examined (80%) had cirrhosis of the liver as did 13 of 21 control rats (62%).
The high frequency of cirrhosis in controls is unusual and makes one wonder
what accounted for this finding.  Other adverse effects of the cadmium treat-
ment include arteriolarsclerosis  in the kidney and enlarged hearts.  Cadmium
accumulated in livers and kidneys of young  rats, but not in animals older
than 2 years.

b. Lew and Clack (1975)  Rat Study
     These investigators administered cadmium sulfate in distilled water
to male specific pathogen-free CB hooded rats.  Doses of 0.35, 0.18, and
0.087 mgAg/week in distilled water were administerd%by stcmach tube for
2 years.  Initially there were 30 animals in each treated group and 90

-------
                                 -24
animals in the control group.  All sick and moribund animals as well as
those surviving for the 2-year period were examined grossly.  Microscope
preparations were made of all tissues of ten high-dose and ten control
animals randomly selected from the 'animals killed at termination of the
experiment.  In addition, sections of all testes and any abnormal tissue
or organ were examined.  Hie survival at 100 weeks for controls, low, medium,
and high dose groups were 67%, 53%, 67%, and 57%, respectively, with some
early deaths from mechanical injury during passage of the stomach tube.
     The incidence of interstitial-cell testicular tumors was 75% in
controls; in the treated rat it was not significantly different.  Careful
observation of prostate glands shewed no gross abnormalities, but foci of
slight hyperplasia, inflammation and descuamation of epithelial cells were
observed in high-dose and control groups.  These changes were not interpreted
as metaplastic or pre-neoplastic.  Tumors of other organs were observed as
follows (number of tumors observed/number of animals in the initial group):
controls, 10/90; low dose, 4/30; nredium dose, 3/30; high dose,. 8/30.  All
tumors were benign except one basal-cell skin carcinoma in the high-dose
group and two sarcomas in the control group^
     The number of liver adenomas in the four dose groups was: controls, 1;
low dose, 2; medium dose, 0; high dose, 3.  The paper did not report which
of the tumors were identified on gross examination and which by random
selection, making it impossible to determine statistical significance.
The Cancer Assessment Group  is seeking additional information  in this area.

-------
                                    -25-
     It is difficult to draw conclusions from this report because: (1) only
10 animals randomly selected from the high dose and control group were
microscopically examined, rather than all aniraals from all dose groups; and
(2) because of the higher mortality in the high-dose group compared to controls,
a correction of the tumor incidence 'data is necessary.  This correction
cannot be done without data for the tice when the tumors occur in individual
animals.

c. Levy et al.  (1975) Mouse Study
     Groups of specific pathogen-free male Swiss mice were given cadmium
sulphate once per week for a period of 18 months.  Three treated groups
of 50 animals each were given  1.75, 0.88, and 0.44 mgAg/week by stomach
tube and the control group of  150 animals received distilled water alone.
     All sick and moribund animals  as well as those  surviving for the 18
month period were examined grossly. Tissue sections of 20 animals frcm
the high-dose group and  20 frcm the control group were examined microscopi-
cally.  These animals were selected randomly from animals with no gross
lesions that had survived until termination of  the experiment.  In addition,
sections were examined microscopically from all tissues appearing abnormal
grossly.  The fraction of animals  surviving at  the time of  terminal  sacrifice
in the control, lew, medium, and high-dose groups were 83%, 94%, 84%, and 84%,
respectively".
     No gross abnormalities were observed in prostate, bladder, kidney, or
testis.  A  small number  of animals in each group had over-distension of
seminal vesicles and  several  showed lesions of  the lung and/or liver. Frcm
the  microscopic sections examined  occasional slight  epithelial hyperplasia
or desquamation was seen in  the prostate of both hi(gh-dose  and control
animals.  The  testes were either normal  or had  slight atrophy of  isolated

-------
                                  -26-
seminiferous tubules, occur ing with equal frequency in control and high-dose
groups.  No treatment-related differences in tumor frequency were observed
for kidney, lung, or liver tumors.
     It is difficult to make quantitative conclusions from this experiment
because (1) not all animals were examined histologically and therefore it
is possible that small tumors not noticed in gross examination remain
undetected; (2) the results of the microscopic examination were only quali-
tatively stated.  Hcwever, there is no fina reason to reject the authors'
conclusion that this treatment caused no carcinogenic effect.
     3.  Sunroary of Animal Experiments
     Although direct injection of cadmium salts intramuscularly and
subcutaneously usually produces malignant tumors at the injection site
and interstitial-cell tumors in the testes, no carcinogenic effect has
been clearly seen in rats or mics following oral administration.  A plausible
explanation for the difference in response is that only about 0.3 to 1% of
the orally-administered cadmium dose is absorbed in chronic studies,
(Fleischer, et al., 1974).  The tumors observed in typical injection
experiments are produced by single doses of 4 to 5 mgAg or repeated doses
of 1.7 mg/kg/week.  The highest oral dose tested by Levy was 1.75 mg/kg/week
in mice and 0.35 mgAg/week in rats, and the oral dose used by Schroeder
in rats was 2.6 mg/kg/week.  If only 1% of this oral dose is absorbed,
the actual exposure of the animals in the oral experiments is only about
1% of  the dose administered in the injection experiments.  Therefore, the
oral dose might  be too snail to produce observable effects. In keeping with
this only in  the Schroeder study was there any indication of a treatment
effect as to  weight  or survival.

-------
                                   -27-
                                                                 DRAFT
                                                        DO NOT QUOTE OR Gill
VII_ RISK ASSESSMENT



                       Introduction and Summary


         It was  noted  in section IV, part 8, that cadmium is carcinogenic


 in experimental animals by  subcutaneous injection.  In the case of human


 exposure to  atmospheric cadmium, epidemiological evidence is accumulat-


 ing from several small-scale  studies which suggests an increase in the


 prostate cancer rate  where  workers were exposed to high levels of


 cadmium dust and/or fumes.


         Two  of  these  studies,  Potts  (1965) and Lemen et al.  (1976), con-


 tain enough  exposure  information  that  crude dose-response relationships


 can be established between  lifetime  exposure  levels and prostate cancer


 death rates.  It was  estimated that  the percentage increase  in the pros-

                                        3
 tate cancer  rate per  increase of  1 ygm/m   of  atmospheric cadmium was


 10.1% based  on the Potts  study and 9.4% based on  the Lemen et al. study.


 The consistencies of  these  estimates should be viewed as merely coinci-


 dental.  Considering the  crudeness of  the  exposure data and  the small


 sample sizes of the study cohorts, it  would not have been surprising to


 find a difference of  a full order of magnitude, even assuming an identi-


 cal effect in the two exposed populations.


         Using the Potts rate estimate  of 10.1% and a 1978 exposure docu-


 ment prepared by Energy and Environmental  Analysis, Inc. (EEAI), for EPA,


 it was calculated that a  population  of about  34 x 10  males  was exposed


 to atmospheric cadmium from point sources  in  sufficient amounts to


 increase their lifetime probability  of death  due  to prostate cancer by

-------
                                   -28-
 more  than 10   .   This  total  exposure  resulted  in an estimate of about



 5  expected deaths per  year,  which  amounts  to approximately  .03% of the



 total prostate  cancer  deaths in  the United States.





                Epidemiological  Studies with Pertinent

                         Exposure  Information




 Potts  (1965)



        Potts describes the  observed mortality in a cohort of men



 exposed,  for at least  ten years, to cadmium oxide dust in the manufac-



 ture of alkaline batteries.  Unfortunately, neither the age structure of



 the worker population nor the duration of  its exposure is given in the



 paper.  As a result, it is not possible to calculate age-exposure-specific



 mortality rates from the supplied information.



        However,  three deaths due to prostate cancer were noted in the



 eight observed deaths in the cohort.  If a sample of eight deaths age-



 matched to the eight deaths  in Potts'  cohort were taken from the United



 States population, the probability of three or more prostate-cancer-



 caused deaths occurring by pure chance in the sample is calculated,  based


                                              -4
on 1973 U.S. vital statistics, to be 2.16 x 10   .   The derivation of this



result is shown in appendix I.  Since this strongly suggests a  cadmium-



caused elevation in prostate cancer, we decided to base part of our  risk



analysis on the proportional mortality of the deceased workers  in Potts'



 cohort.



        Based on results of efforts at monitoring industrial plant



hygiene dating from 1949 and on worker records, it was possible to make



a crude estimate of the lifetime average exposure to atmospheric cadmium



 for each of the deceased.  The bases for these estimates are shown in

-------
                                    -29-
 table  1 .   The  marked  decrease  in  the  average  level  of  exposure over  time




 is  due primarily to major  plant hygiene  improvements in  1950 and  1956.




 The exposure in  the early  time period is taken  as the  geometric mean of



 the quoted measured levels.




        In table 2 the  proportional mortality ratio of prostate cancer




 to  total  deaths  on an age-specific basis is shown for  Potts' cohort, as




 well as U.S. male death rates  based on 1973 vital statistics.  Under the




 following assumptions it is  possible  to  utilize t"he data contained in



 table  2 to estimate the increase  in the  U.S.  prostate  cancer rate per




 change in each ygm of cadmium  in  the  atmosphere.




        1.  Given the same exposure,  x,  the age-specific proportional




 mortality ratio,  p.,  is the  same  for  Potts' cohort  as  for the U.S. male



 population.




        2.  The  increase in  the age-specific  U.S. prostate cancer rate



 is  a. linear function  of the  exposure  x.




        From these assumptions it follows  that :





                  otgjd  + Bx.)
where   a   = the U.S. prostate age-specific death rate for the jth



              interval



        a.  = the total U.S. age-specific death rate for the jth age



              interval



        x.    average lifetime exposure for those in the jth age



              interval



        $   = unknown increase in prostate cancer rate in percent per


                                 3
              change of one ugm/m  of atmospheric cadmium.




Solving for 3x., we have that

-------
                                   -30-
            ex. =  __  -.- i  =Y.

              J    U - pjAa2j     J    ]




so that 3 may be estimated given Y., x., a., ct_ .  by means of fitting a



least squares linear equation through the origin.  If we weight each



point by the value W.,  which in this case is the number of deaths in

                               A

the age interval,  we have that 3 = .5024 (the data utilized to compute


                                                   2
g is shown in table 3).  The goodness of fit in a x  sense is shown in



table 4, where we note that the variation is reduced to less than half



that observed assuming that age or exposure has no effect on the pros-



tate cancer rates.



        At the usual levels of environmental exposures, x, it is gener-



ally the case that a. » 3<*2.x.   If x is continuous over time, a very



close approximation to the increase in the lifetime probability of pros-



tate cancer due to a lifetime exposure x is P ~ PnBx, where Pn is the



lifetime probability of prostate cancer in the U.S. population.   From



1973 vital statistics we calculate that P_ = .0187, so that our predictor



equation for lifetime probability for low-level atmospheric exposures



is





            P  ~ .0187  x .1005 = 1.879 x 10~3x,
             x                                '



                                             3

where x is lifetime average exposure in ugm/m .





Lemen et al. (1976)



        The mortality observed in a cohort of 292 white males engaged in



the production of cadmium metal and cadmium compounds for at least two



years is discussed in Lemen's paper.  These workers were exposed to



cadmium fumes and dust during normal smelter operations.

-------
                                                               DRAFT
                                                      M NOT QUOTE OR CITt
        Ideally the entire exposure pattern for  each  of  the workers, in
conjunction with age-specific expected and observed death rates, would
be utilized for a complete analysis.  Unfortunately,  the only  data
available from the paper regarding exposure durations are for  four
individuals who died from prostate cancer.  Under the assumption that
this is a random sample from the entire cohort we calculate  the lifetime
average exposure as indicated in table 5.   The air exposure  levels  in
each of the time periods is approximated by utilizing information
obtained from industrial hygiene studies conducted in 1947 and 1973.
        It was found that the cohort with 20 or  more  years since first
exposure had 4 observed prostate deaths where only .88 would be
expected.  However, two biasing factors exist that tend  to make the
expected number larger than its best estimate:
        (1)  A U.S. white male death rate was used, while it  is well
    known that factory workers in demanding jobs tend to have  death
    rates lower than the U.S. average.  To correct for this  factor, we
    note that our observed-to-expected ratio in  this  cohort  for total
    nonmalignant neoplasm deaths is f   65/81.8. = .79,  which, assuming
    the same ratio of general health regarding prostate  cancer, would
    give a 21% reduction in the .88 expected cases.
        (2)  A total of 20, or 8% of the cohort, whose vital  status was
    not known were assumed by Lemen to be alive  and were included  in the
    calculations of the expected deaths.  Since  they  could not con-
    tribute to observed deaths, they can be eliminated by assuming  they
    were of the same age structure as the rest of the cohort.   This
    would result in a reduction of 8% in the expected number of cases.

-------
                              -32-
        Accounting for these two factors we calculate that an unbiased



estimate of the standard mortality ratio, R, is





            R   4/(.88 x .79 x .92) = 6.254 .





        Under the assumptions of a linear dose response and equivalence



of part and total lifetime risk, it can be shown that an estimate of



the percentage change in the lifetime probability of contracting a



disease is approximately





            (B/a) = (R - l)/x




                       3

per change in one ygm/m  of lifetime exposure, where x is the average



lifetime exposure in ygm/m .  In this case





            (B/ct) = (6.254 - 1)/56.1 = .0937, or 9.4%.





        The predictor equation in this case for the lifetime probability



of prostate cancer is





        P  ~ .0187 x .Q937x = 1.752 x 10~3x,
         X




where x is the lifetime average exposure to atmospheric cadmium in



Vgm/m  .





            Risk to U.S. Population from Different Sources

                 Releasing Cadmium into the Atmosphere



        The 1978 EEAI document prepared for EPA estimated the human



exposure to cadmium for selected major emission sources.  These data



were summarized  in table 1 of that document, which listed average



exposures from different emission sources and the number of people



exposed from each source.

-------
                               33-
        In table 6 this data is shown along with the average lifetime



probabilities of prostate cancer by source and the expected number of



cases per year calculated from the equations:





            P  = 1.879 x 10~6x
             x
and
            E(D) ~  [1.879 x 10 6  T 70.9]xN = 2.65 x 10 8xN
where,  to correspond  to  the units of exposure"given, x is lifetime aver-



age exposure  in  ng^/m ,  and N  is the number of males exposed.  Based on



1973 vital  statistics, males make up 48.7% of the U.S. population.



Assuming  that exposure is  independent  of  sex, multiplying EEAI's  esti-



mate  of total people exposed by .487  gives  the  number  of males  exposed.



         It  is estimated  that about  5  cases  of prostate cancer would  be



caused by atmospheric cadmium  each  year,  with about 87%  of  that total



due  to municipal incinerators.  This  amounts  to approximately .03% of



the  total yearly deaths  due to prostate cancer,  based  on 1973 vital



statistics.  To place this risk in  perspective,  a comparison using



EEAI's Sources of Atmospheric  Cadmium (p. 16)  can be developed:  based



on a retention rate of  .07 ugm of cadmium per  cigarette smoked, the



 average risk to a person exposed to cadmium from municipal  incinerators



 is equivalent to smoking 1 cigarette every other day.



         If we had used  the estimates derived from Lemen's data rather



 than Potts', all the calculated probabilities  and expected  numbers of



 deaths would have been  reduced by a factor of  only 1.07.

-------
                                                TABLE 1
                                EXPOSURES OVER TIME AND CALCULATION OF
                                 LIFETIME AVERAGE EXPOSURE TO CADMIUM
Age
.Interval
50-54
55-59
60-64

65-70


75-80

Individual ' s
Age at Death
53
59
63
65
65
65

75
78

Working Years of Exposure to Cadmium
1920-1949
(Average Exposure
Level: 7.4 mg/m3)
21
19
21
20
25
23

14
18

1950-1955
(Average Exposure
Level: .5 mg/nr)
6
6
6
6
6
6

0
0

1956-1964
(Average Exposure
Level: .1 mg/m )
8
9
7
5
6
9

0
0

Lifetime Average
Exposure in
pgm/m
660. 8b
262.9
555.6
512.8
638.3
589.3
x = 580.1
303.8
375.7
x = 339.8
    3Average exposure weighted by number of years exposed x (fraction of working year exposed = .22) v
age at death x 103 = lifetime average exposure in ngm/m3.
                                                                   f\
    bExample:  660.8 = (7.4 x 21 + .5 * 6 + .1 * 8) < .22 :  53 x 10 .

-------
                                        TABLE 2
                  DATA USED TO ESTIMATE EFFECT OF CADMIUM EXPOSURE ON
                           INCREASE IN U.S. PROSTATE CANCER
Age Interval
50-54
55-59
60-64
65-69
75-79
Potts' Data
Total Deaths
1
1
1
3
2
Deaths Due to
Prostate Cancer
0
0
0
2
1
P-j = Proportional
Mortality Ratio
0
0
0
.6667
.5000
1973 U.S. Male
Death Rates < 10^
a .
J
Total
1,103.0
1,779.0
2,695.5
3,933.5
8,813.5
a2j
Prostate
Cancer3
5.1
13.6
33.7
72.0
246.0
     ICD 180-187, malignant neoplasms of genital organs, x .942, where
             = 18.830
        •*     19,864

is the fraction of total number of deaths classified 180-187 which are attributed to
ICD 185, malignant neoplasm of prostate, based on 1973 U.S.  mortality data.

-------
                 -36-
             TABLE 3

DATA UTILIZED TO COMPUTE SLOPE OF
      DOSE RESPONSE CURVE g
Age Interval
50-54
55-59
60-64
65-69
75-79
Number of
Deaths in Interval = W.
J
1
1
1
3
2
Lifetime Average
Exposure = x .
660.8
262.9
555.6
580.1
339.8
Y.
J
-1.
-1.
-1.
106.3
33.8
RESULT:  0 =
^    206,485.1
 2   2,054,940.5
                               = .1005

-------
                                                TABLE 4
                                 GOODNESS OF FIT IN x  SENSE OF MODEL
Age Interval
50-54
55-59
60-64
65-69
75-79
Observed Prostate
Cancer Deaths
0
0
0
2
1
Under Assumed Model
Expected Prostate
Cancer Deaths
.238
.174
.418
1.575
1.005
2
X Deviation
.238
.174
.418
.115
.000
Under No-Ef fect-of7Exposure
Assumption7
Expected Prostate
Cancer Deaths
.375
.375
.375
1.125
.75
2
X Deviation
.375
.375
.375
.7777
.08333
                                                                                                             -J
                                                                                                              I
Total
.945
1.986
     e = np =
                 .
                       .1005x.)
               a. + .1005a2>x.
   tt
     e = n x 3/8
                          RESULT:  % reduction in x  is
                                                        1-98
          = 52.42%

-------
                                TABLE  5
            EXPOSURES OVER TIME AND CALCULATION OF LIFETIME

                      AVERAGE EXPOSURE TO CADMIUM
Individual ' s
Age at Death
71
77
79
64
Working Years of Exposure
to Cadmiuma
1925-1947
(Average Exposure
Level: 1.5 mg/m3)
4
5
17
17
1948-1973
(Average Exposure
Level: 1.0 mg/m )
0
8
1
0

Lifetime Average
Exposure in
ygm/m3
18.6
44. 3C
73.8
87.7
                                                   Average = 56.1
     Based on Lemen et al., table 4.


     Average exposure weighted by number of years exposed x (fraction
of working year exposed = .22) T age at death x 1Q~3 = lifetime average
exposure in ygm/m .


    CExample:  44.3 = (5 x 1.5 + 8 x 1.0)  x .22 T 77 x 1Q3.
                                                                                                   oo
                                                                                                   I

-------
 I
01
to
         s
                                                             TABLE 6


                                   EXPECTED NUMBER OF PROSTATE CANCER CASES PER YEAR DUE TO

                                        DIFFERENT SOURCES OF ATMOSPHERIC CADMIUM EXPOSURE
Source
Secondary Copper
i
Secondary Zinc
Municipal Incinerators
Primary Zinc
Primary Lead
Primary Copper
Primary Cadmium
Iron and Steel
Average Exposure
ngm/m^
1.54
.47
7.16
10.
10.
10.
10.
1.9
Lifetime Probability of
Prostate Cancer due
to Cadmium
2.90 x 10~6
8.84 x 10~7
1.35 x io~5
1.88 x 10~5
1.88 x io~5
1.88 < 10~5
1.88 x io~5
3.58 x 10~6
Expected Number of Prostate
Cancer Deaths per Year
due to Cadmium
.20
-
4.53
.01
-
.01
-
.49
                     Total
                                                                                                   5.24

-------
                                 -40-
                              APPENDIX I
          DERIVATION OF PROBABILITY OF THREE OR MORE PROSTATE
               CANCER DEATHS IN U.S. AGE-MATC1ED SAMPLE
        In general, if we wish to calculate the probability of k or more

deaths of a specific type from a series of deaths from m different age

groups, we proceed as follows:
Let     N.  = Number of deaths in jth age group

               m
        N   =  £ N. = Total number of deaths in population of interest
        r.  = Number of deaths in jth age group of type of interest

               m
        r   =  £ r • = Total number of deaths of type of interest
              3=1 J

        o.  = Age-specific death rate for jth age group

        a7 . = Age-specific death rate for jth age group for cause of

              interest

        P.  = a9./a. = Probability, given that an individual in the jth

                       age group died, that death was due to the cause

                       of interest.

Then
                             m
             
-------
                                 -41-
to the cause of interest, where v is all possible combinations of the
r. that sum to r.  It follows that
 3
                        N              k-1
            P(r >_ k) =  I (v,r)
                       r=k             r=0
equals the probability of k or more deaths  due  to  the cause of  interest.


        From section VI, table 2, we  calculate  the P. for prostate death


rates and show  them in appendix  table 1,  along  with  all  the possible


combinations that  could result in two or  less prostate cancer deaths.


For each combination,  4>(v,r)  r = 0,1,2 is calculated where v =  18.  The


probability of  r >^ 3 is simply 1 minus the sum  of  all  these probabili-


ties, or P(r  >_ 3)  = l-(.872032  +  .120665 + .007087) =  .000216.

-------
                 APPENDIX TABLE 1
PROBABILITY OF THREE OR MORE PROSTATE CANCER DEATHS
    OUT OF EIGHT DEATHS IN SPECIFIED AGE GROUPS
          BASED ON 1973 U.S. DEATH RATES



Age
Interval
50-54
55-59
60-64
65-69
75-79


Number
of
Deaths

1
1
1
3
2

Probability
Death Was Due
to Prostate
Cancer
Pj
.004624
.007645
.012502
.018304
.027912

Total Prostate Cancer Deaths r =





RESULT:
P(r > 3) - 1 - (.872032+ .120665

+ .007087)

= .000216
All Possible Combinations That Result in
Two or Less Prostate Cancer Deaths
rj

r=0
0
0
0
0
0
4
o
4
00
to
0
to

^

•
oo
hJ
o
LO
to

r=l
10000
01000
00100
00010
00001
44444
-
44444
o o o o o
O O H-" -C- <-n
f- &* )-> (X O
O ~J O *J O
M 00 O OO OO


'

to
o
Ul

r=2
111100000000
100011100000
010010011000
001001010120
000100101102
44,4.4444,4.4.414.
N)
44-^44-4-44-4-444
oooooooooooo
oooooooooooo
OOOOOOOOOMOO
Ijj {j^ [^J (j^ QQ "vj QQ |— I Jjj ^^ [\^j E__l
f— i i—1 *^j O-? O"i O^ O^ 00 •P'* H^ o^ VO

^


b
o
-J
o
OO
~J











n
O hH
l~tl tj
o M o
rt ^ sf

i P. i±

g.^
flj ^ •nd
JU i i f^
S 3 S3!
rt O cr'
re n M M

-------
                                  -43-
                                 *


 .REFERENCES                                             WJ NOT (JLKjf t UK Ul 1 1

  1. Berg, J.W. and Burbank,  P.,  "Correlations Between Carcinogenic
     Trace Metals in Water Surolies  and Cancer Mortality". Ann.  N.Y.
     Acad. Sci. 199,, 249-261  (1972) .


  2. Energy  and Environmental Analysis, Inc.  "Atmospheric  Cadmium-
     Population Exposure Analysis."  Draft report prepared for  U 5
     Environmental  Protection Agency  17 March  1978.


  3  Energy  and Environmental Analysis, Inc.   "Sources of Atmospheric
     Cadmium. "  Draft  reported prepared for U.S.  Environmental  Protection
     Agency,  27 February 1978.


  4 .  Environmental Protection Agency,  "Scientific  and  Technical
     Assessment Report en Cadmium", EPA-600,/6-75-003,  (Marcn 1975).

  5.  Food and Drug Administration Review, Mutagenicity of Cadmium
     (May 31,  1977).

  6.  Fleischer, M. ,  A. F. Sarofim, D.W. Fassett, P. Hammond, E. T.
     Shacklette, I.  C. T. Nisbet, ard S. Epstein,  " Envirorer-ental
     Impact o£ Cadmium:  A  Review bv the Panel on  Hazardous Trace
     Substances" , . Environmental Healtn Perspectives, 7_, 253-323, (1974).

  7.  Friberg,  L. , Piscatcs, M. , and Nordberg, G.  "Cadmium in the
     Environment" Chemical  Rubber Press, Cleveland (1971).

  8.  International Agency  for Research on Cancer,  "IARC Mcnccrachs
     on the Evaluations  of  Carcinogenic Risk of Chemicals to Man" ,
     U, 39-74 (1976).

  9.  Kipling,  M.D.  and J. A. H. Watsrhouse, "Cadmium and Frostatic
     Carcincma" , Lancet, 730-731, (April 1,1967).

 10.  Knorre,  Von P.," Introduction of  interstitial cell tumors bv
     caamiap  chloride in albino rats1', Arch. Geschwulstforscn, 3S:
     257-263, 1971.
                   -
11.  Lemen, R.A.;0^.  d.  Lee; J. K. Wagoner, H. P. Blejer,  "Cancer
     Mortality Among  Cadmidn Production Workers", Ann. N.  Y. Acad.
     Science, 271,  273-279,  (1976).

.12.  Levy, L. S., Clack , J., "Further Studies on the Effect of Caesium
     on the Prostate  Gland-I", Ann. Occup. Hyg. 17, 205-211,  (1975).

13.  Levy, L. S., Clack, J., and F. J. C. Roe, "Further Studies 'oh  the
     Effect of Cadmium on  the Prostate Glard-IIrt, Ann. Occup.  Eyg.  17
     213-220, (1975).

-------
                                 -44-
 14.    Lucis, Q.J., R. Lucis and K.  Aterman.  "Tmorcenesis bv cadmium",
       Oncology, 2.6:  53-67. 1972

 15.   McMichael, A. J., Andjelkovic,  D.A.  and'Tyrcler, H. A. , ^Cancer
      Mortality Among Rubber Wbrks:   An Ecicemiolcgic Study", Ann.  N.Y.
      Acad. Sci. 271,  124-137  (1976).

 16.  National Cancer Institute  "Cancer Risks and Sates",  (1974)

 17.  Pertel, R. Memo Mutagenicity;  Cadmium  to  Richard Troast,  EPA
     June 30, 1977.

 18.  Potts, C. L., "Cadmium Proteinuria-The Health of Battery Workers
     Exposed to Cadmium Oxide Dust",  Ann. Occup. Hyg.,  8_, 55-61, (1965).

 19.  Reddyr J., D. Svobcda, D.  Azarnoff  and R. Dawar, Cadmium  induced
            cell tumors of rat  testis: morphologic and  cytochemical
     study. J. nat. Cancer Insti. 51:  891-903, 1973.

20.  Sandu, S. Memo (Mutagenic  effects of Cadmium and Validations of
     papers of cadmium)  to Michael  D.  Waters,  EPA. (March 25,  1977)

21.  Schroeder, H.A.,  Balassa,  J. J.,  and Vinton, Jr.,  W. H.,  "Chromium
     Cadmium, and Lead in Rats;  Effects  on  Life Spanr Tumors and Tissue
     Levels", J. Nutrition 86_,  51-66  (1965).

22.  Skog, E. and Wahlberg, J.E.,  "A  Comparative Investigation of the
     Percutaneous Absorption of Metal Compounds in the  Guinea  Pig By
     Means of the Radioactive Isotopes,  "The Journal of Investigative
     Dermatology, 43;  187-192,  1964.

23.  Zyka, V. "Geochemical Environment and  Incidence of Cancer in Kutna
     Hora", Collection of Geological  Sciences, Technology and  Geo-
     chemistry, 11, 155-181, (1973).

-------