PB91-225631
Drinking Water Criteria Document on Radium (Final Draft)
Life Systems, Inc., Cleveland. OH
Prepared for:
Environmental Protection Agency, Washington, DC
14 Jun 91
-------
PB91-225631
TR-1242-85
FINAL DRAFT FOR THE DRINKING HATER
CRITERIA DOCUMENT ON RADIUM
Prepared Under
ICAIR Program No. 1524
for
EPA Contract 68-C8-003303-3279
ERG Subcontract No. LSI-8700
ERG Work Assignment No. 2-05
Life Systems, Inc. Work Assignment No. 251524
for
Drinking Water Standards Division
Office of Ground Water and Drinking Water
Office cf Water
Environmental Protection Agency
and
Bioeffects Analysis Branch
Analysis and Support Division
Office of Radiation Programs
Office of Air and Radiation
Environmental Protection Agency
June 14, 1991
-------
TECHNICAL REPORT DATA
(Heat rtad Inunctions on ifie rtvtne btfore comfit:
1. REPORT NO. 2.
tp-i
x PB91-225631
4. title ano subtitle
Final Draft For The Drinking Water Criteria Document
On Radium
9 REPORT DATE
]/. 1991
6 ȣRFORMlN'G ORGANIZATION CODE
7. AUTHQRISI
Life Systems (Fanny Ennever)
B. »ERFORMING ORGANIZATION REPORT NC.
TO-1242-85
9. PERFORMING ORGANIZATION NAME ANO AOORESS
Drinking Water Standards Div.: OGWDW: EPA
WH550D
401 M St. , SW
Washington. DC 20460
10. PROGRAM ELEMENT NO
ICAIR Proeram No. 1524
11 CONTRACT/GRANT MO.
EPA Contract
68-C8-003303-3279 !
2. SPONSORING AGENCY NAME ANO AOORESS
Drinking Water Standards Division: OGWDW: EPA
WH550D
401 M St. , SW
Washington. DC 20460
13. Tvpg OF REPORT ANO 9E«iOO COV£»£:
Haalph Cr-A rarl a
14. SPONSORING AGENCV CODE
OGWDW
5. supplementary notes
Reviewed by Greg Helms, Neal Nelson, and Jerry Puskin
.abstract
V/This document provides the health effects basis to be considered in establishing
the MC!G for radium. To achieve this objective, data on pharmacokinetics, human
exposure, acute and chronic toxicity to animals and humans, epidemology and the
mechanisms of toxicity vere evaluated. Specific emphasis is placed on literature
data providing dose-response informantion. Thus, while the literature search and
evaluation performed in support of this document was comprehensive, only the reports
considered most pertinent in the derivation of the MCLG are cited in this document.
The comprehensive literature search in support of the analysis of health effects,
exposure and occurrence in this document includes information published up to
January, 1991; however, more recent information may have been added during the
review process.
J
16
17.
KEY WORDS ANC DOCUMENT ANALYSIS
a. DESCRIPTORS
b.lOENTIFlERS/OPE\ sNDEO TERMS
c. CCSati FielJ'Croj;
health criteria document
pharmacokinetics
dose-response
mechanisms of toxicity
MCLG
18. DISTRIBUTION STATEMENT
unlimited
19. SECURITY CLASS "'hi Rtpw.
21. NO OF »Aij£S
186
20. securi ty class r:m tat.
:: po.ci j
I
EPA Pom 2220-1 (R«v. 4-77) PHCviOUS soition il oeiOLft
-------
FOREWORD
Section 2412 (b) (3) (A) of the Safe Drinking water Act, as amended in 1986,
requires the Administrator of the Environmental Protection Agency to publish
maximum contaminant level goals (KCLCo) and promulgate National Primary
Drinking Water Regulations for each contairir.ar.t, which, Lr. the Judgemert of
the Administrator, may have an adverse effect on public health end which is
known or anticipated to occur in public water systems. The HCLG is
nonenforceable and is set at a level at which no known or anticipated adverse
health effects in humans occur and which allows for an adequate margin of
safety. Factors considered in setting the MCIG include health effects data
and sources of exposure other than drinking water.
This document provides the health effects basis to be considered in
establishing the MCLG for radon. To achieve this objective, data on
pharmacokinetics, human exposure, acute and chronic toxicity to animals and
humans, epidemiology and mechanisms of toxicity were evaluated. Specific
emphasis is placed on literature data providing dose-response information.
Thus, while the literature search and evaluation performed in support of this
document was comprehensive, only the reports considered most pertinent in the
derivation of the MCLG are cited in this document. The comprehensive
literature search in support of the analysis of health effects, exposure and
occurrence in this document includes information published up to January,
1991; however, more recent information may have been added during the review
process.
When adequate health effects data exist, Health Advisory values for less than
lifetime exposure {One-day, Ten-day and Longer-term, approximately 10* of an
individual's lifetime) are included in this document. These values are not
used in setting the MCT-G, but serve as informal guidance to municipalities and
other organizations whim emergency spills or contamination situations occur.
James R. Elder
Director
Office of Groundwater and Drinking Water
Office of Water
-------
AUTHORS, CONTRIBUTORS AND REVIEWERS
This document was prepared under a contract with the Office of water,
U.S. Environmental Protection Agency, Washington DC. Primary Agency
contributors and reviewers were Greg Helms (Drinking Water Standards Division,
office of Drinking Water and Ground Water), Neal Nelson (Office of Radiation
Programs) and Jerry Puskin (Office of Radiation Programs).
The following Life Systems, Inc. personnel were involved in the
preparation of this document: Fanny K. Ennver (Principal Author), Betty F.
Neustadter (Contributing Author), William J. Brattin (Reviewer), Joyce K.
Donohue (Program Manager) and Greg E. Schiefer (Task Manager).
-------
TABLE OF CONTENTS
PAGE
LIST OF FIGURES Ui
LIST OF TA31SS iii
I. SUMMARY 5-1
:i. PHYSICAL AND CHEMICAL PROPERTIES II-l
A. Chemical Properties - Elemental II-1
B. Physical Properties XI-3
C. Occurrence II-8
0. Summary II-9
III. TOXICOKINETICS III-l
A. Absorption ......... . XIZ-1
B. Distribution III-2
C. Excretion 111-5
D. Bioaccumulation and Retention III-5
E. Summary 111-12
IV. HUMAN EXPOSURE IV-1
A. Naturally Occurring Levels . IV-1
1. Radium in Water Supplies IV-1
2. Radium in Other Media IV-5
B. Other Sources . 1V-8
C. Sum/nary IV-8
V. HEALTH EFFECTS IN ANIMALS V-l
A. Short-Term Exposure . V-l
B. Longer-Term Exposure v-3
c. Reproductive and Developmental Effects ..... v-3
D. Mutagenicity V-3
E. Carcinogenicity v-3
F. Summary v~~
VI. HEALTH EFFECTS IN HUMANS VI-i
A. Clinical Case Studies Vl-l
B. Epidemiologic Studies VI-2
1. Radium-Dial Painters ...... VI-2
2. Patients Injected with Ra-224 During Medical Treatment VI-12
3. Community Studies VI-16
i
-------
'labia of Contents - continued
PACE
C. High-Risk Populations VI-17
0. Summary vi-18
VII. MECHANISM OF TOXICITY VII-1
A. Effects of Ionizing Energy VII-1
B. Radionuclide Dosimetry vn-3
C. Summary VI1-6
VIII. QUANTIFICATION OF TOXICOLOGICAL EFFECTS VIII-1
A. Noncarcinogenic Effects vni-7
1. One-Day Health Advisory VIII-9
2. Ten-Day Health Advisory VIII-10
3. Longer-Term Health Advisory VIII-11
4. Reference Dose and Drinking Water Equivalent Level . . VIII-13
B. Carcinogenic Effects vill-15
1. Categorization of Carcinogenic Potential VIII-15
2. Quantification of Carcinogenic Risk VIII-15
C. Genetic and Developmental Effects VIII-38
D. Summary VIII-42
IX. UNCERTAINTY ANALYSIS IX-1
A. Range of Assumptions and Models IX-1
1. Noncancer Effects IX-1
2. Cancer Risk IX-3
3. Genetic and Developmental Effects IX-11
B. Parameter Variability IX-13
1. Noncancer Effects IX-13
2. Cancer Risk IX-15
3. Genetic and Developmental Effects IX-20
C. Summary IX-21
X. REFERENCES X-l
ii
-------
LIST OF FICVRES
PAGE
II-l Radioactive Decay Series Showing Precursors and Decay Products of
Radium-226 II-4
II-2 Radioactive Decay Series Showing Precursors and Decay Products of
Radium-223 and Radium-224 ................... II-5
II-3 Radioactive Decay Series Showing Precursors and Decay Products of
Radium-223 II-6
LIST OF TABLES
PAGE
11*1 Physical and Chemical Properties of Selected Radium
Compounds II-2
III-l Retention Parameters for Radioisotopes in the Radium
Decay Series 111-10
IV-l Estimated Cumulative National Occurrence of Radium-226 and
Radium-228 in Community Ground Hater Supplies Based on a
Censored Lognormal Distribution Model IV-3
IV-2 Estimated Cumulative National Occurrence of Radiuci-226 and
Radium-228 in Nontransient Non-Community Ground Water Supplies
Based on a Censored Lognormal Distribution Model IV-4
IV-3 Estimated Cumulative Population (in thousands) Exposed to
Radium-226 and Radium-228 in Drinking Hater Exceeding the
indicated Concentrations from Community Ground Water Supplies IV-6
IV-4 Estimated Cumulative Population (in thousands) Exposed to
Radium-226 and Radium-228 in Drinking Hater Exceeding the
Indicated Concentration Range from Nontransient Non-Community
Ground Water Supplies IV-7
VI-1 Bone Sarcoma Rates among Female Radium Dial Painters
First Employed before 1930 VI-8
vi-2 Head Carcinoma Rates among Female Radium Dial Painters
First Employed before 1930 VI-9
VI-3 Bone Sarcoma Rates among Female Radium Dial Painters
Excluding Workers Possibly Measured Due to Symptoms of
Bone Sarcoma VI-10
VT-4 Bone Sarcomas Rates among Patients Injected with Ra-224 . . VI-14
VIII-1 Observed and Predicted Bone Sarcomas per 1,000 Person-Years
at Risk among Female Radium Dial Painters Excluding Workers
Possibly Measured Due to Symptoms of Bone Sarcoma ..... VIII-22
VIII-2 Risk Estimates for Ra-226 and Ra-228 Based on Linear or
Dose-Squared Extrapolation of Bone Sarcomas and Linear
Extrapolation of Ra-226 Induced Head Carcinomas VIII-26
VIII-3 Annual Organ Doses in the 70th Xear Predicted by the
RABRISK Model from Ingestion of 1 pCi/year of Radium
Isotopes . VIII-28
VIII-4 Organ-Specific Lifetime Cancer Risks Used in the RADRISK
Model from High-LET and Low-LET Irradiation ... VIII-30
VIII-5 Lifetime Cancer Risks from Lifetime Ingestion of 1 pCi/L
Ra-226 or Ra-228 in Drinking Water VIII-34
-------
VIII-6 Drinking Water Concentration# of Ra-226 and Ra-228
Corresponding tc Various Risk Levels Predicted by the
Adjusted RADRISK Model VIII-36
VI11-7 Radiation Risk Factors for Ger.etic and Developmental
Effects VIII-39
VTII-8 30-Year Cumulative Doses to Germ Cells Predicted by the
RADRISK Model from Ingestion of 1 pCi/year of Radium
Isotopes VXXI-40
VIII-9 Germ Cell Mutation Risks from Lifetime Ingestion of
1 pCi/L Ra-226 or Ra-228 in Drinking Water VIH-41
viJi-10 Summary of Quantification of Toxicological Effects for
Radium ........ VIil-43
iv
-------
I. SUMMARY
Radium is a naturally occurring radioactive element that arises from the
decay of primordial uranium and thorium in the earth's crust. The occurrence
of radium in drinking water supplies is governed by the physical and chemical
behavior of uranium, thorium and radium in aquifers and surficial deposits
(Hess et al. 1985). Radium-226 (Ra-226) and radium-228 (Ra-228) are the
isotopes of primary environmental concern. Average population-weighted U.S.
drinking-water concentrations are 0.9 pCi Ra-226/L and 1.4 pCi Ra-228/L
(Longtin 1968, USEPA 1988).
When humans ingest radium, about 20% is absorbed into the circulation
(Maletskos et al. 1966). Radium initially distributes to soft tissues and
bone, but preferentially accumulates in growing bone. Normal processes of
bone remodelling release radium, with an estimated biological half-life of
10 years (Norris et al. 1955, Hrenn et al. 1985a). Excretion is primarily
through the feces (Maletskos et al. 1966). A model for the bioaecumulation
and retention of radium and its decay products, based on the ICRP20 model for
alkaline earth metabolism (ICRP 1973), has been developed as part of the
RADRISX model to predict the consequences of lifetime intake of -radium
(Dunning et al. 1980, Sullivan et al. 1981). The general toxicokinetics of
radium are well-established, based on both human and animal data, but the
exact values of absorption, retention and excretion may vary from individual
to individual, and in a given individual over time. The RADRISK metabolic
model estimates the typical behavior of radium in the human body, but may not
be accurate for every member of the population (USEPA 1989b).
1-1
-------
The occurrence of radiur. grcu.-.i water supplies follows an
approximately I0q-r.0rr.4l distribution, with a mean value of less than 1 pci/L
(USEPA 1988, Longtin 1990). These data can be used to estimate population
exposure from community water supplies and can be extrapolated to estimate
exposure from nontransient non-community water supplies (USEPA 1990a,b).
About 3,000,000 people are exposed to drinking wat«r containing more than
5 pCi Ra-226/L in community water supplies and 48,000 are exposed from
nontransient non-community water supplies. Corresponding numbers for Ra-228
are 1,000,000 and 32,000 people (USEPA 1990a,b). It is estimated that a total
of 890,000 people are exposed to more than 20 pCi Ra-226/L and 164,000 are
exposed to more than 20 pCi Ra-228/L (USEPA 1991a). Some fraction of this
population may be exposed to these levels of both isotopes, as indicated by
co-occurrence analysis (USEPA 1990a,b). Food is another source of radium
exposure, i^nounting to an average of about 1 pCi/day, and often contributes
more than half the daily radium intake (USEPA 1990a,b).
Most studies of health effects in experimental animals following
injection of radium have investigated effects on bone, due to the preferential
accumulation and long-term retention of radium in the skeleton, and subsequent
damage due to radioactive decay. These effects have been demonstrated with
Ra-224, Ra-226 and Ra-228. Effects on bone found at relatively short times
after radium exposure are changes in bone structure {Jee et al. 1969, Momeni
et al. 1976) or hematopoiesis (Schoeters and Vanderborght 1981). When animals
are followed for a considerably fraction of their lifetimes, bone sarcomas are
usually found (Raabe et al. 19S1, Humphreys et al. 1985, Mays et al. 1987).
Such tumors have been observed following a single exposure to radium (Taylor
et al. 1983, Kofranek et al. 1985). Leukemias or lymphomas have been seen
following injection of Ra-224, but the occurrence peaks at low doses, lower
1-2
-------
than the doses that cause osteosarcomas {Humphreys et al. 1985, Mulier et al.
1988). The results of animal studies confirm the association of bone necrosis
and bone sarcomas with radium exposure in humans, and provide support for an
approximately linear relationship between dose and incidence of bone sarcomas
in animals (Wronn et al. 1985b, Mays et al. 1987).
The experience of the radium dial painters, who ingested substantial
amounts of Ra-226 and/or Ra-228, clearly establishes that ingestion of Ra-226
or Ra-228 can cause bone necrosis (Keane et al. 196.3), that ingestion of
Ra-226 can cause bone sarcomas and head carcinomas, and that ingestion of
Ra-228 can cause bone sarcomas (Rowland et al. 1978). Bone is a target organ
of radium toxicity because of the preferential accumulation of radium in bone.
Head carcinomas are believed to result from the accumulation of radon gas, a
decay product of radium, in the cranial sinuses. Head carcinomas do not occur
following ingestion of Ra-228 because its rador. decay product has too short a
half life for substantial accumulation to occur (Rowland et al. 1978). Rates
of other cancers are not substantially elevated among radium dial painters
(Stebbings et al. 1984). Studies of patients injected with Ra-224 show
induction of bone sarcoma but not head carcinomas (Mays and Spiess 1984). In
addition, an elevated rate of breast ar.d liver cancer is found among these
patients (Spiess et al. 1989).
Three studies of populations in the United States exposed to radium in
drinking water have found elevated rates of bor.e cancer (Petersen et al.
1966), blazer, lung and breast cancer (Bean et al. 1982) or leukemia (Lyman
et al. 1985). However, the differing types of cancers that are increased
among the three studies makes attributing these effects to radium
questionable.
1-3
-------
The harmful effects of ionizing radiation are believed to be the result
of damage to tWA and other cellular components caused by chemical reaction
with ionized molecules, particular!/ active oxygen species (Bacq and Alexander
1961, Andrews 1961). The dose of radiation resulting from ingestion of radium
varies from organ to organ. Models have been developed to calculate the rate
at which radiation is absorbed in a given organ, based on the concentration of
radioactive elements in that organ and in surrounding organs, the frequency
with which each element decays, the energy of each decay event and the
fraction of emitted energy that is absorbed in the particular organ (USEPA
1989b).
Quantification of the noncancer health effects of radium ingestion is
based on the dose-response relationship observed by Keane et al. (1933) for
bone necrosis in radium dial painters. Keane et al. (1983) found that below a
total intake to the blood (absorbed dose) of 10 ftCi of either Ra-226 or
Ra-228, observed bone changes were not significantly different from controls.
Reference doses (RfDs) of 4 pCi Ra-226/kg-day and 3 pCi Ra-228/kg-day for
adults were derived from this threshold assuming that average gastrointestinal
absorption is 20* (MaletsJcos et al. 1966), that bone necrosis is a function of
radium intake scaled by body weight, and that the dosimetric adjustment
between instantaneous and lifetime intake is 1.6 for Ra-226 and 1.2 for Ra-228
(Keane et al. 1983). It is very unlikely that radium exposure in drinking
water would result in doses above these RfDs.
Two basic approaches may be used to quantify the carcinogenic effects of
ingestion of Ra-226 and Ra-228 in drinking water: fitting dose-response
models to the incidence of bone and head cancers amorg radium dial painters
exposed to Ra-226 and/or Ra-228 {May3 et al. 1985a, N*S 1986) and using the
1-4
-------
Agency's RADRXSX model eo predict she do»» of isnising radiation deliver ad to
radiosensitive organs by intake of radium and calculating cancer risk based on
a synthesis of human epidemiologic data on the carcinogenic potency of
radiation {USEPA 1969b). Bone sarcoma incidence among dial painters shows a
highly non-linear dose-response relationship (Rowland et al. 1976, HAS 198S).
Using an upper-bound linearized risk coefficient, Kays et al. (198Sa)
calculated risks of 2.1E-5 for Ra-226 and 2.2E-5 for Ra-228 for an intake to
blood of 1 pci/day (equivalent to a drinking water concentration of 2.5 pCi/L,
assuming 20% gastrointestinal absorption and 2 I./day water consumption). The
risks for radium isotopes predicted by the RAORXSK model for lifetime
ingestion of 2 pCl/L in water are 6.1E-6 for Ra-226 and 5.1E-6 for Ra-228 for
total cancers, and 4.8E-6 for Ra-226 and 3.6E-6 for Ra-228 for fatal cancers
(USEPA 1989b). On the advice of the Radiation Advisory Subcommittee of the
USEPA Science Advisory Board (SAB 1990), the RADRISK predictions were
reassessed. Adjustments were made to the predicted rate of leukemias for both
Ra-226 and Ra-228, to the predicted rate of head carcinomas for Ra-226 and to
the predicted rate of bone sarcomas and leukemias for Ra-228. The risks for
radium isotopes predicted by the adjusted RADRISK model for lifetime ingestion
of 1 pCi/L in water are 5.8E-6 for Ra-226 and 5.3E-6 for Ra-228 for total
cancers, and 4.4E-6 for Ra-226 and 3.8E-6 for Ra-228 for fatal cancers.
The Agency considers the RAORISK model to provide the beBt estimate of
cancer risk at low intakes of radium, primarily because of the uncertainty in
using the dial painter data to derive a linear risk coefficient for bone
cancer induction, and the existence of the alternative of using a linear risk
coefficient for Ra-224 derived from epidemiologic data that has been
recommended by two BEIR committees (NAS 1980b, 1988). The dial painter and
other epidemiologic data are used to adjust the predictions of head carcinomas
:-s
-------
and leukemics in the present version of the RADRISX model. The Agency usee
the adjusted RADRISK predictions of fatal cancers as the reference in proposed
regulations limiting radium activity in drinking water. The concentrations in
drinking water corresponding to various levels of excess lifetime risk are
displayed as follows:
Drinking water concentration (pCi/L)
Excess cancer xi.sk Ra-226 Ra-228
Total cancers
lO"4 17 19
10"5 1.7 1.9
10'® 0.17 0. i9
Fatal cancel's
10'4 22 26
10'5 2.2 2.6
10-e 0.22 0.26
Uncertainties in evaluating the health risks of radium in water supplies
were analyzed. One of the largest uncertainties is the shape of the dose-
response curve for bone cancer induction by radium. If the true response were
quadratic (dose-squared), due to a two-hit model of cancer induction (NAS
1980b) or to the promoting effects of bone necrosis (Martland 1931), true
risks could be more than an order of magnitude lower than those calculated.
For other assumptions and parameters, there does not appear to be any
systematic bias towards over- or under-prediction ot risk, except that a
protective value for adult drinking water consumption has been used, that is
higher than the mean by less than a factor of 2. The overall uncertainty
contributed by modelling assumptions and parameter variability may be
1-6
-------
approximately a factor of 5 for calculating radiation dose in target organs
and a factor of 3 for calculating organ-specific risks. These considerations
suggest that the true risk is likely to be within an order of magnitude in
either direction of those calculated for members of the general population.
1-7
-------
II. PHYSICAL AND CHEMICAL PROPERTIES
A. Chemical Properties - Elemental
The chemical properties of radium are similar to other alkaline earth
elements, particularly barium and calcium. Table II-l lists several chemical
and physical properties of radium. Radium exists only in the +2 oxidation
state in solution, and the divalent ion is not easily complexed in water (Ames
and Rai 1979). Some radium salts such as radium carbonate and radium sulfate
are very insoluble in water, but their solubility limits are higher than the
levels of radium that occur naturally in water, which are usually below
10'* mol/L (Langmuir and Hiese 1935]. Thus, the water concentration of radium
appears to be controlled by competing processes of dissolution and sorption.
There will be a large difference in the extent of dissolution of a radium atom
tightly bound in a crystalline lattice, such as in a monazite in granite,
compared to a radium atom adsorbed onto the surface of a sand grain. Sorption
can remove radium from solution, by adsorption and coprecipitation by
scavengers such as iron hydroxide and barite (barium sulfate) (Benes and
Strejc 1986, Benes et al. 1934]. Equilibrium between adsorption and
desorption is quickly established, and the partition coefficient between the
solid and liquid phases depends upon the geochemical affinity of radium
towards the rock surfaces and particulate material in the aquifer (King et al.
1982. Krishna3wami et al. 1982, Benes et al. 19Q5, 19fl6). Radium is most
mobile in aquifers with high concentrations of dissolved solids (Benes et al.
1984, 198S).
II-l
-------
T>bla 11-1 Phytic*! and Ch*alcal Pcopartlaa of S«l«ct«d Radltui Compounds
Radium
Redturn
Radium
Radium
Radium
Radium
Radius
Property
Radium
Bromide
Carbonate
Chloride
Hydroxide
lodata
Nltrste
Sulfate
Chemical foroails
Re
RaBr2
RaCOj
RaClj
Ra<0H)2
RslOj
RaNOj
RaS04
Molecular weight
226.0)
JBVBJ
286.01
296.9)
No dats
475.83
340.04
382.08
Synonyms
No data
No dies
Carbonic
No data
No dats
No dsta
Nitric acid.
Sulfurlc
acid, radius
t
radium salt
acid, radium
salt
salt
CAS number
7440-14-4
100)1-21-9
*116-98-1
10021-66-B
98966-86-0
No data
10213*12-4
7446-16-4
Color
Sllver-
White
White
Kel lovish-
No data
No data
No data
White
vhlte
vhlte
Phyalcsl i(tt«
Solid
Solid
Sol Id
Solid
No daca
Ho data
Solid
Solid
point
700*C
m*c
No data
1000'C
No data
Ho data
No data
No data
Rolling point
-------
B. Physical Properties
Radium has four naturally-occurring isotopes, radium-226 (Ra-226),
radium-228 (Ra-228), radium-224 (Ra-224) and radium-223 (Ra-223). Each is the
product of a radioactive decay series beginning with a long-lived isotope,
uranium or thorium. These isotopes undergo spontaneous radioactive decay.
resulting in che emission of a radioactive particle (alpha or beta particle)
plus radiant energy (gamma rays) and the formation of a radioisotope of a
different element. This process is repeated until a stable isotope of lead is
formed.
Figures IX-1 to IX-3 summarize the decay series of uranium and thorium
isotopes, showing the precursors and decay products of naturally-occurring
radium isotopes. The alpha or beta particles emitted during decay are
indicated, and also the half lives of the decay products. Ra-226 and Ra-228
are the isotopes of primary environmental concern, because the much shorter
half lives of Ra-224 and Ra-223 preclude substantial environmental
accumulation.
A notable difference between Ra-226 and Ra-228 is the half life of their
radon decay products (3.82S days vs. 55.6 seconds). Radon is a noble gas, and
so tends to diffuse from the site where it is formed by decay of radium (NAS
1988). The shorter-lived Rn-220, produced by Ra-228 or Ra-224, quickly decays
into chemirally-reactive elements (polonium, lead and bismuth) and so
diffusion of radioactivity is not as extensive as with Rn-222 produced by
Ra-226 (NAS 1988).
II-3
-------
Uranium-238 Series
Rn-222
3.825 d
a
Adapted from ATSDR 1989
Figure II-1 Radioactive Decay Series Showing Precursors and
Decay Products of Radium-226
II—4
-------
Thorium-232 Series
Th-232
1.4 x 10':y
a
P
!/
Th-22B
1.91 y
Ac-228
6.13 h
a
Rn-220
55.6 s
a
Figure II-2 Radioactive Decay Series Showing Precursors and
Decay Products of Radium—228 and Radium—224
11-5
-------
Uranium-235 Series
U-235
7.04 x 108y
1/
Pa-231
3.27 * 1C4y
Th-231
25.2 h
!/
TTi-227
18.7 d
Ac-227
2V8y
Rn-2t9
3.96 s
a
Po-215
1.8x 10-*s
Figure II-3 Radioactive Decay Series Showing Precursors and
Decay Products of Radium-223
II-6
-------
If radioactive decay occurs in a closed Bystem, the complete series of
isotopes will be produced, each decaying at the sane rate it is produced.
This situation, described as secular equilibrium, is achieved in a period of
time equal to about five half lives of the longest-lived isotope in the
series. Secular equilibrium is attained only when there is no Loss or
addition of an isotope in the decay series. However, in groundwater, it is
common to find a considerable degree of disequilibrium between pairs of
successive elements. Two mechanisms by which disequilibrium or iaotopic
fractionation occurs are chemical differences and alpha recoil. Chemical
differences can affect solubility and sorption, resulting in physical
separation in aquifers (Gilkeson and Cowart 1982, Hess et al. 1985). Examples
are the generation of radon, a noble gas, from radium, a solid, and the
generation of thorium, which is extremely insoluble in oxidizing environments,
from uranium, which is relatively soluble in oxidizing environments (Hess et
al. 1985, Michel 1990}. Alpha particles emitted during decay can damage the
crystal lattice and/or displace decay product atoms (the alpha recoil effect)
(Cilkeson and Cowart 1982, Michel 1990). Both of these recoil effects lead to
a decay product atom that is more prone to dissolution than the original atom
(Hess et al. 1985).
These mechanisms have an important influence on the radiochemical
behavior of radium. The immediate precursor of Ra-226 is Th-230, a highly
insoluble element with a half life of 75,000 years. When Th-230 is produced
by decay of U-234, a soluble element often found in solution, it tends to
adsorb onto the surface of aquifer solids. Dissolution of Ra-226 produced
from Th-230 occurs by desarption or alpha recoil, both of which are enhanced
in radium positioned on the surface of solids (Michel 1990). In comparison,
the thorium precursor of Ra-228, Th-232, is primordial and so will have a
11-7
-------
distribution governed solely by thorium geochemistry, and therefore, Th-232
would not be as likely to occur on surfaces of solids as Th-230 (Michel 1990).
C. Occurrence
The distribution of Ha-226 and Ra-228 in water is a function of tha
thorium and uranium content of the aquifer, the geochemieal setting of the
aquifer solids and the half-life of each isotope. Thorium and uranium have
generally similar behavior, with the important exception that thorium has one
oxidation state and is immobile at low temperatures. Therefore, thorium
distribution is controlled by primary geochemieal processes (such as magmatic
crystallization) and secondary physical processes (such as sedimentary
enrichment of thorium-bearing minerals in placer deposits} (Hess et al. 1985}.
Uranium has two oxidation states, and the +6 (uranyl) ion can form highly-
soluble complexes which can be transported long distances in oxidizing
groundwater before being removed by adsorption or reduction to the +4 state
(Hess et al. 1985). These factors account foe local variations from the
average crustal Th/U activity ratio of 1.2 to 1.5 (Hess et al. 1985). Other
processes described in the previous section cause isotopic fractionation of
uranium and thorium decay products in aquifers. This may include separation
of rh-228 from Ra-226, resulting in some difference in the distributions of
Ra-228 and Ra-224 (see Figure II-2). Overall, although there is some
correlation anong uranium, thorium, Ra-224p Ra-226 and Ra-228 contents of an
aquifer, the only way to determine the radium contents precisely is to measure
each separately.
The occurrence of Ra-226 and Ra-223 in ground/ surface and treated water
has been aaaeaacd (Rieta et al. 1937, Cecn et al. 198S, Longtin 1988, OSEPA
11-8
-------
1988). The radium content of surface waters is usually very low (Heao at al.
19S5, CTSEFA 1586). Hess et al» (1985) estimated that approximately 200-
500 public water supplies contain more than 5 pCL/L total radium after
treatment, and the mean radium activity in these supplies is about 10 pCi/L.
The National Inorganics and Radionuclides Survey> which covered nearly one
thousand drinking ¦vtatec sources in Che United states, found maximum activities
to be 15 pCi Ra-22S/L and 12 pCi Ra-228/L, with estimated population-weighted
averages of 0.9 pCi Ra-226/i and 1.4 pCi Ra-223/L (Longtin 1938, US2PA *938).
Approximately 600 systems were estimated to have more than 5 pCi Ra-226/L and
approximately S00 were estimated to have more than 5 pCi Ra-228/L (USEPA
1991a).
D. Summary
Radium is a naturally occurring radioactive element that arises from the
decay of primordial uranium and thorium in the earth's cruBt. The occurrence
of radium in drinking water supplies is governed by the physical and chemical
behavior of uranium, thorium and radium in aquifera and surficial deposits
(Hess et al. 1965). Figures II-l to TI-3 su^narire the decay series of
uranium and thorium isotopes, showing the sources and decay products of radium
isotopes. Ra-225 and Ra-228 are the isotopes of primary environmental
concern. Average populat..on-weighted U.S. drinking-water contents are 0.9 pCi
Ra-226/L and 1.4 pCi Ra-220/1 (Longtin 1988, USSPA 1938).
IX-9
-------
III. TOXICOKINETICS
A. Absorption
Maletskos et al. <1966) performed a careful study of the absorption of
radium from the gut in humans. A group of elderly, male and female volunteers
either ingested 0.1 to 4 jjCi radium as Ra-224 sulfate or were administered
0.9 vCi carrier-free fta-224 in saline intravenously. Hetabolle studies were
conducted over a period of 21 days, including measurements of blood, urine,
feces and breath samples, and of the whole body and the upper 20% of the body.
The activity of Ra-224 in blood, urine, feces and whole body following oral
and intravenous administration were assumed to be equal on the basis of
retained radium (Maletskos et al. 1966f. This allows the uptake from the gut
to the blood to be calculated from curves of blood concentration as a function
of time obtained after absorption is complete, about four days in these
experiments. The various curves derived for an individual subject gave
generally consistent estimates of gut uptake of radium, and ranged ftom 0.1 to
0.3 with an average of 0.2 in 5 subjects, with a low value of 0.01 obtained in
a sixth subject who was later determined to have achlorhydria (absence of
stomach acid) [Kaletskos et al. 1965}.
Maletskos et al. (1966) compared the average value of 0.2 to the value of
approximately 0.15 calculated from data on human body burden compared to food
or water content of Ra-226 (Stehnay and Lucas 1956). HaletsJcos et al. (1966)
also pointed out that much lower values are found in experiments on animals,
but noted that Che value of 0.3 used in 3CRP (1959) was derived from aniaa.1
experiments. They also noted chat total absorption could be at most a factor
of 5 higher than the average '.'alue of 0.2 found among their elderly
-------
volunteers. Maletskos et al. (1966) concluded that the actual absorption of
dial paint by the young female dial workers would be within a factor of 2-3 of
their measured value of 0.2. The form of radium used in these experiments was
designed to mimic dial paint,- but the observed high absorption fraction
indicated that absorption was not limited by the low solubility of radium
sulfate (Maletskos et al. 1966). Although the ingested activities of Ra-224
were relatively high, on a gram basis the levels are comparable to
environmental levels of Ra-226 and Ra-228. Thus, these results should be
relevant to absorption of radium from environmental media.
B. Distribution
Radium distributes to soft tissue and to bone, approximately following
the distribution of calcium in the body. Maletskos et al. (1966) did not
3tudy the distribution of radium in detail in their experiments described
above on injected and ingested radium, but did find that intravenously
injected radium was initially cleared very rapidly from the blood, with a half
time of about 2 days. Maletskos et al. (1969) analyzed evidence concerning
the dependence of the distribution of radium between soft tissue and bone on
time since injection. Radium is released from soft tissue more quickly than
from bone, so that eventually most of the radium retained in the body is in
bone (NAS 1988). Ingested radium is estimated to distribute about 85% to bone
and 15% to soft tissue {UNSCEAR 1972). Bone formation occurs by apposition of
new bone on existing bone surface, and the radium content of bone will reflect
the radium activity in the blood at the time that the bone is being formed.
With continuous intake, the radium content of bone is likely to be relatively
uniform, in contrast to short-term, high-level exposure, such as in the dial
painters, in which case bone formed during the time of exposure would have
XII-2
-------
much higher radium content, forming so-called "net spars" ir. the bone (NAS
1935}. Because of its short half-life. Ra-;:4 will decay while still on the
bone surface, and not be buries ir. the bane volume as are Ra-225 and Ra-228.
Radium is incorporates into the hydroxyapatite crystals of the bone
essentially interchangeably with calcium, but subsequent growth, diffusion and
recrystalliratior. tend to eliminate radius preferentially {Muth and Globel
1932). The skeletal ouroer. zz raoium nas oeer. estimates to se 4C to 50 times
the daily radium intake (FRC 1?£1, UNSCEAS 19'TZ). Wrenr. et al. (19S5a)
estimated that radium is removed froir. bone by biological processes with a
biological half-life of about 1C years. Measurements of the radium/calcium
ratio in bone froir. autopsy samples generally indicate a constant ratio with
age, and the dietary intake of radium appears to be roughly proportional to
calcium intake, about 1 pCi radium per cram, of calcium (Hallden et al. 2.962,
Wrer.r. et al. 1985a). The constant radium/calcium bone ratio implies that, in
children, the greater radium incorporation ir. growing bone is counterbalanced
by an enhanced rate of removal due to bone remodelling and resorption (Wrenr.
et al. 1985a). Muth and Globel (1983) found peak concentrations 2-3 times
average below age 1 and at age 10 in German autopsy samples, which they
attributed to a greater incorporation of radium during periods of rapid bone
growth. Because the results of Muth and Globel (19S3) are not normalised to
calcium, but to fresh or ashed bone weight, direct comparison with the results
cf other investigators is not possible.
Norris et al. (1955) measured the retention of Ra-226 in the human body
following injection, and found that retention declines as a power of time
since exposure. The relationship derived by Norris et al. (1955) iB:
-------
R(t) - R(0) x 0.54 x t'° "
(III-l)
where R(t) is whole-body radium content at time t after absorption of R(0)
microcuries of radium. Finkel et al. (1969a) repeated the analysis on 8 of
the patients studied by Norris et al. (1955), and found the sane form of
dependence, but the following range of coefficients:
This variation, which can amount to a factor o£ more than 7 in extrapolated
initial body burdens (Finkel et al. 1969a), is probably mainly attributable to
differences in bone remodelling due to diet, exercise and/or hormonal status.
The effect of this variation on quantification of the toxicological effects of
radium is addressed in Chapter IX.
There is some evidence that placental transfer of radium is limited.
Martland and Martland (1950) found very low body burdens of radium in
17 children of 10 mothers who had worked as dial painters. Rajewsky et al.
(1965) found that the Ra-226 content of human fetal bone was independent of
gestational age and equal to the average Ra-226 content of adult bone.
Wilkinson and Hoecker (1953) found no detectable radioactivity in rat
placentas or fetuses 5 days after a single intraperitoneal injection of
15-45 fjg Ra-226 in pregnant rats on gestation day 15.
R(t) » R(0) x 0.18 x t-0 "
(III-2)
R(t) = R(0) x 0.89 x t'0M
(III-3)
III-4
-------
C. Excretion
Maletskos et al. (1966) determined urinary and fecal excretion of Ra-224
in elderly humans following ingestion or intravenous injection. Following
injection, 45% to 70% of radium was excreted within S days, with fecal
excretion averaging 30 times higher than urinary excretion. Both fecal and
urinary excretion were proportional to the blood level of radium (Maletskos
et al. 1966). Fecal excretion was higher following ingestion than following
injection, due to excretion of unabsorbed radium, but urinary excretion was
comparable after accounting for 20% absorption (Maletskos et al. 1966). An
initial rapid excretion, primarily by the fecal route, has also been observed
in humans exposed to radium by inhalation or intravenous injection (Looney
et al. 1956).
0. Bloaccumu1at ion and Retention
Information on radium toxicokinetics has been integrated into an overall
model for radium uptake and retention that is the biokinetic portion of the
RADRISK model used by USEPA (Sullivan et al. 1981, Dunning et al. 1984). The
radium toxicokinetic model is derived from the ICRP20 model for alkaline earth
metabolism (XCRP 1973). The ICRP task group developed the model using human
and animal data on alkaline earth intake, retention and excretion (ICRP 1973).
Adams et al. (1978) made a functional fit to the tabulated retention values in
ICRP20, while holding constant the area under the retention curve. Since this
retention function avoided the necessity of evaluating the incomplete gamma
function in the ICRP20 model, it has been used in USEPA biokinetic modeling
(Sullivan et al. 1981, Dunning et al. 1984).
III-5
-------
Parameters needed for the biokinetic model are absorption of the element
from the intestine (f,), fractional uptake into various tissues (f2') and
retention of the element in tissues. Values for f, and f2* are needed only
for radium, but values for retention must be specified both for radium and for
its decay products, because the RADRISK model differs from ICRP models in
general in the treatment of radioactive daughters produced in organs and
tissues (Dunning et al. 1984). ICRP models assume that all radioactive decay
prpducts generated in the body follow the metabolic behavior of the precursor
radionuclide (ICRP 1979). In contrast, the RADRISK model assumes that the
retention of a decay product in an organ or tissue is governed by the
metabolic properties of that element. The.afore, to calculate the ionizing
energy deposition in tissues associated w.^h intake of radium, not only is the
metabolic model for radium used, but retention parameters for all radium decay
products must be evaluated and energy deposition summed within tissues to
calculate the total absorbed radiation dose in the various tissues.
Calculating the radiation dose from deposited and retained radium requires use
of toxicokinetic models for all the elements in the radioactive decay chains
of Ra-226 and Ra-228 (see Figures II-l and II-2).
The gastrointestinal tract model used co evaluate ingestion exposure to
radium has four compartments: stomach, small intestine, upper large intestine
and lower large intestine. Rates of transfer, expressed as organ volumes
transferred per day, are: stomach to small intestine, 24/day; small intestine
to upper large intestine, 6/day; upper large intestine to lower large
intestine, 1.85/day; and removal from lower large intestine, 1/day (Sullivan
et al. 1981). Radium is absorbed to the blood only from the small intestine,
and the value of f, for radium used in the RADRISK model is 0.2 (see
Section VIII.A). After absorption to the blood, radium is distributed among
III-6
-------
bone and other tissue, with values of f,' of 0.46 for bone and 0.54 for all
other tissue, based on estimates of the long-term behavior of radius in the
body (ICRP 1979). Radium is assumed to he distributed uniformly in all
tissues other than bone.
The calculation of organ and tissue retention of radium and its decay
products accounts both for biological removal by exchange and excretion and
for physical removal by radioactive decay. In general, retention is defined
as the amount of an element remaining in a tissue or organ at some time after
the intake of the element. Removal of the element by biological processes is
assumed to be first order, so that the amount removed at any given time is
directly proportional to the axour.t in ri-.e tissue a- -ihat, lime, This
assumption means that biological removal can be characterized by a biological
half life, TB, which is the time taken to remove half of the concentration
initially present. Experimental studies have shown that many elements have a
number of different half lives in a given tissue or organ, each corresponding
to a different fraction of the element initially present. For example, radium
is estimated to have five compartments in bone, with biological half lives
ranging from 0.023 days to 9,600 days (Adams et al. 1978). The numerical
values for the number of compartments and their biological half lives are
usually determined in studies where there is a known intake of a radionuclide.
The whole body or specific organs or tissues are radioassayed, to determine
the amount of activity present, at various times after intake. This data is
used to develop a retention curve from which the numerical values of
parameters can be obtained. Metabolic studies to determine retention
parameters are usually done in experimental animals. However, for some
radionuclides there is metabolic data from both humans and animals.
-------
For a given element i, retention in organ s at time t is given by:
RltE(t) = 2 Flj,exp(-(Xl" +
j-1
where:
Rlte a the effective retention of element i in organ s at time t
(dimensionless).
Lla 3 the number of biological compartments for element i in organ s,
up to a maximum of 5 (dimensionless).
FtJI = the fraction of element i residing in the jth compartment in
organ s (dimensionless).
XtR = the rate of removal of element i by radioactive decay in
(days)"', calculated as (In 2)/T*, where is the radioactive
half life of element i in days.
X1]tB = the rate of biological removal of element i from compartment j
of organ s in (days)'1, calculated as (In 2)/TlJt8, where TlJtB is
the biological half life of element i in compartment j of organ
s in days.
t = time in days.
III-8
-------
Table III-l lists the reter.ricr. parameters 5;r ;:-.e elements in the radium
decay series: thorium, actinium, radium, radon, polonium, bismuth, lead and
thallium. The RADRISX model actually also Includes the small contributions
from minor branching products (USEPA 1909b). For example, in the 0-239
series, the RADRISK model includes astatine-218, representing 0.019% of Po-218
decays, and thallium-210, representing 0.021% of Bi-214 decays (HAS 1988).
The compartments for other tissue in the radium biokinetic model add up to
less than 1.0 to allow for a compartment representing prompt loss (Adams at
al. 1978). In the radon biokinetic model, 70% is lost very quickly from bone,
while 30% is retained indefinitely (i.e., biological removal is negligible
compared to radioactive decay).
The concentration of radium and its decay products in the body is
calculated for a constant intake of radium of 1 pCi/yr (Dunning et al. 1980,
Sullivan et al. 1981). The rate of change of the concentration of radium in a
given organ is calculated as the difference between the inflow from the
gastrointestinal tract and the loss through biological removal and radioactive
decay. The rate of change of the concentration of decay products is
calculated as the difference between production by radioactive decay of its
immediate precursor
-------
Table III-l Retention Parameters for Radioisotopes
in the Radium Decay Series
Number of
Biological
half life,
tiemenc
Thorium
utuan
Bone
uamDaii-iuButa
1
rractiwn
1.0
aavB
8,000
nBierencBs
ICRP 1979
Liver
1
1.0
700
Other tissues
1
1.0
700
Actinium
Bone
1
1.0
3,500
Adams et al
1978
Liver
1
1.0
3,500
Spleen
1
1.0
3,500
Other tissues
1
1.0
3,500
Radium
Bone
5
0.525
0.023
Adams et al
0.435
3.6
1978, ICRP
0.022
1,300
1979
0.00875
3,500
0.013
9,600
Other tissues
5
0.16
0.05
0.54
1.0
0.11
35
0.046
200
0.009
1,400
Radon
Bone
2
0.70
2.65 x 10'4
Bernard and
0.30
indefinite
Snyder 1975
Other tissues
5
0.874
2.65 x lO'4
0.0913
0.0031
0.0198
0.0288
0.00863
0.146
0.00612
0.963
Polonium
Bone
1
1.0
50
ICRP 1979
Liver
1
1.0
50
Kidney
1
1.0
50
Spleen
1
1.0
50
Other tissues
1
1.0
50
continued-
111-10
-------
Table III-l - continued
Element
Bismuth
Lead
Thallium
Organ
Kidney
Other tissues
Bone
Liver
Kidney
Other tissues
Kidney
Other tissues
Number of
Compartments
1
1
Fraction
0.60
0.40
0.60
0.40
0.60
0.15
0.25
0.80
0.18
0.02
0.80
0.18
0.02
0.80
0.18
0.02
1.0
1.0
Biological
half life,
0.60
5
0.60
5
12
180
12,000
12
180
12,000
12
180
12,000
12
180
12,000
7
7
Referencea
Adams et al.
1978
Adams et al.
1978
Adams et al.
1978
Adapted from Sullivan et al. 1981
III-ll
-------
The RADRISK model assigns the metabolic behavior of radium decay products
on the basi9 of the elemental form of the product, in contrast to the ICRP
model which treats all decay products the same as the precursor radionuclide.
This difference can have a substantial impact on dose estimates (Dunning
et al. 1984). The assumption that retention of decay products is different
from retention of the precursor radionuclide is clearly appropriate for soft
tissues where elemental form will govern metabolic behavior in the aqueous
phase in tissues. In skeletal tissues, however, the decay productB may be
produced and trapped in the crystal lattice of the hydroxyapatite. They
would, in this situation, continue to have the retention characteristics more
like the precursor radionuclide (Neuman and Neuman 19S8, McLean and Budy
1964). That portion of the decay products created in skeletal tissues outside
the crystal lattice would be expected to behave as their elemental
characteristics direct them, when Ra-226 transforms to Rn-222 within a
crystal, 70% of the Ra-222 escapes from the crystal {McLean and Budy 1964);
very little of other radon isotopes produced in the bone crystal can escape
before decaying radiologically. Uncertainties in the calculation of organ
distribution of ingested radium are discussed in Chapter IX.
E. Summary
When humans ingest radium, about 20% is absorbed into the circulation
(Maletskos et al. 1966). Radium initially distributes to soft tissues and
bone, but preferentially accumulates in growing bone. Normal processes of
bone remodelling release radium, with an estimated cio^ogical half-life of
10 years (Norris et al. 1955, Wrenn et al. 1985a). Excretion is primarily
through the feces (Maletskos et al. 1966). A model for the bioaccumulation
and retention of radium and its decay products, based on the ICRP20 model for
111-12
-------
alkaline earth metabolism (ICRP 1973), has been developed as part of the
RADRISK model to predict the consequences of lifetime intake of radium
(Dunning et al. 1980, Sullivan et al. 1981). The general toxicokinetics of
radium are well-established, based on both human and animal data, but the
exact values of absorption, retention and excretion may vary from individual
to individual, and in a given individual over time. The RADRISK metabolic
model estimates the typical behavior of radium in the human body, but may not
be accurate for every member of the population (USEPA 1989b). The influence
of uncertainty in toxicokinetics on the quantification of toxicological
effects of radium is discussed in Chapter IX.
111-13
-------
IV. HUMAN EXPOSURE
A. Naturally Occurring Levels
1. Radium in Water Supplies
Several surveys of radium occurrence in U.S. water supplies have been
performed. USEPA (1936] summarized surveys in New England, the Coastal Plain
and Piedmont regions of the east coast, and a nation-wide survey in 27 states
by the 0£fice of Radiation Programs. The most recent survey, the National
Inorganics and Radionuclides Survey (NIRS) was performed by the Office of
Drinking Water in the mid-1980's (USEPA 1988). The NIRS results were
considered the most suitable for deriving estimates of national radium
occurrence because the NIRS sampling program was designed to reflect the
national distribution of community ground water supplies (Longtin 1988, USEPA
1988) and both Ra-226 and Ra-228 were surveyed. This survey included
additional sampling and data analysis tc aLLow for estimation of Ra-228
distribution (Radium-228 Occurrence Model Study (ROMS)) (USEPA 1988, Longtin
1990), since Ra-228 is found at low (but not insignificant) frequencies in
public water supplies (USEPA 1990b). The NIRS found mean concentrations of
about 0.4 pCi Ra-226/L and 0.7 pCi Ra-228/L, with concentrations exceeding
5 pCi/L in 1.6% and 0.8% of supplies, respectively (USEPA 1938, 1990a,b,
Longtin 1990). Mean concentrations based on only positive values were 0.9 pci
Ra-226/L and 2.1 pCi Ra-228/L, and maximum detected concentrations were 15 pCi
Ra-226/L and 12 pCi Ra-228/L (USEPA 1988, Longtin 1990). No similarly-
designed surveys were available to estimate radium occurrence in non-community
water supplies (e.g., schools) or private wells (USEPA 1990a,b).
IV-1
-------
Occurrence distributions were derived from the NIRS data for four size
categories of population served: very small (25 - 500), small (501 - 3,300),
medium (3,,301 - 10,000) and large/very large (>10,000). A log-normal
distribution was found to describe the occurrence data, and mean and standard
deviation estimates were derived, talcing into account the samples reported as
below detection limits (0.18 pCi Ra-226/L and 1 pCi Ra-228/L) (tJSEPA 1990a,b).
The resulting occurrence estimates are presented in Table IV-1 for community
ground water supplies. Because the ROMS data are included in these occurrence
estimates, the data for Ra-228 occurrence are influenced by the addition of
200 samples in areas where the concentration of Ra-228 was expected to be
above average due to geologic formations. On the assumption that these
distributions by size category were similar for community ground water
supplies and non-community water supplies, the values in Table IV-2 were
derived for non-community water supplies serving nontransient populations.
Approximately 1% of community water supplies and 0.6% of nontransient non-
community water supplies are estimated to have Ra-226 and/or Ra-228
concentrations greater than 5 pCi/L (USEPA 1990a,b). It is estimated that
67 systems have more than 20 pCi Ra-226/L and that 3& have more than 20 pCi
Ra-228/L (USEPA 1991a).
Co-occurrence analysis of the data indicate that Ra-226 and Ra-228 co-
occur about 10% of the time, about twice the frequency expected for a random
phenomenon (USEPA 1990a). When these isotopes were detected at concentrations
>3 pCi/L, co-occurrence was greater than 50% (USEPA 1990a).
Approximately 76 million people in the United States are served by ground
water supplies. The populations exposed to various levels of radium were
estimated by combining the occurrence estimates in Tables IV-1 and IV-2
IV-2
-------
Table IV-1 Estimated Cumulative National Occurrence of Radium-226 and Radium-228 in Community
Ground Water Supplies Based on a Censored Lognormal Distribution Model
Systea Size
Nuaber of
Nuaber oT
supplies with
concentratIons
JpCl/Ll of:
(populatIon
Systeas
Radlua-226
Hadlua -2S>B
served)
In the It.S.
iO. IS
>2
>3
>4
>5
> B
>10
*1
>3
~ >5
Very Snail
25-100
17,079
6,571
523
278
171
114
46
29
1,711
235
84
101-500
15,354
470
270
250
153
103
42
26
1,552
215
76
Saall
501 -1.000
5,038
1,885
134
69
41
27
11
7
514
139
70
1,001-3,300
5,185
1,940
137
71
43
28
11
7
510
134
68
Uedlua
3,301•10,000
2,308
982
233
167
129
105
66
52
492
95
49
Larae/Verv Larae
10,001-25,000
823
380
109
81
65
54
36
30
180
38
20
25,001-50,000
278
128
37
27
22
18
12
10
60
12
6
50,001-75.000
77
36
10
8
6
5
3
3
1
75,001 -106,000
17
8
2
2
1
1
1
1
13
2
100,001-500,000
39
18
5
4
3
3
2
1
4
1
0
500,001-1,000,000
4
2
1
O
0
0
0
0
8
2
0
)1,000,600
0
0
0
0
0
0
0
0
0
0
0
Total
46,202
17,858
1,661
957
634
458
230
166
5,044
873
374
lower Bound
15,641
832
327
122
25
0
0
2,522
1
0
Upper Bound
20,075
2,491
1,587
1,146
891
533
421
7,575
1,745
900
Note: lower and upper bounds are based upon 95% confidence Intervals.
Adapted froa USEPA (1990a,b)
-------
Table IV-2 Estimated Cumulative national Occurrence of Radium-226 and Radium-228 in Nontransient
Non-community Ground Water Supplies Based on a Censored Lognormal Distribution Model
Systea Size
Nuaber of
Nuaber of
supplies with concentrations (pCl/L)
of:
(population
Systeas
Radlua-226
Kadlua-228
served)
In the U.S.
10.18
>2
>3
>4
>5
>8
}10
tl
p
>5
Very Saall
25*160
13,899
5,348
425
226
139
93
38
24
1,568
230
83
101-500
3,964
1,525
121
64
40
27
11
7
400
56
20
Saall
501¦1.000
1,151
441
31
16
9
6
2
1
118
32
16
1,001-3,300
205
77
5
3
2
1
0
0
20
6
3
lledlua
3,301-10,000
49
21
3
4
3
2
1
1
10
2
1
Laroe/Verv Larae
10,001 25,000
4
2
1
0
0
0
O
0
0
0
0
25,001-50,000
11
5
1
1
1
1
0
0
3
0
0
50,001-73.000
6
3
t
1
0
0
0
0
1
0
0
75,001-106,000
2
1
0
0
0
0
0
0
0
0
0
100,001-500.000
7
3
1
1
1
1
0
0
1
0
0
500,001-1,060,000
0
0
0
0
0
0
0
0
0
0
0
)1,000,600
0
0
0
0
0
0
0
0
0
0
0
Total
19,298
7,416
591
316
195
130
52
33
2,121
326
123
lower Bound
6,359
217
42
0
0
O
0
938
0
0
Upper Bound
8,473
965
590
411
307
165
123
3,303
720
345
Hote: Lower and upper bounds are based upon 95\ confidence Intervals.
Adapted froa USEPA (1990a,b)
-------
with the number of people served (USEPA 1990a,b). These estimates are given
in Tables IV-3 and IV-4. The number of people exposed to more than 5 pCi
Ra-226/L is approximately 3/100,000 from community water supplies and 48,000
from nontransient non-community water supplies, and the number exposed to more
than 5 pCi Ra-228/L is approximately 1,000,000 from community water supplies
and 32,000 from nontransient non-community water supplies (USEPA 1990a,b). A
total of 890,000 people are estimated to be exposed to more than 20 pCi
Ra-226/L and 164,000 to be exposed to more than 20 pCi Ra-228/L (USEPA 1991a).
2. Radium in Other Media
Radium has been detected in several kinds of food in the U.S. (USEPA
1990a,b). From the sparse data available, it has been reported that eggs,
pasta, bread and other bakery products, and potatoes are the major sources of
Ra-226 in the diet. It is estimated that U.S. adults have an average dietary
intake of Ra-226 and Ra-228 of about 1 pCi/day (USEPA 1990a,b). The values
for the distribution of total daily intake indicate a median value between 1
and 2 pCi/day for both Ra-226 and Ra-228. Drinking water supplies generally
contribute less than 50% of the total daily intake of radium. The median
values for percent drinking water contribution are 25% and 49% for Ra-226 and
Ra-228, respectively (USEPA 1990a,b).
Radium is present in air at extremely low levels. Outdoor air levels of
about 1.6E-5 pCi Ra-226/m3 and 2E-3 pCi Ra-228/mJ have been reported (USEPA
1990a,b). Therefore, respiratory intake of radium is considered negligible
relative to intake fro.n food and drinking water.
IV-5
-------
Table IV-3 Estimated Cumulative Population (in thousands) Exposed to Radium-226 and Radium-228 in
Drinking Water Exceeding the Indicated Concentrations from Community Ground Water Supplies
Systea SIze Nuaber of Population (thousands! Served by Water Supplies In concentrations (oCl/LI of:
(population People Radlua-226 Systea Size Hadlua-228
cervcd) (thousands! *0. IB >2 >3 )4 )5 )B >10 (pop, servedl tl >3
Very Saall
25-100
101-500
950
3,850
366
1,481
29
118
15
63
10
39
6
26
3
10
2
7
25-500
485
67
24
Saall
501-1.000
1,001-3,300
3.910
10,000
1,463
3,741
104
265
53
137
32
82
21
54
8
21
5
11
501-3,300
1,375
364
181
Medlua
3,301-10,000
13,310
5,665
1,343
961
745
605
380
300
Larae/Verv Larae
10,001-25,000
<5,001-50,000
50,001-75,000
75,001-100,000
100,001-500,000
500,001-1,000,000
)1,000,000
13,110
9,540
4,770
1,360
7,600
2,760
0
6,057
4,408
2,204
628
3,511
1,275
0
1,731
1,260
630
180
1,004
365
0
1,295
942
471
134
751
O
0
1,038
755
378
108
602
0
0
867
631
315
90
503
0
0
578
421
211
60
335
0
0
472
343
172
49
273
0
0
>3,300
10,210
1,942
854
Total
71,160
30,799
7,028
4,822
3,787
3,118
2,027
1,635
12,070
2,373
1,059
Lower Bound
Upper Bound
19.796
41,802
0
14,415
0
11,320
0
9,661
0
8,519
0
6,484
0
5,674
HA*"
NA
NA
NA
NA
NA
Note: Lower and upper bounds are based upon 95% confidence Intervale.
Adapted froa USE PA (1990a,b)
HA • not available
-------
Table IV-4 Estimated Cumulative Population (in thousands) Exposed to Radium-226 and Radium-228 in
Drinking Water Exceeding the Indicated Concentration Range from Nontransient
Non-Conununity Ground Hater Supplies
Systea Size
Nuaber of
Population
(thousands)
Served bv Water Supplies In
concentrations
(PC1/L) of
(populatIon
People
Hadlua-226
Systea Size
Radlua-22B
served)
(thousands)
£O.IB
)2
>3
) 4
? 5
>8
HO
(pop. served)
" >5
Verv Saall
25-fOO
101-500
615
937
237
361
19
29
10
15
6
9
4
6
2
3
1
2
15-500
157
21.5
7.7
Saall
501-1.0M
1,001-3,300
1,015
351
380
131
27
9
14
5
8
3
6
2
2
0
1
0
501-3,300
138
37
20.6
Uedlua
3,301-10,000
273
116
28
20
15
12
8
6
La roe/Verv Lai ae
10,001-25,000
25,001-50,000
50,001-75.000
75,001-100,000
100,001-500.000
500,001-1,060,000
>1,000,600
71
275
398
160
1,096
0
0
33
127
184
83
506
0
0
9
36
53
0
145
0
0
O
27
39
0
108
0
0
O
22
0
0
87
0
0
O
18
0
0
0
0
0
0
0
0
0
0
0
0
O
0
0
0
0
0
0
>3,300
290
10.7
3.0
Total
5,211
2,158
054
238
151
48
14
10
585
69
31
lover Bound
Upper Bound
1,585
2,731
O
731
0
569
0
450
0
323
0
241
0
216
NA'"
HA
HA
HA
HA
HA
Note: Lover and upper bounds are based upon 95% confidence Intervals.
Adapted froa USEPA (1990a,b)
HA - not available
-------
B. Other Sources
Elevated levels of radium in soil, water and ambient air can be produced
by contamination from coal burning and uranium mining or milling operations
(ATSDR 1989). The mean concentration in coal is about 1 pCi Ra-226/g and
concentrations reported on fly-ash range from 1 to 10 pCi Ra-226/g and 1.8 to
3.1 pCi Ra-228/g (ATSDR 1989). Deposition of this fly ash may cause elevated
levels of radium in local areas. Levels up to 8.1 pCi Ra-226/g have been
detected in soil in industrial areas (ATSDR 1989). Surface runoff or leachate
from uranium mine tailings and the release of ore-processing effluents may
cause radium contamination of local water resources (ATSDR 1989). Disposal of
waste from thorium or radium processing can also cause contamination of soil,
groundwater and/or structures (USEPA 1990c). These sources can result in high
exposures to populations in the vicinity, particularly if homes are located
near contaminated soil, or if a contaminated aquifer is used as a source of
potable water.
C. Summary
The occurrence of radium in ground water supplies follows an
approximately log-normal distribution, with a mean value of less than 1 pCi/L
(USEPA 1988, Longtin 1990). These data can be used to estimate population
exposure from community water supplies and can be extrapolated to estimate
exposure from nontransient non-community water supplies (USEPA 1990a,b).
About 3,000,000 people are exposed to drinking water containing more than
5 pCi Ra-226/L in community water supplies and 48,000 are exposed from
nontransient non-community water supplies. Corresponding numbers for Ra-228
are 1,000,000 and 32,000 people (USEPA 1990a,b). It is estimated that a total
IV-8
-------
of 890,000 people are exposed to more than 20 pCi Ra-226/L and 164,000 are
exposed to more than 20 pCi Ra-228/L (USEPA 1991a). Some fraction of this
population may be exposed to these levels of both isotopes, as indicated by
co-occurrence analysis (USEPA 1990a,b). Food is another source of radium
exposure, amounting to an average of about 1 pci/day, and often contributes
more than half the daily radium intake (USEPA 1990a,b).
IV-9
-------
V. HEALTH EFFECTS IN ANIMALS
A. Short-Term Exposure
Several groups have investigated the effects of short-term exposure to
radium in experimental animals. All such studies used the injection route of
exposure. A number of repocta from these groups ace summarized in this
section; a more complete review is provided in ATSDR (1989}.
Jee et al. (1969) injected 230 adult beagle dogs intravenously with a
single dose of Ra-226 or Ra-228, with doses for each isotope equal to 0,
0.057, 0.17, 0.34, 1.1, 3.2, or 10.0 /iCi/kg. Changes in bone vasculature were
evaluated from micrographs of bone sections taken after injection of metatarsi
with India-ink and gelatin. The lowest doses causing significant reductions
in vascularization of compact and spongy bone were 1.1 ftCL Ra-226/kg at
1,900 days after injection and 0.17 /jCL Ra-228/kg at 2,500 days after
injection. In a later report, Taylor et al. (1976) described the minimum
radiographic lesions as foci of decreased density, and reported that such
skeletal lesions appeared at the minimum tumor-inducing dose.
Momeni et al. (1976) injected 14-month old beagle dogs intravenously with
3 semimonthly doses of Ra-226, with total doses of 0, 0.024, 0.064, 0.376,
1.12, 3.36, or 10.0 jjCi/kg. Radiographs of skeletons were obtained at yearly
intervals for all 324 dogs, and bone changes were scored for 20 skeletal areas
on a scale of 0 to 6. A dose-dependent increase in skeletal lesions was
found, with endosteal or periosteal cortical sclerosis and thickening being
the most sensitive indicator of damage, and fractures and trabecular
coarsening appearing at higher dose levels. Dogs injected with 10.0 /jCi/kg at
V-l
-------
2 or 4 months of age exhibited lesions earlier after injection than dogs
injected with this dose at 14 months of age. At the lowest dose showing an
appreciable effect, 0.376 pCi/kg, the rate of skeletal changes increased
linearly with age, while at higher doses, the rate leveled off and then
decreased with age.
Raabe et al. (1981) reported on the causes of death among the beagle dogs
treated as described in the experiments of Momeni et al. (1976).
Nonneoplastic causes of death included radiation-induced nephritis, pathologic
fractures and osteodystrophy. The lowest doses at which these effects
occurred were not reported.
Schoeters and Vanderborght (1981) injected male inbred BALB/c mice
intraperitoneally with single dose of Ra-226 (230 or 660 kBq Ra-226/kg) and
sacrificed 5 from each dose group at 4 hours and at 1, 3, 10, 24, 100 and
300 days after injection. Control mice were sacrificed on days 0 and 300. At
the higher dose, the number of hemopoietic stem cells was depressed up until
day 100, but showed recovery to near control values at day 300.
Schoeters et al. (1983) injected a single dose of Ra-226 (4.4, 10.7, or
24.8 kBq Ra-226 per mouse) intraperitoneally in male C57B1 mice, and blood was
drawn at 200, 400 and 530 days after injection. In the highest dose group,
the number of peripheral white blood cells was reduced at 400 and 530 days
after injection.
V-2
-------
B. Longer-Term Exposure
No data were located concerning noncancer effects of radium exposure for
periods longer than 90 days in experimental animals.
C. Reproductive and Developmental Effects
No data were located concerning reproductive or developmental effects of
radium exposure in experimental animals.
D. Mutagenicity
No data were located concerning the mutagenicity of any radium isotope.
Ionizing radiations, including alpha particles, are known to cause mutations,
chromosomal aberrations and transformation of mammalian cells (NAS 1988).
E. Carcinogenicity
Several groups have investigated the carcinogenic effects of exposure to
radium in experimental animals. Most such studies used the injection route of
exposure. A number of reports from these groups are summarized in this
section; a more complete review is provided in ATSOR (1989).
Evans et al. (1944) administered Ra-226 to 5 albino male rats by medicine
dropper daily for 20 days (Series 1), to 3 white male rats by medicine dropper
daily for 10 days (Series II), to 3 male Wistar rats subcutaneously for
10 days every other day (Series III) and to 13 male Wistar rats by medicine
dropper daily for 20 days (Series IV). All 5 rats in Series I and 9 out of
v-3
-------
13 rats in Series iv developed osteosarcoma about a year after oral exposure.
The lowest dose producing a confirmed osteosarcoma was a total of 22 pq
Ra-226.
Finkel et al. (1969b) exposed a total of 3,210 CFl/Anl female mice to
Ra-226 by a single intravenous injection at doses from 0.05 to 120 pCi
Ra-226/kg. Animals were X-rayed after death for diagnosis of osteosarcoma,
which was confirmed by histologic examination if necessary. The time of
radiographic appearance of bone tumor9 was back-calculated from measured tumor
volume. A significant increase in the incidence of bone sarcomas over
controls was observed in the lowest dose group, 0.05 ^Ci Ra-226/kg. Up to
doses of 1.25 pCi Ra-226/kg, there was no relationship between dose and tumor
appearance time, but as the dose increased from 2.5 to 120 pCi. Ra-226/kg, the
average time to tumor appearance dropped from 639 to 329 days.
Taylor et al. (1972) exposed 169 beagle dogs to Ra-226 or Ra-228 by a
single intravenous injection at 16 to 17 months of age at doses ranging from
0.00074 to 10.4 fjCL Ra-226/kg and from 0.018 to 8.5 ^Ci Ra-228/kg. The
injection solutions for Ra-228 contained 0.6% or 3% Th-228. After follow-up
times of about 10 years, bone sarcomas were found at doses of 0.17 ^Ci Ra-226
and above and at doses of 0.15 pCi Ra-228/kg and above. Intraocular melanomas
were induced at doses between 0.062 and 1.1 pCi Ra-226/kg and between
0.018 and 0.15 ^Ci Ra-228/kg.
Mays et al. (1987) reported results for the Taylor et al. (1972) beagles
after more than 15 years of follow up. Bone sarcomas were induced at doses of
0.022 pCi Ra-226/kg and above, and at doses of 0.050 pCi Ra-228/kg and above.
V-4
-------
The dose-response data followed a linear, no-threshold model, with a
particularly good fit for Ra-226.
Raabe et al. (1981) exposed 243 beagle dogs to Ra-226 by eight
intravenous injections at two-week intervals starting at 435 days of age at
doses ranging from 0.024 to 10 /jCi Ra-226/kg. Premature deaths and incidence
of bone sarcomas both showed a clear dose-response relationship with radium
exposure, with primary bone cancers causing death in 2/36 dogs given 0.064 pCi
Ra-226/kg but in none of the 45 dogs given 0.024 piCi Ra-226/kg, the lowest
dose group or in the controls. Deaths from malignant lymphomas and manmiary
carcinomas occurred in all groups, including controls, with no clear
association with radium dose.
Taylor et al. (1983) exposed 59 male and 62 female C57BL/DO (Black) mice
and 71 male and 62 female CS7BL/Do (Albino) mice to Ra-226 by a single
intraperitoneal injection at doses ranging from 0.057 to 9.88 fjCL Ra-226/kg.
Bone tumors were detected by X-ray of dead mice and were classified
histologically. The lowest doses causing a bone sarcoma were
1.03 pCi Ra-226/kg in male Black mice, 3.1 fiCi Ra-226/kg in female Black mice,
9.88 pCi Ra-226/kg in male Albino mice and 3.23 yCL Ra-226/kg in female Albino
mice.
Schoeters et al. (1983) exposed 557 male C57B1 mice to Ra-226 by a single
intraperitoneal injection of 0.12, 0.29 or 0.67 ^iCi Ra-226 per mouse. Dead
mice were autopsied, examined histologically and X-rayed for diagnosis of
osteosarcomas. Non-thymic lymphomas were found in 8 of the 40 mice receiving
V-5
-------
the highest dose, but the incidence was not reported in lower dose groups or
in controls. A dose-related increase in bone sarcomas was found, with a six-
fold increase in incidence over controls at the lowest dose, 0.12 pCi Ra-226
per mouse. Assuming a body weight of 30 g, this dose is 4 yCi Ra-226/kg.
Kofranek et al. (1985) exposed 431 female ICR mice to Ra-226 by a single
intraperitoneal injection at 10 weeks of age at doses ranging from 1.7 to
25 fjCL Ra-226/kg. Dead mice were X-rayed and examined histopathologically.
The lowest dose inducing bone sarcomas was the lowest dose tested, 1.7 nCi
Ra-226/kg.
Kofranek et al. (1985) exposed 200 female ICR mice to Ra-224 by
150 repeated intraperitoneal injections starting at 10 weeks of age at total
doses of 17 or 50 pCi Ra-224/kg. Dead mice were X-rayed and examined
histopathologically. Bone sarcomas were induced at both doses.
Humphreys et al. (1985) exposed 488 male CBA/H mice to Ra-224 by a single
intraperitoneal injection or by eight injections twice a week at total doses
ranging from 1.9 to 59 pCi Ra-224/kg. Blood and tissue samples were collected
and X-rays were taken of mice at sacrifice. Osteosarcomas were diagnosed by
x-ray and histology, and myeloid leukemias were diagnosed by excess of
immature granular cells at hematopoietic sites, abnormal sites and peripheral
blood. Osteosarcomas were induced at doses of 7.4 fjCL Ra-224/kg and above by
single injection and at doses of 1.9 fjCi Ra-224/kg and above by multiple
injection. The incidence of myeloid leukemia showed a maximum at intermediate
doses (peak at 15 yCL Ra-224/kg), and the ratio of myeloid leukemia to bone
sarcoma at this peak was 1.5, while this ratio was 0.09 at the maximum dose,
V-6
-------
59 pCi Ra-224/kg. These results indicate that the ratio of induced leukemias
and osteosarcomas varies with dose, at least in this system.
Muller et al. (1988) exposed 500 female NMRI mice to Ra-224 by a single
intraperitoneal injection or by 72 injections twice a week, both at total
doses equal to 0.5 uCL Ra-224/kg. Mice were observed for 360 days, and
complete) autopsies were performed on all dead or moribund animals. No
osteosarcomas appeared in either treated group or in controls. -Malignant
lymphoblastic lymphomas appeared at a total incidence of 13% among mice
treated with multiple injections, but no lymphomas appeared among mice given
the same total dose as a single injection.
F. Summary
Most studies of health effects in experimental animals following
injection of radium have investigated effects on bone, due to the preferential
accumulation and long-term retention of radium in the skeleton, and subsequent
damage due to radioactive decay. These effects have been demonstrated with
Ra-224, Ra-?.26 and Ra-228. Effects on bone found at relatively short times
after radium exposure are changes in bone structure (Jee et al. 1969, Womeni
et al. 1976) or hematopoiesis (Schoet?rs and Vanderborght 1981). When &nimal9
are followed for a considerably fraction of their lifetimes, bone sarcomas are
usually found {Raabe et al. 1981, Humphreys et al. 1985, Mays et al. 1987).
Such tumors have been observed following a single exposure to radium (Taylor
et al. 1983, Kofranek et al. 1985). Leukemias or lymphomas have been seen
following injection of Ra-224, but the occurrence peaks at low doses, lower
than the doses that cause osteosarcomas (Humphreys et al. 1985, Muller et al.
1988). The results of animal studies confirm the association of bone necrosis
7-7
-------
and bona sarcomas with radium exposure in humans, and provide support for an
approximately linear relationship between dose and incidence of bone sarcomas
in animals (Wrenn et al. 1985b, Mays et al. 1987). Uncertainty concerning
this dose-response relationship is discussed in Chapter IX.
V-6
-------
VI. HEALTH EFFECTS IN HUMANS
A. Clinical Cage Studies
Fatal cases of jaw necrosis and aplastic anemia among women employed as
dial painters in New Jersey were an early indication of the hazards associated
with ingestion of Ra-226 and/or Ra-228 (Martland 1931). From 1922 to 1928,
12 deaths occurred due to these causes among radium dial painters and chemists
in New Jersey (Martland 1931). By 1931, an additional effect of radium
exposure, bone sarcoma, was recognized, with 3 cases diagnosed between 1924
and 1931 (Martland 1931). This identification o£ bone sarcoma as a
consequence of radium exposure has been extensively confirmed by epidemiologic
studies of dial painters and others exposed to Ra-226, Ra-228 or Ra-224 (see
Section VI.8. below).
Sharpe (1974) reported the case histories of 42 workers employed in the
New Jersey radium dial painting industry who survived more than 25 years after
radium exposure. These workers represent about 4% of the people identified as
having worked in the radium industry in New Jersey. Malignant tumors
developed in 24 cases (57%). Sharpe (1974) compared specific cancer death
rates to average 1950 and 1960 Unites States death rates, and concluded that
rates of paranasal sinus, pulmonary and reticuloendothelial (Including
multiple myeloma) malignancies were elevated. Other health effects suggested
as possibly related to radium exposure were bone lesions and hearing loss.
Females exposed to radium had a decreased number of children compared to wives
of males exposed to radium, and one female dial painter had a child with a
congenital heart defect. The number of cases was too small to attribute
reproductive or developmental effects to radium exposure (Sharpe 1974).
VI-1
-------
B. Epidemiologic Studies
A substantial amount of epidemiological data are available concerning the
health effects of exposure to ionizing radiations, particularly x-rays and
gamma-rays (NAS 1930b, USEPA 1989b). The principal data concerning human
effectB of exposure to r^ium come from epidemiological stucUes of workers,
mainly women, employed as radium dial painters {Rundo et al. 1966) and of
patients injected with Ra-224 for the treatment of ankylosing spondylitis
(spinal arthritis) and tubercular infection of the bone (Mays and Spiess
1984). Studies of these populations and community studies comparing cancer
rates as a function of radium content of drinking water supplies are presented
below.
1. Radium-Dial Painters
The luminizing industry used radium-containing paint to make watch dials
with numerals that would glow in the dark. Prior to the mid-1920s, little
attention was paid to limiting the radium exposure of workers. A common
practice was sharpening the tip of the paint-laden brush by twisting it in the
corner of the mouth, which led to considerable ingestion of paint among many
workers (Martland 1931, Sharpe 1974). Occupational exposures in the radium
dial industry continued after 1930 but at substantially lower levels
(Stebbinys et al. 1984). The health effects relevant to assessing risks of
low-level exposure to radium are discussed below.
VI-2
-------
Bone necrosla
Keane et al. (1983) investigated the prevalence of bone changes in radium
dial workers compared to matched controls with no radium exposure. Radium
exposure was expressed a9 intake to the blood {i.e., absorbed dose) in
luCL Ra-226 or Ra-228, based on measured residual body content and backward
extrapolation using the Norris retention function (Norris et al. 1955). Bone
changes were evaluated by examination of radiographs (X-rays) of exposed and
control women. Changes considered characteristic of radium deposition were
foci of necrosis associated with either decreased or increased bone density.
A total score for bono necrosis was calculated for 20 areas of the skeleton.
Keane et al. (1983) found that below a calculated total intake to blood of
10 ^iCi of either isotope, the frequency and severity of bone changes was not
significantly different in the exposed and matched control group. They also
found that below a total intake to blood of 100 pCi of either isotope, all
bone changes were mild or minimal. Keane et al. (1983) reported that earlier
investigators in the 1960's had stated that the clinical significance of bone
changes, such as susceptibility to fracture, were considered to begin between
mild and moderate degrees, thus raising the question of whether the mild
changes seen at intakes between 10 and 100 pCi should be considered adverse.
However, radiographs are not sensitive to histologic changes in bone in
radium-exposed animals (Park et al. 1972, Momeni et al. 1976), and no evidence
was located to indicate that Keane et al. (1983) or earlier investigators had
considered possible effects of mild lesions on bone growth, bone healing or
calcium homeostasis. Thus, all detected lesions will be considered adverse,
and the NOAEL identified by this study is a systemic intake to the blood of
10 pCi of Ra-226 or Ra-228, the level at which bone necrosis detected by
X-rays is significantly different from controls.
vi-3
-------
The subjects in the study by Keane et al. (1983) ingested radium over a
very short period of time {median duration o£ intake was about a year for Ra-
226 and half a year for Ra-228). However, because of the dosimetry of radium
(preferential concentration in the bone and subsequent long-term radioactive
decay)t the effect of a given total ingested amount is thought to be relative-
ly independent of the time-course of exposure. Keane et al. (1983) estimated
the radiation received by cells near bone surfaces (to a depth of 10 fjmj for a
single intake and for an equal intake spread over 50 years. The ratio of the
radiation dose for a single and a 50-year intake was estimated to be 1.6 for
Ba-226 and 1.2 for Ra-228. Because of the near independence of radiation dose
to duration of intake, the NOAEL of 10 (jGL radium for noncancer health effects
identified by the Keane et al. (1983) study will be used to assess limits on
exposure on all time scales for protection against noncancer effects.
Mutagenicity
Huller et al. (1966) examined bone marrow cells from 5 controls and
16 individuals exposed to Ra-226 and/or Sr-90. Exposed individuals had
significantly greater incidences of aneuploid cells and colls with chromosomal
aberrations.
Boyd et al. (1966) examined chromosomes of peripheral lymphocytes of
62 British radium dial painters and 57 control women. A significcnt dose-
related increase in chromosomal aberrations was found.
Hoegerman et al. (1973) examined chromosome? of peripheral lymphocytes of
19 dial painters and 5 controls. A weak positive correlation was found
between body burden of radium and the frequency of chromosomal aberrations.
VI-4
-------
Reproductive and Developmental Effecta
Polednak {1980} investigated the fertility of female Illinois radium dial
painters employed between 1916 and 1929 with measured radium body burdens.
The index of fertility was the live-birth rate, defined as the number of
reported live births divided by the nuntber of years of marriage prior to
surgical menopause or age 45. Originally 274 white females were in the study,
and this number was reduced to 199 after excluding women first marrying before
beginning radium exposure, women with unknown dates of first marriage and
women never marrying or being married for less than 3 years. Ovarian doses
were calculated based on internal and external irradiations. There was no
significant difference in the percentage of childless women or rate of miscar-
riages among the four ovarian dose groups, but the live birth rate was signi-
ficantly lower for ovarian doses above 20 rem compared to doses below 20 rem.
The decrease was related to radium intaks but not duration of employment. No
information was available to determine whether the decrease in live-birth rate
was attributable to contraceptive practices or other confounding variables.
Cancer
It has long been recognized that two types of cancer with very low
spontaneous rates, bone sarcomas and head carcinomas, are elevated in exposed
radium dial workers (Martland 1931, Evans at al. 1944, Sharpe 1974). The
discussion in this section will focus on studies that have made quantitative
evaluations of the incidence of tumors as a function of ingested dose.
Systematic investigation has led to the identification of 4,863 radium
dial painters, using both objective {e.g., employee lists) and non-objective
VI-5
-------
(e.g./ self-referral) means. Of the total cohort chat was identified, 803
(17%) have not been located, 2,144 (44%) have been located but their radium
exposure has not been measured, and 1,916 (39%) have been located and measured
(Rundo et al. 1986). Because of the persistence of Ra-226 in the skeleton,
the body burden of Ra-226 can be reliably measured from whole-body gamma
radiation, exhaled radon gas, or in some cases tissue samples, even many years
after cessation of exposure. The initial intake of Ra-226 is back-
extrapolated from the estimated rate of radium elimination from the human body
(see Section XXX.8.) (Norris et al. 1955). However, the rate of elimination
of Ra-226 from the human body is variable, which can lead to differences in
initial estimated dose of approximately a factor of 7 (Finkel et al. 1969a).
Because of its shorter half-life, Ra-228 disappears soon after intake, and
usually was not detected in women measured many years after exposure.
However, the initial intake of Ra-228 can be extrapolated from the calculated
intake of Ra-226 together with information of varying reliability concerning
the isotope ratio at the place of employment, which varied both over time and
among companies (Sharpe 1974, Stebbings et al. 1984, Rundo et al. 1986).
Calculated doses are generally expressed as intake to the blood (i.e.,
absorbed dose), but have also been converted to ingested dose using the
absorption factor of 0.2 found by Maletskos et al. (1966).
The rate of bone sarcomas and head carcinomas is significantly elevated
in the dial painter cohort. A total of 64 bone sarcomas have been observed in
the 4,032 located members of this cohort, 44 in workers with measured body
burdens, and the remaining 20 in unmeasured workers (Rundo et al. 1986). A
total of 24 head carcinomas have been observed, 19 in measured and 5 in
unmeasured workers (Rundo et al. 1986). No estimates are available of the
number of cancers among the 803 identified dial painters that have not been
VI-6
-------
located/ and most quantitative evaluations have considered the measured
workers only.
Rowland et al. (1978) considered a cohort of 759 measured women first
employed before 1930. In this cohort, 38 bone sarcomas and 17 head carcinomas
were observed, and all observed cancers were in the high-do3e groups. These
results are summarized in Tables VI-1 and VI-2. All dose groups below 100 nCi
have been combined, and no bone sarcomas were found among these lower-dose
groups. Rundo et al. (1S86) reported that through 1984, no bone sarcomas or
head carcinomas have so far appeared in the lower-dose groups of measured dial
painters. The rate of bone sarcomas and head carcinomas is significantly
elevated, as less than one case of either bone sarcoma or head carcinoma would
be expected in this entire population, based on national incidence data
(Rowland et al. 1978).
Rowland et al. (1983) considered a subgroup of the dial painters that
excluded all workers who may have had their radium intake measured as the
result of early symptoms of bone cancer. This subpopulation, designated
"Subgroup B", is epidemiologically more acceptable because there is no bias in
the cohort towards a higher incidence of bone sarcomas (MAS 1988). A total of
13 bone sarcomas occurred in Subgroup B, while 42 bone sarcomas occurred in
the entire cohort of female dial painters first employed before 1950, Group A.
The occurrence of bone sarcomas in Subgroup B is shown in Table VI-3. Dose
groups below 250 pCi have been combined, and bone sarcomas were observed only
in the three highest dose groups.
VI-7
-------
Table VI-1 Bone Sarcoma Rates among Female Radium Dial Painters
First Employed before 1930
Estimated Systemic Intake
(uCi Ra-226 + 2.5 uCi Ra-22Bi
Number
Examined
Bone
Sarcomas
Bone Sarcomas per
1,000 person-years
at risk
22,500
16
4
18.3
1,000-2,499
22
15
36.8
500-999
18
8
13.1
250-499
32
9
7.21
100-249
27
2
1.72
99-S0.5
644
0
0
Adapted from Rowland et al. 1978
VI-8
-------
Table VI-2 Head Carcinoma Rates among Female Radium Dial Painters
First Employed before 1930
Estimated Head Carcinomas per
Systemic Intake Number Head 1,000 person-years
fuel Ra-226l Examined Carcinomas at risk
21,000 10 3 18.3
500-999 11 2 7.27
250-499 25 5 6.16
100-249 31 5 5.29
50-99 23 1 1.12
25-49 34 1 0.75
24-20.5 615 0 0
Adapted from Rowland et al. 1978
VI-9
-------
Table VI-3 Bone Sarcoma Rates among Female Radium Dial Painters
Excluding Workers Possibly Measured Due to Symptoms of Bone Sarcoma
Bone Sarcomas per
Estimated Systemic Intake
(uCL Ra-226 + 2.5 uCi Ra-22St
Number
Examined
Bone
Sarcomas
1,000 person'
at risk
22,500
2
0
0
1,000-2,499
6
3
46.2
500-999
11
5
41.3
250-499
24
5
14.6
249-SO.S
1,214
0
0
Adapted from Rowland et al. 1983
vi-io
-------
Rowland et al. (1978) compared the relative effectiveness of Ra-226 and
Ra-228 in inducing bone sarcomas, and concluded that a mierocurie of Ra-228
was 1.7 to 2.77 times as effective as a mierocurie of Ra-226. In addition,
Rowland et al. (1978) demonstrated that the Incidence of head carcinomas was
associated only with exposure to Ra-226, not with exposure to Ra-228. This
association would be expected if the accumulation of radon gas in the mastoid
air cells and paranasal sinuses is important in the etiology of these tumors
(NAS 1988). Both Ra-226 and Ra-228 have a radon decay product, but the half
life of Rn-220, the decay product of Ra-228, is only 54.5 seconds (see Figures
1 to 3), too short for substantial diffusion to air cells to take place (NAS
1988). The relative potency of the two isotopes of radium depends upon the
biological effect being considered. Besides relative potencies of 1:2.5 for
bone sarcoma and 1:0 for head carcinoma (Rowland et al. 1978), other values
that have been derived are 1:1 for bone necrosis, based on empirical evidence
(Keane et al. 1983), and 1:6 for irradiation of soft tissues, based on
assuming that all Rn-222 and no Rn-220 escapes from soft tissue and there is
no translocation of decay products between tissues (Keane and Schlenker 1983).
No conclusive evidence has been found for a statistically significant
increase in cancers other than bone sarcomas and head carcinomas in cohorts of
radium dial painters (Stebbings et al. 1984). Marginally significant
increases in breast cancer and multiple myeloma have been noted, but the
incidences are better correlated with duration of employment (a surrogate for
external dose of gamrca irradiation) than with calculated radium intake
(Stebbings et al. 1984). These and other authors have pointed out the
difficulties in establishing radium exposure as a cause of such late-
appearing, common cancers, since there are many other potentially confounding
risk factors (Finkel et al. 1969a, Stebbings et al. 1984).
vi-n
-------
Leu/temia rates were elevated in certain subgroups of radium dial painters
(e.g., in pre-1930 female workers in an Illinois plant) but overall no
statistically significant elevation of leukemias was found (SpLers et al.
1983, Stebbir.gs et al. 1984}. Mays et al. (1985a) calculated that the
observed ratio of induced bone sarcomas to leukemias is at least ten to one
hundred at high doses of Ra-226 and/or Ra-228 {Maya et al. 1985a); at lower
closes neither bone sarcomas nor excess leukemias are observed (Rundo et al.
1986). The lack of an Increase in leukemias is unexpected because the
accumulation of radium in bone would be expected to provide substantial
irradiation of potentially leukemcgenic cells (Kays et al. 1985a), and
external irradiation has clearly been established as a cause of leukemia in
humans (MAS 1930b). A number of possible explanations were summarized by Hays
et al. (1985a) to account for the lack of observable increase in leukemias
among persons exposed to radium, including nonuniformity of irradiation,
lethality of irradiation to target cells, low frequency of leukemogenic cells
in irradiated regions, and overestimation either of the risk coefficient for
low-LET radiation or of the relative effectiveness of high-LET radiation.
Data on patient3 treated with thorium suggest that the risk coefficients for
high-LET and low-LET leukemia induction may be approximately equal (Mays
et al. 1985a, NAS 1988).
2. Patients Injected with Ra-224 During Medical Treatment
Epidemiologic studies of patients treated with injections of Ra-224 for
ankylosing spondylitis (spinal arthritis) or tubercular infection of the bone
also provide information on the health effects of internally-deposited radium
{Mays and Spiess 1964, Mays et al. 1985b, Spiess et al. 1989). The half-life
of Ra-?24 is only 3.64 dayB (see Figure II-2), and so most Ra-224 decays while
V7.-12
-------
on the surface of bone, in contrast to Ra-226 and Ra-228 which are buried in
the bone volume by new bone growth (NAS 1938).
Noncancer Effects
Mays et al. (1985b) briefly discuss non-malignant diseases observed at
increased incidence in patients injected with Ra-224. These diseases include
exostoses (benign bone growths) and severe growth retardation in children and
tooth breakage, kidney disease, liver disease and cataracts in both children
and adult9. No dose-re9ponse information was provided for these non-malignant
lesions, but average injected dose was approximately 300 /iCi Ra-224 (Mays
et al. 1985b). Radiographic analysis for bone necrosis has not been reported
for these patients.
Cancer
Sr>iess et al. (1989) summarized follow-up data on 899 located German
patients injected with Ra-224 shortly after World War II. Table VI-4 presents
the rate of bone sarcomas among these patients as a function of skeletal dose
in rads. Doses were substantially higher in juveniles than in adults.
Appearance times of bone sarcomas ranged from 3.5 to 33 years, averaging about
11 years for both juveniles and adults. The last bone sarcoma appeared in
1983. Only 0.2 to 0.3 cases of bone sarcoma would be expected in the total
cohort. The risk of bone sarcoma appeared to be independent of age at
exposure to Ra-224 but was increased by protraction of the dose (Mays and
Spiess 1984).
VI-L3
-------
Table VI-4 Bone Sarcomas Rates among Patients
Injected with Ra-224
Age at First
Injection,
years
I-5
6-10
II-15
16-20
21-29
30-39
40-49
50-59
60-73
Average
Skeletal
Dose, rads
1,674
1,204
1,010
734
280
202
169
175
133
Number
Examined
34
71
51
62
170
191
200
93
26
Bone
Sarcomas
9
14
8
4
9
3
5
1
0
Bone Sarcomas
per million
rads
158
164
195
88
189
78
148
61
0
Adapted from Mays and Spiess 1984
VI-14
-------
Spieas et al. (1989) reported that the incidence of breast cancer was
increased among 107 females injected as juveniles for treatment of
tuberculosis (8 observed vs. 0.6 to 0.9 expected). Exposure to other
carcinogens, delayed childbirth, obesity and family history ware possible
confounding factors, but were not considered to account for the high relative
risk found. A significant excess of liver cancers was found in the whole
cohort (6 observed vs. 1.1 to 1.2 expected). No cases of head sinus
carcinomas or malignant melanomas were observed.
Spiess et al. (1989) reported that leukemias occurred in 6 patients in
their cohort, 4 in ankylosing spondylitis pfttients and 2 in tuberculosis
patients. The alpha particle dose to red m.irrow among all patients was
estimated to average roughly 200 rads. The expected number of leukemias among
the 396 ankylosing spondylitis patients was 5, based on the elevated rate of
leukemia among ankylosing spondylosis patients attributed to treatment with
pain-relieving drugs such as phenylbutazone. The expected number of leukemias
among the 504 tuberculosis patients is 1, based on German national rates. One
of the 2 leukemias found among the tuberculosis patients occurred in a
juvenile patient and was a type of leukemia not associated with radiation
exposure. Thus, only one of these leukemia cases is likely to be attributed
to Ra-224 exposure, the one occurring in an adult treated for tuberculosis.
Mays and Spiess (1984) report that 13 bone sarcomas occurred among adults
treated for tuberculosis. Thus, in this cohort, the ratio of leukemias to
bone sarcomas is between 0 and 1/13.
VI-15
-------
3. Community Studies
Petersen et al. (1966) studied mortality statistics of almost one million
rural residents of Iowa and Illinois who had an average of 4.7 pCi Ra-226/L in
their drinking water. Compared to controls with <1 pCi Ra-226/L in drinking
water, fatalities from bone malignancies were elevated, but the statistical
significance was marginal (P<0.08). No dose-response analysis was reported.
The BEIR IV committee of the National Academy of Sciences pointed out that the
fatality rate was also elevated in the city of Chicago, with 0.03 pCi Ra-226/L
in the water supply, and that confounding factors of population mobility and
variations in actual radium intake made evaluation of the study difficult (NAS
1988).
Bean et al. (1982) studied residents of small communities in Iowa and
found that the incidence of 4 out of the 10 cancers investigated increased
with increasing radium content of the water supply. These cancers were
bladder and lung cancer for males and breast and lung cancer for females.
Bone cancer, stomach cancer and leukemia were not investigated because of the
small numbers of observed and expected cases. Differences in smoking rates
were not considered to be responsible for the effects in men, but correlation
with indoor radon concentrations could not be ruled out (Bean et al. 1982).
The BEIR IV committee pointed out that including populations in similar
communities with drinking water from surface water sources as a low-dose group
substantially weakened the dose-response relationships found by Bean et al.
(1982), and also that these associations were not confirmed by the studies of
radium dial painters (NAS 1988).
VI-16
-------
Lyman et al. (1983) investigated the correlation between radium content
of groundwater and leukemia incidence in 27 Florida counties. When counties
were divided into 4 groups on the basis of the number of private wells with
radium content over 5 pCi/L, a small but consistent excess of leukemias was
found in high-exposure areas, but there was no evidence of a dose-response
trend (Lyman et al. 1985). Rates of colon, lung and breast cancer and of
lymphoma showed no consistent excess (Lyman and Lyman 1986). The significance
of these results has been debated, particularly in regard to the lack of an
observed increase in leukemias among radium dial painters (Stebbings et al.
1986, Polednak 1986, Tracey and Letourneau 1986, NAS 1988).
C. Hiah-Risk Populations
Women may be more sensitive than men to induction of bone sarcoma by
Ra-226 or Ra-228. Rowland et al. (1983) compared the observed number of bone
sarcomas among males exposed to radium with the expected number predicted from
the dose-response relationship for female dial painters. For these analyses,
exposure was expressed as uptake in (/jCi Ra-226 + 2.5 fjCL Ra-228).
Significantly fewer bone sarcomas than predicted were found among males, and
most of the deficit occurred among males first expou^J after age 35. In
contrast, there was good agreement between observed and predicted incidence
among non-dial painting females exposed to radium. This analysis suggests
that women may be more sensitive than men to induction of bone sarcoma by
Ra-226 or Ra-228. However, Mays and Spiess (1984) report no sex difference in
bone cancer induction by Ra-224.
There is conflicting evidence concerning the possible greater sensitivity
of children to radium. Rowland et al. (1983) reported that among 75 female
VI-17
-------
dial painter9 first exposed to Ra-226 and/or Ra-228 between ages 10 and 14,
4 bone sarcomaB were observed/ while 1.7 are predicted based on the entire
cohort. However, for 4 other girls exposed to Ra-226 and/or Ra-228 between
ages 10 and 14, the 1 observed bone sarcoma closely matches the 1.1 predicted.
The bone ratio of Ra-226 to calcium with age has been fourd to be nearly
independent of age by most investigators (Hallden et al. 1963, Rajewsky at al.
1965), which implies that in the growing bones of children, greater radium
uptake is counterbalanced by greater radium release during remodelling (Wrenn
et al. 1985a). For the short-lived Ra-224, most decay would occur prior to
release from remodelling, and Mays and Spiess (1984) calculated that the
amount of Injected Ra-224 that is absorbed in bone and subsequently decays is
60% between ages 1 and 15, 40% between ages 16 and 20 and 20% after age 21.
With this assumption, the rates of bone sarcoma induction in children and
adults calculated by Mays and Spiess (1984) were not substantial different.
On the other hand, an excess of breast cancer was found only among females
exposed to Ra-224 as juveniles (Spiess et al. 1989), and children are more
susceptible to non-cancer effects of Ra-224 exposure such as tooth breakage,
growth stunting and exostoses (Mays et al. 1985b). Mays et al. (1985a) point
out that in calculating cancer risk from lifetime exposure to radium, any
increased risk during the childhood period would have relatively little effect
on total cancer risk, due to the short duration cf childhood. The uncertainty
in the sensitivity of children to radium is discussed further in Chapter IX.
0. summary
The experience of the radium dial painters clearly establishes that
ingestion of Ra-226 or Ra-228 can cause bone necrosis (Keane et al. 1983),
that ingestion of Ra-226 can cause bor.e sarcomas and head carcinomas, and that
Vl-lfl
-------
ingestion of Ra-228 can cause bone sarcomas (Rowland et al. 1978). Bone is a
target organ of radium toxicity because of the preferential accumulation of
radium in bone. Head carcinomas are believed to result from the accumulation
of radon gas, a decay product of radium, in the cranial sinuses. Head
carcinomas do not occur following ingestion of Ra-228 because its radon decay
product ha3 too short a half life for substantial accumulation to occur
(Rowland et al. 1978). Rates of other cancers are not substantially elevated
among radium dial painters (Stebbings et al. 1984). Studies of patients
injected with Ra-224 show induction of bone sarcoma but not head carcinomas
(Mays and Spiess 1984). In addition, an elevated rate of breast and liver
cancer is found among these patients (Spiess et al. 1989).
Three studies of populations in the United States exposed to radium in
drinking water have found elevated rates of bone cancer (Petersen et al.
1966), bladder, lung and breast cancer (Bean et al. 1982) or leukemia (Lyman
et al. 1985). However, the differing types of cancers that are increased
among the three studies makes attributing these effects to radium
questionable.
The uncertainty concerning the health effects of radium exposure in
humans and its influence on the quantification of toxicological effects is
addressed in Chapter IX.
VI-19
-------
VII. MECHANISM OF TOXICITY
A. Effects of Ioni2ino Energy
The potential for Ionizing radiation to cause tissue damage and cancer
was established within X5 years after the discovery of X-rays (USEPA 1989b,
Upton 1975). Extensive studies have been done on radiation effects in man and
animals following exposure to X-rays, gamma rays, radium, radon, thorium and
other sources of ionizing radiations.
Ionizing radiation is not an organ- or tissue-specific carcinogen.
Storer (1975) stated: "Ionizing radiation in sufficiently high dosage acta as
a complete carcinogen in that it serves as both initiator and promoter.
Further, cancers can be induced in nearly any tissue or organ of man or
experimental animals by the proper choice of radiation dose and exposure
schedule." The following tissues have shown the induction of cancers in man
by radiation in one or more epidemiological studies: thyroid, female breast,
lung, bone marrow (leukemia, multiple myeloma), stomach, liver, large
intestine, small intestine, brain, salivary glands, bone, esophagus, urinary
bladder, pancreas, rectum, lymphatic tissues, skin, pharynx, uterus, ovary,
mucosa of cranial sinuses and kidney (USEPA 1989b).
The damage caused by ionizing radiation involves a number of rapid
radiochemical reactions. Absorption of energy frou alpha and beta particles
induces ions and excited radicals In matter, particularly in aqueous
solutions. Absorption of energy from X- and gamma-radiations by atoms results
in the ejection of electrons from the atoms. These ejected high speed
electrons then produce ionizations in the same manner as the alpha and beta
VII-l
-------
particles. These types of interactions occur in about the first 10"'4 seconds
(Bacg and Alexander 1961, Andrews 1961),
The ions and excited radicals induced in cells (aqueous solutions) occur
both in organic molecules of the cell contents and in the water Burrounding
them. In addition, specific compounds, such as H202, and radicals, such as HOj
are formed in the water phase. Further chemical interactions can then occur
with proteins, sugars, lipids, enzymes and more importantly, ONA and RNA.
Types of damage to DNA and other organic molecules include bond scission,
chain scission, cross-linking, single and double strand breaks and development
of adducts. These interactions are generally complete by about 10'4 seconds
after initial exposure (Bacq and Alexander 1961, Andrews 1961).
Ionizing radiation delivered at sufficiently high dose and dose rate
leads to cellular death and eventually death of a living organism. At lower
doses and dose rates ionizing radiations induce cancer. The ionizing
radiations can inhibit mitosis and cause chromosome aberrations including
nondisjunction, clastogenesis and point mutations (Bacq and Alexander 1961,
Andrews 1961). These effects may be demonstrable minutes to years after
exposure.
while carcinogenesis and mutagenesis has not been investigated for every
radionuclide, the results of studies that have been performed are consistent
with expected effects of ionizing radiation. Mammalian cells in culture have
been transformed; chromosome aberrations have been observed in cultured
peripheral lymphocytes; even activation, by frame shift due to a single base
deletion in an oncogene, has been reported (USEPA 1989b). Induction of
VII-2
-------
enzymes of unscheduled DNA repair, perhaps signaling error-prone repair, has
also been noted (Tuschl et al. 1980, 1983, Olivieri et al. 1984).
Continuing epidemiological studies have shown increased cancer mortality
in persons exposed to x-rays, gamma rays and various internal emitters such as
radon, radium and thorium, at various ages {including in vttero) (USEPA 1989b).
Additional support comes from the studies of mice, rats, hamsters, guinea
pigs, cats, dogs, sheep, cattle, pig? and monkeys that have demonstrated
increased incidence of cancer following exposure to a source of ionizing
radiation (USEPA 1989b).
B. Radionuclide Dosimetry
Ionizing radiation is known to induce cancer in many organs and tissues,
and so risk estimates for internally deposited radionuclides are based on the
distribution of the radionuclides in tissues and the ionizing radiation dose
associated with that distribution. The toxicokinetic portion of the RADRISK
model used to estimate distribution is described in Section IIZ.O. As
explained there, the radionuclide concentration in various organs is
calculated for a lifetime intake of 1 pci per year. For all organs or tissues
where there is no organ-specific model, the activity in the organ compartment
"other" is considered to be uniformly distributed in all body tissues not
specified.
The rate at which radiation is absorbed in a given organ depends on the
concentration of each radioactive element in that organ and in surrounding
organs, the frequency with which each element decays, the energy of each decay
event and the fraction of emitted energy that is absorbed in the particular
VII-3
-------
organ (Sullivan et al. 1981). The concentrations of radionuclides resulting
from radium intake are calculated using the biofcinetic model described in
Section IXI.D. The frequency and energy of the radionuclide decay events are
well-known physical properties. Tor alpha and beta particles, the specific
absorbed fraction (see below) is assumed to be the inversely proportional to
the mass of a given target organ; cross-irradiation from sources outside the
target tissue is assumed to be negligible except for sources tissues in close
proximity (USEPA 1989b», [e.g., source skeletal tissue and target and active
bone marrow and endosteal cells (Thorne 1977, ICRP 1980)]. Absorption of
photons {i.e., gamma irradiation) is calculated by the Monte Carlo method
(CJSEPA 1989b).
The equation used to calculate the rate of radiation absorption from
element i in organ X is:
M
D1 (X;t) = 2 Alk(t> SL(XHfk} (VII-1)
k=l
where:
DJX;t) =» the absorbed dose rate to organ X at time t due to radionuclide
i in source organs Y,, The first source organ Y, is
organ X.
M ® the total number of source organs contributing irradiation to
organ X.
Au(t) * the activity at time t of radionuclide i in source organ
-------
St(X*-YJ B the dosimetric source (S) factor for irradiation of organ
X by source organ
The S factor is calculated by the following equation:
N
r f. E, *.(X-Y„)
m»l
(VII-2)
where:
c ¦ a constant that depends on units of dose, energy and time being
used.
N » the total number of decay events m (e.g., alpha decay, beta
decay).
f. = intensity of decay event m (number per disintegration).
E. = average energy of decay event m in Mev.
$a{X«-¥„) = specific absorbed fraction, ie., the fraction of emitted energy
from 9ource organ Y, absorbed by target organ X per gram of
tissue in organ X.
The S factor is similar to the SEE factor used by the ICR? in ICRP
Publication 30. However, the SEE factor includes a quality factor for high-
LET radiation whereas the RADRISK calculations carry forward high-LET and low'
LET dose rates as separate entities (Dunning et al. 1984, USEPA 1989b).
VII-5
-------
The absorbed dose rate for each target organ is calculated by summing
equation VII-1 over all radionuclides i. The output front these calculations
is a list of organs with doses of high-LET and low-LET radiation at various
ages corresponding to ingestion of 1 pci/year of Ra-226 or Ra-228. The
RAERISK model combines these calculated organ-specific doses with risk
coefficients derived from quantitative evaluation of human cancer risk from
various types of ionizing radiations (see Section VIII,0) to estimate organ-
specific and total risk corresponding to ingestion of Ra-226 or Ra-228.
Uncertainties in dosimetric calculations are discussed in Chapter IX*
C. Summary
The harmful effects of ionizing radiation are believed to be the result
of damage to DNA and other cellular components caused by chemical reaction
with ionised molecules, particularly active oxygen species (Sacq and Alexander
1961, Andrews 1961). The dose of radiation resulting from ingestion of radium
varies from organ to organ. Models have been developed to calculate the rate
at which radiation is absorbed in a given organ, based on the concentration of
radioactive elements in that organ and in surrounding organs, the frequency
with which each element decays, the energy of each decay event and the
fraction of emitted energy that is absorbed in the particular organ (CTSEPA
1989b). Uncertainties in dosimetric calculations are discussed in Chapter IX.
VII-6
-------
VIII. QUANTIFICATION OF TOXICOLOGICAL EFFECTS
General Approach for Drinking Water contaminants
This introductory section summarizes the general approach used to
evaluate the ha2ard of contaminants in drinking water. This text appears in
all criteria documents to provide information relevant to the basic issue of
how toxicological effects are quantified.
The quantification of toxicological effects of a chemical consists of
separate assessments of noncarcinogenic and carcinogenic health effects.
Chemicals which do not produce carcinogenic effects are believed to have a
threshold dose below which no adverse, noncarcinogenic health effects occur,
while carcinogens are assumed to act without a threshold.
Quantification of Noncarcinogenic Effects
In the quantification of noncarcinogenic affects, a Reference Dose (RfD),
(formerly called the Acceptable oaily intake (ADI)), is calculated. The RfD
is an estimate of a daily exposure to the human population that is likely to
be without appreciable risk of deleterious health effects, even if exposure
occurs over a lifetime. The RfD is derived from a No-Oboerved-Adverse-Effeet
Level (NOAEL), or Lowest-Observed-Adverse-Effect Level (LOAEL), identified
from a 3ubcfcronic or chronic study, and divided by an uncertainty factor(s).
The RfD is calculated as follows:
RfD ¦ m mg/kg bw/day
Uncertainty Factor(s; *
VIII-1
-------
Selection of the uncertainty factor to be employed in the calculation of
th9 RfD is based on professional judgment, while considering the entire data
base of toxicological effects for the chemical. In order to ensure that
uncertainty factors are selected and applied in a consistent manner, the
Office of Hater (OW) employs a modification to the guidelines proposed by the
National Academy of Sciences (NAS 1977, 1980a) as follows:
• An uncertainty cactor of 10 is generally used when good chronic, or
subchronic human exposure data identifying a NOAEL are available and
are supported by good chronic, or subchronic toxicity data in other
species.
• An uncertainty factor of 100 is generally used when good chronic
toxicity data identifying a NOAEL are available for one or more
animal species (and human data are not available), or when good
chronic, or subchronic toxicity data identifying a LOAEL in humans
are available.
• An uncertainty factor of 1,000 is generally used when limited or
incomplete chronic, or 3ubchronic toxicity data are available, or
when good chronic, or subchronic toxicity data identify a LOAEL, but
not a NOAEL for one or more animal species are available.
The uncertainty factor used for a specific risk assessment is based
principally on scientific judgment rather than scientific fact and accounts
for possible intra- and interspecies differences. Additional considerations
not incorporated in the NAS/OW guidelines for selection of an uncertainty
factor include the use of a less than lifetime study for deriving a RfD, the
VI11-2
-------
significance of the adverse health effect and the counterbalancing of
beneficial effects.
From the RfD, a Drinking Water Equivalent (DWEL) can be calculated. The
DWEL represents a medium specific (i.e., drinking water) lifetime exposure at
which adverse, noncarcinogenic health effects are not anticipated to occur.
The DWEL assumes 100% exposure from drinking water. The DWEL provides the
noncarcinogenic health effects basis for establishing a drinking water
standard. For ingestion data, the DWEL is derived as follows:
DWEL - (PEP) x (Bo<*Y We^t j,n Kqi „ /L
Drinking Water Volume in L/day
where:
Body weight = assumed to be 70 kg for adult
Drinking water volume = assumed to be 2 liter? per day for an adult
In addition to the RfD and the DWEL, Health Advisories (HAs) for
exposures of shorter duration (One-day, Ten-day and Longer-terra) are
determined. The HA values are used as informal guidance to municipalities and
other organizations when emergency spills or contamination situations occur.
The HAs are calculated using a similar equation to the RfD and DWEL; however,
the NOAELs or LOAELs are identified from acute or subchronic studies. The HAs
are derived as follows:
HA = !*" = «*/!.
(UF) x ( L/day) "
vii:-3
-------
Using the above equation, ehe following drinking water HAs are developed
for noncarcinogenic effects:
1. One-day HA for a 10-kg child ingesting 1 L water per day.
2. Ten-day HA for a 10-kg child ingesting 1 L water per day.
3. Longer-term HA for a 10-kg child ingesting 1 L water per day.
4. Longer-term HA for a 70-kg adult Ingesting 2 L water per day.
The One-day KA calculated for a 10-kg child asEumes a single acute
exposure to the chemical and is generally derived from a study of less than
7 days duration. The Ten-day HA assumes a limited exposure period of
1-2 weeks and is generally derived frcm a study of less than 30-daya duration.
A Longer-term HA is derived for both the 10-kg child and a 70-kg adult and
assumes an exposure period of approximately 7 years {or 10% of an individual's
lifetime). A Longer-term HA is generally derived from a study of subchronic
duration (exposure for 10% of animal's lifetime).
Quantification of Carcinogenic Effects
The EPA categorizes the carcinogenic potential of a chemical, based on
the overall weight-of-evidence, according to the following scheme:
• Group A: Human Carcinogen. Sufficient evidence exists from
epidemiology studies to support a causal association between
exposure to the chemical and human cancer.
-------
• Group B: Probable Huinan Carcinogen. Sufficient evidence of
carcinogenicity in animals with limited (Group Bl) or inadequate
(Group B2) evidence in humans.
• Group C: Possible Human Carcinogen. Limited evidence of
carcinogenicity in animals in the absence of human data.
• Group D: Not classified as to Human Carclnooenicitv. Inadequate
human and animal evidence of carcinogenicity or for which no data
are available.
• Group E: Evidence of Woncarclnooenicltv for Humans. No evidence of
carcinogenicity in at least two adequate animal tests in different
species or in both adequate epidemiologic and animal studies.
If toxicological evidence leads to the classification of the contaminant
as a known, probable or possible human carcinogen, mathematical models are
used to calculate the estimated excess cancer risk associated with the
ingestion of the contaminant in drinking water.
Quantification of Cancer Risk for Chemicals
For chemicals, the data used in these estimates usually come from
lifetime exposure studies in animals. In order to predict the risk for humans
from animal data, animal doses must be sorverted to equivalent human doses.
This conversion includes correction for noncontinuous exposure, less than
lifetime studies and for differences in size. The factor that compensates for
tho size difference is the cube root of the ratio of the animal and human body
VII1-5
-------
weights. It is assumed that the average adult human body weight is 70 kg and
that the average water consumption of an adult human is 2 liters of water per
day.
For contaminants with a carcinogenic potential, chemical levels are
correlated with a carcinogenic risk estimate by employing a cancer potency
(unit risk) value together with the assumption for lifetime exposure via
ingestion of water. The cancer unit risk for chemicals is usually derived
from a linearized multistage model with a 95% upper confidence limit providing
a low dose estimate; that is, the true risk to humans, while not identifiable,
i9 not likely to exceed the upper limit estimate and, in fact, may be lower.
Excess cancer risk estimates may also be calculated using other models such as
the one-hit, Weibull, logit and probit. There is little basis in the current
understanding of the biological mechanisms involved in chemically-caused
cancer to suggest that any one of these models is able to predict risk more
accurately than any others. Because each model is based upon differing
assumptions, the estimates which were derived for each model can differ by
several orders of magnitude.
The scientific data base used to calculate and support the setting of
cancer risk rate levels for chemicals has an inherent uncertainty due to the
systematic and random errors in scientific measurement. In most cases, only
studies using experimental animals have been performed. Thus, there is
uncertainty when the data are extrapolated to humans. When developing cancer
risk rate levels, several other areas of uncertainty exist, such as the
incomplete knowledge concerning the health effects of contaminants in drinking
water, the impact of the experimental animal's age, sex And species, the
nature of the target organ system(s) examined and the actual rate of exposure
VIII-6
-------
of the internal targets in experimental animals or humans. Dose response data
usually are available only for high levels of exposure, not for the lower
levels of exposure closer to where a standard may be set. When there is
exposure to more than one contaminant, additional uncertainty results from a
lack of information about possible synergistic or antagonistic effects.
Quantification of Car.cer Risk for Radionuclides
For radionuclides, human epidemiologic data rather than animal
experiments form the basis of the cancer risk rate levels, which are derived
as be9t estimates rather than as 95% upper confidence limits of the unit risk
from a linearized multistage model. The true risks to humans may be higher or
lower than the predicted risks, but the overall uncertainty is probably less
than an order of magnitude. Because human data are used, the individual sites
of cancer are predicted, as well as the total risk. Therefore, projections
can be made both of cancer incidence and of cancer fatality, which are related
for a given organ to site-specific survival data, which ranges from 90%
survival (10% mortality) for thyroid cancer to 0% survival (100% mortality)
for liver cancer (USEPA 1989b).
A. Noncarcinooenic Effects
The only r.oncancer health effect of radium ingestion with an adequately
characterized dose-response relationship in humans is bone necrosis (Keane
et al. 1983). Other noncancer effects appear to be less senuitive indicators
of radium toxicity in humans (Rundo et al. 1986). Animal experiments have
confirmed that bone necrosis is an adverse effect of radium exposure (Taylor
et al. 1976, Jee et al. 1969, Momeni et al. 1976).
-------
Keane at al. (1933) used ICRP models for the metabolism and dosimetry of
Ra-226 and Ra-228 to predict that the dose of radiation to the endosteal layer
is nearly independent of time-course of exposure to Ra-226 or Ra-228.
Assuming this calculation is correct, the data reported by Keane ee al. (1983)
may be used to derive Health Advisories for all durations as well aa a
Drinking Water Equivalent Level for lifetime exposure.
Keane et al. (1983) found that below a total intake"to the blood
(absorbed dose) of 10 pCl of either Ra-226 or Ra-228, observed bone changes
were not significantly different from controls. To use this value to
calculate HAs and DWELs, the following aasumptions will be made:
• The average gastrointestinal
Maletskos et al. (1966) will
(10 fiCi)/0.2 ¦ 50 fjCL Ra-226
absorption factor of 20% measured by
be used. This results in a NOAEL of
or Ra-228 for adultn.
• Based on the near constancy of the bone radium/calcium ratio with
age (Hallden et al. 1963, Wrenn et al. 1985a), children are assumed
to absorb radium at the same rate as adults from food or water.
• Bone necrosis is assumed to be a function of radium intake scaled by
body weight/ a surrogate for skeletal weight. For a child weighing
10 kg, the NOAEL is (50 ^Ci) x {10 kg/70 kgl * 7 ^Ci.
• . a adjustment for duration of exposure will be made for time periods
up to 10% of lifetime. For lifetime values, the dosimetric ratios
of instantaneous to lifetime intake calculated by Keane et al.
(1983) will be used (1.6 for Ra-226 and 1.2 for Ra-228).
vin-a
-------
HA and DWEL values £or Ra-226 and Ra-220 ace calculated using these
assumptions in the following sections. The uncertainties associated with
these assumptions and the effect on calculated HA and DWEL values are
discussed in Chapter IX.
1. One-Day Health Advisory
The One-day HAs for the 10-kg child are calculated a» followst
» - ,10) u Wd^'llly) ¦ <>•' "ci «*-22t'L
• 700,000 pCi Ra-226/L
One-day m = ,10, {I ffL??*"'*.,, " °'7 «cl
a 700,000 pCi Ra-228/L
where:
7 pCi Ra-226 = NOAEL for children, based on absence of bone necrosis
in adults exposed to Ra-226 by ingestion.
7 fiCi Ra-228 = NOAEL for children, based on absence of bone necrosis
in adults exposed to Ra-228.by ingestion.
10 = uncertainty factor; this uncertainty factor was
chosen In accordance with NAS/OW guidelines in ..hich
a NOAEL from a study in humans is employed.
1 L/day 3 assumed water consumption by a child.
Vm-9
-------
1 day - assumed duration of water consumption.
2. Ten-Day Health Advisory
The Ten-day HAs for the 10-kg child are calculated as follows:
*«<-<"* • no !i gynsw,- °-7 "ci
¦ 70,000 pCi Ra-226/L
» • (10, !I Wd»»"uo'a.y)- °-7 "C1
¦ 70,000 pel Ra-228/L
where:
7 ijCL Ra-226 - NOAEL for children, based on absence of bone necrosis
in adults exposed to Ra-226 by ingestion.
7 pCi Ra-228 « NOAEL for children, based on absence of bone necrosis
in adults exposed to Ra-228 by ingestion.
uncertainty factor; this uncertainty factor was
chosen in accordance with NAS/OW guidelines in which
a NOAEL from a study in humans is employed.
assumed water consumption by a child.
assumed duration of water consumption.
1 L/day =
10 days »
VIII-10
-------
3. Longer-Term Health Advisory
The Longer-term HAS for the 10-kg child are calculated as follows:
Longer-term HA . ,10, d.ys, * 274 "L *-""<¦
(rounded to 300 pCi Ra-226/L)
L=„,"-ten. » ¦ ,10) days) - "4
(rounded to 300 pCi Ra-228/L)
where:
7 fjCL Ra-226 ¦ NOAEL for children, based on absence of bone necrosis
in adults exposed to Ra-226 by ingestion.
7 fjCL Ra-223 « NOAEL for children, based on absence of bone necrosis
in adults exposed to Ra-228 by ingestion.
10 a uncertainty factor; this uncertainty factor was
chosen in accordance with NAS/OW guidelines in which
a NOAEL from a study in humans is employed.
1 L/day 3 assumed water consumption by a child.
2,557 days ®
assumed duration of water consumption (7 years,
365.25 days/year).
-------
The Longer-term HAs for a 70-kg adult consuming 2 L/day of water are
calculated as follows:
Long«r-t«tm HA - (10| * »« P« *•-««/!.
(rounded to 1,000 pCi Ra-226/L)
Longer-term KA - (10)
-------
4. Reference Dose and Drinking Water Equivalent Level
The DWEL is derived as follows:
Step 1: Determination of RfD
R£D ¦ UQL Rfl"22§) (lt§) j nci Ra-226/Jca-dav
RfD (10) (70 kg) (25,568 days) 4,47 pCl Ra 226/kg day
(rounded to 4 pCi Ra-226/kg-day)
B«n 150 uCl Ra-228) (1.2) . p / Vrr-»H*v
RfD (10) (70 kg) (25,568 days) ° 3,36 pCi /*?"***
(rounded to 3 pCi Ra-228/kg-day)
where:
50 /jCi Ra-226 ¦ NOAEL for adults, based on absence of bone necrosis
in humans exposed to Ra-226 by ingestion.
50 pCi Ra-228 » NOAEL for adults, based on absence of bone necrosis
in humans exposed to Ra-228 by Ingestion.
1.6 ¦ calculated dosimetric ratio of effectiveness for
instantaneous vs. 50-year exposure to Ra-226 (Keane
et al. 1983).
1.2 ¦ calculated dosimetric ratio of effectiveness for
instantaneous vs. 50-year exposure to Ra-228 (Keane
et al. 1983).
VIII-13
-------
10 ¦ uncertainty factor; this uncertainty factor was
chosen in accordance with NAS/OW guidelines in which
a NOAEL from a ntudy in humans is employed.
70 - assumed body weight of an adult.
25,568 days * assumed duration of exposure (70 years, 365.25 days
per year).
Step 2: Determination of DtTELs
DHEL - PO Hal . 14(J pC1 M.jj6/L
(rounded to 100 pCi Ra-226/L)
OKEt - " f? M . 105 pci B.-J28/L
(rounded to 100 pCi Ra-228/L)
where:
4 pCi Ra-226/kg-day » RfD for Ra-226
3 pCi Ra-228/Jcg-day = RfD for Ra-223
70 kg = assumed weight of adult.
2 L/day a assumed water consumption by a 70-kg adult.
VIII-14
-------
B. Carcinogenic Effacta
1. Categorization of Carcinogenic Potential
Radium is classified as a Group A carcinogen, a human carcinogen (USEPA
1991b). This classification is based on the known human carcinogenicity of
ionizing radiation, supported by sufficient evidence from epidemiologic and
experimental studies establishing the cancer-causing potential of Ra-224,
Ra-226 and Ra-228.
2. Quantification of Carcinogenic Risk
Two basic approaches may be used to assess the risk from low doses of
Ra-226 and Ra-228. The first is to fit dosa response models to the
epidemiological data concerning the incidence of bone and head cancers as a
function of estimated intake of Ra-226 and Ra-22S among radium dial painters
(Mayn et al. 198Sa, NAS 1988). The second is to quantify the dose of ionizing
radiation delivered to radiosensitive organs by intake of radium and to
calculate cancer risk based on a synthesis of human epidemiologic data on the
carcinogenic potency of radiation (USEPA 1989b). This second approach is
implemented by the RADRXSK model, the Agency's dosimetry/risk model approach.
The RADRISK model, described in Sections IIX.D and VII.B, uses a metabolic
kinetic model to predict the dose of radiation received by body organs due to
ingested radium. risk of cancer is calculated based on epidemiologic
data for exposure of humans to several types of ionizing radiation, including
patients injected with Ra-224 (USEPA 1989b). Each approach is summarized
below.
VIII-15
-------
A third possible approach, which was described in the Advanced Notice of
Proposed Rulemaking (USEPA 1986), is the use of ICRP effective dose
equivalents. The effective dose equivalent approach is quite similar to the
RADRXSK model, differing primarily in the details of che conversion of organ-
specific doses to lifetime excess cancer risks (USEPA 1989b). The effective
dose equivalent approach is not used in this document to calculate risks.
For purposes of risk assessment, the Agency assumes a linear relationship
between doee and cancer risk. For radiogenic cancers, this approach has been
specifically recommended by the Radiation Advisory Committee of the USEPA
Science Advisory Board (SAB 1987). Data from both human and animal studies
generally support a linear relationship between radiation dose and cancer risk
(USEPA 1989b). However, an exception is bone sarcoma risk among dial
painters, which is best fit by a quadratic (dose-squared) response to the
estimated dose (Rowland et al. 1978, 1983, NAS 1988). In contrast, both head
carcinoma risk among dial painters exposed to Ra-226 and bone sarcoma risk
among patients exposed to Ra-224 show a linear response (NAS 1988). In
addition, animal experiments using Ra-226 or Ra-228 show a linear response of
bone sarcomas (Finkel et al. 1969b, Wrenn et al. 1985b, Mays et al. 1987).
Linear coefficients for bone sarcoma risk from Ra-226 and Ra-228 ingestion can
be derived as upper bounds on the dial painter data (Mays et al. 1985a,
Schlenker 1982), as described below. However, rather than using an upper
bound, the Agency's approach to radiogenic risk is to use best estimates.
Therefore, radium risks are evaluated using the RADRISK model, which
incorporates linear risk factors derived as best estimates from a synthesis of
epidemiologic data on human responses to several types of ionizing radiations,
including patients injected with known amounts of Ra-224. The dial painter
data are used in interpretation of the risks predicted by the RADRISK model,
VIII-16
-------
Co incorporate a head carcinoma risk for Ra-225 and to adjust for the over-
prediction of leukemias by the RADRISK model. These issues are discussed in
greater detail in the remainder of this section and also in Chapter IX.
Radium exposure in animals causes bone sarcomas, and possibly leukemiaa
or lymphomas. Quantitative evaluation of animal carcinogenicity studies
requires using models to estimate the skeletal doses resulting from single or
multiple injections of radium isotopes {Taylor et al. 1983). A few
investigators have compared the carcinogenic potency of radium in animals and
in humans, and have concluded that humans are more sensitive than rats (Evans
et al. 1944), that humans are less sensitive than beagle dogs (Raabe et &1.
1981) or that uncertainties about relative dosimetry due to differences in
bone structure prevent establishing a quantitative relationship between
potency in humans and mice (Finkel et al. 1969b).
Bone Sarcomas and Head Carcinomas Among Dial Painters
The principal population which forms the basis of quantitative
epidemiological studies of Ra-226 and Ra-228 exposure is a group of
approximately 2,000 persons, primarily women, who worked as radium dial
painters in the early part of this century and whose body burdens of radium
have been estimated (Rundo et al. 1986). Two types of cancer, both with very
low spontaneous incidences, are clearly elevated in this cohort: bone
sarcomas and head carcinomas (Rundo et al. 1986). Quantitative analysis is
subject to uncertainties concerning incidence, because of incomplete
identification of the cohort and selection bias in workers possibly measured
due to radium-related symptoms (NAS 1988). Estimated doses are also
uncertain, because measured body burden of Ra-226 at the time of first
VIII-17
-------
radioassay must be back extrapolated to initial intake of Ra-226 and Ha-228
(Sharpe 1974). The effects of these uncertainties are discussed in more
deta.iL in Chapter IX.
Recent analyses of the dial painter data have concentrated on
establishing dose-response relationships which can be used to estimate risks
at low (environmental) exposures to radium (NAS 19BB). Moat analyses indicate
that a quadratic response be3- describes the relationship in the observed
intake ranges, and only an upper limit an the coefficient for a linear
response at low intakes can be derived from the data (NAS 1988).
Some investigators have found that at lower radium dosfts, not only does
the Lifetime risk of tumors decrease, but also the time between radium
exposure and development of cancer may increase (Evans et al. 1972). This has
been suggested to result in a "practical threshold" (Evans et al. 1972), the
dose at which the median tumor appearance time would exceed the human lifetime
(Raabe et al. 1930, 1983). However, the weight of evidence does not support
such a practical threshold (Mays 1993). Minimum tumor appearance times in
humans do not decrease with dose the way median or maximum tumor appearance
times may, and bane sarcomas have occurred in humans at doses barely above the
postulated practical threshold (Hays 1983). In animal studies, the time to
tumors doubles with each 10-fold decrease in dose only at high and moderate
dose rates, but not at low tiose rates jMays 1988). The BEIR IV committee
pointed out that in dial painters, epidemiologic data are not sufficient to
establish the difference between a zero risk and a risk just below that
observable in the exposed cohort (NftS 1986). The effect of uncertainty
concerning the existence of a practical threshold on rieic estimates is
discussed in Chapter IX.
VIII-13
-------
Mess quancj.taci.ve analyses c£ risks of radi-m exposure hav« treated bone
sarcomas and head carcinomas separately. Art ex rep tier, is the study by Evans
et al. (1972), which found that when incidence was expressed as percent
cumulative incidence of both tumors, there was no excess risk below an average
skeletal dose of 1,CG0 rad, and a constant excess risk of 28 ± 6\ for average
skeletal doses between 1,0C0 and 50,000 rad. Such an unusual step-like dose-
response relationship indicates that the two types of cancer are probably best
analyzed separately.
Bone sarcomas. Several investigators have analysed the dose-response
relationship for bone sarcomas caused by radium exposure. Mays and Lloyd
(1972) expressed dose as average skeletal dose in rads, weighting Ra-226 and
Ra-228 equally, and fit a linear relationship giving a tumor incidence of
4.6E-5 per rad. A skeletal dose of 1 rad corresponds to lifetime intake of
2 L/day of water containing approximately 10 pCi Ra-225/L (NAS 1988). The
BEIR IV committee pointed out that the tumor incidence at low doses predicted
by Mays and Lloyd (1972) is below the spontaneous incidence of about 10'1,
that statistical analysis of the results of Mays and Lloyd (1972) indicates
that the response is non-linear with a high probability, and that a linear
analysis significantly overpredicts the observed lcw-dese tumor incidence (NAS
1988)
Rowland and co-workers have performed several analyses of the
epidemiological data with dose calculated as total intake of radium to blood,
weighting Ra-228 2.S times higher than Ra-226 (Rowland et al. 1978, 1983).
Total intake to blood in microcuries is back-calculated from current
measurements of Ra-226, the ratio of Ra-228 to Ra-226 at the facility and the
Norris retention function (Norris et al. 19SS). For estimates of dose, intake
VIII-19
-------
is assumed to have been constant throughout the period of employment
(Stebbings et al. 1984). Ingestion is converted to intake to blood using a
gastrointestinal absorption factor of 20% (Maletskos et al. 1966, 1969).
Incidence data in this cohort are listed in Tables VI-1 and vt-3. Ho analysis
was done of the uncertainty in incidence or dose in this cohort. Rowland et
al. (1978, 1983) found the base fit to tht bone sarcoma data using all
measured female dial painters to be the following responsei
I - 7.OE-0 Dv' exp(-0.0CU DJ (VIII-1)
where:
X - Yearly excess incidence (bone sarcomas per person-year).
DA » Total lifetime intake to the blood (;/Ci Ra-226 plus 2.5 pCi
Ra-228J.
Linear functions provided very poor fits to the data, rejected at the P<0.05
level. Rowland et al. (1933} demonstrated that three or more additional bone
sarcomas would have to be discovered among low-dose dial painterB for a linear
fit not to be rejected. Rowland et al. (1983) also demonstrated that changing
the intake ranges produced acceptable linear fits only under special
conditions.
Rowland et al. (2983) derived a subgroup of dial painters by excluding
all women who developed cancer within two years of being measured for radium
body burden. The data from this cohort, Subgroup B, are listed in
VllI-20
-------
Section vr.B.l, Table v;-3. two i-jr.etians gave *ccepsaile fits to these data:
I • 1.8E-7 0/ exp(-0.0015 D,J (V::i-2>
i - 2.0S-5 o, (v:x:-3>
where * are1 D4 are defined as above (Rowland et al. 1983). To illustrate the
uncertainty about the dose-response relationship in this cohort. Table VIH-1
lists the data from Table VI-3, aIon? with the prediction* (torn
aquations IVIII-2} and (V13:~3).
Maya et al. (1935a) analysed the predicted bone-sarcoat iiicider.ee among
low-doee dial painters in the lar?er cohort based on equation (VIII-3}, and
concluded chat if the linear coefficient were reduced by a factor of 2, to
1E-5, the prediction could not be rejected at the P<0.05 level, schlenker
(1962) used the same definition of dose as Rowland et al. (1978) and analysed
the upper 68th and 95th percentile dose-response functions in the low-dose
regions. Both the 68th and 95th percentile functions had quadratic terms, but
they also included linear terns, which would dominate at low doses. The
coefficients for the Du terms were 6.1E-6 (68th percentile) and 1.6E-S (95th
percentile) (Schlenker 1982).
-------
Table vi::-l Observed and Predicted Bone Sarcomas per 1,000 Person-Years
at Risk among Female Radium Oial Painters
Excluding Workers Possibly Measured Due to Symptoms of Bone Sarcoma
Estimated Systemic
Intake {^Ci Ra-226
~ 2.S iiCi Ra-22Bl
£2,500
1,000-2,499
500-999
250-499
249-30.5
3one Sarcomas nar 1.000 P«r«en-Y»»r
Prrtlctrt
Equation VXXX-2
{Quadratic /Exponential >
a 20.0
46.2 38.4
41.3 27.3
14.6 14.3
0 S.3
Equation vxn-3
lLinear \
57.4
37.0
12.6
7.5
6.3
Adapted from Rowland et al. 1983
VIII-22
-------
To summarise, most analyses ci ber.e sarroea incidence found that a
quadratic relationship was the bast fit to the data, and that a linear
relationship is only plausible as an upper bound response at low doses.
Coefficients derived for the linear term range from about 6E-6 to 2K-S (#jCi
Ra-226 ~ 2.5 pCi Ra-228)'.
Head carcinomas. Only one separate analysis of head carcinomas has been
performed (Rowland et al. 1978). These investigators demonstrated that the
incidence of head carcinomas was clearly associated with exposure to Ra-226
and not exposure to Ra-228. This association is expected if the mechanism of
induction of sinus and mastoid cancer is accumulation and subsequent decay of
radon in air spaces in the head, because the radon decay product of Ra-226 is
too short lived for substantial accumulation in air cells to occur (KAS 1968).
Rowland et al. (1978) found the following linear relationship:
I - 1.6E-S (VIII-4)
where:
I » Yearly excess incidence (head carcinomas per person-year).
Dt a Total lifetime intake to the blood (juCi Ra-226).
Relationships with a quadratic term also provided acceptable fits to the head
carcinoma data (Rowland et al. 1978), and equation (VIII-4) may underpredict
the incidence at long times after first exposure (Rowland et al. 1978).
VIII-23
-------
relation te low do.^ jfay« •: al. C90Sa), at the request of
use?A, analyzed the risks corresponding to lifetime daily intake of 1 pel of
radium to tha blood, using the analyses of Rowland et al. (1978, 1983),
including the assumed 2.5-fold higher potency of fta-228 compared to Ra-226.
Lifetime intake of 1 pci of radium to the blood corresponds to e drinking-
water concentration of 2.5 pCi/L, assuming 2 I./day consumption of drinking
water and 20t gastrointestinal absorption. Mays et al. (198Sa) derived a
linear coefficient of 1E-S per pCi based on lack of statistical rejection of
predicted cancer incidence among low-dose members of the dial-painter cohort.
This results in the following risk equation:
I • 1E-5 (VIII-5)
where:
1 » Yearly excess incidence (bone sarcomas per person-year).
Dt » Total lifetime intake to the blood (jjCi Ra-226 plus 2.5 **Ci
Ra-228).
Hays et al. (1985a) used this relationship and equation VXII-4 for head
carcinoma risk, and assumed the exposure period was 65 years for bone sarcomas
<75 year lifetime minus a 10-year latency period) and 70 years for head
carcinomas (75 year lifetime minus a 5-year latency period). They calculated
the average yearly risk rate to be half of the risk rate at the end of the
exposure period and the cumulative risk to be the exposure period times the
average yearly risk rate. A more precise way to convert yearly excess cancer
incidence to lifetime risk is to use life-table methods (HAS 1988, USEPA
V1I1-24
-------
1989b). It should also be noted that this procedure is different fran the
approach used In Section v:::.A. to extrapolate nencar.cer effects seen in dial
painters to those expected in the general population froa radium exposures of
varying durations.
Mays et al. (1985a) calculated the total risks from Ra-226 and Ra-228 to
be nearly equal, because the induction of head carcinomas by Ra-226 nearly
balanced the 2.5-fold greater induction of bone sarcomas by Ra-228. Their
calculated risks are 2.1E-5 for Ra-226 and 2.2E-5 for Ra-228 for an intake to
blood of 1 pci/day. To provide perspective on the uncertainty associated with
epidemiologic extrapolation, Table VIIX-2 lists calculations based on both the
estimate provided by Mays et al. (1985a) and the quadratic response for bone
sarcomas. These predictions are presented as the levels of radium in water
corresponding to lifetime risks of 10"* to 10'*, assuming 2 L/day water
consumption and 20% gastrointestinal absorption.
3ADRISK Model
The USEPA Office of Radiation Programs (ORP) has developed a set of
rr.odels to provide consistent, detailed evaluation of the human cancer risks
posed by radioactive elements (Dunning et al. 1980, Sullivan et al. 1981,
USEPA 1989b). An outline of the methodology is provided in Sections rit.D.
(biokinetic model) and VII.B (dosimetric model). Important features are the
use of a metabolic compartmental model to calculate absorption, distribution,
metabolism and excretion of radium and its decay product, the separate
calculation of absorbed dose rate for low-LET and high-LET radiation, the use
vn:-25
-------
Table v:::-2 Risk Z3ZL~i'.ea is: Ra-226 a.-.d Ra-228
Based on Linear or Dose-Squared Extrapolation of Bone Sarcoma*
and Linear Extrapolation of Ha-226 Induced Head Carcinomas
Drir.teino water activity leval faCl/Li
<*»
Excess tez22£ B4=22S
Cancer Risk U,fi«ftg"! Linear-8' fttfltigKiC "
10'4
10'*
10®
12 '
1.2
0.12
21
2.1
0.2:
li
i.i
0.11
250
79
25
Assuming 2 L/day water car.aun-.ption and 20% gastrointestinal absorption,
so that lifetime intake to the blood is 0.256 pCi per pCi/L in water.
(8' Average lifetime risks calculated by Mays ee al. (1985a) using a linear
risk coefficient for both bone sarcomas and head carcinomas.
(e| Average lifetime risks calculated with a linear risk coefficient for head
carcinomas and assuming a quadratic response for bone sarcomas. Predicted
risk is virtually all due to head carcinomas.
-------
of the orgar.-«p«ci*ic risk coefficient! derived by the BLIR 121 coanitcee (HAS
1980b) based on a synthesis of epidemiologic information on radiogenic
cancers, and the use of life-table methods to convert age* and aex-epecific
yearly risks to total lifetime risk.
The dose rates calculated by the RADRISK model for variouo organs are
listed in Table VIII-3 for Ra-226 and Ra-228. These values are the annual
dcse rates in year 70 resulting from a lifetime constant intake of 1 pCi/year.
The calculated dose rate in a given year for continuous intake is numerically
equal to the integrated lifetime dose for the cumulative intake until that
year (ICRP 1971). The RADRISK model actually calculates the annual organ-
specific dose rates at various times during the life span, not just the dose
in year 70, for purposes of assessing lifetime risk (Sullivan et al. 1981,
Dunning et al. 1980, USEPA 1989b).
The RADRISK model u».ss age- and sex-specific risk coefficients for
various organs, based on quantitative evaluation of data on human cancer risk
following exposure to several types of ionizing radiations (USEPA 1989b).
Organ-specific risk coefficients were derived from values darived by the BEIR
III committee of the National Academy of Sciences (NAS 1980b). The specific
values used were mortality risk predicted by the linear-linear model with a
10-year minimum latency and a 30-year follow-up. Life-table methods based on
U.S. mortality using 1969-1971 and a birth male:female ratio of 1.0511,
constrained by total risk at all sites (USEPA 1989b). This procedure was used
to derive relative risk coefficients for fatal cancers except for leukemia and
bone cancer, for which absolute risks with a 2-year minimum induction period
and a 25-year expression period were calculated from the BEIR ZZZ absolute
risk coefficients (NAS 1980b). The leukemia and bone
VIII-27
-------
Table VIII-3 Annual Organ Doses in the 70th Year
Predicted fay the RADRXSK Model
from Ingestion of 1 pCi/year of Radium Isotopes
R.-226
Ra-228
Oroan
LET
frad/vr\ /foci/vn
f rad/vr)/(oCi/vrl
Pulmonary lung
Low
1.94E-9
4.13E-9
High
1.6SE-8
1.64E-8
Stomach wall
Low
1.31E-9
3.55E-9
High
1.67E-8
1.63E-8
Intestine
Low
2.06E-8
4.33E-8
High
2.37E-8
1.63E-8
Kidneys
Low
2.00E-9
4.80E-9
High
1.65E-8
1.66E-8
Liver
Low
1.50E-9
4.42E-9
High
1.61E-8
1.95E-8
Breast
Low
2.27E-9
4.74E-9
High
1.66E-8
1.64E-8
Pancreas
Lew
2.13E-9
4.42E-9
High
1.66E-8
1.64E-8
Red marrow
Low
6.00E-6
5.03E-8
High
1.12E-7
4.64E-6
Bone surface
Low
8.62E-8
1.72E-7
High
I.36E-6
4.72E-7
Thyroid
Low
1.49E-9
4.10E-9
High
1.66E-8
1.64E-8
All other
Low
1.94E-9
4.42E-9
tissue
High
1.66E-8
1.64E-8
VIII-28
-------
cancer risk coefficients were combined using a relative biological
effectiveness of 8 for alpha particles. Mortality to incidence ratios were
derived from values provided by SE'R III (NAS 19S0b, USEPA 1989b).
Risks for high-LET radiation are derived from risks for low-LET radiation
by multiplying by a relative biological effectiveness of 8 (EJSEPA 1989b).
This is the appropriate value for comparing high-LET risks to the low-LET
risks derived by VSEPA (1989b). ICRp recommends a value of 20 when
multiplying by ICRP-derived low-LET risks (ICRP 1977). The ICRP-derived low-
LET risks incorporate a 2.5-fold reduction for low doses and dose rates (ICRP
1977). However, the USEPA (1989b) low-LET risk numbers do not incorporate
this 2.5-fold reduction, because human data on breast and thyroid cancer do
not support such a reduction and because there is an essentially equal
reduction in predicted low-LET risk due to the use of a linear rather than a
linear-quadratic risk projection model (USEPA 1989b). Human data on alpha
emitters also support a linear risk projection model (except for bone sarcomas
among radium-exposed watch dial painters). The appropriate relative
biological effectiveness to use with VSEPA (1989b) risk coefficients for
low-LET radiation is 20/2.5 = 8.
Averaged risk coefficients used in the RADRISK model are listed in
Table VIII-4 for high-LET and low-LET radiation in units of risk per rad for
fatal and total cancers. The RADRtSK model uses age- and sex-specific organ
dose rates and risk coefficients together with life-table methods to calculate
lifetime risk by organ resulting from high-LET and low-LET radiation^ The
risks for radium isotopes predicted by the RADRISK model for lifetime
VIII-29
-------
Table VIII-4 Organ-Specific Lifetime Cancer Risks
Used in the RADRISK Model
from High-LET and Low-LET Irradiation
Rlsk/rad
oyqan I*II_ Fatality Incidence
Pulmonary lung
Low
7.0E-5
7.5E-5
High
5.7B-4
6.0E-4
Stomach wall
Low
4.6E-5
6.0E-5
High
3.7S-4
4.8E-4
Intestine
Low
2.3E-5
4.3E-5
High
1.8E-4
3.4E-4
Kidneys
Low
1.8E-5
4.3E-5
High
1.4E-4
3.4E-4
Liver
Low
5.0E-5
5.0E-5
High
4.0E-4
4.0E-4
Breast
Low
5.5E-5
1.4E-4
High
4.4E-4
1.1E-3
Pancreas
Low
3.5E-5
3.8E-5
High
2.8E-4
3.1E-4
Red marrow
Low
4.SE-5
4.5E-5
High
3.6E-4
3.6E-4
Bone surface
Low
2.5E-6
2.5E-6
High
2.0E-5
2.0E-5
Thyroid
Low
6.4E-6
6.4E-5
High
5.1E-5
5.1E-4
Esophagus
Low
9.1E-6
9.1E-6
High
7.3E-5
7.3E-5
Lymphoma
Low
1.4E-5
1.9E-5
High
1.1E-4
1.5E-4
All other tissue
Low
1.9E-5
3.4E-5
High
1.6E-5
2.7E-4
Adapted from USEPA (1989b) Tables 6-6 and 6-7.
VIII-30
-------
ingestion of 1 pCi/L in water (equal to ingestion of 51,000 pCi in 70 years)
are S.1E-6 for Ra-226 and 5.1E-6 for Ra-228 for total cancers, and 4.8E-6 for
Ra-226 and 3.6E-6 for Ra-228 for fatal cancers (CSEPA 1969b).
For the evaluation of the risks of radium in drinking water supplies, the
Agency has reassessed the RADRISK predictions. As pointed out by the
Radiation Advisory Committee of the OSEPA Science Advisory Board (SAB 1990),
the rate of leukemias predicted by the RAOR1SK model is not consistent with
epidemiologic data showing no substantial increase in leukemia incidence among
dial painters exposed to Ra-226 and/or Ra-228 or among patients injected with
Ra-224 (see Section vi.B.l). The Agency has recognized the need to adjust the
high-LET leukemia risk for internally-deposited radionuclides from the value
of 3.6E-4 per rad for both incidence and fatality (see Table VIII-4).
The BEIR IV committee estimated a high-LET leukemia risk coefficient as
5.0E-5 to 6.0E-5 per rad based on thorotrast data (NAS 1988). Mays et al.
(1985a) estimated the high-LET leukemia risk to be 4.02-5 per rad, also based
on thorotrast data. For this document a leukemia risk estimate for high-LET
of 5.0E-5 per rad will be used. Other methods of estimating the high-LET
leukemia risk coefficient are discussed in Section IX.B.2.
The high-LET leukemia risks predicted by the RADRISX model for lifetime
ingestion of 1 pCi/L in water are 1.5E-6 for Ra-226 and 7.6E-7 for Ra-228.
Multiplying these risks by the ratio of the new and original high-LET leukemia
risk coefficients (5.0E-5/3.6E-4) yields adjusted high-LET leukemia risks of
2.1E-7 for Ra-226 and 1.1E-7 for Ra-228. The adjusted risks for lifetime
ingestion of 1 pCi/L in water are 4.9E-6 for Ra-226 and 4.5E-6 for Ra-228 for
total cancers, and 3.5E-6 foe Ra-226 and 2.9E-6 for Ra-228 for fatal cancers.
VIII-31
-------
Another comment made by the Radiation Advisory Committee of the USEPA
Science Advisory Board (SAB 1990) was that the RADRISK model, like the ICRP
model from which it is derived, does not include a separate compartment for
paranasal sinuses and mastoid air cells, and therefore does not predict head
carcinomas resulting from trapped Rn-222 gas produced by decay of Ra-226. No
detailed dosimetry has been developed for these tissues (MAS 1988).
Therefore, the risk of head carcinomas is estimated based on the ratio of head
carcinomas to bone sarcomas among the dial painters. Xn the total cohort,
which was exposed to both Ra-226 and Ra-228, there are 85 suspected or
confirmed bone sarcomas and 37 suspected or confirmed head carcinomas (Hays
1988). Thus, head carcinoma incidence is a minimum of 40% of bone sarcoma
incidence among people exposed to Ra-226. The linearized low-dose
extrapolations of Mays et al. (1985a) per yCi intake to the blood for Ra-226
alone are 1.6E-5 for head carcinomas and 1.0E-5 for bone sarcomas. This
dosimetry gives a ratio of head carcinomas to bone sarcomas of 160%. As a
representative value, the risk of head carcinomas is assumed to be equal to
the risk of bone sarcomas among populations exposed to Ra-226, which is 9.4E-7
for both total and fatal cancers for lifetime exposure to 1 pCi/L of Ra-226 in
drinking water. Adding in this risk results in a total risk for lifetime
ingestion of 1 pCi/L of Ra-226 in water of 5.8E-6 for total cancers and
4.4E-6 for fatal cancers.
A final comment made by the Radiation Advisory Committee of the USEPA
Science Advisory Boa^'d (SAB 1990) was that the risks predicted by RADRZSK
model were lower for Ra-228 than for 3a-226, in contrast to empirical and
dosimetric estimates that Ra-228 would be 1.5 to 2.8 times more potent than
Ra-226 in inducing bone sarcomas (Rowland et al. 1978, NAS 1988). Upon
investigation, it was determined that the dose from Ra-228 to the bone marrow
V1II-32
-------
and bone surface calculated by the RfCRISK model were lean than half of the
doses calculated with the ICRP 30 model (ICRP 1979) or by Adams et al. (1978).
the doses for otiier tissues did agree in the cwo modal3, as did the doses for
bone marrow and bone surfaces for Ra-226. Since the RADRISK model La baaed on
similar data and assumptions as the ICRP 30 modeL and the model of Adans
et al. (1973), all calculated doaes should agree. the source of the
discrepancy for bone narrow and bone surface doses for Ra-228 has not yet been
located. To adjust for this discrepancy, the bone marrow and bone surface
doses for Ra-228 from the ICRP 30 modeL were substituted into the RADRISK
model. The rCRP 30 doses were higher by 2.46 for bone marrow and by 2.23 for
bone surfaces. These factors were applied to the RADRISK predictions for
Ra-228 for leukemias and bone sarcomas. The low-LET risk, foe leukemia
increases from 1.0E-B to 2.5S-7 and the adjusted high-LET risk increases front
l.iE-7 to 2.75-7, for a total leukemia risk of 5.2E-7. The bone sarcoma risk
increases from 4.4E-7 to 9.9E-7. The total adjusted risk frost lifetime
ingestion of 1 pCL/L of Ra-228 is 5.3E-6 for total cancers and 3.3EHS for
fatal cancers.
lbs adjustments to tbe RACF-ISP. pcedizticas far Ra-22£ and Ra-220 are
summarised In Table Vlll-5^ showing the corrections for high-LET leukemia risk
for Ra-226 and Ra-22B> for head carcinomas for Ra-226 and for bone marrow and
bone surface dcees for Ra-223. The net result of the adjustments is a change
of 4-S%. The risks for radium isotopes predicted by the adjusted RMRISK
model for lifetime ingestion of 1 pCi/z, in water are 5.8S-6 for Ra-226 and
5.3E-6 for Ra-228 for total cancers, and 4.4E-6 for Ra-226 and 3.8E-6 for
Ra-223 for. fatal cancers.
v:::-33
-------
Table VIH-5 Lifetime Cancer Risks from
Lifetime Ingestion of 1 pCi/L Ra-226 or Ra-228 in DrinkLng Water
Tyoe
Ra-
•226
Ra-22B
Fatal
Total
P^al
Total
1. Predicted by the RADRISK
Model
Bone sarcoma
9.4E-7
9.4E-7
4.4E-7
4.4E-7
Leukemia, high-LET
1.5E-6
1.5E-6
7.6E-7
7.SE-7
Leukemia, low-LET
9.6E-8
9.6E-8
1.0E-7
1.0E-7
Ml other
3t9E-S
3'9E"S
Total
4.8E-6
6.3E-6
3.6E-6
5.1E-6
2. Using corrected high-LET
leukemia risk"1
Bone sarcoma
9.4E-7
9.4E-7
4.4E-7
4.4E-7
Leukemia, high-LET
2.IE-7
2.12-7
1.1E-7
2.1E-7
Leukemia, low-LET
9.6E-8
9.6E-8
1.0E-7
1.0E-7
All other
2r?S-S
i-SW
2.3B-§
3.flE-6
Total
3.5E-6
5.1E-6
2.9E-6
4.4E-6
3. Adjusting for head carcinomas from
Ra-226""
Bone sarcoma
9.4E-7
9.4E-7
4.4E-7
4.4E-7
Kead carcinoma
9.4E-7
9.4E-7
0
0
Leukemia, high-LET
2.1E-7
2.1E-7
1.1E-7
1.1E-7
Leukemia, low-LET
9.6E-8
9.6E-8
1.0E-7
1.0E-7
All other
2.3E-6
3.8E-6
Total
4.4E-6
6.OE-6
2-9E-6
4.4E-6
4. Adjusting for corrected bone marrow
and bone
surface doses for
Ra-228(e
Bone sarcoma
9.4E-7
9.4E-7
9.7B-7
9.7E-7
Head carcinoma
9.4E-7
9.4E-7
0
0
Leukemia, high-LET
2.1E-7
2.1E-7
2.6E-7
2.6E-7
Leukemia, lcw-LET
9.6E-8
9.6E-8
2.6E-7
2.6E-7
All other
2.3E-6
3.3E-6
2 -3E-6
3r«;-e
Total
4.4E-6
6.OE-6
3.BE-6
5.3E-6
Based on a high-LET leukemia risk coefficient of 5.0E-5 (Kays et al.
1985a, NAS 1988) rather than 3.6E-4 (see Table VIII-4), predicted riska
are reduced by a factor of (S.0E-5/3.6E-4).
(S| Head carcinoma risk is assumed to be equal to bone sarcoma risk for Ra-226
due to acsumulation of Rn-222 gaa; Rn-220 is assumed to be too short-lived
to accumulate.
ls| Based on higher red marrow and bene surface doses for Ra-226 predicted by
the ICRP 30 model (3CRP 1979), leukemia risks are multiplied by 2.46 and
bene sarcoma risks are multiplied by 2.23.
VIII-34
-------
Levels of radium in drinking water corresponding to lifetime riaka of
10'* to 10"s are given in Table Vlii-6, baaed on the adjusted RADRISK model.
The risk estimates in Tables VIH-2 and VIII-6 are quite close, except for the
quadratic extrapolation of bone Bareoma risk of Ra-228. Mays et al. (198Sa)
have shown that their linearized risk estimate for Ra-226 expressed per rad o£
endosteal dose is very similar to the BEIR III estimate used in the RADRISK
model. Mays et al. (1985a) assumed a wet bone weight of 2.6 kg for the
Reference Woman, an equilibrium skeletal content of radium equal to 25 times
the daily ingestion, and ratios of endosteal and skeletal dose per rad of 7.S
for Ra-224 and 0.63 for Ra-226; their calculated risk coefficient was 26 bone
sarcomas per 10* person-rad of endosteal dose, while the BEIR III estimate was
27 bone sarcomas per 10* person-rad of endosteal dose (Kays et al. 1985a). As
pointed out by the BEIR IV committee, the induction of a fibrotic layer on the
bone surface by exposure to Ra-226 and Ra-228 (Lloyd and Henning 1983) would
influence the dosimetry and cancer rl3k of these isotopes, particularly at
high intake levels (NAS 1988). Although the unit risk of bone sarcoma in the
RADRISK model is nearly identical to that in Mays et al. (1985a), the
calculated lifetime risk is not the same. The primary cause of the difference
is that Mays et al. (1985a) simply used an equilibrium factor of 25 for radium
in bone, while the calculation of concentrations and doses in the RADRISK
model used a metabolic model for radium and its decay products based on a
synthesis of human and animal toxicokinetic data (USEPA 1989b).
It is also important to note that there are differences among the
specific types of cancer giving rise to the overall risks predicted i.i
Tables VIII-2 and VtII-6. The risks considered by Mays et al. (1985a) include
only bone sarcoma and head carcinoma, but the RADRISK model predicts cancers
At several other sites as a consequence of radium exposure, based on data on
VIII-35
-------
Table VTtl-6 Drinking Water Concentrations of Ra-226 and Ra-229
Corresponding to Various Risk Levels Predicted by the
Adjusted RADRISX Model
Excess Cancer Drinking water activity level foCl/M"'
KM Ra-229
Total cancers
10-4 17 19
10'1 1.7 1.9
10'e 0.17 0.19
Fatal cancers
10'4 22 26
10® 2.2 2.6
10* 0.22 0.26
'"Assuming 2 L/day water consumption and 20% gastrointestinal absorption.
VIII-36
-------
radiogenic cancers in humans (USEPA 1989b). As discussed above, the high-LET
leukemia risk coefficient has been adjusted downwards from the original
RADRISK predictions, based on epidemiologic data on thorotrast patients. For
soft-tissue cancers, no adjustment has been made, because the epidemiologic
data on soft-tissue cancers among radium dial painters are not sufficiently
sensitive to derive quantitative estimates of soft-tissue cancer risks
(Stebbings et al. 1984).
Choice of Model for Aoencv Risk Assessment
Two methods for estimating the risk at low doses, linearized
extrapolation of dial painter data and the adjusted RADRISK model, give
generally similar results (Tables VIXI-2 and VIII-6). The Agency considers
the RADRISK model to provide the best estimate of cancer risk of low-level
intake of radium, primarily because of the uncertainty in using the dial
painter data to derive a linear risk coefficient fcr bone cancer induction,
and the existence of the alternative of using a linear risk coefficient for
Ra-224 derived from epidemiologic data that has been recommended by two BEIR
committees (NAS 1980b, 1988). The dial painter and other epidemiologic data
are used to adjust the predictions of head carcinomas and leukemias in the
present version of the RADRISK model. The Agency will use the adjusted
RADRISK predictions of fatal cancers as the reference in proposed regulations
limiting radium activity in drinking water. Uncertainties in the estimates of
cancer risks are discussed in Chapter IX.
VIII-37
-------
C. Genetic and Developmental Effects
Exposure to ionizing radiations can cause mutations in germ cells
(oocytes and spermatogonia) leading to hereditary defects, and can damage
embryos and fetuses exposed jjj utero leading to pre-implantation loss, mental
retardation or malformations {USEPA 1989b). These effects are primarily
established in experimental animals exposed to gamma irradiation, but a risk
of mental retardation and microcephaly has been established among Japanese
A-bomb survivors exposed ifl utero (USEPA 1989b). No direct evidence exists
that exposure to radium causes hereditary genetic defects or teratogenesis
(see Sections V.C and Vl.B.l). However, estimates of genetic risks can be
made by extrapolation of observed effects of low-LET irradiation. The Agency
has reviewed several estimates of the risk factors for low-LET and high-LET
irradiation and has developed the risk factors shown in Table VIII-7 (USEPA
1989b). These risk factors are based primarily on evaluations by the National
Academy of Sciences (NAS 1980b, 1988). Risk factors were not developed for
teratogenic or somatic risk for high-LET irradiation because of the lack of a
plausible dosimetry for the developing fetus.
Since no data are available to estimate the teratogenic or somatic risk
from high-LET irradiation, the assessment of risk from radium exposure is
limited to genetic defects. The 30-year cumulative doses to ovaries, testes
and average gonadal tissue from intake of radium are listed in Table VHI-8 in
units of rad/(pCi/year). A drinking water concentration of 1 pCi/L results in
ingestion of 730 pCi/year. The calculated germ cell mutation risks are shown
in Table VIII-9. These risks are about 5% of the predicted cancer risk for
Ra-226 and about 3% for Ra-228. Uncertainty in the estimates of genetic and
developmental effects are discussed in Chapter IX.
vm-38
-------
Table viii-7 Radiation Risk factors for Genetic
and Developmental Effects
Sndoolnt
Low LET
Teratogenic
Severe mental
retardation
Somatic
Fatal cancers
Genetic
Severe hereditary
defects, all
generations
High LET
Genetic
Severe hereditary
defects, all
generations
Significant
Exposure Period
Weeks 8 to 15
of gestation
Weeks 0 to 40
of gestation
30-year
reproductive
generation
30-year
reproductive
generation
Risk oar rad
Beat Estimate
4.0E-3
6.0E-4
2.SE-4
Ranoe
6.9E-4
2.5E-3 - 5.5E-3
l.BE-4 - 1.8E-3
S.0E-5 - 1.1E-3
1.6E-4 - 2.9E-3
Adapted from CTSEPA (1999b) Table 6-27.
The range assumes a linear, non-threshold dose response. However, it is
plausible that a threshold may exist for teratogenic effects.
VIII-39
-------
Table viii-9 30-^ear Cumulative Doses to Germ Cells
Predicted by the RADRISK Model
from Ingestion of 1 pCi/year of Radium isotopes
IfET ,
Ra-226
frad)/fDCi/vr»
Ra-228
/radWfoCi/vr)
Ovaries
Low
6.14E-8
1.83E-7
High
4.49E-7
4.44E-7
Testes
Low
2.72E-8
1.10E-7
High
4.49E-7
4.44E-7
Average
Low
4.43E-8
1.46E-7
High
4.49E-7
4.44E-7
VIII-40
-------
Table VXII-9 Cerai Cell Mutation Risks from
Lifetime Ingestion of I pCi/L Ra-226 or Ra-228 in Drinking Hater
Quantity
Dose in rad/(pCi/L)">
Risk per pCi/L(#l
LET
Ra-228
Low
3.2E-5
1.1E-4
High
3.3E-4
3.2E-4
Low
8.4E-9
2.8E-8
High
2.3E-7
Total
2.3E-7
2.5E-7
Calculated as average risk in Table vill-8 multiplied by
710 (pCi/yr)/(pCi/L).
(6) Calculated from the previous entries multiplied by the coefficients in
Table VIII-7.
VIII-41
-------
D. Summary
summarizes the HA and DWEL values calculated on the basis
end points, and the estimated excess cancer risk calculated
RADRISK model.
Table VIII-10
of noncarcinogenic
using the adjusted
VIII-42
-------
Table VIII-10 Summary of Quantification of
Toxicological Effects for Radium
VAiaa
One-Day HA for 10-kg child
Ten-Day HA for 10-kg child
Longer-Term HA for 10-kg child
Longer-Term HA for 70-kg adult
DWEL (70-kg adult)
Excess cancer risk
10'4
10'5
10'#
Drinking water
concentration (cCL/L\
Ra-226
Ra-228
Reference
700,000
700,000
Ke&ne
et
al.
(1983)
70,000
70,000
Ke&ne
et
al.
(1983)
300
300
Keane
et
al.
(1933)
lr000
1.0QQ
Ksane
et
al.
(X9B3|
100
100
Keane
et
al.
(1983)
20
20
USEPA
(1989b)
2
2
USEPA
(1989b)
0.2
0.2
USEPA
(1989b)
VIII-43
-------
IX. UNCERTAINTY ANALYSIS
The health effects following exposure to high levels of radium and
ionizing radiations in general are very well established from both human and
animal data. Host of the uncertainty arises in extrapolating effects observed
at high levels to risks expected at low (environmental) levels of exposure.
The remainder of this section briefly describes the assumptions and parameters
contributing uncertainty to the quantification of the toxicological effects of
radium in drinking water.
A. Ranoe of Assumptions and Models
1. Noncancer Effects
The key assumptions made in evaluating the threshold for bone necrosis
are identification of the NOAEL, quantification of radium intake and
adjustment for duration of exposure and age at exposure.
Identification of the NOAEL
The NOASL was identified as an intake to the blood of 10 /iCi, the level
at which bone lesions found in x-rays differed significantly between dial
painters and unexposed controls (Keane et al. 1983). This level of bone
lesion was classified as mild or minimal, and was considered to be of
questionable clinical relevance
-------
On the other hand, bone effects were only identified by X-ray of relatively
young women, and no information was available concerning possibly more
sensitive adverse bone effects such as post-menopausal osteoporosis, delayed
bone healing or altered calcium homeostasis. The uncertainty factor of 10 is
used to account for possibly more sensitive members of the population.
However, if more subtle effects occur at exposures more than 10 times less
than the NOAEL identified by examination of X-rays, this would lead to a
decrease in the calculated HA and DWEL values.
Quantification of Radium intake
The radium intake of the dial painters in the study by Keane et al.
(1983) was back-calculated from measured body burden of Ra-226 many years
after exposure. Known inter-individual variations in radium retention can
lead to differences of up to a factor of 7 in initial intake calculated from
measured body burden (Finkel et al. 1969a). Intake of Ra-228 was estimated
from information of varying accuracy on the isotope ratio of the paint used at
a particular facility (Sharpe 1974). Thus, although the true uncertainty in a
given estimate of intake is difficult to quantify, it could easily be a factor
of 5 for intakes primarily of Ra-226, and considerably higher for intakes
primarily of Ra-228. This uncertainty could result in either an increase or a
decrease in calculated HA and DWEL values.
Adjustment for Duration of Exposure
The adjustment for duration of exposure was done using the dosimetric
calculations performed by Keane et al. (1983). This calculation used the ICRP
model for metabolism of radium and assumed that the sensitive endosteal layer
IX-2
-------
was between 0 and 10 pm of mineral bone. An instantaneous intake was
calculated to deliver a dose that was 1*6 tines higher for Ra-226 and
1.2 times higher for Ra-226 than the sane intake spread over 50 years (Keane
et al. 1983). Uncertainties in this result derive from individual variations
in radium uptake to and elimination from bone, microdosimetry of radium in
bone and the location of the sensitive layer of cells. The overall
uncertainty is difficult to quantify, and could lead to either an increase or
decrease in calculated DWEL values.
Adjustment for Age at Exposure
Adjustment for age at exposure was done assuming that gastrointestinal
absorption was equal in children and adults and that the NOAEL identified
among young women would scale with body weight. For some elements, such as
lead and cadmium, absorption is known to be higher in children than adults.
There is some evidence that children may absorb more radium than adults
(Stehney and Lucas 1956, Muth and Globel 1963). The maximum uncertainty that
this could introduce is a factor of 5, if children were to absorb 100% of
ingested radium. The assumption that radium effects would scale with body
weight is plausible, but no information otv botxe necrosis in children exposed
to Ra-226 or Ra-228 is available to verify this assumption. Uncertainty
concerning adjustment for age at exposure could lead to a decrease in
calculated HA value9 by up to a factor of 5.
2. Cancer Risk
The key assumptions made in evaluating the cancer risk of radium
ingestion ars the carcinogenicity of radium ingestion, the shape of the dose-
IX-3
-------
response curve, the choice of the method of low-dose extrapolation and
assumptions used in model formulation.
r.f gginoqenlcitv of Radium Ingestion
there is very little uncertainty concerning the capability of ionizing
radiation in general and radium in particular to cause cancer in humans at
high doses, by various routes of exposure. This low level of uncertainty is
reflected in the classification of radium as a Group A carcinogen, a human
carcinogen (USEPA 1991b).
Shane of the Doae-Resoonae Curve
The Agency uses a linear dose-response relationship to assess cancer
risk. For radiogenic cancers, most data are consistent with a linear
relationship (USEPA 1989b). However, the best fitting models for bone
sarcomas among dial painters as a function of estimated radium intake
(AiCi Ra-226 + 2.S nCL Ra-228) are quadratic with an exponential decrease at
high intakes. Linear fit3 are rejected except in restricted subgroups which
are more suitable for epidemiological analysis (Rowland et al. 1978, 1983).
Head carcinomas appear to have a linear response Co Ra-226 intake, although a
quadratic response is not rejected (Rowland et al. 1978). For bone sarcomas
among patients injected with known amounts of Ra-224, various dose-response
functions fitted to the data all are linear or contain a linear term (NAS
1988). The Agency has chosen to use a linear extrapolation of risks from low-
level radium exposure, based on the recommendations of the BEIR committees and
of the Radiation Advisory Committee of the USEPA Science Advisory Board. If
the true bone sarcoma response were quadratic rather than linear, risks at low
IX-4
-------
intakes would be lower than predicted, and the drinking water levels
corresponding to various lifetime excess cancer risks would be higher than
those presented in Table vin-10. The BEIR IV committee pointed out that
uncertainty in the shape of the dose-response curve results in great
quantitative uncertainty of risk at low levels of radium exposure (HAS 1988).
A related issue in low-dose extrapolation is a postulated "practical
threshold" for bone sarcoma risk, arising from an increased median tumor
appearance time with decreased do9e (Evans et al. 1972). The "practical
threshold" would be the dose at which the median tumor appearance time would
exceed the human lifetime (Raabe et al. I960, 1983). As discussed in
Section VI11.B.2, the weight of evidence does not support the postulated
practical threshold of 80 rads in humans, although the epidemiologic data are
not sensitive enough to rule out a lower threshold (Mays 1988, NAS 1988). If
a practical threshold does exist, and if this threshold is above environmental
exposure levels, then the risks from Ra-226 and Ra-228 exposure would be less
than predicted, and possibly zero.
Method of Extrapolation
The Agency has chosen to use the RADRISK model to extrapoLate cancer
risks from radium ingestion, adjusted for over-prediction of leukemias and
lack of prediction of head carcinomas. An important reason for this choice is
that it allows the incorporation of a linear risk coefficient for bone sarcoma
induction that is recommended by the BEZR committees of the National Academy
of Sciences. An alternative risk analysis was performed by Kays et al.
(19B5a), with a linear risk coefficient derived from the epidemiologic data on
dial painters that is quite similar, in terma of risk per endosteal dose, to
IX-5
-------
the coefficient recommended by the 8EIR in committee. Not surprisingly, the
resulting calculated risks agree to within a factor of 2 {see Tables V2II-2
and VIII-6). The main reasons that similar risk, coefficients result in
different calculated doses are the different ways of calculating radium
concentration in bone and of converting yearly risk to lifetime risk.
The RADRISX model was developed by USEPA specifically for the purpose of
assessing risk of lifetime exposure to radionuclides by members of the general
population. Uncertainty in using a mathematical model can be divided into
model formulation, discussed below, and parameter variability, discussed in
the next section.
Dosimetric Model Formulation
The RADRISK model calculates the dose to target organs from ingestion of
radium using ICR? models for alkaline-earth metabolism, with certain
modifications (see Section III.D), together with calculations of absorbed
dose, also based on ICR? models with minor modifications (see Section vil.B).
USEPA (1989b) contains an extensive discussion of the uncertainty associated
with the ICRP models (pp. 5-20 to 5-23), including the use of the "Reference
Han" concept, the gastrointestinal tract model (absorption only from the small
intestine, no endogenous secretion, exponential removal from each sequential
compartment), and the ICRP 30 biokinetic models (distribution within
compartments, growth of radioactive decay products) and the ICRP radioactive
dose models (identification of source and target regions, spatial
relationships between source and target regions). Specific uncertainties
associated with radium isotopes include calculation of total radium bone
content, distribution of radium bone content and possible shielding of
IX-S
-------
intestinal tract cells by the mucus lining (OSEPA 1989b). In addition, the
ICRP model does not evaluate dose to epithelial cells of the cranial sinuses
or risk of head carcinoma separately from the overall skeletal risk, and so
these risks were calculated separately for Ra-226. The magnitude and
direction of these additional uncertainties were not evaluated in USEPA
(1989b). Overall, considering both model formulation and parameter
variability, it was estimated that the uncertainty in dose was a factor of 3
or less in major organs and a factor of 5 or more in other tissues for Ra-226
and Ra-228 (USEPA 1989b).
Compared to external x- or gamma radiation, which gives a fairly uniform
dose, internal emitters, particularly alpha emitters, give an irregular,
rather uncertain dose distribution. The ICRP and RAORISK pharmacokinetic
models assume uniform distribution of radionuclides in organ or tissues. This
approximation is not necessarily true. While many studies have shown uneven
distribution of radionuclides in tissues, there is not sufficient data on any
radionuclide to establish a detailed pharmacokinetic model using non-uniform
tissue or organ nuclide distribution. Inadequacies in the pharmacokinetic
model will be reflected in the calculated dose and risk.
Bone exemplifies the dosimetry problems associated with alpha emitting
isotopes. The short range of alpha particles in tissue and the high rate of
deposition of energy make the distance from source to target cells a very
important consideration. In bone, target cells for osteosarcoma induction are
considered to be, and the dose is estimated for, cells up to 10^ from bone
surfaces* The selection of the thickness of the layer in the calculation of
bone surface dose is somewhat uncertain. Harley and Pasternack (1976)
calculated alpha dose rate for radium-226 and its decay products in uniformly
IX-7
-------
labelled bona in the tissues outside the mineral bone. Noting the report of
Lloyd (1970} of a 5^m thick layer of osteoid between mineralized bone surfaces
and cellular layers, Harley and Pasternack (1976) adopted a distance of lO^un
from mineral bone as appropriate for calculating endosteal dose. On the other
hand, Scott (1967) Indicated cell nuclei at a distance of 10-lSpm from
mineralized bone surface in rats, and James and Taylor (1971) calculated dose
out to a distance of 20^m from bone surface. The International Commission on
Radiological Protection (ICRP) recommended calculating endosteal doae rate
over a distance 5-lOym from the trabecular surface (ICRP 1968). Kays and
Tueller (1964) calculated doses for Ra-226 out to 70^im from bone surface and
stated that the average dose rate for soft tissues from 0-10 M® "as used
because: (a) osteosarcoma induction appeared to occur in cells adjoining bone
surfaces and (b) dose-rates were maximum at bone surfaces. They also reported
the suggestion of Owen (1963) that cells located more than one-cell layer away
from bone surfaces were at risk. Vaugan (1972) had 3hown a 5 to 20pa thick
cell layer in rabbits and Lloyd and Henning (1983) demonstrated a single cell
layer 0.6 to 28.1 pim thick in human bone. Thus, the assumption that the
endosteal dose should be calculated for a 10
-------
complicates direct application of a risk coefficients derived from X-ray or
gamma exposure with a dose calculated using a pharmacokinetic model and
increases the uncertainty in the assessment of risk from ingested radium.
The risk estimates based on coefficients derived from epidemiological
studies o£ persons exposed to whole body or partial body external radiation
(X- or gamma rays) do not always closely correspond to risk estimates derived
from epidemiological studies of ingested or injected radium. The problem may
relate either to the pharmacokinetic-dosimetric calculations or to special
problems in the epidemiology of persons ingesting or injected with radium.
Efforts to determine the reason(s) for discrepancies should address the
epidemiological, radiobiological and other parameters of the studies and
continue until there is resolution of the questions.
Epidemiologic studies of ingested or injected radium typically involve
activity concentrations far in excess of expected environmental exposures.
The pharmacokinetic model estimated for radium, based primarily on studies of
persons Injected with radium (ICRP 1979), differs appreciably -rom a model
based on environmental exposure to radium (Stehney and Lucas 1956). In
addition, peritrabecular fibrosis has been noted in many species, including
humans, following elevated doses of radiation to bone (Ackerman and Spjut
1962, Clarke 1962, Jee et al. 1969, Sharpe 1976, Wilson et al. 1976, Lloyd and
Henning 1983). This peritrabecular fibrosis may occur within about 150 days
post exposure, following an abortive attempt by osteoblasts to lay down bone
(Jee et al. 1969). At the same time, the microvasculature of bone and bone
marrow are disrupted. Jee (1971) reported a 50% destruction of marrow
vasculature within days of injection of 10 ^tCi/kg of Ra-226. The extensive
pathology reported by Jee (1971) following radiation exposure of bone implies
IX-9
-------
chat the metabolism and physiological responses in bone following high doses
of radium are abnormal. Thus, it is questionable whether bone with abnormal
physiology and metabolism is suitable for estimating pharmacokinetics, and
whether epidemiology of persons with such internal pathology can be compared
to environmental situations. The circumstances influencing pharmacokinetics,
physiology and pathology must be carefully evaluated for epidemiological
studies of internally deposited radionuclides. In the case of radium/ CJSEPA
judged that risk from environmental exposure levels were better estimated with
the RADRISK model using a bone sarcoma risk coefficient derived from Ra-224
data rather than by applying data from epidemiological studies of persons
exposed to extremely high activity concentrations of radionuclides. In fact,
the risks estimated by the two approaches differ by less than a factor of 3
(see Tables VIII-2 and VIII-6), giving an indication of the magnitude of the
uncertainty introduced by the choice of method of low-dose extrapolation.
Oroan Rl3k Model Formulation
The assumptions used in calculating organ risk per unit dose in the
RADRISK model were developed by the Agency with extensive consideration of
work by ICRP and NAS and of advice from the Radiation Advisory Committee of
the USEPA Science Advisory Board. Uncertainties in the model formulation are
addressed in USEPA (1939b), Chapter 6. These uncertainties include the shape
of the dose-response curve, the effect of dose rate for low-LET and high-LET
irradiations, the risk projection models (use of relative risk for all cancers
except use of absolute risk for leukemia and bone cancers), the use of
lifetable methods to assess lifetime risk, the use of a single value for
relative biological effectiveness of high-LET irradiations for all organs and
the implications of the revised dosimetry for low-LET irradiations from the
IX-10
-------
Japanese A-bomb survivor data. The dosimetry for external X-rays and gamma
rays is well established and has minimal uncertainty due to the nature of the
radiation. The magnitude of other uncertainties was not addressed, but it was
concluded that the overall projections of the RADRISK model were not biased
low in a serious way for radium (USEPA 1989b).
Use of the RADRISK model to estimate risk from ingested radium assumes
that such ingestion may lead to cancers in any radiosensitive organ. Dial
painters exposed to Ra-226 and/or Ra-228 by ingestion have not been shown to
have excess cancers except bone sarcomas and head carcinomas; however,
existing studies do not have sufficient sensitivity to exclude increases in
late-appearing, common cancers such as lung or breast (Stebbings et al. 1984).
Increases in liver and breast cancer incidence has been observed in patients
injected with Ra-224 (Spiess et al. 1989). Thus, although there is no direct
evidence that ingestion of radium in drinking water may cause cancers in all
radiosensitive organ, this assumption is supported by the general
understanding of the carcinogenicity of ionizing radiation. The maximum
uncertainty that would be introduced if this assumption is incorrect Would be
the ratio of bone sarcomas plus head carcinomas to total radium cancer risk
predicted by the adjusted RADRISK model, which is between a factor of 2 and 5
(see Table VIII-5).
3. Genetic and Developmental Effects -
No direct evidence is available concerning genetic or developmental
effects of radium exposure in humans. Assumptions used to evaluate these
risks are: that developmental effects cannot be calculated due to the lack of
a plausible dosimetry for the developing fetus for internally-deposited
IX-11
-------
radionuclides, that genetic effects of high-LET radiation can be calculated
from genetic effects of low-I,ET radiation, that genetic effects of low-LET
radiation are a stochastic (non-threshold) response to irradiation of gonads
and that the radiation dose to ovaries and testes from radium ingestion is
appropriately calculated in the RADRISK model.
Lack of Quantification of Developmental Effects
Risk coefficients have been derived for developmental effects of external
low-LET exposure (USEPA 1989b). The most substantial developmental risk is
mental retardation from exposure during gestation weeks 8 to 15 (USEPA 1989b).
For continuous exposure, assuming a uniform dose rate to all tissues, the risk
per offspring corresponds to about 10% of the risk of genetic defects due to
30-year exposure of a parent. If the dose to the developing fetus from
internally deposited radium were similar to the dose to germ cells, then
genetic plus developmental risks would be only about 10% higher than genetic
risks alone. Developmental effects may exhibit a threshold for induction
(USEPA 1989b), in which case there would be no increase in the calculated
risks.
Basing Genetic Effects of Hiah-LET Radiation on Effects of Low-LET Radiation
High-LET radiation has been shown to damage genetic material in somatic
cells of exposed humans (USEPA 1989b). This provides evidence supporting the
likelihood of genetic damage to germ cells from high-LET radiation, but does
not provide a basis for independent assessment of this risk. The assumption
that high-LET effects can be extrapolated from low-LET effects based on
Relative Biological Effectiveness (RBE) is commonly made in radiological risk
rx-12
-------
assessment, but the degree of uncertainty contributed by this assumption is
difficult to quantify.
Non-Threshold Nature of Genetic Damage
The radiation protection community is generally in agreement that the
probability of genetic change is linearly proportional to radiation dose and
that no threshold exists for initiation of heritable damage to germ cells
(USEPA 1989b). If such a threshold does in fact exist, and if this threshold
exceeds the doses from environmental levels of radon exposure, then no genetic
risks would be expected.
Estimation of Dose to Ovaries and Testes
The discussion in the previous section on the formulation of the
dosimetric component of the RADRISK model would also apply to the estimation
of dose to germ cells. The factor of 5 overall uncertainty for minor organ
systems discussed above would be an appropriate estimate for the dosimetry
uncertainty for Ra-226 and Ra-228.
B. Parameter Variability
1. Noncancer Effects
The key parameters used in evaluating the threshold for bone necrosis are
the extent of gastrointestinal absorption, the uncertainty factor used to
derive the RfD, and the values used for body weights and drinking water
consumption in deriving the HA and DWEL values.
IX-13
-------
Gastrointestinal Absorption
A gastrointestinal absorption value of 0.2 was used, based on the careful
work of Maletskos et al. (1966). Those authors compared this value to those
derived by others based on absorption of nominally more soluble forms of
radium in animals or humans. Maletskos et al. (1966) concluded that the
actual absorption of dial paint by the young female dial workers would be
within a factor of 2-3 of their measured value of 0.2. This conclusion is
likely to apply to the absorption of radium by other members of the general
population as well. Risks are linearly proportional to the assumed absorption
rate, 90 that if the true absorption were twice as high, the risks would be
twice as high, and HA and DWEL values would be half as high.
Uncertainty Factor
An uncertainty factor of 10 was used to derive the RfD from the NOAEL, to
account for sensitive members of the human population. This value may be
overly conservative, because the population from which the NOAEL was derived
was composed of females, who are generally more sensitive to effects on bone
than males. However, this increased sensitivity i9 most apparent after
menopause, and the population was apparently studied at a fairly young age.
If a lower uncertainty factor were used, the HA and DWEL values would be
higher.
Values Used for Body Weight and Drinking Water Consumption
Standard values were used for body weights and drinking water consumption
(a 70-kg adult drinking 2 L/day and a 10-kg child drinking 1 L/day). The body
IX-14
-------
weights are means for adults and children aged about 18 months {USEPA 1989a),
and the drinking water consumption rates represent reasonable, protective
values (CJSEPfl 1999a). The effect of using these values is that most members
of the population (children or adults) will receive a dose of radium that is
less than one-tenth of the RfD from water containing radium at the HA or DWEL
limit, but some fraction {those that consume the most water) will receive a
dose higher than one-tenth of the RfD.
2. Cancer Risk
The key parameters used in evaluating the cancer risk of radium ingestion
are the drinking water ingestion rate, the gastrointestinal absorption, age-
dependent metabolic parameters, the retention of radium decay products, the
values of organ-specific risk coefficients and the relative biological
effectiveness of alpha irradiation.
Drinking Water Ingestion Rate
As discussed in the previous section on noncancer risks, an adult
drinking water ingestion rate of 2 L/day is a reasonable maximum value. The
mean rate of total fluid consumption for adults is approximately 2.1 L/day,
with approximately 1.2 L/day from tap-water based drinks (Ershow and Cantor
1989). A few adults may habitually drink much more than 2 L/day of tap water,
and some adults may drink essentially no water from the tap. The risks of
radium ingestion are linearly proportional to drinking water ingestion rate,
and the use of 2 L/day provides a standard value for which cancer risks are
calculated for drinking water contaminants. The risks to a typical member of
the population may be approximately 60% lower than those calculated.
IX-15
-------
Gastrointestinal Uptake
A3 discussed in the previous section on noncancer risks, the value of 0.2
used for gastrointestinal absorption is likely to be correct within a factor
of 2 to 3.
Aae-Deoendent Metabolic Parameters
Age-dependent metabolic parameters include organ size, organ uptake
fractions and effective organ half lives. USEPA (1989b) did not specifically
address the uncertainty contributed by age-dependent factors for radium in the
RADRISK model. Hcwaver, since risks are calculated for lifetime ingestion,
overall uncertainty will be less than the uncertainty in childhood riskat due
to the relatively short duration of childhood.
Retention of Radium Decay Products
Retention of radium decay products is evaluated based on the elemental
form of the decay product and on the radiologic half life in each isotope (see
Section XXI.D). The biological half life of radon in bone compartments used
in these calculations (see Table IXX-1) is based on data on retention of
Rn-222 (Sullivan et al. 1981). Some uncertainty is introduced in using these
parameters to assess retention of Rn-220, the radon decay product of Ra-228.
The retention parameters used imply that about 30% of Rn-222 decays in bone
(70* is lost) and about 40% of Rn-220 decays in bone (60% is lost). Other
investigators have assumed that 85% or 100% of Rn-220 decays in bone (Rowland
et al. 1978, ICR? 1979). Since decay products contribute substantially to the
total risks calculated for Ra-226 and Ra-228, the magnitude of the uncertainty
IX-16
-------
introduced by using the Rn-222 retention parameters for Rn-220 can be
estimated to be up to a factor of 2.
Organ-Speelfle Rlak Coeffieianta
As outlined in Section VII.B, orjan-specific riak coefficients used in
the RADRISJC model were derived primarily from relative risk values presented
by the BETR ill committee of the National Academy of Sciences (NAS 1980b).
The linear-linear model with a 10-year minimum latency and a 30-year follow-up
was used. NCHS 1970 mortality data and a birth maleifemale ratio of 1.0511
were chosen, and the model was constrained by total risk at all sites IUSEFA
1989b). For leukemia and bone cancer, absolute risks with a 2-year minimum
induction period and a 25-year expression period were calculated from the BEIR
III absolute risk coefficients, using a relative biological effectiveness of 8
for alpha particles. Mortality to incidence ratios were derived from vclues
provided by BEIR III. USEPA (1989b) estimated that uncertainty in these
parameters from low-dose extrapolation was a factor of 2 or less, from the
time and age dependent factors was a factor of 2, from extrapolation to the
U.S. population was less than 10% and from dosimetry and sampling errors was
roughly 10%.
Revised Hloh-LET Leukemia Risk Coefficient
In the case of leukemia prediction, as shown in Section VIZI.B.2, the
high-LET risk coefficient was adjusted from 3.6E-4 to 5.0E-5 per rad. The
high-LET leukemia risk coefficient based on thorctrast patients was estimated
to be 5E-5 to 6E-5 per rad by the BEIR IV committee (NAS 1988) and 4E-5 by
Mays et al. (1985a). This adjustment was made in response to comments by the
IX-17
-------
SAB that leukenia9 had not been observed among radium dial painters. Three
other methods of deriving a high-LET Leukemia risk coefficients from the
incidence of leukemias among patients injected with Ra-224 are outlined below.
These calculations are baaed 0:1 the assumption that all observed leukemias
among these cohorts are related to exposure to Ra-224, but some might be
induced by analgesic drugs also given to these patients {Hays et al. 1986).
Mav9 et al. (1986). Mays et al. (1986) reported 3 excess cases of
leukemia in 681 adults with an average skeletal dose of 206 rad, following
treatment with Ra-224. The endosteal dose is about 8.9 times the average
skeletal dose for Ra-224 (Spiesa and Mays 1970; Marshall et al. 1978). The
marrow dose is about 0.09 times the endosteal dose (ICRP 1979; USEPA 1989b).
Therefore, the marrow dose Ls (8.9)x(0.09)=0.8 times the average skeletal
dose. The marrow dose in this cohort is (0.8)x{206 rad}=165 rad and the total
cohort irradiation i3 (16S rad)x(681 persons)=1.12E5 person-rad. Three
leukemias in this cohort gives an incidence of 2.7E-5 per rad.
Wick et al. <19861. Wick et al. (1986) reported 7 excess bone marrow
failures and one excess leukemia in 1,501 adults with an average skeletal dose
of 56 rad, following treatment for ankylosing spondylitis with Ra-224. The
marrow dose in this cohort is (0.8|x(56 rad)=45 rad and the total cohort
irradiation is {45 rads)x
-------
this cohort is (0.8)x(204 rad)«163 rad and the total cohort irradiation is
(153 rade)x(7Qe persona)«i.isss peraon-rad. For S bone marrow failures, the
incidence is 4.3E-5 per rad.
These marrow risks for Ra-224 range from 1.5E-5 to l*0E-4 and agree
fairly well with the risks ranging from 4E-5 to 6E-5 calculated for thorotrast
(Mays et al. 1965a, NAS 1988). This agreement supporta the use of an estimate
of 5.OE-S leukemiaa per rad as a reasonable estimate of the effects of alpha
particle irradiation of bone marrow, and suggests that the uncertainty in this
estimate is a factor of 2 to 3.
Head Carcinoma Risk Coefficient
No detailed dosimetry has been developed for Rn-222 gas that accumulates
in paranasal sinuses and mastoid air cells from decay of internally deposited
Ra-226 (NAS 1988). The risk of head carcinomas associated with Ra-226
ingestion was estimated based on the ratio of head carcinomas to bone sarcomas
among the dial painters exposed to both Ra-226 and Ra-228. In the total
cohort, the ratio of head carcinomas to bone sarcomas is 37/85 ¦ 0.44 (Maya
1983}. In the linearized Ra-226 low-dose extrapolations of Kays et al.
(1985a), the ratio of risk coefficients for head carcinomas to bone sarcomas
is 1.6E-S/1.0E-5 =» 1.6. The ratio assumed in this document was 1.0, near the
midpoint of these two estimates. The uncertainty in this estimate is
approximately a factor of 2.
IX-19
-------
Relative Biological Effectivenes9 of Aloha Irradiating
The RADRISK model uses a value of 8 for the relative biological
effectiveness of alpha irradiation in all tissues. This value was derived
from the value of 20 recommended by ICRP {1980) when comparing high-LET
radiations to low-LET radiations at low doses and dose rates. The Iow-let
effectiveness at low doses and dose rates was reduced by a factor of 2.5 in
the ICRP formulation. Since the USEPA linear risk model for low-LET radiation
does not use a dose rate reduction factor, the appropriate relative biological
effectiveness is equal to 20/2.5 = 8 (USEPA 1989b). If the value for relative
biological effectiveness were to be changed, calculated risks from radium
ingestion would change less than linearly, because a substantial portion of
the risk is due to bone sarcomas, leukemias and head carcinomas (see
Table VIII-5), which are calculated from a high-LET risk coefficient that
would not be affected by a change in the relative biological effectiveness.
3. Genetic and Developmental Effects
The risks of genetic and developmental effects of radium were analyzed
using th« radiation dose calculated with the RADRISK model. The same
parameters discussed in the previous section also influence the predicted
genetic risk: drinking water ingestion rate, gastrointestinal absorption,
age-dependent metabolic parameters, retention of radium decay products, values
of organ-specific risk coefficients (in this case, mutation risk in gonads)
and the relative biological effectiveness of alpha irradiation in causing
genetic damage. Only the latter two parameters are specific to genetic
effects.
IX-20
-------
Mutation Risk Factors
The Agency has derived risk factors for genetic effects of low-LET
irradiations, based on estimates by expert committees evaluating data
primarily derived from animals exposed to low-LET radiation (USEPA 1989b).
The Agency identified six sources of uncertainty in the estimates for low-LET
radiation risk factors: species used for extrapolation, genetic loci used for
extrapolation, dose-rate reduction factor, relative sensitivity of oogonia and
spermatogonia, spontaneous mutation rate and doubling dose in man. The true
risk was estimated to be within a factor of 4 of the predicted risk for
estimates of mutations in the first generation, with additional uncertainty
introduced by extrapolation to the equilibrium (all-generation) rate of
genetic effects (USEPA 1989b).
Relative Biological Effectiveness of Aloha Irradiation
The Agency based the RBE of alpha radiation in causing genetic effects on
the conclusions of the BEIR IV committee that the RBE was IS for chromosomal
aberrations and 2.5 for all other effects (NAS 1988). The overall average RBE
is then 2.7, which was rounded to 3 (USEPA 1989b). The uncertainty in the
selection of RBE was estimated to be a factor of 5 (USEPA 1989b).
C. Summary
One of the largest sources of uncertainty in assessing the risks of
radium in drinking water is contributed by uncertainty in the shape of the
dose-response curve for bone cancer induction by radium. If the true response
were quadratic (dose-squared), due to a two-hit model of cancer induction (NAS
IX-21
-------
1980b) or to the promoting effects of bone necrosis (Martland 1931), true
risks could be more than an order of magnitude lower than those calculated.
For other assumptions and parameters, there does not appear to be any
systematic bias towards over- or under-prediction of risk, except that a
protective value for adult drinking water consumption has been used, that is
higher than the mean by less than a factor of 2. The overall uncertainty
contributed by modelling assumptions and parameter variability may be
approximately a factor of 5 for calculating radiation dose in target organs
and a factor of 3 for calculating organ-specific risks. These considerations
suggest that the true risk is likely to be within an order of magnitude in
either direction of those calculated for members of the general population.
IX-22
-------
X. REFERENCES
Ackerman LV, Spjut HJ- 1962. Turners of bone and cartilage. Washington, DCs
Armed Forces Institute of Pathology.
Adams N, Hunt BW, Reisland JA. 197B. Annual limits of intake of
radionuclides for workers. Harwell, Didcot: National Radiological Protection
Board. NRPF-R82.
Aiata EM, Singly JE, Trussell AR, et al. 1987. Radionuclides in drinking
water: An overview. J AWWA 79:144-152.
Ames LL, Rai D. 197S. Radionuclide interactions with soil and rock media. Vol
1: processes influencing radionuclide mobility and retention, element
chemistry and geochemistry, conclusions and evaluation. Las Vegas, NV: U.S.
Environmental Protection Agency, Office of Radiation Programs. EPA
520/6-78-007. p3-142-3-149; 4-1-4-6.
Andrews HL. 1961. Radiation biophysics. Edgewood Cliff, NJ: Prentice-Hall.
ATSDR. 1969. Agency for Toxic Substances and Disease Registry.
Toxicolcgical Profile for Radium. Draft for public comment. Atlanta, GA:
Agency for Toxic Substances and Disease Registry.
Sacq ZM, Alexander P. 1961. Fundamentals of radiobiology, 2nd ed. New York:
Pergaraon Press.
Bean JA, Isaacson P, Hahne RM, Kohler J. 1982. Drinking water and cancer
incidence in Iowa: II. Radioactivity in drinking water. Am. J. Epidemiol.
116:924-932.
Benes P, Strejc P. 1936. Interaction of radium with freshwater sediments and
their mineral components: IV. Waste water and riverbed sediments. J.
Radioanal. Nucl. Chem. 92:407-422.
Benes P, Borovec 2, Strejc P. 1986. Interaction of radium with freshwater
sediments and their minera1, components: III. Muscovite and feldspar. Journal
of Radioanalytical and Nuclear Chemistry 98:91-103.
Benes P, Borovec Z, Strejc P. 1985. Interaction of radium with freshwater
sediments and their mineral components: II. Kaolinite and montmorillcnite.
Journal of Radioanalytical and Nuclear Chemistry 89:339-3Sl.
Srnes P, Strejc P, Lukavec Z. 1984. Interaction of radium with freshwater
sediments and their mineral components: I. Ferric hydroxide and quartz.
Journal of Radioanalytical and Nuclear Chemistry 82:275-285.
Bernard, SR, Snyder, WS. 1975. Metrbolic models for estimatation of interal
radiation exposure received by human subjects from the ingestion of noble
gases. Health Physics Division Progress Report. Oak Ridge, TN: Oak Ridge
National Laboratory. June 20.
X-l
-------
Boyd JT, Court Brown WM, Vennart J, Wookcock GE. 1966. chromosome studies on
women formerly employed ag luminous-dial paintBrs. Br. Med. J. 1:377-382.
Cech I, Lemma M, Kreitler CW, Prichard HM. 1988. Radium and radon in water
supplies from the Texas Gulf coastal aquifer. Water Res. 22:109-121.
Clarke WJ. 1962. Comparative histopathology of Pu-239, Ra-226, and Sr-90 in
pig bone. Health Phys. 8:621-627.
Clifford DA. 1990. Removal of radium from drinking water. In: Cothern CR,
Rebers PA, ed9. Radon, radium and uranium in drinking water. Chelsea, HI:
Lewis Publishers, pp. 225*247.
Crawford-Brown DJ. 1990. The price of confidence: The rationality of radium
removal from drinking water. In: Cothern CR, Rebers PA, eds. Radon, radium
and uranium in drinking water. Chelsea, MI: Lewis Publishers, pp. 213-223.
Dunning DE, Leggett RW, Sullivan RE. 1984. An assessment of health risk from
radiation exposures. Health Phys. 46:1035-1051.
Dunning DE, Leggett RW, Yalcintas MG. 1960. A combined methodology for
estimating dose rates and health effects from exposures to radioactive
pollutants. Oak Ridge, TN: Oak Ridge National Laboratory. TM-7105.
Ershow AG, Cantor KP. 1989. Total water and tapwater intake in the United
States: Population-based estimates of quantities and sources. Report
prepared under National Cancer Institute Order #263-MD-310264.
Evans RD, Harris RS, Bunker JWM. 1944. Radium metabolism in rats and the
production of osteogenic sarcoma by experimental radium poisoning. Am. J.
Roentgenol. Radium Therapy. 52:353-373.
Evans RD, Keane AT, Shanahan MM. 1972. Radiogenic effects in man of long-
term skeletal alpha-irradiation, in: Stover BJ, Jee wss, eds. Radiobiology
of plutonium. Salt Lake City, UT: The J. W. Press, pp. 431-463.
Finkel AJ, Miller CE, Hasterlik RJ. 1969a. Radium-induced malignant tumors
in man. In: Mays CE, Jee WSS, Lloyd RD, Stover BJ, Doughtery JF, Taylor GN,
eds. Delayed effects of bone-seeking radionuclides. Salt Lake City, UT:
University of Utah Press, pp. 195-225.
Finkel MP, Biakis BO, Jinkins PB. 1969b. Toxicity of radium-226 in mice.
In: Ericson A, ed. Radiation induced cancer. Vienna, Austria:
International Atomic Energy Agency, pp. 369-391.
FRC. 1961. Federal Radiation Council. Background material for the
development of radiation protection standards. FRC Report No. 2. Washington
DC: U.S. Department of Health Education and Development, USPHS.
Gilkeson RH, Cowart JB. 1982. A preliminary report on '"uranium series
disequilibrium in groundwater of the Cambrian-Ordovician aquifer system of
Northeastern Illinois. In: Perry EC Jr., Montgomery CW, Eds. Isotope studies
of hydrologic processes. DeKalb, IL: Northern Illinois University Press, pp.
109-113.
X-2
-------
Hallden $A, Fi9enne IK, Karley JH. 1963. Radium-225 in human diet arid bone-
Science. 140:1327-1329.
Harley Nil, Pasternack 05. 1975. A comparison of the dose to cells on
trabecular bone surfaces from Piutonium-239 and Radium-225 baaed on
experimental alpha absorption mea.3>ir«iet\ts* Healer. Phys. 33:35-46.
Hesfl CI, Ki.cir.al J, Korcon TR, PriOiird Km, C&niglio WA. 1935. The occurrence
of radioactivity in public water supplies in the United Statea. Health. Shys.
43;553-586.
Boegerman ST, Cuiwiins HT, Greco I, Smith J, Sor.za-Novera j, 1973. Chromosome
breakage in patients exposed to Laternai emitters. Radiological and
Environmental Research Division Annual Report. July 1972-June 1973. Center
for Human Radiobiology. ANL-S06O-I1, p. 81-90.
Humphreys ER, Loutit JF< H»jor IR, Stones VA. 1905. The induction by tt*Ra of
myeloid leukaemia and osteosaxcoaa is\ male CBA mica- int. J. Radiat. Biol.
47:239-247,
tCRP. 1959. International Commission on Radiological Protection. Report of
the ICR? committee II on permissible dose for internal radiation* New York:
Pergamon Press (also reprinted as Health Miys. 3:1~3Q, 1960).
ICJtP. 1968. International Cswfcissiaa on Radiological Protection. A review
of the radiosensitivity of the tissues La bone. ICRP publication 11. He*»
York: Pergainon Press.
ICR?. 1971. International Commission on Radiological Protection. The
assessment of internal contamination resulting from recurrent or prolonged
uptakes. ICRP Publication 10A. New York: pergatfon Press.
ICRP. 1973. International Commission on Radiological Protection. Alkaline
earth metabolism Lti adult man. ICRP Publication 20. Mew forks Patrgamen
Press.
ICRP. 1975. International Commission otv radiological Protection. Report of
the taflk group on reference man, ICRP publication 23, Hew York; pergarfion
Pre3s.
IC-RP. 1977, International Commission on Radiological Protection. Report of
the task group on reference man, ICRP Publication 25. Hew York: pergamon
press.
ICRP. 1979. International commission on Radiological Protection. Limits for
intakes of radionuclides by workers. ICRP publication 30. New York:
pergamon Press.
ICRP. 1980. international commission on Radioingica.1 Protection. Limits for
intakes of radionuclides by workers. ICRP publication 30, Part 2. Oxford:
pergatnon Press,
James KG, Ta-yler DM. 1971. DTPA therapy for chelation of "'pu in bone: The
influence of bone remodelling. Health Phys. 21t31-S'3.
X-3
-------
sjee WSS. 1971. Bone-seeking radionuclides and bones. Ins BerdjiB CC.
Pathology of irradiation. Baltimore, MO: Williams 4 Wilkins Co., pp.
186-212.
jee ws, Hartley MH, Dockum NL, Yee J, Kenner GH. 1969. Vascular changes in
bones following bone-seeking radionuclides, in: Mays CW, Jee WS, Lloyd RD, at
al.,eds. DeLayed effects of bone-seeking radionuclides. Univ. of Utah Press,
Salt Lake City., pp. 437-455.
Kalin H. 1988. Long-term ecological behaviour of abandoned uranium mill
tailings. 3. Radionuclide concentrations and other characteristics of
tailings, surface waters, and vegetation. Ottawa, Ontario, Canada:
Environment Canada. EPS 3/HA/4.
Keane AT, Kirsh IE, Lucas HF, schlenker RA, Stehr.ey AF. 1963. Non-stochastic
effects of SMRa and i2SRa. in the human skeleton. In: Biological effects of
low-level radiation. Vienna, Austria: International Atomic Energy Agency,
pp. 329-350.
Keane AT, Schlenker RA. 1983. Soft tissue dosimetry of radium in humans-I.
Health Phys. 44:81-89.
King PT, Michel J, Hoore MS. 1982. Ground water geochemistry of iURa, JMRa
and iJ,Rn. Geochim. Cosmochim. Acta 46:1173-1182.
Kofranek V, Sedlak A, Bubenikova 0, Dvorak V. 1985. Late effects of ""Ha,
"*Ra and JS9Pu in female mice of the ICR strain. Strahlentherapie [Sonderb.]
80j 88-92.
Krishnaswami S, Graustein WC, Turekian XX, Dowd JF. 1982. Radium, thorium
and radioactive lead isotopes in groundwaters: Application to the in situ
determination of adsorption-desorption rate constants and retardation factors.
Water Resources Res. 18:1663-1675.
Langmuir 0, Riese AC. 1985. The thermodynamic properties of radium.
Geochim. Cosmoschim. Acta 49:1593-1601.
Lloyd E. 1970. Argonne National Laboratory Report, ANL 7760, Part II (cited
in Harley and Pasternack 2976> -
Lloyd EL, Henning CB. 1983. Cells at risk for the production of bone tumors
in radiura exposed individuals: an electron microscope study. Health Phys.
44(Supp 1):135-148.
Longtin JP. 1988. Occurrence of radon, radium, and uranium in groundwater.
J. AWWA (July):84-93.
Longtin J. 1990. Occurrence of radionuclides in drinking water, & national
study. In: Cothern CR, Rebers PA, eda. Radon, radium and uranium in drinking
water. Chelsea, HI: Lewis Publishers. 97-139.
Looney WB, Hursh JB, Archer VE, Steadman LT, colodzin M. 1956. A summary of
radium and thorium excretion in humans. Proceedings of the international
conference on the peaceful uses of atomic energy: biological effects of
radiation, Geneva, 11 •.55-64.
X-4
-------
Luz A, Muller WA, Gossner W, Hug O. 1976. Estimation of tumour risk at low
dose from experimental results after incorporation of short-lived bone-seeking
alpha emitters ?MRa and "7Th in mice. In: Biological and environmental
effects of low-level radiation. Vol. II. Vienna, Austria: International
Atomic Energy Agency, pp. 171-181.
Lyman GH, Lyman CG. 1986. Leukemia and groundwater contamination [Letter].
JAMA 256:2676-2677.
Lyman GH, Lyman CG, Johnson W. 1985. Association of leukemia with radium
groundwater contamination. JAMA 254:621-626.
Maletskos CJ, Keane AT, Telles NC, Svans RD. 1966. The metabolism of
intravenously administered radium and thorium in human beings and the relative
absorption from the human gastrointestinal tract of radium and thorium in
simulated radium dial paints. In: Radium and mesothorium poisoning end
dosimetry and instrumentation techniques in applied radioactivity. MIT-952-3.
Cambridge, MA: Massachusetts Institute of Technology, Physics Department, pp.
202-317.
Maletskos CJ, Keane AT, Telles NC, Evans RD. 1969. Retention and absorption
of 22'Ra and 134Th and soma dosimetric consideration of "*Ra in human beings.
In: Mays CE, Jee WSS, Lloyd RD, Stover BJ, Doughtery JF, Taylor GN, eds.
Delayed effects of bone-seeking radionuclides. Salt Lake City, UT:
University of Utah Press, pp. 29-49.
Martland HS. 1931. The occurrence of malignancy in radio-active persons.
Am. J. Cancer 1S:2435-2516.
Martland HS, Martland HS Jr. 1950. Placental barrier in carbon monoxide,
barbiturate and radiusi poisoning. Am. J. Surg. September: 270-279.
Marshall JH, Groer ?G, Schlenker RA. 1978. Dose to endosteal cells and
relative distribution factors for Radiutn-224 and Plutonium-239 compared to
Radium-226. Health Phys. 35:91-101.
Mays CW. 198C. Alpha-particle-induced cancer in humans. Health Phys.
55:637-652.
Mays CW, Lloyd RD. 1972. Bone sarcoma incidence vs. alpha particle dose.
In: Stover 8J, Jee WSS, eds. Radiobiology of plutonium. Salt Lake City, UT:
The J. W. Press, pp. 409-430.
Mays CW, Lloyd RD, Taylor GN, Wrenn ME. 1987. Cancer incidence and lifespan
vs. alpha-particle dose in beagles. Health Phy3. 52:617-624.
Kays CW, Rowland RE, Stehney AF. 1985a. Cancer risk from the lifetime intake
of Ra and U isotopes. Health Phys. 48:635-647.
Mays CW, Spiess H. 1984. Bone sarcomas in patients given radium-224. In:
Boice JD, Fraumeni JF, eds. Radiation carcinogenesis; epidemiology and
biological significance. New York, NY: Raven Press, pp. 241-252.
X-5
-------
Maya cw, Spiesa H, Chmelevsky D, Kellerer A. 198Sb. Bone sarcoma cumulative
tumor rates in patients injected with M*Ra. Strahlentherapie [Sonderb.j
80:27-31.
Mays CW, spiess H, Chmelevsky D, Kellerer A. 1986. Bone Barcoma cumulative
tumor rates in patients injected with "*Ra. Ins Gossner W, Gerber GB, Hagen
U, Luz A, eds. The Radiobiology of Radium ar.d Thorotrast. Baltimore: Urban
and Schwarzenberg, Baltimore, pp. 27-31,
Kays CW, Tueller AB. 1964. Determination of localized alpha-dose IV in aoft
tissue near radioactive bone. In: Research in Radiobiology. University of
Utah Report COO-229, pp. 199-205
McLean FC, Budy AM. 1964. Radiation, Isotopes and Bone. New York: Academic
Press.
Michel J. 1990. Relationship of radium and radon with geologic formations.
In: Cothern CR, Rebers PA, eds. Radon, radium and uranium in drinking water.
Chelsea, MI: Lewis Publishers, pp. 83-95.
Momeni MH, Williams JR, Jow N, Rosenblatt LS. 1976. Dose rates, dose and
time effects of #0Sr & and JsaRa on beagle skeleton. Health Phys.
30:381-390.
Muller J, Klener V, Tuscany R, Thomas J, Brezikova D, Houskova M. 1966.
Study of internal contamination with strontium-90 and radium-226 in man in
relation to clinical findings. Health Phys. 12:993-1006.
Muller WA, Linzner U, Luz A. 1983. Early induction of leukemia (malignant
lymphoma) in mice by protracted low alpha-doses. Health Phys. 54:461-463.
Muth H, Globel B. 1983. Aged dependent concentration of "#Ra in human bone
and some transfer factors from diet to human tissues. Health Phys. 44(Suppl.
1):113-121.
NAS. 1977. National Academy of Sciences. Drinking water and health.
Washington, DC: National Academy of Sciences, pp. 489-493.
NAS. 1980a. National Academy of Sciences. Drinking water and health.
Washington, DC: National Academy of Sciences, pp. 24-37.
NAS. 1980b. National Academy of Sciences - National Research Council. The
effects on populations of exposure to low levels of ionizing radiation, report
of the Comriittee on Biological Effects of Ionizing Radiations (BEIR III).
Washington, DC: National Academy of Sciences.
NAS. 1988. National Academy of Sciences - National Research Council. Health
risks of radon and other internally deposited alpha-emitters, report of the
Committee on Biological Effects of Ionizing Radiations (BEIR IV). Washington,
DC: National Academy of Sciences.
ncrp. 1987. National Council on Radiation Protection and Measurement,
jxposure of the population in the United States and Canada from natural
background radiation. Bethesda, KD: National Council on Radiation Protection
and Measjirfyaeat. NCRP Report No. 94,
X-6
-------
Neuman WR, Newman MW. 1958. The chemical dynamics of bone mineral. Chicago:
University of Chicago Press.
Norris WP, Speckman TW, Gustafson PF. 1955. Studies of the metabolism of
radium in man. Am. J. Roentg. Rad. Therapy Nuclear Med. 73:785-802.
Olivieri G, Bodycote J, Wolff s. 1984. Adaptive response of human
lymphocytes to low concentrations of radioactive thymidine. Science
223:594-597.
Owen M. 1963. Cell population kinetics of an osteogenic tiasue: I and II,
Brookhaven National Laboratory Reports BNL-6737 (1963J and BNL-6736 (Cited in
Mays and Tueller 1964).
Park HZ, Whitson SW, Jee WS. 1972.
bone. In: Stouer BJ, Jee WS, eds.
Salt Lake City, UT, pp. 305-323.
Vascular theory of radiation injury to
Radiobiology of Plutonium. J.W. Press,
Petersen NJ, Samuels LD, Lucas HF, Abrahams SP. 1966. An epidemiologic
approach to low-level 2JBRa exposure. Public Health Rep. 81:805-814.
Polednak AP. 1986. Leukemia and radium groundwater contamination [Letter].
JAMA 255:903-904.
Polednak AP. 1980. Fertility of women after exposure to internal and
external radiation. J. Environ. Pathol. Toxicol. 4:457-470.
Raabe OG, Book SA, Parks NJ. 1980. Bone cancer from radium: Canine dose
response explains data for mice and humans. Science 208:61-64.
Raabe OG, Book SA, Park3 NJ. 1983. Lifetime bone cancer dose-response
relationships in beagles and people from skeletal burdens of "#Ra and ®°Sr.
Health Phys. 44(Suppl. 1):33-48.
Raabe OG, Book SA, Parks NJ, Chrisp CE, Goldman M. 1981. Lifetime studies of
226 Ra and 90 Sr toxicity in beagles- a status report. Radiat. Res.
86:515-528.
Raabe OG, Parks NJ, Book SA. 1981. Dose-response relationships for bone
tumors j.n beagles exposed to i!9Ra and ®°Sr. Health Phys. 40:863-880.
Rajewsky B, Belloch-Zimmermann V, Lohr E, Stahlhofen-W. 1965. "*Ra in human
embryonic tissue, relationship of activity to the stage of pregnancy,
measurement of natural 2MRa occurrence in the human placenta. Health Phys.
11:161-169.
Rowland RE, Stehney AF, Lucas HF, Jr. 1978. Dose-response relationships for
female radium dial workers. Radiat. Res. 76:368-383.
Rowland RE, Stehney AF, Lucas HF, Jr. 1983. Dose-response relationships for
radium-induced bone sarcomas. Health Phys. 44(Suppl. 1):15—31.
X-7
-------
Rundo J, Keane AT, Lucas HF, Schlenker RA, Stebbings JH, Stahney AF. 1986.
Current (1984) status of the study of mRa and l,aRa in humans at the Center
for Human Radiobiology. In: Gossner w, Gerber GB, H&gn V, Luz A, eds. the
Radio biology of Radium and Thorotrast. Baltimore: Urban C Schwarzenberg.
Strahlentherapie [Sonderb.] 80:14-21.
SAB. 1987. Science Advisory Board to the U.S. Environmental Protection
Agency, Radiation Advisory committee, Drinking Water Subcommittee. Review ot
the Office of Drinking Water's assessment of radionuclides in drinking water
and four draft criteria documents: Man-made radionuclide occurrence/ uranium,
radium, radon. July 1907. SAB-RAC-87-035.
SAB. 1990. Science Advisory Board to the U.S. Environmental Protection
Agency, Radiation Advisory Committee, Radionuclides in Drinking Water
Subcommittee. Review of the Office of Drinking Water's criteria documents and
related reports for uranium, radium, radon and manmade beta-gamma emitters.
Draft dated December 30, 1990.
Schlenker RA. 1982. Risk estimates for bone. In: Critical issues in
setting radiation dose limits. Proceedings No. 3. Bethesda, MS: National
Council on Radiation Protection and Measurements, pp. 153-163.
Schoeters GE, Luz A, Vanderborght OL. 1983. "*Ra induced bone-cancers; the
effects of a delayed Na-alginate treatment. Int. J. Radiat. Biol.
43:231-247.
Schoeters GE, Vanderborght LJ. 1981. Temporal cr.d spatial response of marrow
colony-forming cells (CFU-s and CFU-c) after M#Ra incorporation in BALB/c
mice. Radiat. Res. 88:251-265.
Scott BL. 1967. J. Cell Biol. 35:115 (cited in Harley and Pasternack 1976}.
Sharpe WD. 1974. Chronic radium intoxication: clinical and autopsy findings
in long-term New Jersey survivors. Environ. Res. 8:243-383.
Sharpe WD. 1976. Chronic radium intoxication: morphology of bone and marrow
infarcts. In: Jee WSS, ed. The Health Effects of Plutonium and Radium.
Salt Lake City, UT: The J.W. Press, pp. 4S7-483.
Sikov MR, Mahlum DD. 1976. Influence of age and phy3iochemical form on the
effects of 239 Pu on the skeleton of the rat. In: Jee, WSS, ed. The Health
Effects of Plutonium and Radium. Salt Lake City, OT: The J.W. Press, pp.
33-47.
Spiers FW, Lucas HF, Rundo J, Anast GA. 1983. Leukemia incidence in the U.S.
dial workers. Health Phys. 44(Supp. l):65-72.
Spiess H, Mays CW. 1970. Bone cancers induced by JJ*Ra (ThX) in children and
adults. Health Phys. 19:713-729.
Spiess H, Mays CW, Chmelevsky D. 1989. Malignancies in patients injected with
"4Ra. In: Taylor DM, Mays CW, Gerber GB, Thomas RG, eds. Risks from radium
and thorotrast. London, England: British Institue of Radiology.
X-8
-------
Stebbing3 JH, Lucas HF, stehney AF. 1984. Mortality from cancers of major
sices in female radium dial workers. Am* J. Ind. Med. 5:435-459.
Stebbing3 JH, Lucaa HP, Toohey RE. 1966. I-eukemia and radium groundwater
contamination {Letter], JAMA 255:902.
Stehney AF, Lucas HF. 1956. Studies on the radium content of humans arising
from the natural radium of their environment. Proceedings of the
International conference on the Peaceful Uses of Atomic Energy, Volume 11,
Biological Effects of Radiation, Geneva, pp. 49-54.
Storer JB. 1975. Radiation carcinogenesis, in: Becker FF, ed. cancer: A
comprehensive Treatise; Vol. 1: Etiology: Chemical and Physical
Carcinogenesis. New ¥ork: Plenum Press, pp. 453-433.
Sullivan RE, Nelson NS, Ellet WH, Dunning DE, Jr., Leggett RW, Tfalcintas MC,
Eckerman KF. 1981. Estimates of health risk from exposure to radioactive
pollutants. Oak Ridge, TN: Oak Ridge National Laboratory. TM-7745.
Taylor GN, Dougherty TF, Mays CW, Lloyd RD, Atherton DR, Jee WSS. 1972.
Radium-induced eye melanomas in dog3. Radiat. Res. 51:361-373.
Taylor DM, Bensted JPM. 1969. Long-term biological damage from Plutoniura-239
and Americium-241 in Rats. In: May CW et al, eds. Delayed effects of bone-
seeking radionuclides. Salt Lake City, (JT: University of Utah Press, pp.
357-370.
Taylor GN, Jee ws, Mays CW. 1976. Some similarities of radium and plutonium
toxicity in the beagle and man. In: Jee wss, ed. The health effects of
plutonium and radium. Salt Lake City, UT: J.W. Press, pp. 523-536.
Taylor GN, Mays CW, Lloyd RD, Gardner PA, Talbot LR, McFarland SS, et al.
1983. Comparative toxicity of "®Ra, aBPu, M,Am, i4'cf, and !S2Cf in C57BL/DO
black and albino mice. Radiat. Res. 95:584-601.
Thome MD. 1977. Aspects of the dosimetry of alpha-emitting radionuclides in
bone with particular emphasis on iS,Ra and i3SPu. Phys. Med. Biol. 22:36-46.
Tracey 3L, Letourneau EG. 1986. Leukemia and radium groundwater
contamination (Letter). JAMA 255:3365-3366.
Tuschl H, Altmann H, Xovac R, Topaloglore A, Egg D, Gunther R. 1980. Effects
of low-dose radiation on repair processes on human lymphocytes. Radiat. Res.
81:1-9.
Tuschl H, Kovac R, Altmann H. 1983. UDS and SCE in. lymphocytes of persons
occupationally exposed to low levels of ionizing radiation. Health Phys.
45:1-7.
UNSCEAR. 1972. United Nations Scientific Committee on the Effects of Atomic
Radiation. Ionizing radiation: Levels and effects, volume I: Levels.
Geneva: United Nations. U.N. publication #.72.IX.17.
X-9
-------
Upton AC. 197S. Physical carcinogenesis: radiation - history and aources.
In: Becker FF, ed. Cancer: A comprehensive Treatise; Vol. 1: Etiology:
Chemical and Physical Carcinogenesis. New York: Plenum Press, pp. 387-403.
USEPA. 1906. U.S. Environmental Protection Agency. Office of Drinking
Hater. Water pollution control; National Primary Drinking Water Regulations;
radionuclides; advance notice of proposed rulemaking. Fed. Regist., Sept. 30,
1906, 51 34636-34862.
USEPA. 1988. U.S. Environmental Protection Agency. Regarding distribution
tables for the National Inorganics and Radionuclides Survey (KIRS) results.
Memorandum from Jon Longtin, WSTB to Arthur Perler, STB. February 23, 1988.
USEPA. 1989a. U.S. Environmental Protection Agency. Office of Health and
Environmental Assessment. Exposure factors handbook. Washington! DC: U.S.
Environmental Protection Agency. EPA/600/3-89/043.
USEPA. 1989b. U.S. Environmental Protection Agency, Office of Radiation
Programs. Risk assessment methodology, environmental impact statement for
NESHAPS radionuclides. Vol. I. Background Information document. Washington
DC: United States Environmental Protection Agency. EPA 520/1-80-005.
USEPA. 1990a. U.S. Environmental Protection Agency, office of Drinking
Water. Occurrence and exposure assessment for radium-226 in public drinking
supplies. Prepared by Wade Miller Associates, Inc. for U.S. EPA Office of
Drinking Water under contract 68-03-3514. April 30, 1990. Washington, DC:
U.S. Environmental Protection Agency.
USEPA. 1990b. U.S. Environmental Protection Agency, Office of Drinking
Water. Occurrence and exposure assessment for radium-228 in public drinking
water supplies. Prepared by Wade Miller Associates, inc. for U.S. EPA Office
of Drinking Water under contract 68-03-3514. November 30, 1990. Washington,
DC: U.S. Environmental Protection Agency.
USEPA. 1990c. U.S. Environmental Protection Agency, office of Solid Waste
and Emergency Response and Office of Radiation Programs. Assessment of
technologies for the remediation of radioactively contaminated Superfund
sites. Appendix B: Radioactive waste superfund site description.
Washington, DC: U.S. Environmental Protection Agency. EPA/540/2-90/001.
USEPA. 1991a. U.S. Environmental Protection Agency. Office of Ground Water
and Drinking Water. Regulatory impact analysis of proposed national primary
drinking water regulations for radionuclides. Prepared by Wade Miller
Associates, Inc. for U.S. EPA Office of Ground Water and Drinking Water under
contract 68-03-3514. May 15, 1991. Washington DC: U.S. Environmental
Protection Agency.
USEPA. 1991b. U.S. Environmental Protection Agency. Office of Research and
Development. Health effects assessment summary tables. Annual FY-1991.
Washington, DC: U.S. Environmental Protection Agency. OEER 9200.6-303
(91-1).
Vaugan J. 1972. Bone surfaces: What are they? In: Stover BJ and Je® WSS,
eds. Radiobiology of Plutonium. Salt Lake City: The J.W. Press, pp.
323-332.
X-10
-------
Watson AP, Etnier EL, McDowell-Boyer LM. 1984. Radium-226 in drinking water
and terrestrial food chains: Transfer parameters and normal exposure and
dose. Nuclear Safety. 25:815-829.
Wick RR, Chmelevsky D, Gossner W. 1986. "
------- |