U.S. Environmental Protection Agency
NATIONAL WORKSHOP ON
RADIOACTIVITY IN DRINKING WATER
Workshop Chairman: C. Richard Cothern
Health Effects Branch
U.S. Environmental Protection Agency
Washington, DC
Workshop Coordinator: Sheri E. Marshall
Dynamac Corporation
Rockville, MD
-------
REGULATORY DEVELOPMENT OF THE INTERIM
AND REVISED REGULATIONS FOR RADIOACTIVITY
IN DRINKING WATER - PAST AND PRESENT
ISSUES AND PROBLEMS
William L. Lappenbusch
C. Richard Cothern
-------
.S. Environmental Protection Agency
^NATIONAL WORKSHOP ON
RADIOACTIVITY IN DRINKING WATER
Committee Issue Papers
-------
REGULATORY DEVELOPMENT OF THE INTERIM AND REVISED
REGULATIONS FOR RADIOACTIVITY IN DRINKING WATER -
PAST AND PRESENT ISSUES AND PROBLEMS*
William L. Lappenbusch and C. Richard Cothern
Office of Drinking Water
U.S. Environmental Protection Agency
Washington, D.C. 20460
Abstract
Developing the Revised Regulations for Radioactivity in
Drinking Water under the Safe Drinking Water Act requires
information from all areas and disciplines related to this
endeavor. As one step in the regulatory process, the back-
ground and history of that process as it applies to radio-
activity in drinking water is described. The issues involved
in developing the Revised Regualtions are detailed in the
following areas: monitoring and sources of exposure, dose
evaluation, health effect, engineering, economics and
general policy development. This paper thus was prepared
for use at the National Workshop for Radioactivity in Drinking
Water held at Easton, MD, May 24,26, 1983.
INTRODUCTION
The 93rd Congress passed Public Law 93-523 known as the
Safe Drinking Water Act (SDWA) , on December 16, 1974. This
law has subsequently been amended on November 16, 1977 (Public
Law 95-190), September 6, 1979 (Public Law 96-63) and December
5, 1980 (Public Law 96-502). The purpose of the Act was to
*The views expressed in this paper are those of the authors
and do not necessarily reflect the views and policies of the
U.S. Environmental Protection Agency.
-------
2
amend the Public Health Service Act to assure that the public
is provided with safe drinking water.
Section 1412 of the SDWA instructed the Environmental
Protection Agency (EPA), among other things, to propose and
promulgate first the interim regulations and later, the
revised regulations for hazardous constituents in drinking
water (organics, inorganics, microorganisms and radionuclides).
The EPA was instructed by Congress to work with the National
Academy of Sciences (NAS), or another equivalent body, and
propose recommended maximum contaminant levels (RMCLs) for
undesirable pollutants in drinking water as part of the
revised regulations. The RMCLs are to be based on health
and to take into consideration the impact of:
"(A) The existence of groups or individuals in the population
which are more susceptible to adverse health effects than the
normal healthy adult;
(B) The exposure to contaminants in other media than drinking
water (including exposures in food, in the ambient air, and
in the occupational settings) and the resulting body burden
of contaminants;
(C) Synergistic effects resulting from exposure to or
interaction by two or more contaminants;
(D) The contaminant exposure and body burden levels which
alter physiological- function or structure in a manner reasonably
suspected of increasing the risk of illness."
-------
3
Congress also instructed EPA to promulgate revised national
primary drinking water regulations which specify a Maximum
Contaminant Level (MCL), or to require the use of treatment
techniques for each contaminant for which an RMCL is established
if monitoring is not technically and economically feasible.
An MCL is a standard based on health, taking cost occurrence
and practicality into account. The health effects include
both carcinogenic, non-carcinogenic, fatal, and non-fatal
effects; the costs include those directly due to health
effects as well as those needed to cover monitoring and
treatment. As an adjunct to formal MCL's for drinking water
contaminants, the Office of Drinking Water (ODW) sometimes
develops Health Advisories (HA). These guidance levels have
no legal standing; however, they can provide useful information
to assist public water systems when an unregulated contaminant
is detected.
The NAS did not provide EPA with RMCLs for most of the
radionuclides that may occur in drinking water as dictated
by Congress (they did provide some discussion for uranium).
Therefore, this National Workshop for Radioactivity in Drinking
Water was particularly important in the potential standard-
setting process because it provided a public forum for
discussion of a large number of technical issues that will
be considered by EPA in its rulemaking activity. The Office
of Drinking Water assembled knowledgeable experts at this
workshop to address issues relating to radioactivity in
drinking water in such areas as (1) occurrence, (2) sampling
-------
4
and analytical methods, (3) metabolism and dosimetry of high
linear energy transfer (LET) radioisotopes, (4) health effects
and risks due to the ingestion of naturally occurring alpha
particle emitters, (5) preliminary thoughts about radon
ranging from intake(s), uptake(s) and resulting bioeffects,
(6) treatment, waste management and associated costs, and
(7) compliance and policy issues. The information will be
evaluated very carefully by EPA as the ODW prepares the
standard-setting criteria for radioactivity in drinking
water.
INTERIM REGULATIONS AND CONTROL MEASURES
A. OVERVIEW OF INTERIM REGULATIONS (NIPDWR)
On August 14, 1975, EPA proposed the interim primary
regulations for radioactivity in drinking water. The National
Interim Primary Drinking Water Regulations (NIPDWR) for
radionuclides were promulgated in their final form on July 9,
1976 (Federal Register, Vol. 41, No. 132, pgs. 28404-28409).
In 1979, a Variance and Exemption (V&E) report was prepared
for all regulated constituents in drinking water including
radionuclides (USEPA79). Also in 1979, suggested guidance
based primarily on health was provided for natural uranium
to the State of Colorado, Congressmen, Bureau of Indian
Affairs, other governmental bodies and private citizens upon
their request (La79). A more formalized Health Advisory for
uranium was submitted to EPA's Science Advisory Board (SAB)
in 1983.
-------
5
The interim regulations address natural and man-made
radioactivity as shown in the flow diagrams in Figures 1, 2
and 3 and described in more detail below. For natural radio-
activity, control measures were established for gross-alpha-
particle-activity, Radium-226 (Ra-226) and partially for
Radium-228 (Ra-228). Natural uranium and radon were specifically
and intentionally excluded.
GROSS-APLHA-PARTICLE-ACTIVITY, RA-226 AND RA-228
I. Maximum Contaminant Level (MCL)
A. Ra-226 and Ra-228 = 5 pCi/1 where detection limit =
1 pCi/1
B. Gross-alpha-particle-activity (including Ra but
excluding U and Rn) = 15 pCi/1 where detection limit =
3 pCi/1
II. Monitoring procedures (for all community water supplies -
ground and surface)
A. Initial sampling
1. Sampling initiated June 24, 1979
Analysis completed by June 24, 1980
2. Compliance based on analyses of an annual composite
of four consecutive quarterly samples or average
of analyses of four consecutive quarterly samples
3. Overall method of compliance:
a. Perform gross-alpha-particle-activity analysis.
b. If gross-alpha-particle-activity > 5 pCi/1,
perform Ra-226 analysis.
c. If Ra-226 > 3 pCi/1, perform Ra-228 analysis.
4. At discretion of State, data collected within one
year prior to the effective date of the regulations
(June 24, 1977) could have been substituted for
the "initial sampling".
-------
6
B. Follow-up sampling
1. Suppliers must repeat the complete initial sampling
process once every four years unless radioactivity
is < 1/2 MCL's whereby a single sample once every
four years is sufficient at the discretion of the
State.
2. Suppliers must monitor more frequently if contamination
is likely only when the State orders them to do so.
3. Suppliers must repeat the complete initial sampling
process again and within one year following (a)
introduction of new water source, (b) contamination,
(c) major change in distribution system, or (d) major
change in treatment processing.
4. Suppliers must monitor source water in addition to
free-flowing tap water if two or more sources
exist with significantly different levels of
radioactivity.
5. Suppliers need not monitor for Ra-228 if analyzed
in the initial sampling.
6. Suppliers must complete annual monitoring of any
water supply where Ra-226 > 3 pCi/1 or when ordered
by the State.
7. When the average annual MCL for Ra or gross alpha
particle activity is exceeded, supplier must
notify State and public. Quarterly sampling is
required until problem is resolved.
BETA AND GAMMA RADIOACTIVITY FROM MAN-MADE RADIONUCLIDES
I. MCL
A. Four mrem/year to total body or critical organ with a
variety of detection limits.
II. Monitoring procedures (for all surface water supplies
supplying 100,000 people or more or those designated by
the State as being impacted by a nuclear facility)
A. Initial sampling
1. Completed by June 24, 1979.
2. Compliance based on analysis of a composite concen-
tration of four consecutive quarterly samples or
an average of analyses of four quarterly samples.
-------
7
3. Overall method of compliance:
a. Perform gross beta analysis.
b. If gross-beta-particle-activity < 50 pCi/1,
then must check for 3h < 20,000 pCi/1 and
90gr < 8 pCi/1. If both 3h and §0sr are
present, sum annual dose equivalent to bone
marrow (see figure 2).
c. If gross-beta-particle-activity > 50 pCi/1,
then radiochemical analysis is required and
critical organ dose must be calculated. Dose
must not exceed four mrem/year.
4. Suppliers may be required by the State to conduct
additional monitoring to determine the concentration
of man-made radioactivity in principal watersheds.
5. Suppliers using only ground water may, at the
discretion of the State, be required to monitor
for man-made radioactivity.
6. Data collected within one year prior to the
effective date of the regulations (June 24, 1977)
may be substituted for the initial sampling.
B. Follow-up Sampling
1. Suppliers must monitor via initial sampling
techniques at least once every four years.
C. Special requirements for man-made radionuclide
contamination of drinking water by nuclear facilities
(see f igure 3).
1. Suppliers must initiate quarterly monitoring
(average of analyses of three monthly samples or
analysis of a three month composite sample) of
gross beta particle activity.
a. If gross beta particle activity > 15 pCi/1,
then analyze for 89sr and 134cs. If gross
Beta > 50 pCi/1, then conduct radiochemical
analysis and calculate total body or critical
organ dose. Dose must not exceed four mrem/year.
2. Suppliers must collect and analyze a composite of
five consecutive daily samples each quarter for
!31i. state may require more frequent monitoring
when 131l is present in finished water.
3. Suppliers must monitor annually for 90gr and
whereby four quarterly samples are analyzed either
as a composite or individually and averaged.
-------
8
4. State may allow substitution of environmental
surveillance data taken in conjunction with a
nuclear facility for direct monitoring by the
supplier, himself.
D. If average annual NCL for man-made radioactivity is
exceeded, supplier must notify State and public.
Sampling must then occur on a monthly basis until
problem is resolved.
In response to the proposed interim regulations of
August 14, 1975, five major issues surfaced and were considered
in promulgating the NIPDW regulations in 1976:
(1) The number and location of the public water systems
impacted by the proposed maximum contaminant levels for
radionuclides.
(2) The number and location of water supplies requiring radium
analysis at the proposed 2 pCi/liter gross-alpha-particle-
activity screening level.
(3) The estimated preliminary assessments of the costs and
technology for radium removal.
(4) The validity and appropriateness of an aggregate dose
method for setting maximum contaminant levels.
(5) The acceptability of a maximum contaminant level for
radium of 5 pCi/liter as opposed to a higher or lower level.
With respect to the number of public water systems
impacted, EPA estimated at that time that approximately 500
of the nation's community water systems would exceed the 5
pCi/1 MCL for radium. The proposed 2 pCi/1 gross-alpha-
particle-activity screening level to determine if analysis
for Ra-226 was needed was increased to 5 pCi/1 to avoid a
relatively large number of water supplies from conducting a
-------
9
rather expensive radium analysis. EPA did recommend but not
require, however, that in localities where Ra-228 may be
present in significant quantities, the State establish a
screening level no greater than 2 pCi/1. Treatment methods
including ion exchange, reverse osmosis and lime-softening
were recommended for consideration as possible methods for
lowering the concentration of radium in drinking water.
At that time, operating data from municipal water treatment
systems indicated that average radium removal efficiency via
the ion exchange cycle ranged from 93-97%. Concern was
expressed that operating personnel at some treatment plants
would be exposed to radiation. The EPA made a limited evalua-
tion of exposures to operating personnel working in the
vicinity of ion exchange units and determined that their
exposure levels could be in the range of 25-100 mrem/yr.
This was well below the Federal occupational guide for radiation
workers of 5,000 mrem/yr. Concern about inadequate waste
management practices also became an issue and EPA committed
itself to address this problem area by the time the revised
regulations were proposed.
EPA considered the question of whether or not a small
community would be required to adhere to the interim regulations
even if the aggregate dose for that particular community is
small. It was decided that the individual risk rates are
useful tools to protect public health in small communities,
and population risk values are useful in determining overall
national priorities in standard-setting. EPA also chose to
-------
10
assume a linear, non-threshold relationship between the
magnitude of the radiation dose received at environmental
levels of exposure and ill health. This policy was adopted
in conformity with the generally accepted prudent assumption
that there is some potential ill health attributable to any
exposure to ionizing radiation and that the magnitude of this
potential ill health is directly proportional to the magnitude
of the dose received. In adopting this general policy, EPA
recognized the inherent uncertainties, especially at low
radiation doses. Furthermore, EPA acknowledged that at
environmental levels it may well be impossible to statistically
prove via epidemiological studies that radiation in drinking
water causes cancer. The Agency nevertheless believed that
the policy was a prudent one.
The 5 pCi/1 total radium concentration was accepted as
the most appropriate level to protect public health considering
cost and feasibility. Using the National Bureau of Standards'
Handbook 69 (NBS63), it was calculated that if one consumed
two liters/day at 5 pCi Ra-226/1 over a lifetime, the radiation
dose to the bone would be approximately 150 mrem/yr. Further-
more, using the BEIR I report (NAS72), an excess cancer risk
rate of 100 cases/10® people exposed/lifetime was estimated.
It should be remembered, however, that animal studies (Do69)
showed that Ra-228 has a measured biological effectiveness
over twice as great as that of Ra-226, so that the calculated
excess cancer risk rates for Ra-226 may well have underestimated
the total risk of 5 pCi/1 of Ra-226 and Ra-228 combined.
-------
11
B. CONTROL MEASURES
The variance and exemption (V&E) levels for man-made
radioactivity in drinking water were prepared in 1979 and
are essentially the same guidances as that finalized on
April 5, 1979, following the Three Mile Island (TMI) incident.
It is unlikely that variances nor exemptions will be necessary
because of the remoteness of an accident. Even then only
variances for tritium (3h) should be granted. Following the
TMI accident, two offices within EPA, namely the ODW and the
Office of Radiation Programs (ORP), jointly issued "Drinking
Water Alert Levels":
I. Screening Level
Gross beta with iodine precipitated 40 pCi/1
Gross beta without iodine precipitated 100 pCi/1
(separate radioiodine tests may be needed)
II. Alert Levels
50 mrem/year
(12 times EPA 4 mrem/year standard*)
10 mrem dose commitment for any one day
*(See National Interim Primary Drinking Water Regulations,
USEPA76)
Current Federal Guidance for transient rate of intake
provides limitations on food and water intake that are compar-
able to an annual dose equivalent of 50 mrem/year and contain
a recommendation that for transient situations the dose
should be averaged over one year (26 Federal Register 9057,
1961). IT IS HEREIN RECOMMENDED THAT THIS ANNUAL DOSE RATE,
50 MREM/YEAR TO ANY ORGAN, BE USED AS AN ALERT LEVEL FOR
RADIOACTIVITY IN FINISHED DRINKING WATER.
-------
12
For continuous intake, concentrations of a single
man-made radionuclide yielding 50 mrem/year to any specific
organ can be found by multiplying by 12 the concentration
listed in reference USEPA76 or USEPA81. When more than
one radionuclide is present, instructions for limiting their
sums are given in reference USEPA7 6.
For continuous intake, an annual dose rate of 50 mrem
results from each of the following concentrations.
For drinking waters having concentrations of radionuclides
above the alert level, a determination on use should depend
on radiochemical identification. IT IS FURTHER RECOMMENDED
THAT THE INDIVIDUAL DOSE COMMITMENT FROM ONE DAY'S INTAKE
(2 LITER) SHOULD NOT EXCEED 10 MILLIREM. A 10 mrem dose
commitment results from the following concentrations:*
Proper use of the alert level should keep annual doses
below 50 mrem/year and is applicable to a clean-up operation
lasting as long as one year.
In the case of natural radioactivity, EPA used the
guidance provided the President in 1961 which was specified
by the Federal Radiation Council (FRC) for transient intakes
(FRC61). For Ra-226 and Ra-228 the V&E level is 10 pCi/1.
This level was chosen since the upper limit of the FRC Range
~Concentrations causing a 10 millirem dose commitment are
900 times those given in Appendix B of NIPDWR (USEPA76).
Strontium-90
Strontium-89
Cesium-137
Iodine-131
96 pCi/1
960 pCi/1
2400 pCi/1
36 pCi/1
Strontium-90
Strontium-89
Cesium-137
Iodine-131
7,200 pCi/1
72,000 pCi/1
180,000 pCi/1
2,700 pCi/1
-------
13
II guide was 20 pCi/day via food and water. Above this
range, evaluation and application of additional control
measures is always necessary (26 FR 9057, 1961). Provided
that a comparable intake of radium via the food pathway is
unlikely, exemptions for water supplies containing less than
10 pCi/1 or a dose equivalent of 300 mrem/yr would be compatible
with FRC Guides. Occassionally, exemptions for concentrations
exceeding 10 pCi/1 for strictly limited times, may be acceptable.
Exemption, but not variance, is provided for gross alpha
up to 30 pCi/1 using the same justification as provided for
radium.
In granting V&E schedules for compliance, the following areas
are considered: the source of exposure, extent to which the
MCL is exceeded, type of radionuclide and amount present,
number of people exposed, duration of exposure, both past
and future if no remedial action should occur, other
sources of exposure, and other sources of water. When
treatment methods are available, compliance schedules should
encourage early installation of treatment processes or
encourage the water supplier to find and use an alternate
source of water.
Guidance concerning the health effects of uranium in
drinking water was developed after receipt of requests from
the State of Coloradok in 1979, and several other governmental
bodies for health advice on uranium in drinking water. A
suggested guidance of 10 pCi/1 or 20 pCi/day of natural
uranium was proposed (La79).
-------
14
The suggested guidance for uranium was calculated using
the present interim standard for radium as a comparable
level. The radium level was used for comparative purposes
because (1) no acceptable uranium chemical toxicity data were
available, (2) a lifetime number as opposed to a 1-day,
10-day or longer-term transient exposure would be needed
since uranium is a natural, not a man-made constituent and
lifetime exposures are commonplace, (3) the radiotoxicity of
radium is rather well known, and (4) both radium and uranium
are bone seekers with similar distribution patterns in bone.
In order to calculate the suggested guidance for uranium,
the concentration of natural uranium, consumed over a lifetime
in two liters of drinking water/day which would result in
approximately the same dose and risk to the bone and bone
marrow as radium was determined. Using ICRP # 2 (ICRP 59)
and a composite risk estimation (relative and absolute)
established by the Office of Radiation Programs, USEPA, it
was estimated that 5 pCi Ra-226/1 results in a 150 mrem/yr
bone dose which could cause an estimated 0.7-3 bone and other
cancers per million people exposed per year which is equivalent
to an excess cancer risk rate of about 100 cases per million
people exposed (10~^ risk rate) per lifetime. It was deter-
mined that exposure to 10 ,pCi natural uranium/1 resulted
essentially in the-same dose and risk of cancer; thus, the
suggested guidance of 10 pCi/1.
Since that time, the Office of Drinking Water has reviewed
its approach and estimates of risk using a modified ICRP 30
-------
15
I
model (Su81). Assuming uptake values of fi = 0.2 and f2 =
0.2 as suggested by ICRP 30 (ICRP 79), a quality factor of
20 for high-LET alpha radiation and assuming that the activity
of U-238 and U-234 are equal, an endosteal bone and red bone
marrow dose equivalent of 100 mrem/yr was calculated for the
lifetime consumption of 10 pCi Unat/1 at an ingestion rate of
2 1/day (Co83a). Using EPA's linear hypothesis, this dose
rate might cause 34 excess cases/million people exposed/lifetime
(Table 1). These calculations of dose equivalents and risk
rates for uranium remain essentially equivalent to those
recently calculated for radium, namely 92 mrem at the 70th
year and a cancer risk rate of 44 excess cases/lifetime/million
people exposed to 5 pCi Ra-226/1 at 2 1/day (Table 2). Thus,
the comparison of 5 pCi Ra/1 and 10 pCi U/l in terms of dose
equivalency and cancer rates remain essentially the same.
The National Academy of Sciences (NAS8 3) has recommended
a chronic Suggested-No-Adverse-Response-Level (SNARL) for
uranium based only on chemical toxicity. They assumed a
minimum-observed-effect-level of 1 mg/kg/day, an uncertainty
factor of 100 and that a 70 kg adult consumes 2 1/day of water
which provides 10% of the daily uranium intake. They calculate
the chronic SNARL as follows:
1 mg/kg x 70 kg x 0.1 = 35 micrograms/1 iter
100 x 2 liters
At equilibrium they observe that this would be equivalent to
11.6 pCi U/l. The EPA does not necessarily endorse this
calculation.
-------
16
REVISED REGULATIONS
Under the SDWA, one of the significant steps in the
regulatory process is the promulgation of the Revised
National Primary Drinking Water Regulations (NPDWR). These
revised regulations will contain several parts including an
MCL based on possible adverse health effects and feasible
treatment technology while taking cost into consideration.
These regulations shall be amended whenever changes in
technology, treatment techniques and other means permit
greater protection of the health of persons, but in any
event such regulations will be reviewed at least once every
three years.
It should not be surprising if the revised drinking
water regulations for radioactivity differ from the interim
regulations. This "open-minded" approach at this juncture
would seem warranted, not because the interim regulations
are good or bad, too strong or too weak, but because;
(1) our understanding of the occurrence of Ra-226, gross-
alpha-particle-activity, and uranium in drinking water
has been substantially enhanced,
(2) our concern for radon has increased because of its presence
and its quite significant population risk, and
(3) the nuclear industry's ability, responsibility and effort
to properly build, operate, decommission/store nuclear
facilities, devices and/or their by-products is becoming
more evident.
-------
17
The ODW, EPA, via this workshop activity identified
problem areas that need resolution prior to the completion
of our criteria for development of the revised regulations.
What follows is a list of some of the issues and areas of
concern to EPA. The issues discussed here are divided into
the five areas of exposure, dose, health effects, engineering
and economics and policy.
A. MONITORING AND SOURCES OF EXPOSURE
In order to develop regulations for radioactivity in
drinking water, ODW must determine the sources of exposure
and body burdens from natural and man-made radionuclides in
drinking water, food and air as well as the concurrent exposure
in the workplace. Exposure to other contaminants that would
be expected to result in similar or other mechanisms of stress
should also be identified, quantified or at least estimated.
Sampling and analytical methods are needed to adequately
monitor the occurrence of radionuclides in drinking water.
This area may seem straight-forward and complete but that
is not necessarily the case. Several issues need exploration
including (1) use and abuse of "screen" monitoring,
(2) frequency of monitoring for select radionuclides,
(3) need for inclusion of additional analytical techniques,
(4) identification of the appropriate detection limits, (5)
impact of high dissolved solids on alpha counting, (6) use
or misuse of the fluorometric method for analyzing uranium,
and (7) the design of our Laboratory Certification program.
-------
18
Naturally occurring radionuclides like U-238, U-234,
Ra-226, Rn-222f Pb-210 and Po-210 and daughter products are
part of The Uranium Series. Th-232, Ra-228, Th-228, Ra-224,
Rn-220 and daughters are members of The Thorium Series and
also contribute to the body burden. Another set of naturally
occurring radionuclides is called The Actinium Series, but
this Series appears to be less significant in our discussions
here. Many of these important radionuclides exist in drinking
water primarily via ground as opposed to surface water sources.
The abundance of isotopes in natural uranium is such
that U-238, U-235 and U-234 are present 99.27, 0.72 and
0.006%, respectively. The mass ratio of U-234 to U-238 is
small but at equilibrium their activities are equal. The
hexavalent state is particularly important in water because
almost all tetravalent compounds are practically insoluble.
Using a recent ORNL Report (Dr81) and the paper by
Cothern and Lappenbusch (Co83b) entitled "Occurrence of
Uranium in Drinking Water in the U.S.", EPA estimates that:
(1) Average natural uranium concentration in surface, ground
and domestic water is 1, 3 and 2 pCi/1, respectively.
Domestic sources are those that could be used for drinking
water purposes, but are not necessarily now in use.
(2) The highest population weighted averages are geographically
lying between the States of Montana and Texas, and California
and Kansas. South Dakota, Nevada and New Mexico have
average values of 6.7, 4.3 and 2.9, respectively. The
-------
19
States of California, Wyoming, Texas, Arizona and Oklahoma
all seem to have values between 2.5 and 2.7.
(3) The population weighted average for the U.S. is 0.8 pCi/1.
(4) The U-234/U-238 ratio for most water supplies seems to be
between 1 and 3, with surface waters in the U.S. being
closer to equilibrium than that of ground water. The
ratio can range up to 10-15.
(5) Of the 60,000 some community water supplies, 25-650 would
exceed a uranium concentration of 20 pCi/1, 100-2,000
would exceed 10 pCi/1 and 2,500-5,000 would exceed 5 pCi/1.
This estimate is based on those supplies designated as
domestic supplies.
(6) The average U.S. citizen exposed to uranium ingests 1,460
pCi/year (85%) via drinking water and 240 pCi/year (15%)
via food. The concentration of uranium in the atmosphere
is responsible for only a minor part of one's intake.
Using the compliance monitoring data specified in the
interim regulations as a reference point, it appears that
some 500 community drinking water supplies exceed the 5 pCi
Ra/1. This estimate is based primarily on Ra-226 data rather
than Ra-228 data since the interim regulations did not require
monitoring for Ra-228 unless Ra-226 levels exceeded 3 pCi/1.
About 170 public water supplies are known to have Ra-226
concentrations greater than 5 pCi/1 and about 350 more are
expected from those that have not reported yet (Co83c).
Problem areas include Illinois, Iowa, North Carolina, South
Carolina, Georgia and a few other states. It should also be
-------
20
pointed out that many homes use private wells which are not
subject to NIPDWR for their water supply and a significant
number contain Ra-226. Thus, the human exposure to radium
in drinking water will be larger than that indicated here.
Human exposure to radon comes from the earth's crust
and is in our ambient outdoor environment. It may become
concentrated in homes. One vehicle for transport into the
house is via the water distribution system. The drinking
water concentrations in the U.S. range from a few pCi/1 to a
few 100,000 pCi/1 in the Northeast, with the average somewhere
between 500-5,000 pCi/1. In areas where radon is high,
private wells probably deliver more radon into the household
than do community water supplies.
Locations of the larger deposits of thorium in the United
States has been published by Staatz, et al. in 1979 (St79)
and the U.S. Geological Survey appears to be studying its
presence in the U.S. Significant deposits have been found
in Montana, Idaho, Colorado, Wyoming, California, and Alaska
in the West; Illinois and Wisconsin in the Midwest; and
North Carolina, South Carolina, Florida and Georgia in the
East. However, thorium is highly insoluble in water and
its occurrence is thus limited.
Our occurrence data base for Ra-228, the U-234/U-238
ratio, Rn-220 and Rn-222 and their daughters Th, Pb-210 and
Po-210 is insufficient. There seems to be two methods ODW
can use to fill in the data gaps: (1) more monitoring or
-------
21
(2) predictive modelling. Perhaps strategic monitoring to
support and verify predictive models would enable us to
successfully estimate the amount of one or more members of
the Uranium Series or the Thorium Series when our understanding
of one or more other members of each series is known via
existing data. For example, if we know how much U-238 is
present, could we predict how much Ra-226 and/or Rn-222
should be there as well? This type of predictive modelling
will require a substantial understanding of geochemistry,
aquifers, etc. Furthermore, could we predict qualitatively
and quantitatively Ra-228 in water supplies? The reward for
accurate predictive modelling could be high and could eliminate
substantial monitoring costs for those water suppliers not
predicted to have a particular radionuclide(s).
ODW needs to determine the relative source contribution
from food, air and water for radium, uranium and radon. The
indoor exposure of radon via the water distribution system
needs delineation. Occupational exposures must also be
taken into consideration. Good occurrence data or estimates
will be required to predict population risk where population
risk equals the product of the population concentration times
the risk/concentration or better yet the population dose
times the risk/dose. Population risk estimations are useful
in setting priorities for standard-setting but are not necessarily
useful in determining whether a small community should undertake
remedial action because of a contaminated water supply.
-------
22
B. DOSE EVALUATION
An examination of the pharmacokinetics and dosimetry of
both natural and man-made radionuclides is required because
of (1) the introduction of ICRP Report No. 30 (ICRP79) in
place of ICRP Report # 2 (ICRP59), (2) both uranium and
radon (and their daughters) are under consideration for
regulation, (3) concentrations associated with a specific
dose rate (i.e., 4 mrem/yr) must be established for the
man-made beta and photon emitters, (4) variance and exemption
levels must be proposed for radionuclides in drinking water,
and (5) protective action levels (PAGs) will ultimately be
required, if not sooner, should an uncontrollable situation
arise unexpectedly.
Several points need clarification and delineation:
/
(1) fi (gut to blood) and f2 (blood to bone) for Ra-226 vs.
Ra-228 (and daughters)
(2) fi (gut to blood) and f2 (blood to kidney and bone) for
U-238 vs. U-234
(3) differential deposition of uranium and radium in the bone
(4) impact of nutrition on uranium and radium deposition
(5) differential retention and biological half-life of uranium
and radium in critical organs (bone and bone marrow)
(6) differential dose rate of uranium and radium to the bone
and bone marrow
(7) different pharmacokinetic patterns and dose estimation
following experimental vs. environmental exposures
-------
23
(8) appropriateness of extrapolating animal pharmacokinetics
data to humans
(9) intake, uptake, distribution, retention and metabolism of
radon and its daughters via ingestion, inhalation and
their combination
(10) impact of other sources (house air, environmental air,
occupational setting) upon the total committed dose via
drinking water
(11) relationship between exposure and organ dose and dose
rate as complicated by age
(12) impact of tobacco smoking on the lung dose via radon
daughters
C. HEALTH EFFECTS
In order to properly identify the RMCL, the adjusted
Acceptable Daily Intake (ADI) for non-carcinogen bioeffects
and non-threshold carcinogenic effects, the health effects
data base must be reexamined. The extrapolation process
must be reviewed and the major issues associated with making
these scientific judgments must be discussed. Points for
consideration/discussion include:
(1) Identification and validation of the linear dose response
curves for natural radioisotopes (i.e., Ra-226, Ra-228,
U-238, U-234, radon and its daughters), man-made beta
and photon emitters
(2) Comparative toxicities (Ra-226 vs. Ra-228, U-238 vs. U-234,
radium vs. uranium)
-------
24
(3) Use or abuse of the threshold and "practical" threshold
concepts
(4) Suitability of the BEIR III (NAS80) report for use in
determining the health effects of radionuclides in water
(5) Chemical toxicity of uranium vs. its radiotoxicity
(6) Calculation of an adjusted ADI -for natural uranium using
non-carcinogenic data
(7) Validity of safety factors
(8) Use of NAS1 toxicological analysis of uranium
(9) Recognition and calculation of the impact of multiple
stressors
At this point, ODW favors the following positions on
those issues stated above; however, ODW is open to constructive
suggestions towards possible modification:
(1) For standard-setting purposes, the linear, non-threshold
dose/response curve is most prudent
(2) Ra-228 is 2-3 times more radiotoxic than Ra-226. U-238
and U-234 are roughly equivalent in toxicity. Uranium
is one-half as toxic as radium
(3) The threshold concept is not prudent for calculating
excess cancer risk rates, especially for genotoxic,
human carcinogens. The "practical" threshold concept
has not been proven or shown to be a viable concept
(4) BEIR III is a useful guide for appreciating the health,
risks of radioactivity
-------
25
(5) Chemical and radiotoxicity of uranium may be of the
same order of magnitude numerically but the organ systems
affected are much different
(6) An Adjusted ADI cannot be calculated for uranium, based
on chemical toxicity data until better dose/response data
become available
(7) Use of NAS' safety factors for non-carcinogenic data
would be appropriate; however, consideration must also be
given to bioaccumulation, experimental design and multiple
stress interaction
(8) NAS' estimate of the toxicity of natural uranium should be
reconsidered
(9) When more than one carcinogen is present, additivity must
be assumed especially if the same organ system is impacted
directly or indirectly
In regards to multiple stress toxicology and environmental
standards/health advisories, two basic issues need attention
(Fig 4).
(1) Dose/response studies need to be designed multifactorially
c
and must account for chemical, biological and physical
stressors; all acting concurrently prior to, during and
after the experimental exposure or insult
(2) Health effects guidance and the control of multiple
stressors (inorganics, organics or radionuclides) in
drinking water should appreciate the concepts of additivity,
synergism and antagonism
-------
26
Co-insult studies should be preceeded by a detailed
appreciation of "baseline stress", where "baseline stress"
means the normal quantifiable and transitory changes that
occur in an animal who has been subjected only to necessary
environmental stimuli. "Baseline stress" takes into account
the normal quantifiable changes in blood hormone levels, WBC
and RBC density in the peripheral blood, or mitotic indices,
weight of an organ, or any quantifiable parameter which may
change as a result of the time of day, sex or age, etc. An
understanding of fluctuation in measurable parameters in
the experimental animal is imperative to obtaining good data
(La77). Also significant is gaining an understanding of
corresponding variability of similar parameters for humans.
To illustrate the point (Fig. 5), male Chinese hamsters
were subjected to various x-ray doses at five different times
of the day or night (La72). In this study, LD50/30 values
for animals exposed at 9:00 AM, 1:00 PM, and 3:00 PM were
essentially identical while LD50/30 11:00 AM and 1:00 AM
were found to be significantly different: that is, 823 rads
and 954 rads, respectively. This paper also reported that
there is a true linear correlation between LD50/30 values
and mitotic indexes and that radiosensitivity for the Chinese
hamster is dependent upon the time of the day. This study
clearly pointed out the importance of circadian rhythm on
radiosensitivity. The effect of circadian rhythm on radio-
sensitivity is not unique to terrestrial organisms. Some
-------
aquatic animals including the rough-skinned newt (Taricha
granulosa) are circadium rhythm dependent as far as radio-
sensitivity is concerned (La70).
D. ENGINEERING AND ECONOMICS
Lime softening, ion exchange and reverse osmosis have
been demonstrated as methods for removing radium from drinking
water (USEPA77). Although EPA is currently conducting
laboratory and field studies of methods for removing uranium
from drinking water, these studies are incomplete. Methods
being investigated for uranium removal include anion exchange,
lime softening, and reverse osmosis. It thus remains to be
seen if a method for uranium treatment is both available and
can be operated at a reasonable cost. Two methods appear
possible for the removal of radon from drinking water; viz,
aeration and adsorption by granular activated carbon. More
information is needed on the feasibility and cost of these
methods. Possible methods for removing man-made radionuclides
from drinking water are mixed bed ion exchange resins and
reverse osmosis. The feasibility and costs involved in these
and other possible methods need to be determined.
Removal of radioactivity from drinking water produces
waste in the form of brine from ion exchange, lime softening
sludge and the reject stream from reverse osmosis. Some
issues that relate to these and other potential wastes are:
(1) Is disposal of such wastes already prohibited by some States
(2) How is the waste to be classified - high level, low level,
source material, other?
-------
28
(3) How is this problem related to the Resource Conservation
and Recovery Act (RCRA) and Superfund regulations and
activities?
(4) What criteria should be used to protect the public?
(5) What criteria should be used to protect workers at the
treatment plant?
(6) How should the radioactive waste be disposed of?
(7) What costs are incurred in waste management practices
involved in the treatment of drinking water for radioactivity?
E. GENERAL DEVELOPMENT
The ODW would encourage input in the area of interpretive
issues for the control of radioactivity in drinking water.
Issues of particular importance include:
(1) Should Health Advisories be developed and if so, under
what conditions?
(2) How can predictive modelling which would estimate that
radionuclides would not be expected in geographic
portions of the U.S. be used in implementation and
design of regulations?
(3) What factor should be used to determine appropriate
monitoring frequencies?
(4) Since uranium is both chemical and radiotoxic, how does
one quantifiably determine an RMCL?
(5) Should ODW assume linearity between dose and effect?
(6) Does the concept of de minimus risk have application in
the design of drinking water regulations as developed
under the Safe Drinking Water Act?
-------
(7) Should the health basis of standards be based on individual
risk, population risk, or both?
(8) Should natural radionuclides be regulated individually
and/or via the umbrella concept as man-made radionuclides
presently are controlled in the NIPDWR of 1976?
(9) If a composite RMCL approach is acceptable, how should
the question of synergism be addressed?
(10) Since Congress instructed EPA to take other sources of
exposure into account when setting drinking water RMCL
standards, should EPA set aside a specific exposure
allotment for drinking water after weighing, of course
the dose from air, food and the occupational setting?
(11) Should exposure of treatment plant operators be a factor
in the setting of standards for radioactivity in drinking
water?
(12) How should the drinking water standards consider waste
management problems?
(13) How should public notices be presented in view of the
public's perception of radioactivity?
(14) What are the training needs at national, state and local
levels?
(15) What elements and input should be considered in analyzing
the risks of radioactivity in drinking water and determining
what an acceptable risk is? (see the following discussion).
The estimation of risk for drinking water contaminants
requires the knowledge of the occurrence, the population
-------
30
exposed and the individual risk. Figure 6 shows how this
information inputs to the risk estimation process.
In general, there are limits to knowledge about any
subject and the current state of knowledge about the health
effects of contaminants in drinking water has its limits.
Since it is not possible to estimate risk with 100% accuracy,
it is often expressed as a range of values.
Some of the complexities involved in determining an
acceptable risk rate level are whether the risk involved is:
voluntary or involuntary, ordinary or catastrophic, natural
or man-made, immediate or delayed, continuous or occasional,
controllable or uncontrollable, old or new, clear or unclear,
necessary or a luxury, temporary or permanent, fatal or
debilitating, curable or uncurable, equitable or unequitable,
reversible or irreversible, long or short biologic half life.
The way in which the public perceives all of these variables
contributes to the complexity of choosing an acceptable risk
rate level. Although we are generally a risk aversive nation,
there is a wide variety of attitudes, responses and value
judgments to these complexities.
Several Federal agencies and offices have determined
risk rate levels to be used for standard setting purposes.
The EPA's Office of Water Regulations and Standards considers
risks in the range of 10"""' to 10"^ cases per lifetime as
a target. The U.S. Supreme Court in its recent decision
involving benzene (July 2, 1980) stated that "if the odds are
one in a thousand ... a reasonable person might well consider
-------
31
the risk significant and take appropriate steps to decrease
or eliminate it." In the same decision, they also stated
that "If, for example, the odds are one in a billion that a
person will die from cancer by taking a drink of chlorinated
water, the risk clearly could not be considered significant."
These observations have been taken from dicta and thus are
out of context, but they do give an idea of the possible
bounds. The Food and Drug Administration considers that It)"6
cases/lifetime a "virtually safe" level.
Only a few federal standards allow lifetime risk levels
as high as 10"^ cases per lifetime, according to conservative
high dose/low dose extrapolation models. The MCL for trihalo-
methanes in drinking water and the vinyl chloride air standard
are among those that do. When the radium MCL was promulgated,
it was thought to be of that same order. After careful
examination of recent model calculations, however, the excess
risk of about 0.4 x 10~4 cases/lifetime may be more accurate.
In view of the uncertainties in scientific data and the
variety of value judgments involved in evaluating risks, it
appears that there is no single, universal and systematic
method for determining acceptable risk. From the above
discussion, standards have been set by government that result
in individual risks in the range of 10~4 to 10~® cases/lifetime.
For humans the most acceptable risk rate level is zero. But
we do not live in a risk free society and thus it is possible
to estimate a number that could be considered in the standard
setting process to protect humans from unnecessary risk.
-------
32
The above discussion concerns the individual risk rate
level. Although it is important to develop such a value for
the individual it is important to know the risk that the
entire population is exposed to, commonly known as population
risk. The population risk can be calculated by multiplying
the population dose/concentration by the risk per dose/concen-
tration .
DISCUSSION AND CONCLUSION
From the list of issues in the areas involved in the
development of the Revised Regulations for Radioactivity in
Drinking Water, it should be clear that they are many, varied
and interrelated. It should be realized that the issues
described here are those that relate to the scientific,
technological areas with some discussion of the cost involved.
In the process of regulatory development, inputs from several
other areas will be needed. Some of these additional areas
include: more economic input including cost/benefit analysis,
psychological analysis of perception, fear and other emotionally
related phenomena, political aspects on the Federal, State
and local levels, social and moral aspects and others. Thus
the information provided by this national workshop on Radio-
activity in Drinking Water is only part of that needed to
develop and promulgate the Revised Regulations for Radioactivity
in Drinking Water.
-------
33
References
Co83a Cothern C.R., Lappenbusch W.L. and Cotruvo J.A., to
be published, "Health Effects Guidance for Uranium in
Drinking Water", Health Physics, 44, 377-384.
Co83b Cothern C.R., and Lappenbusch W.L., 1983, "Occurrence
of Uranium in Drinking Water in the U.S.", Health
Physics, 45, 88-99.
Co83c Cothern C.R., and Lappenbusch W.L., to be published in
Health Physics, "Compliance Data for the Occurrence
of Radium and Gross Alpha Particle Activity in
Drinking Water Supplies in the United States", 361-367.
Do69 Dougherty T.F. and Mays C.W., "Bone Cancer Induced
by Internally Deposited Emitters in Beagles,"
Radiation Induced Cancer, IAEA-SM-118/3, International
Atomic Energy, Vienna, Austria.
Dr81 Drury J.S., Reynolds S., Owen P.T., Ross R.H. and
Ensminger J.T., EPA 570/9-81-001, EPA, Report by
Health and Environmental Studies Program, Information
Center Complex, Oak Ridge National Laboratory, Oak
Ridge, TN, 37830, January 1981. Available from NTIS,
U.S. Department of Commerce, Springfield, VA, 22161.
The first volume is a summary and is identified by
the above number. The other three volumes of the
1980 page report list the data by the location and
are identified by ORNL/EIS-192.
FRC 61 Federal Radiation Council, 1961, "Background Material
for the Development of Radiation Protection Standards,"
Report No. 2, Superintendent of Documents, U.S.
Government Printing Office, Washington, D.C.
ICRP 59 International Commission for Radiation Protection,
1959, "Report of Committee II on Permissible Dose
for Internal Radiation," ICRP publication 2 (New
York, NY, Pergamon Press).
ICRP 79 International Commission for Radiological Protection,
1979, "Limits for Intakes of Radionuclides by
Workers," Annals of the ICRP, ICRP Publication 30,
Volume 2, No. 3/4 (Oxford: Pergamon Press).
La70 Lappenbusch W.L., Effect of Circadian Rhythm on the
Radiosensitivity of the Rough-Skinned Newt (Taricha
granulosa). J. Rad. Res. 2(3-4):134-137.
La72 Lappenbusch W.L., Effect of Circadian Rhythm on the
Radiation Response of the Chinese Hamster (Cricetulus
griseus). Radia. Res. 50( 3):600-610.
-------
34
La77 Lappenbusch W.L., 1972, Review of Research Papers
and Salient Points to Consider when Conducting
Radiation and Multiple Stress Studies, AIHA Meeting,
New Orleans, LA.
La79 Lappenbusch W.L., 1979, Personal Communication,
letter to Mr. Frank Rozich, Director, Water Quality
Control Division, Colorado Department of Health,
4210 East 11th Ave., Denver, Colorado, 80220.
LOC 82 Library of Congress, 1982, "A Legislative History of
the Safe Drinking Water Act, Serial No. 97-9, Prepared
by the Environment and Natural Resources Policy
Division of the Congressional Research Service,
Washington, D.C.
NAS 72 National Academy of Sciences, 1972, "The Effects on
Populations of Exposure to Low Levels of Ionizing
Radiation", Report of the Advisory Committee on the
Biological Effects of Ionizing Radiation (often
called BEIR I), Washington, D.C. 20006.
NAS 80 National Academy of Sciences, 1980, "The Effects on
Populations of Exposure to Low Levels of Ionizing
Radiation; Report of the Advisory Committee on
Biological Effects of Ionizing Radiation", BEIR III,
Washington, D.C.
NAS 83 National Academy of Sciences, 1983, Drinking Water
and Health, Vol. 5, pgs. 147-157, Washington, D.C.
20006.
NBS 63 National Bureau of Standards, 1963, "Maximum
Permissible Body Burdens and Maximum Permissible
Concentrations of Radionuclides in Air and Water for
Occupational Exposure", Handbook 69, U.S. Department
of Commerce, Washington, D.C.
St79 Staatz M.H., Armbrustmacher T.J., Olson J.C.,
Brownfield I.K., Brock M.R., Lemons J.F. Jr., Coppa
L.V. and Clingan B.V., 1979, "Principal Thorium
Resources in the United States", Geological Survey
Circular 805, U.S. Geological Survey, 1200 South
Eads Street, Arlington, VA 22202.
Su81 Sullivan R.E., Dunning D.F. Jr., Nelson N.S., Ellett
W.H., Leggett R.W., Yalcintas M.G. and Eckerman K.F.,
1981, "Estimates of Health Risk from Exposure to
Radioactive Pollutants", Oak Ridge National Laboratory
Report ORNL/TM-7745, Oak Ridge, Tennessee 37830.
-------
35
USEPA 76 U.S. Environmental Protection Agency, 1976, "National
Interim Primary Drinking Water Regulations", Office
of Water Supply, Report EPA-570/9-76-003, Washington,
D.C. 20460.
USEPA 77 U.S. Environmental Protection Agency, 1977, "Manual
of Treatment Techniques for Meeting the Interim
Primary Drinking Water Regulations", Water Supply
Research Division, Municipal Environmental Research
Laboratory, EPA-600/8-77-005, Cincinnati, Ohio 45268.
USEPA 79 U.S. Environmental Protection Agency, 1979, "Guidance
for the Issuance of Variances and Exemptions", Office
of Drinking Water (WH-550), Washington, D.C.
USEPA 81 U.S. Environmental Protection Agency, 1981, "Radio-
activity in Drinking Water", Office of Drinking
Water (WH-550), Report EPA-570/9-81-002, Washington,
D.C. 20460.
-------
Figure
1
2
3
4
5
36
Figure Captions
Flow chart for gross-alpha-particle-activity
monitoring (U.S. EPA, Las Vegas, Environmental
Monitoring and Support Laboratory). Mote that
the interim regulations do not require that
radon and uranium be measured if the gross
alpha activity is greater than 15 pCi/1.
Flow chart for gross-beta-particle-activity
monitoring for a water source not designated
as being contaminated by effluents from nuclear
facilities serving more than 100,000 persons as
designated by the State.' (U.S. EPA, Las Vegas,
Environmental Monitoring and Support Laboratory)
Flow chart for monitoring drinking water samples
near a nuclear facility. (U.S. EPA, Las Vegas,
Environmental Monitoring and Support Laboratory)
Interrelationship between factors in a multiple
stress situation.
Survival of 80-day-old, young adult, male
Chinese hamsters, weighing 20-26 g each, after
exposure to Various x-ray doses at five different
times of the day or night. Percentage survival
was calculated using data through Day 30.
Sample size ranged from 10 to 30 animals point,
(reproduced from La72)
Flow chart showing the input of information for
the risk estimation process.
-------
NO
NO
YES
YES
NO
NO
YES
YES
NO
YES
COMPLIANCE
NON-COMPLIANCE
MEASURE
Ra-226
MEASURE
Ra-228
IS Ra-226
PLUS Ra-228
> 5pCi/l
IS Ra-226
> 3pCi/l
MEASURE
RADON &
URANIUM
IS ALPHA
> 5pCi/l
IS ALPHA
> 15pCi/l
MEASURE
GROSS ALPHA
Is ALPHA
MINUS
RADON &
URANIUM
ALPHA
> 15 p C i /1
-------
ANALYZE
TO IDENTIFY
RADIONUCLIDES.
DETERMINE
COMPLIANCE
ANNUAL DOSE
YES FROM
RADIONUCLIDES
FOUND IS
> 4 mrem/yr
MEASURE
GROSS BETA
I
YES
IS BETA
> SOpCI/l
NO
MEASURE
TRITIUM AND
Sr-90
NO
COMPLIANCE
IS
TRITIUM
> 20,000 pCi/l
f NO
IS
Sr-90
> 8pCi/l
f no"
ANNUAL
DOSE FROM
TRITIUM Sr-90
IS > 4 mrem/yr.
YES
YES
YES
NO
NON-COMPLIANCE
Figure 2
-------
MONITOR QUARTERLY
MONITOR
ANNUALLY
YES
r
MEASURE
GROSS BETA
~
Is BETA
> 50pCI/l
ANALYZE
TO IDENTIFY
RADIONUCLIDES,
DETERMINE
COMPLIANCE
WITH 141.16
NOi Is BETA
> 15pCi/l
J j YES
ANALYZE
FOR
Sr-89, Cs 134
*
ANNUAL
DOSE FROM
YES RADIONUCLIDES
FOUND
IS
> 4 mrem/yr
NO
NO
COMPLIANCE
COMPLIANCE
Is Cs-134
> 80pCI/l
MEASURE
Is 1-131
>3pCI/l
Is Sr-89
80pCI/l
YES
ANNUAL
DOSE FROM
Sr-89-Cs-134
> 4 mrem/yr
YES
YES NO
COMPLIANCE!
NON-
COMPLIANCE
MEASURE
TRITIUM
AND Sr-90
T
Is TRITIUM
> 20,000
pCI/l
t NO
Is
Sr-90
> 8 pCI/l
j NO
YES
YES
ANNUAL
DOSE FROM
Sr-90 H-3
>4 mrem/yr
f NO ""
COMPLIANCE
NON-
COMPLIANCE
YES
-------
BIOLOGICAL
STRESSORS
CHEMICAL
STRESSORS
V
Absorption
Adsorption
Ingestion
Inhalation
Injection
Body
Burden
PHYSICAL
STRESSORS
~
Diagnosis
treatment
Additive, Synergistic
or
Antagonistic Bioeffects
(Acute or Chronic)
-------
100
90
80
70
60
50
40
30
20
10
0
o 9 A.M
~ 11 A.M
1 P.M
3 P.M
1 A.M
9
800
850
900
950
1000
1050
RADIATION DOSE (RADS)
-------
POPULATION ^
OCCURRENCE
POPULATION
CONCENTRATION
INDIVIDUAL
RISK RATE
POPULATION
RISK
NUMBER OF LIVES
SAVED BY
ELIMINATING
CONCENTRATIONS
THAT LEAD TO A
RISK OF 10-4 10-5
AND 10*® PER LIFETIME.
-------
Table
1
2
37
Table Captions
Dose Equivalent and Risk Rate Estimate for the
Ingestion of Uranium
Dose Equivalent and Risk Rate Estimate for the
Ingestion of Radium
-------
38
Table 1
Dose Equi-
I so tope Organ - Risk3 valent Rate*3
238U endosteal
bone 8.9 84
bone marrow 12.0 5
all others 10.3
234u endosteal
bone 11.1 103
bone marrow 16.6 7
all others 9.3 -
a ¦ Number of premature deaths per lifetime per million
people exposed for ingestion of 20 pCi/day (50% via
each isotope)
b ¦ Dose equivalent rate in mrem/yr at 70 years from
ingestion of 20 pCi/day
Source: USEPA, Office of Drinking Water. Health Advisory for
Uranium.
Table 2
Dose Equi-
Isotope Organ Riska valent Rate^
226na endosteal
bone 8.5 83
bone marrow 30.8 8.6
all others 5.1
a ¦ Number of premature deaths per lifetime per million
people exposed for ingestion of 10 pCi/day
b - Dose equivalent rate in mrem/yr at 70 years from
ingestion of 10 pCi/day
Sources USEPA, Office of Drinking Water. Health Advisory for
Uranium
-------
NO
NO
YES
YES
NO
NO
YES
YES
NO
YES
COMPLIANCE
NON-COMPLIANCE
MEASURE
Ra-226
IS Ra-226
PLUS Ra-228
> 5pCi/l
MEASURE
Ra-228
IS Ra-226
> 3pCi/l
MEASURE
RADON &
URANIUM
IS ALPHA
> 5pCi/I
IS ALPHA
> 15pCi/l
MEASURE
GROSS ALPHA
Is ALPHA
MINUS
RADON &
URANIUM
ALPHA
> 15pCi/l
-------
NO
YES
YES
YES
NO
YES
YES
NO
IS BETA
> 50pCi/l
NON-COMPLIANCE
COMPLIANCE
MEASURE
GROSS BETA
MEASURE
TRITIUM AND
Sr-90
Sr-90
> 8pCi/l
TRITIUM
> 20,000 pCi/l
ANNUAL
DOSE FROM
TRITIUM Sr-90
IS > 4 mrem/yr.
ANALYZE
TO IDENTIFY
RADIONUCLIDES
DETERMINE
COMPLIANCE
ANNUAL DOSE
FROM
RADIONUCLIDES
FOUND IS
> 4 mrem/yr
Figure 2
-------
MONITOR QUARTERLY
MEASURE
GROSS BETA
YES
r
I
Is BETA
> 50pCi/l
ANALYZE
TO IDENTIFY
RADIONUCLIDES,
DETERMINE
COMPLIANCE
WITH 141.16
ANNUAL
DOSE FROM
YES RADIONUCLIDES
FOUND
IS
> 4 mrem/yr
NO
NO
NO
COMPLIANCE
Is BETA
> 15pCi/l
j YES
ANALYZE
FOR
Sr-89, Cs 134
~
Is Cs-134
> 80pCi/l
NO
MEASURE
Is 1-131
>3pCi/|
YES NO
YES
Is Sr-89
80pCi/l
YES
ANNUAL
DOSE FROM
Sr-89-Cs-134
> 4 mrem/yr
compliance:
YES
NON-
COMPLIANCE
MONITOR
ANNUALLY
MEASURE
TRITIUM
AND Sr-90
Is TRITIUM
^ 20,000
pCi/l
~ NO
Is
Sr-90
>8 pCi/l
YES
YES
NO
ANNUAL
DOSE FROM YES
Sr-90 H-3
>4 mrem/yr
IHo
COMPLIANCE
NON-
COMPLIANCE
-------
BIOLOGICAL
STRESSORS
CHEMICAL
~
Absorption
Adsorption
Ingestion
Inhalation
Injection
Burden
PHYSICAL
STRESSORS
~
Diagnosis
treatment
Additive, Synergistic
or
Antagonistic Bioeffects
(Acute or Chronic)
-------
100
90
80
70
60
50
40
30
20
10
0
o 9 A.M.
~ 11 A.M.
1 P.M.
3 P.M.
1 A.M
0
800
850
900
950
1000
1050
RADIATION DOSE (RADS)
-------
OCCURRENCE
POPULATION ^
POPULATION
CONCENTRATION
INDIVIDUAL
RISK RATE
POPULATION
RISK
NUMBER OF LIVES
SAVED BY
ELIMINATING
CONCENTRATIONS
THAT LEAD TO A
RISK OF 10"4. 10-5
AND 10"® PER LIFETIME.
-------
Table 1
DOSE EQUIVALENT AND RISK RATE ESTIMATE
FOR THE INGESTION OF URANIUM
Isotope Organ Rfckf yatent Ftete'b
238y ENDOSTEAL
BONE 8.9 84
BONE MARROW 12.0 5
ALL OTHERS 10.3 -
234y ENDOSTEAL
BONE 11-1 103
BONE MARROW 16.6 7
ALL OTHERS 9 3 -
a - NUMBER OF PREMATURE DEATHS PER LIFETIME PER MILLION PEOPLE
EXPOSED FOR INGESTION OF 20 pCi/day (50% via each isotope)
b = DOSE EQUIVALENT RATE IN mrem/yr. AT 70 YEARS FROM INGESTION
OF 20 pCi/day
Source: USEPA, OFFICE OF DRINKING WATER. HEALTH ADVISORY FOR URANIUM.
-------
Table 2
DOSE EQUIVALENT AND RISK RATE ESTIMATE FOR THE
INGESTION OF RADIUM
Isotope Organ Risk8
— —2—. —— valent Rate"
226Ra ENDOSTEAL
BONE 8.5 83
BONE MARROW 30.8 8.6
ALL OTHERS 5.1 -
a = Number of premature deaths per lifetime per million people exposed for
ingestion of 10 pCi/day
b = Dose equivalent rate in mrem/yr at 70 years from ingestion of 10 pCi/day
Source: USEPA, OFFICE OF DRINKING WATER. HEALTH ADVISORY FOR URANIUM.
-------
REEGULATORY CONSIDERATION IN
RADIATION PROTECTION
William A. Mills
-------
REGULATORY CONSIDERATION IN RADIATION PROTECTION
William A. Mills
U.S. Nuclear Regulatory Commission
ABSTRACT
A regulatory scheme is suggested that identifies regions of "unacceptable" and
"safe" as the upper and lower bounds and "operational" in the continuum
between. These regions are associated with levels of annual risk of cancer
death for a given level of lifetime exposure between 100 mrem/yr and 1 mrem/yr,
upper and lower bounds, respectively. Concern is expressed with establishing
public health standards at ALARA levels, which result in lower standards for
reference, and views are presented on several issues of interest in regulations
for protection of public from radioactivity in drinking water. Based on the
regulatory scheme suggested, author concludes that existing standards for
drinking water appear to be lower than need be.
INTRODUCTION
Wise members of Congress write perfect environmental laws under which
Solomon-like decisions are made by regulators to provide an environmental
paradise. If you believe this statement is true then what I have to say will
have little, if any, meaning to you. I will assune that most of you believe
this statement is at least partially false.
In this light I would like to share with you some thoughts on regulations
that I have formulated during my years of being involved in attempting to set
Federal radiation protection standards. I offer these thoughts totally as my
own and in no way should they be considered as views of others within the
Nuclear Regulatory Commission. My comments will be long on philosophy and
short on facts; this is because I believe our current philosophy on radiation
protection is in need of repair. We have a wealth of scientific facts behind
us as evidence by the outstanding reports of this workshop. Simply stated my
request or plea, if you will, is that we apply common sense in establishing
standards; an environmental paradise has never existed nor can we produce one.
-------
2
Protecting public health does not mean zero risk; even for vaccines
against diseases some risk is involved with the vaccine itself. I believe that
the public does understand not having protection with zero risk and on a purely
health basis is willing to accept some environmental risks. The question is
then how much risk.
My concern is that with our desire to protect the health of our children,
grandchildren, and future generations (that is provide them a very high degree
of protection) we are in fact jeopardizing their economic safety. As we
establish lower and lower radiation protection standards, we establish new
reference points to be considered. What has happened is that "as low as
reasonably achievable (ALARA)" has become not the operational and judgmental
radiation protection tool that was originally intended, but in fact our
standard. If it is possible to set a lower standard, do it, appears to be our
current philosophy—reasonableness or common sense appears to be less in
evidence. Are we than under the banner of conservative health protection
running up a much heavier economic burden for our offspring?
Shown below are three principal elements ususally considered by a public
health agency in meeting its responsibilities to protect health against adverse
effects from a given environmental agent.
o HEALTH
o BENEFIT
0 COST
The public first concerns itself with whether the contaminating agent is
"safe" or "unsafe" for health. If the risk is so high that the agent is
considered unhealthy by any reasonable standards, then the debate is likely to
be moot and cost is considered irrelevant. The same is true, I believe, when
any health risk is trival or unquestionably safe—the public does not want to
see its money spent on insignificant risks.
-------
3
Benefit is not generally considered in the standard setting process when
the word is used in the context of "raison d'etre" of the activity producing
the radiation exposure or of the source itself. Most often the benefit is a
given, such as having drinking water or the need for electrical generation.
Thus, for the purpose of this workshop no further discussion of this element is
warranted.
Given then that the public will not accept an "unhealthy or unsafe"
environment and that a benefit of the source exists, we are faced with
consideration of cost. Some would argue that a cost effectiveness analysis or
an evaluation of dollars spent to avoid a health detriment is a proper
trade-off approach, but it too must be judgmental and rely on common sense. If
we regulators were Solomons, it might be a tool, complete in itself. But we
are not so wise, and we often bias our analyses and evaluations to meet our
prior objectives. We have a tendency to overestimate risk and underestimate
cost and those we regulate often do the reverse. So it becomes a numbers game
on costs and health risks—trusting that the final position is closer to the
correct one that benfits the U.S. citizenry.
Congress has not explicitly stated that the health risk from drinking
water containing radioactivity must be zero and the public appears willing to
accept some risk. I suggest we use these assumptions in establishing standards
that are based on our judgment of what risk is safe and what risk is unsafe.
We cannot give unequivocal answers to what are the risks because, as Dr. Mays
(attending this workshop) stated a few years ago, "science is untidy" (Ma78),
but I believe this is recognized by a public that still ranks scientists high
in their confidence ratings.
In making these decisions, ICRP (1977) has provided the following guidance
which can be helpful:
1. It has established a risk based system in which the health risk resulting
from total body exposure is the point of reference. By weighting the risk
to specific organs from internally deposited radionuclides against this
reference level, total effective whole body risk from both internal and
-------
4
external exposures can be summed. For drinking water, our concerns are
limited only to internal emitters, and the weighting factors allow us to
estimate the risk from intakes of multiple radionuclides in order to
obtain an "effective total body" risk estimate.
2. It has also stated that risks to the public that are of the order of 10"5 y"1
or 10"® y"* appear to be in line with risk generally accepted in everyday
living.
If we accept ICRP's guidance for a risk based system, we can propose a
regulatory system for protecting the public that in general prescribes what
constitutes both "safe" and "unsafe" levels of risk. Between these two bounds
of "safe and "unsafe" is the range for decision in which some considerations of
cost of controls and other factors (non-health) become of interest. The
regulatory scheme shown below based on levels of annual risk of dying of a
radiation induced cancer from continuous radiation exposure at a given
associated annual dose equivalent rate depicts this point.
_5 UNACCEPTABLE REGION
10" /yr (Risk Generally too High for Lifetime Exposure) 100 mrem/yr
OPERATIONAL REGION
Practice of ALARA orother judgments
10~^/yr 1 mrem/yr
"SAFE" REGION
(Risk low enough to be of no regulatory concern)
The associated dose equivalent levels provided are based upon the NAS/BEIR
Committee (NAS80) estimate of lifetime risk of approximately one hundred cancer
deaths per 106 rads from low LET radiation for continuous lifetime exposure to
one rad per year (linear-quadratic dose response model). I would note that the
BEIR Committee also states that it "...does not know whether dose rates of
gamma or x-rays of about 100 mrads/yr are detrimental to man."
Levels about 100 mrem/yr (approximately equal to natural background) are
not necessarily unacceptable if they exist for only short periods of time. For
example, the limit of 500 mrem/yr for the maximum exposed individual approved
-------
5
by President Eisenhower in 1960 as Federal Guidance (FRC60), and reaffirmed by
NCRP in 1971 (NCRP71) and by ICRP in 1977 (ICRP77) is consistent with this
approach. In situations where the source of radiation remains under control or
will decay with a rather short half-life (days) 500 mrem/yr is not an
unreasonable limit. The risk to health would still be quite small. However,
in the case of drinking water these conditions are not applicable, particularly
for natural radioactivity in drinking water where the source is not under
control in a practical sense and, for the most part, the nuclides have long
half-lives.
The levels in the operational region do not necessarily consistitute
limits. It is a region in which health is generally the more important factor
-5 -1
near the 10 yr risk level and cost to reduce health risk is a consideration
of increasing importance with decreasing risk. It is the region in which
judgment plays the major role in decisions that involve factors other than
-5 -6 -1
health. It is consistent with the 10 to 10 y value recommended by ICRP
and a philosophy of ALARA for public health reason. At a level of risk of
10"7 y"1 the health risk and cost of control are no longer a consideration,
the risk is insignificant.
If you accept this regulatory scheme, a gene-ic standard for limiting
radioactivity in drinking water and applicable to all radionuclides
collectively is warranted. Such an effective dose equivalent limit based on
health risk would convey to users of the water that the limits are established
because higher risks are unacceptable (unsafe) and steps are required to reduce
those risks. If the level of radioactivity in the drinking water is below the
limit and in the operational region, doing something about it might be
considered but would not be mandatory.
Although I believe the numerical values I have shown in this scheme are
reasonable, I am not suggesting that the values be adopted per se for
radioactivity in drinking water. I do believe, however, that a risk based
scheme as proposed can be used to address any risk from radioactivity in
drinking water generically. It would eliminate the need to have separate
standards for man-made contaminants and for radium and uranium.
-------
6
I would like to conclude by commenting on some very specific items of
interest in considering drinking water standards.
1. Population vs. maximum individual risk. When the probability of exposure
is one, or when the probability is high, as it is when using water with
any radioactivity for human intake, then individual risk appears to be the
reasonable criterion for protection. Protecting the individual assures
protection for all. Some standard developers believe that societal risk
is more important and that the population size served by the water supply
is an overriding factor in protection. This idea was soundly rejected by
public comments on EPA's proposed interim drinking water standards for
radioactivity. The public wants equal protection in the water it
consumes. Population collective dose is useful to provide a health
perspective on risk but even this use should be limited in time and space.
In determining either individual or population risk it is important to not
compound uncertainties, i.e., multiply, conservative values by conservative
values. If prudent public health policy requires a "safety factor" or
"ample margin of safety" this can be introduced at the end of the
calculation, after arriving at a best estimate. Too often we maximize all
assumptions on use of water, periods of such use, etc., which results in
unrealistic conclusions.
2. Radon. How big a problem is exposure to radon in water? Are we convinced
that lung cancer risk warrants a standard for radon? I am not so
convinced to date. Nevertheless, if the plan is to establish a standard
for radon in water, I suggest that consideration be given to considering
the risk to the non-smoker as the basis and that realistic assumptions be
made concerning exposure conditions. I specify the non-smoker for two
reasons: (1) the existing confounding smoking factor in miner data has
been inadequately delineated in determining risk coefficients, and (2) in
applying these risk coefficients when estimating risk from exposure to
radon decay products for the general population the larger lung cancer
rate for smokers in the population results in overestimates of radiation
-------
7
risk for the non-smoker. Consideration should be given to basing our
standards on lung cancer rates for the non-smoking population only.
3. Use of Drinking Water Standards. Application of maximum contaminant
limits (MCL's) established under the Safe Drinking Water Act appears to be
inappropriate for application broadly to water use that does not require
drinking water quality. It is especially important to keep in mind how
and why the standards were developed. For example, using MCL's as
standards for disposal of uranium mill waste is, in my view, a misuse
because MCL's apply at the tap and not at the source. Again the tendency
is to use the lowest standards available as a reference point without
reexamining either the reasonableness of the standard or its basis.
In conclusion I would encourage EPA to use the wealth of information
provided by the outstanding scientists contributing to this workshop and to
develop more reasonable standards for drinking water. Within the regulatory
scheme I have suggested, the existing standards appear to be lower than they
need to be.
References
FRC60 Federal Radiation Council, 1960. Radiation Protection Guidance
for Federal Agencies. Federal Register, 52, 5/18/60.
ICRP 77 Recommendations of the International Commission on Radiological
Protection, 1977: Publication 26. Pergamon Press, New York.
NAS 80 National Academy of Sciences, 1980. "The Effects on Populations
of Exposure to Low Levels of Ionizing Radiation," 1980. National
Academy Press, Wash., D.C.
NCRP 71 National Council on Radiation Protection and Measurements, 1971.
"Basic Radiation Protection Criteria," NCRP Report No. 39. Bethesda,
MD.
Ma 78
Mays, C.W., 1978. "Introduction to the Proceedings.of the
International Symposium on Biological Effects of Ra and
Thorotrast," Health Physics, 35:1.
-------
COMMITTEE ON TREATMENT, WASTE MANAGEMENT,
AND COST
Chairman: George W. Reid
Recorder: Tom Sorg
Committee Members: William Brinck
Floyd Galpin
Steven Hathaway
Peter Lassovszky
Roy Reuter
Richard E. Rozelle
-------
TREATMENT, WASTE MANAGEMENT AND COST FOR RADIOACTIVITY
IN DRINKING WATER
George W. Reid
Regents Professor/Director
Bureau of Water and Environmental Resources Research
The University of Oklahoma
202 West Boyd, Room 301
Norman, Oklahoma 73019
Peter Lassovszky
Office of Drinking Water (WH-550)
U.S. Environmental Protection Agency
Washington, D.C. 20460
AND
Steven Hathaway
Municipal Environmental Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The processes (costs, efficiencies and reliability) used
to treat drinking water to remove radioactive contaminants
of concern and the disposal of wastes generated by the treatment
processes are analyzed and discussed. The study was limited
to uranium, radium, and radon. Initially concepts of water
and waste treatment in terms of their applicability to the
drinking water industry were established. The alternative
processes for treatment of radium, uranium, radon, water and
sludges were described and evaluated, in terms of cost, efficiency,
reliability, process control and feasibility.
-------
INTRODUCTION
The substance of this paper can be thought of as defining
a series of adjustments between the occurrence of radionuclide
in water and user safety and health requirements shown in Figure
1. These adjustments, treatments from occurrence to inges-
tion or other exposure, are not thought of in isolation, that is
selections are dependent on levels defined by both source and user
committees.
In general the concern is with feasible treatment and/or
management of small ground water supplies for reduction/or removal
of naturally occurring radioisotopes. Though this is the primary
thrust, treatment on larger scale is also important for surface
waters. The primary natural isotopes selected for study are radi-
um, uranium, and radon. Manmade isotopes and other natural iso-
topes, thorium, and polonium, though important, were not considered
at this time. Also of concern is whether or not, and under what
conditions treatment to reach a defined level is feasible.
There are also questions of level and risk.
-2-
-------
Removed waste issues addressed include RCRA and super-
fund, legal bases in states, sludge classification, storage
dilution, isotopic dilution, gasification, thickening, ultimate
disposal and recovery.
Background
The National Interim Primary Drinking Water Regulations
(NIPDWR) established maximum contaminant levels (MCLs) for natural
and manmade radionuclides (Na76). The MCLs for naturally occurr-
ing radioactivity include radium 226 and radium 228 and gross
alpha particle activity. The MCL for manmade radionuclide con-
taminants include beta and photon emitters. EPA is in the process
of revising the NIPDWR, and the Agency is considering the possi-
bility to propose MCLs for uranium and radon.
Radium, uranium and radon are naturally occurring ele-
ments and are primarily present in ground waters. The radiation
problems in ground waters are attributed to the leaching of radium
and uranium from rock-bearing strata. Elevated radium and uranium
levels in surface waters have been attributed to mining operations.
The occurrence of radium in surface and ground waters has been
investigated in a number of studies (Mc60, Sc76, Ka77). Kaufmann
et jil. (Ka77) noted that the radium content of surface waters
ranges from 0.01 to 0.1 pCi/L, while some ground waters may
contain as much as 100 pCi/L. Drury (Dr81) estimated that of the
60,000 community drinking water supplies in the United States,
-3-
-------
100-2,000 would exceed a uranium concentration of 10 pCi/L and 2,500
to 5,000 would exceed 5 pCi/L. Radon is an inert noble gas and is
a product of the decay of radium 226. Some amount of radon can
be found in all ground waters, however, a number of studies
investigating the occurrence of this element indicate large
variations in its concentration (Ku54, Sm61, A175, Oc76). Duncan
et al. (Du76) reporated that in some areas radon concentration
levels in drinking waters may range between 1,000 and 30,000 pCi/L.
The exposure from radon in drinking water to humans is by inhala-
tion and ingestion. However, Partridge et al. (Pa79) found that
airborne radon released during normal household activities such as
showers, diswashers, etc., poses a greater potential health risk
than radon ingested with water.
Manmade radionuclides do not occur naturally in drinking
water. These radioisotopes may find their way into the water from
several sources. Other than those that may result from nuclear
weapon fall-out, the most likely sources of manmade radionuclides
are from nuclear power plants, research and industrial facilities,
radioactive storage and waste disposal sites. There may be a
number of radioisotopes that may be present in manmade nuclear wastes.
The composition of radionuclides depend on the type of activity the
wastes originate from. In an event of a spill or fallout, the
presence of manmade radionuclides will be primarily in surface
waters. Leaching of contaminants from waste disposal sites would
-4-
-------
result in ground water contamination, and would eventually result
in contamination of surface waters.
The naturally occurring radionuclides are alpha emitters,
such as, uranium, radium, 226 and 228, and radon 222, while in
general, the manmade radioisotopes are beta and gamma emitters. In-
gested and inhaled radiation can be a serious health risk and in
general, the absorbed dose and biological effect from alpha radi-
ation is estimated to be twenty times more damaging than from beta
and gamma radiation.
There are several different treatment methods that can be
used for removing radionuclides from drinking water. It must be
understoood that each radionuclide has its own specific treatment
method. The property of radioactivity has little or no effect on
the extent of removal by a particular treatment method. Therefore,
if there are a number of different radionuclides present in the
water, a specific treatment process may not be applicable to remove
all of them. In those instances, multiple treatment techniques may
be required.
Prior to implementing any kind of treatment, a utility
should evaluate the options that are available to reduce the identi-
fied radionuclides that are present in the drinking water. The
major considerations in selecting the appropriate options are
technical feasibility, economics, capital and operating costs. In
-5-
-------
some instances treatment may not be the best option. A proximity
of an alternate water supply with suitable water quality may make
it more economical to the utility to switch to that water source
instead of implementing treatment. Purchase of water from another
utility may be another feasible alternative to treatment. Imple-
mentation of point-of-use treatment rather than centralized treat-
ment may be another possibility. Nevertheless, whatever option is
selected by the water supplier, it must be based upon a case-by-
case technical evaluation of the system's entire process and the
assessment of the economic factors involved. In selecting an
appropriate treatment method, the major factors of consideration
should include quality of the source water, extent of contamination
by radionuclides of concern, economies of scale, ability of the
community to absorb the costs, disposal of wastes produced during
the treatment processes. The effectiveness of some of the treat-
ment techniques identified below have been proven in actual prac-
tice. The other treatment technologies which are discussed have
been evaluated in the laboratory and pilot plant studies. The
result of these studies indicate that these technologies may be
potentially viable methods to remove radionuclides from drinking
water.
GENERAL CONCERNS ABOUT TREATMENT TECHNOLOGIES
Treatment technologies are derived from water works prac-
tice, and perhaps will affect not more than 600 of some 60,000 pub-
lic water supply systems in the U.S.A. The degree of removal of a
-------
specific contaminant from source waters is limited by the efficiency
of the particular treatment applied. Treatment of source water?
containing high concentration levels of contaminants will produce
poorer quality of processed waters compared to waters containing
lower levels of contaminants prior to treatment. This is il-
lustrated in Table 1. Treatment can be thought of as a tree
diagram, in which, one process is layered on another or on itself
to achieve increasing efficiencies. It is basically a problem of
solid concentration, such that the cost to remove from 30% to 90*
efficiency is the same as from 90% to 95%, as 95% to 97.5%, etc.,
that is the solids are doubled from 97.5% to 95% as from 95% to
90%, etc. So if one can achieve 90% efficiency, the next 5% will
cost him as much as the previous efficiencies, to go to 97.5%
would be another increment. VThere high source concentration of
contaminants exist or the maximum allowable level of contaminant
iin the product water is very low, it may be necessary to use
several stages of treatment. In that case the cost of producing
acceptable water multiplies, and other options, dilution, split-
ting, blending, point-of-use treatment, etc., become important.
There are also unanswered questions as to treatment induced
conflicts, such as sodium discharge from ion exchange columns
and heart disease, chlorine for disinfection, radioactivity
and cancer, etc.
-7-
-------
Radium Removal Processes
Processes that are effective in removing radium from
drinking water include lime softening, cation exchange, reverse
osmosis and selective absorption. The efficiencies of these
processes have been domonstrated by full scale operating facili-
ties, pilot plants and laboratory studies. (See Table 2)
Cation Exchange
Ion exchange using natural or synthetic resins to re-
place calcium and magnesium ions with sodium ions to soften the
water is a widely practiced technology. If radium is present in
the water it will be removed with the hardness since radium is a
divalent cation and similar in chemistry to calcium and magnesium.
As illustrated in Table 3, experience in Iowa and Illinois indi-
cates that a well operated ion exchange softening plant is capable
of removing 85-97% of the radium from drinking waters (Be76, Br78,
Sc76). Since radium removal still takes place after the resin
is exhausted for hardness, regeneration of the resin to achieve
good hardness removal will assure good radium removal also (Ep77).
There are several advantages and disadvantages to the
use of ion exchange softening systems. One major advantage is that
ion exchange is readily adaptable for small systems. The units
are commonly available, relatively simple and easy to operate. In
addition, ion exchange systems require relatively small space. In
-8-
-------
areas where the water is hard, ion exchange will perform a dual
purpose. It will remove hardness along with the radium. However,
an area of concern associated with the use of sodium ion exchange
softeners is that as a result of treatment the sodium content of
the water increases. The National Academy of Sciences fNa77) con-
cluded that in areas where the sodium content of the water is
already high, this may be a potential problem for people with
restricted diets and hypertension. Ciccone (Ci83) noted that
in order to avoid the addition of sodium, potassium chloride may
be substituted as a regenerant. However, potassium chloride may
be as much as five times more expensive than sodium chloride.
Another alternative to sodium cation exchange is hydrogen ion
exchange to remove radium from drinking water is being investi-
gated at the University of Illinois (CrftfM. Rorg (RoRl) notes
that potential saving with this process may be realized if sul-
furic acid is used as a regenerant since the stoichiometric
amount of sulfuric acid required for regeneration of the resin
is considerably less than the sodium chloride required for
regenerating a sodium ion exchange system. On the other hand,
the hazards associated with the handling of the acid used for
regeneration may preclude the use of hydrogen exchange resins
for home or point-of-use applications.
Another disadvantage associated with the use of
softeners is that the softened water can be corrosive to the
distribution system. In some instances the radium content
-9-
-------
of the untreated water may be low enough to allow blending with the
treated water, thus producing a less corrosive water. Where the
radium content of the untreated water is excessively high, treated
water and unheated water cannot be blended. The treated water
in this instance may require the addition of corrosion inhibitors.
Although large ion exchange softening systems are technically
feasible, the economies of scale make them less attractive com-
pared to lime softening. Most of the larger water supply systems
use surface waters that may require pretreatment such as filitra-
tion before ion exchange could be used.
Reverse Osmosis
Reverse Osmosis (RO), a relatively new technology is
being commonly used in areas where the water has high total dis-
solved solids concentration levels (TDS). The RO process utilizes
a membrane which allows the passage of the water but rejects the
dissolved salts. Only a portion of the water is treated. Pressure
is required to force the water through the membrane. The removal
efficiency of RO is better at higher pressures. The process is
continuous and the rejected concentrate is discharged as a continu-
ous stream. Experience in Florida by Sorg (S08O) and Iowa by
Bennet (Be76) indicates that RO is highly effective in removing
radium from drinking water. As shown in Table 4, the process can
be over 90% efficient.
-10-
-------
The economic advantages are most favorable when using
RO for the removal of radium if the drinking water contains other
contaminants to be removed at the same time. Space requirements
for RO are small compared to equivalent capacity lime softening
or ion exchange systems. Sorg (S08O) notes that RO units for
small systems are prefabricated and pretested at the factory so
their installation can be completed in a short time.
Compared to other treatment techniques RO is relatively
expensive to operate due to high energy requirements for pressure
pumps. In addition, where feed waters contain suspended solids,
iron, manganese and scale-forming compounds, pretreatment (lower-
ing of the pH) of the water is required before RO can be used.
Since RO also removes the hardness from the water, stabilization
of the water may be required to make it non-corrosive.
Lime Softening
Lime softening is a commonly used process to remove hard-
ness from the water. Lime is added to the water to neutralize the
carbonic acid and to form insoluble calcium carbonate and magnesium
hydroxide which will precipitate out during the removal process.
Experience in the field and the laboratory (Br78, EP77) shows that
lime softening can remove 80%-90% of the radium provided the pro-
cess pH is maintained above 10. The results of field studies are
shown in Table 5.
-11-
-------
Although small lime softening systems may be technically
feasible, this treatment technology is best suited for large capa-
city plants. Lime softening systems require more complicated
equipment and demand more operating supervision. Lime softening
is also well suited to the treatment of waters containing iron
and turbidity since these contaminants are also reduced.
Adsorption Processes
Adsorption processes for radium removal are in the
development stage with the exception of greensand filtration which
has been used extensively to remove iron and manganese from water.
However, because of their potential for removal of radium, these
technologies should be given consideration as potentially viable
technologies. One of the adsorption processes utilizes the capa-
bility of manganese dioxide (Mn02) to adsorb metal ions, while the
other process involves a radium selective complexer developed by
DOW Chemical Corporation.
MnO? Adsorption. Paw water containing iron and manga-
nese can be treated by passing it through greensand filters after
oxidation. It has been observed in Iowa and Illinois (Fp77) that
potassium permanganate added to the water as an oxidant, improves
radium removal efficiency. However, as shown in Table 6, the radium
removal efficiency is limited to not more than 56%. The removal
of the radium is attributed to the manganese dioxide (Mn02)
formed during the oxidation. The Mn02 is known to be an effective
adsorber of many metal ions. In order to increase the efficiency
-------
of the process and to determine its technical and economic feasi-
bility, EPA has funded additional research to conduct pilot and
field studies in Iowa (Rc81).
Because of low radium removal efficiency, the use of
iron removing greensand filters is limited to instances where
radium levels do not exceed twice the MCL for radium. However,
when this technology is used, no sodium is added to the water as
a result of treatment. Iron, manganese, and radium are all
removed by this method.
Additional field studies by Cook (C068) indicate that
manganese dioxide impregnated fibers may remove radium more ef-
fectively than greensand filters continuously fed with potassium
permanganate.
The treatment consists of passing water through a filter
vessel loaded with manganese dioxide impregnated fiber media.
Once the fiber media is exhausted, it is replaced with a new one.
Since the preparation of the fibers is not simple, it may pose
operational problems in small system application where no qualified
operators are available. No backwashing or regenerating of the
system is required, thus eliminating the need for disposing the
wastewater discharge. Results of the studies by Cook (C068)
indicate that this process may be able to remove radium from
-13-
-------
relatively large amounts of water. However, the practical perfor-
mance of the units may be adversely affected by the washout of
loosely held manganese oxide from the fiber.
DOW Radium Selective Complexer (RSC). This material
was originally developed by DOW Chemical Company to remove radium
selectively from minewater effluents. The RSC system is basically
a water softener unit in which the normal exchange media is replaced
with the RSC. Boyce (Bo82) reported that the RSC has a very high
loading capacity for radium. Initial field test data indicated
that the RSC reduced radium levels ranging from 20 to 45 pCi/L in
the water to below 3 pCi/L for approximately four months without
any need for regeneration. Similar results were obtained in Brady
and Bellville, Texas, where that the RSC was capable of reducing
radium levels from 30 pCi/L to below 5 pCi/L without any need
for regeneration (Te82). The potential for the RSC as an effective
treatment method to control radium in drinking water is consider-
able. Due to its simplicity the process could be adopted to both
small and large water supply systems which need to treat the
water for radium removal. Because of the RSC's high loading
capacity for radium the sorbent units may be kept in service for
a long time before replacing the sorbent media. The operational
requirements for backwashing and regenerating the units would be
eliminated as well as the problems associated with the disposal of
waste streams. Since hardness is not removed during the process,
post treatment for corrosion control would also be eliminated.
-14-
-------
However, by eliminating the need for regeneration, the buildup
of radium in the RSC may be a potential concern. Therefore,
additional studies are needed to evaluate the possible hazards
associated with the radiation fields resulting from the accumu-
lation of radium in the adsorbent bed.
Direct Precipitation by a Soluble Barium Compound
Kerr-McGee Corporation commenced in 1973 examining
methods of radium removal from a by-product stream in order to
permit its use directly as a fertilizer solution (Sh83). Regu-
lations subsequently were established for point source discharges
of uranium minewater requiring the control of soluble radium-226
at a level of 3 pCi/L on the average with a maximum of 10 pCi/L.
Kerr-McGee uranium mines, expected to be operating in the 1980's,
could be expected to discharge approximately 10,000 gallons per
minute cumulatively. It became imperative to determine technology
which permitted the reliable control of radium to these established
levels. As a result of work previously done with radium solutions
in AEC processes and a classical analytical procedure for the
removal of radium, direct addition of a soluble barium solution to
process by-product and minewater effluents, containing excess sul-
fate ion, promised the possibility of radium control at the
desired levels. Over a period of time, Kerr-McGee evolved a direct
precipitation procedure employing soluble barium compounds, i.e.,
barium chloride, carbonate, nitrate, which when dissolved in water
-15-
-------
at the appropriate concentration, could be added to a radium con-
taining solution of 50 to 200 pCi/L by a metering pump at the ap-
propriate rate. This addition at the metered rate resulted in the
precipitation of radium as a barium-radium sulfate precipitate
which is then permitted to grow in the fashion postulated by
Ostwald, resulting in particles of sufficient size to be removed
by simple settling.
The design of the settling ponds was determined empiri-
cally to be approximately as follows. The pond would be designed
with a length-to-width ratio of approximately 4:1, inflow and out-
flow controlled to secure approximately 1 foot per minute velocity
through the pond between inlet and outlet with a total residence
time of 72 hours. Measurements of the pond's efficiency normally
disclosed that the efficiency of pond in achieving these condi-
tions was in the range of 70% to 80%. Thereby, providing a true
residence time of approximately 50 hours. This design provided
the ability to reduce radium contents to required levels or below.
Radium concentrations on the order of 1 pCi/L or less have been
routinely maintained by treatment systems of this type.
This technique of radium control has been successfully
applied on an experimental basis to solutions containing 60,000
pCi/L successfully reducing terminal concentration to less than
5 pCi/L.
-16-
-------
Probably this technique could be applied to a drinking
water supply at a cost somewhat less than current water treatment
techniques such as reverse osmosis, ion exchange, etc., as reported
in the current literature.
The technique does require careful control of barium
concentrations injected into the water and routine analysis of
of the treated water to assure achievement of target concentra-
tions .
Uranium Removal Methods
Due to the absence of standards limiting the presence of
uranium in drinking waters, development of treatment techniques
to remove uranium from water has been aimed at recovery operations
from mine process waters and effluents. However, recent studies
by Drury (Dr82) and Bondietti, White and Lee (Le82) at the Oak
Ridge National Laboratory (ORNL) and by Hathaway (Ha82) indicate
that a number of treatment techniques have the potential to reduce
and maintain uranium levels in drinking waters at or below 10 pCi/L.
These treatment techniques include anion exchange, lime softening,
reverse osmosis and under certain conditions conventional coagula-
tion using alum or iron salts.
Anion Exchange
Because of their effectiveness and relatively low operat-
ing costs, anion exchange resins have been used to recover uranium
-17-
-------
from mine process waters. Laboratory and field studies by Bondi-
ette, White, Lee (Le82), and Hathaway (Ha82) indicate that anion
exchange technology is feasibile to remove uranium from drinking
waters. The results of laboratory studies by Bondietti, White and
Lee (Le82) (Figure 2), indicate that anion exchange resins have
a large adsorption capacity and high selectivity for uranium.
Bench scale studies at EPA by Hathaway (Ha82) (Figure 3) show
that when using actual drinking water contaminated with 300
microgram uranium/L, over 9,000 bed volumes were treated before
uranium was detected in the effluent. Field studies by EPA
confirm the findings of these laboratory studies. These studies
involved the evaluation of the performance of twelve 1/4 cubic
foot anion exchange systems installed in New Mexico, Colorado,
and Arizona at sites where the uranium levels in the untreated
water were in excess of 10 micrograms/liter. The preliminary
results of these field studies are shown in Table 7. Because of
the high loading capacity of the anion resins for uranium, these
units are especially suitable for point-of-use applications
where on-site regeneration is not feasible. For centralized
treatment, the resin may be regenerated and recycled by back-
washing it with sodium chloride solution.
Activated Alumina
The EPA has also evaluated activated alumina as a re-
moval media for uranium. Small columns packed with Alcoa acti-
vated alumina (129 mL) were used to treat raw water containing
-18-
-------
300 micrograms uranium/liter. Figure 4 shows the general removal
efficiency found in the laboratory study. Subsequent regeneration
and treatment revealed that approximately 5,000 bed volumes could
be treated to total exhaustion. No pretreatment of the raw water
was practiced in this study.
Lime Softening
Laboratory studies by Bondietti, White, and Lee (Le82)
indicate that lime softening may be effective in removing uranium
only if the pH of the water is maintained between 10.6 and 11.5.
At these pH levels 85% to 90% removal of uranium may be achieved
as shown in Table 7. The removal efficiency of uranium may be
improved to 99% by the addition of 80 to 120 ppm magnesium carbo-
nate to the water, provided the lime doseages are maintained at
or above 100 mg/l as illustrated in Table 8. At less than the
above dosages of lime and magnesium carbonate, the uranium removal
efficiency will be reduced drastically. Use of modified lime
softening techniques (magnesium addition) to remove uranium is
especially suitable in communities where lime softening is already
used to treat the water.
Conventional Coagulation Techniques
Limited information is available regarding the effective-
ness of coagulation techniques to remove uranium from drinking
waters. Laboratory studies by Bondietti, White and Lee (Le82)
-19-
-------
evaluating the effectiveness of iron salts and alum indicated that
uranium removal efficiencies are very sensitive to pH. The removal
efficiency at a specific pH level depends on the prevailing charge
on the floe and the uranium species present. Table 9 compares the
charges of alum and ferric salt floes with the suspected uranyl
species present in the water at various pH levels. Because of
the unstable nature and solubility of alum, coagulation may be
the most effective at a pH level of 6 as shown in Table 10.
Iron salts may be used at a wider pH range because they are more
stable. Use of coagulant aids such as polymers may improve the
efficiency of treatment and may reduce the alum dosage require-
ments. Practical plant operating data reported from Arvada,
Colorado, for the year 1982 shows that raw water uranium levels
varied from 24 micrograms/liter to below detectable levels depend-
ing on season. Effluent uranium was always 3 micrograms/liter or
less. The treatment method was coagulation with 12 — 25 mg/L
alum, 0.2 mg/L polymer, and direct filtration of the coagulated
water. Because of limited practical experience dealing with the
application of coagulation technology to remove uranium from drink-
ing waters, additional studies are needed to determine the practi-
cality of this technology. However, in instances where coagula-
tion is already used for water treatment, and uranium removal
from the water is desirable, the utility should investigate the
possibility of modifying the process before adding on a different
-20-
-------
treatment process. The selection of appropriate coagulant or co-
agulant aid, the determination of actual dosages and pH adjustment
must be done through onsite laboratory and pilot plant testing
with actual raw water. In general, coagulation treatment is more
suitable for larger systems where trained operators are available.
However, pre-fabricated package treatment plants requiring limited
operator attention are available for smaller communities.
Reverse Osmosis
There is no practical experience involving the utili-
zation of RO technology to remove uranium from drinking waters.
However, studies evaluating RO performance in removing uranium
from minewaters solutions (Sa76) indicate that RO can be over
90% effective. Results of limited bench scale studies by Bondi-
etti, White and Lee (Le82) seem to verify these results. Since
RO is effective in removing most inorganic contaminants from
water it can be expected that RO may be a practical option as a
treatment method to remove uranium from drinking waters.
Electrodialysis (ED)
Another through membrane process, ED, modified by
Ionic Inc., to create polarity reversing ED is competitive with
RO (Re83). Studies at the Teton-Nedco Leuenberger Research and
Pilot in Casper, Wyoming, by Garling, indicate comparative cost
-21-
-------
for O&M, and pretreatment. The unit operated 86* of the time,
at SI.02/1000 gallons (See Table 11) plant capacity was 30,000
gallons per day.
Radon Removal Methods
There are two methods available that may effective in
removing radon from drinking water. These methods are adsorbtion
by granular activated carcon (GAC) and aeration.
Granular Activated Carbon
The effectiveness of GAC as an adsorbent of noble gases
is well known. GAC has been used to remove radon from water and
to concentrate it for analytical measurement. However, with the
exception of studies performed at the University of Maine, there
is no documented evidence of utilizing GAC as a continuous pro-
cess to remove radon from drinking waters (Lo82). Because of its
short half life of 3.82 days it has been found that large portions
of the adsorbed radon decay within the GAC bed before breakthrough.
Thus, in effect the GAC bed in the process acts as a storage vessel
for the radon. Because of the decay, effective life of the GAC
bed is extended many times over the life indicated by the adsorp-
tion isotherm. The removal efficiency of radon from the water is
governed by the design of the GAC bed. It has been estimated that
a 1.0 to 2.5 cubic feet GAC contactor is capable of removing 96%
of the radon supplied to a single family home unit. Because of the
-22-
-------
decay of radon and radon daughters accumulated within the GAC
contactor, concern has been expressed about gamma radiation
hazard. Another potential concern regarding the use of GAC unit
is the fate of radon daughters in the GAC bed. However, prelimi-
nary tests with a commercial unit indicate that even after four
months of operation no radon daughters were detected in the ef-
fluent water. Due to their simplicity, use of GAC units to
remove radon in small water supply systems are attractive. No
moving parts are needed to operate the units. In areas where
turbidity is present, occasional backwasing of the GAC units may
be required.
Aeration
Laboratory and field studies performed at the University
of Maine by Lowry (Lo82) indicate that a well designed 35 to 40
gallon diffused aeration tank is capable of removing more than 95%
of the radon. A unit of this size is suitable to meet the needs
of a single family home. Hinckley (Hi82) reported that evaluation
of spray aeration systems by the State of Maine indicates that 93%
of the radon may be removed from the water. In general, spray
aeration to remove radon from water is best suited for larger water
supply systems where the radon levels are high. When utilizing
aeration techniques to remove radon from water, care should be
exercised to assure that humans are not exposed to the radon
released into the air.
-23-
-------
Disposal of Water Treatment Residuals
An important consideration in determining the appropri-
ate method of water treatment for radionuclide removal, and the
associated costs, is the disposal of the treatment residuals. This
appears to be a more significant potential problem in the case of
small groundwater supplies remote from surface streams. Table 15
summarizes the methods of disposal of treatment wastes as identifid
in the EPA 1977, "Manual of Treatment Techniques for Meeting the
Interim Primary Drinking Water Regulation," EPA-600/877-005 (Ep77).
These methods are considered to also be generally applicable for
the disposal of wastes generated in processes used for radionuclide
removal. However, there are a number of special considerations
that deserve specific mention.
A. Residuals from Radium Treatment
V
1. Dried Sludges and Other Solids
Present data indicates that sludges resulting from lime
softening processes and coagulation methods are not expected to
exceed 5 to 10 picocuries per gram of sludge. This should not be
a problem in the usual methods of sludge disposal. However, for
exceptional cases that may fall outside this range several caution^
are appropriate.
-24-
-------
Sludges containing radium concentrations of greater than
5 picocuries per gram, in large volumes, must be restricted in waste
disposal practices where they could be incorporated in building
materials or where they would serve as the main constituent of
landfill when future construction on the site is a possibility.
Such restrictions will limit the possibility of excessive radon
emanation into living spaces. Radium concentrations up to 100
picocuries per gram could probably be suitably disposed of in
sanitary landfills where they are relatively minor constituents
of the total waste in the facility. Higher concentrations of
radium should be considered for designated hazardous or radio-
active waste disposal sites. However, wastes of this concentra-
tion probably would not result from present conventional water
treatment processes used for radium removal. Other processes now
in the developmental stage, such as manganese fiber filters, may,
however, produce higher specific activity residuals.
2. Ion Exchange
Backwashes of cationic exchange resins used for radium
removal can usually be disposed of by discharge to surface streams
if there is sufficient dilution to handle the salinity. In most
cases special consideration of the radium content should not be
necessary but may require special permit authorization. It is
noted that the public perception that it is a radioactive discharge
may also create problems.
-25-
-------
Reliability of Treatment Methods
From existing information, Reid and Law (Re83) estimated
the reliability of the processes being used to remove Ra U, and Rn.
This information is presented in Tables (12, 13, and 14). Reason-
ably sufficient data exists only on coagulation, softening, EDR
and RO which indicate that these processes are more reliable.
Costs Associated with the Removal of Radionuclides
The capital, operating and maintenance costs (O&M), and
the costs for producing 1,000 gallons of water by various treat-
ment techniques versus the size of treatment facility are shown in
Figures 5, 6, and 7. The treatment techniques considered include
anion exchange, cation exchange, reverse osmosis and conventional
coagulation techniques using alum as a coagulant. Also included
are costs for modifying lime softening and alum coagulation. In
instances where one of these processes are already in use, modifi-
cation of the existing process to remove radionuclides may be the
most economical solution. Costs associated with the use of manga-
nese dioxide fibers and the RSC sorbent are not illustrated in the
Figures due to the lack of sufficient data. The cost of removing
uranium by anion exchange cannot be represented by Figures 5, 6,
and 7. The high capacity of strong base anion resins to remove
uranium would substantially change the cost curves due to less
frequent regeneration and low operating cost. No actual costs
-26-
-------
have been developed for this process, These costs were developed
for new systems designed for capacities in excess of one million
gallons per day (MGP). They were determined on the basis of work
developed by the Municipal Environmental Pesearch Laboratory, n.P.
Environmental Protection Agency, Cincinnati, Ohio, using the
computer model written for EPA by Gumerman et al^ (CUT79). The
costs for new systems having capacities less than 1 MOD were
developed by v.j. Ciccone and Associates in coordination with
EPA and several engineering, design and construction firms. The
capital costs were based upon an extrapolated Engineering News
Pecord average construction cost index for 20 U.S. cities.
Operating costs were based upon a Producer Price Index for 19R2
dollars. For the purpose of estimating costs per 1,000 gallons
of water treated, the capital costs were amortized over 20 years
at 12* interst rate. Figures 5, 6, and 7 represent average
estimated costs. The actual costs involving particular treatment
facilities may greatly vary due to site specific conditions such
as water quality, geographical and local conditions and waste
disposal requirements. Costs for sorbents to remove radium are
not included.
In some instances point-of-use treatment to remove radio-
nuclides may be more economical than centralized treatment. The
costs per customer may be significantly lower for small communities
since only a portion of the water intended for consumption may
require treatment. For ion exchange, small water softeners may be
-27-
-------
adopted. Home softeners are frequently purchased or leased to
soften hard water. Ciccone (Ci83) estimated that a commonly used
home softener system amy cost $829.00 per unit. These units may
be utilized to remove radium. If regeneration at site is practi-
cal, the operational cost for salt usage may be $5.00 per month
or $60.00 per year. Cartidge size units for ion exchange cost
about $450.00. Replacement costs for cartridges including labor
is estimated to be $138.00 per year. Removal of uranium in a
home system may be considerably less expensive than removing radi-
um. It is estimated that a 1/4 cubic foot unit could treat
20,000 gallons of uranium contaminated water. A 1/4 cubic foot
unit with strong base resin would cost about $200.
Similar cartridge size units for RO are available for
point-of-use. The average cost of a small RO unit for point-of-use
is estimated to be $780.00. Operating costs for an RO unit
include electricity and labor to replace the membrane element
once per year. Replacement cost for the cartridge element includ-
ing labor is estimated to be $146.00 per year. Costs to remove
radon from drinking water in small household application have
been determined by Lowry (Lo82). Comparison of costs to remove
low, medium and high levels of radon to produce 200 gallons of
water per day are shown in Table 16.
-28-
-------
INFORMATION NFFDS
A. Treatment Technology
General information gaps, where further research, and data
gathering needed to evaluate practical applications of treatment
technologies to remove radionuclides from drinking water have
been identified by the committee. These are listed below:
1. Information is needed on the operation and costs of the
magnesium dioxide impregnated fiber and the barium chloride
coagulation methods for radium removal.
2. Information is needed on full-scale operation and costs
of electrodialysis and radium selective complexer (RSC) for
radium removal.
3. Information is needed on full-scale operation and costs
of anion exchange, lime softening, reverse osmosis, electro-
dialysis and conventional coagulation for uranium removal.
4. Information is needed on full-scale operation and costs
involving GAC and aeration treatment for radon removal.
5. Information is needed on direct radiation from treatment
processes and devices used for radium and radon removal.
fi. Information is needed on reliability of processes in usp.
7. It would be advisable for EPA to develop a standard method-
ology for considering calculating and integrating in an objective
-29-
-------
manner the aspects of treatment cost, treatment performance pa-
rameters (reliability, operator skill, etc.), feasibility and
availability of processes that are used for recommending all MCL's
balancing the reasonableness of cost and availability of treatment
against the health risks and address, or encourage others to fill
in obvious gaps of information.
P. Waste Disposal
Oeneral information gaps, where further research and data
gathering are needed to evaluate potential waste disposal pro-
blems have been identified by the committee. These are listed
below:
1. Methods for disposal of ion exchange recharge
brines, where an adequate dilution surface
stream does not exist, need to be developed.
These may include various forms of evaporation
or other means of concentration, with subse-
quent disposal of solid residuals.
2. There is a lack of information on the specific
activities (i.e., concentrations) of uranium
that may occur in the residuals from uranium
removal systems. The backwash from anion
exchange resins is a particular case where more
detailed data is necessary.
-30-
-------
3. A primary method of choice for the removal of
radon from water appears to be aeration. The
potential for creating a source of radon ema-
nation at a water treatment facility which in
turn may require control under the Clean Air
Act needs to be examined.
The problem of waste disposal for small systems may be
especially acute. A large number of small communities discharge
their sewage into individual septic tanks. These tanks may not
be able to handle the high salinity wastes generated by ion
exchange units. In order to solve the waste disposal problem,
the community may elect to use selective sorbents or high loading
capacity resins. Upon exhaustion, the sorbent or ion exchange
media may be removed from the contractors and disposed as a solid
waste. In some instances the resin may r>e regenerated at a
central site. This method of treatment and disposal is especially
suited for point-of-use application.
Because of the accumulated radionuclides in the sorbent
or the resin, special care or procedures may have to be used in
handling and disposal. It has been shown that gamma radiation in
the surface of GAC bed used to remove radon may reach levels as
high as 16 millirem/hours. Thus, before disposal the OAC bed
should be left idle to allow most of the radon to decay. It has
been estimated that after 12 days 90% of the radon will decay.
Waste disposal requirements and limitations may considerably
-31-
-------
influence the feasibility and costs associated with the overall
treatment. Therefore, the utility should evaluate thoroughly
the available means of waste disposal associated with particular
treatment methods.
-32-
-------
REFERENCES
A175 Aldrich, L.K., Sasser, M.K. and Conners, D.A., 1975, Evalu-
ations of Radon Concentrations in North Carolina Ground
Supplies, Department of Human Resources Division of Facility
Services, Radiation Protection Branch, P.O. Box 12200,
Raleigh, North Carolina 27605.
Be76 Bennet, D.L., Bell, R.C., and Markwood, I.M., 1976, Determi-
nation of Radium Removal Efficiencies in Illinois Water Sup-
ply Treatment Processes. U.S. Environmental Protection
Agency, Technical Note ORP/TAD-76-2.
Bo82 Boyce, T.D., and Boom, S., 1982, Removal of Soluble Radium
from Uranium Minewaters by a Selective Complexer. Presented
at the SMEW—AIME Annual Meeting, Dallas, Texas, February 14-
18.
Br78 Brinck, W.L., Schliekelman, R.J., Bennet, D.L., Bell, C.,
and Markwood, I.M., Radium Removal Efficiencies in Water
Treatment Processes. Jour. AWWA 70, 1978.
Ci81 Ciccone, V.J. and Associates, 1981, Technologies and Costs
for the Removal of Radium from Potable Water Supplies,
Report prepared for Environmental Protection Agency (Draft
Report)
Co68 Cook, L.M., 1968, Advanced Technology for Radium Removal
from Drinking Water: The Flatonia Water Treatment Project,
Environmental Protection Agency Contract No. 68-01-3985.
Cr80 Removal of Barium and Radium from Groundwater, 1980, Univer-
sity of Illinois Cooperative Agreement No. CR 808912-01-1
with U.S. Environmental Protection Agency.
Dr81 Drury, J.S., Reynolds, S., Owen, P.T., Ross, R.H., and
Ensminger, J.T., 1981, Uranium in U.S. Surface, Ground and
Domestic Waters, Volume 1, EPA-570/9-81-001, ORNL/EIS-192 VI.
Dr82 Drury, J.S., Michelson, D., and Ensminger, J.T., 1982, Methods
of Removing Uranium from Drinking Water: Vol. I. A Literature
Survey, EPA-570/9-82-003, ORNL/EIS-194.
Du76 Duncan, D.L., Gessel, T.F., and Johnson, Jr., R.H., 1976,
Radon 222 in Potable Water, Proceedings of the Health Physics
Society 10th Midyear Symposium: Natural Radioactivity in
Man's Environment.
-------
Ep77 Environmental Protection Agency, Manual of Treatment
Techniques for Meeting the Interim Primary Drinking Water
Regulations, EPA-600/8-77-005, May 1977.
Gu79 Gumerman, R.C., Culp, R.L., and Hansen, S.P., 1979, Estima-
ting Water Treatment Costs, Volumes 1-3, U.S. Environmental
Protection Agency, EPA-600/2-79-162, a,b,c.
Ha82 Hathaway, S.W., 1982, "Current Research on Uranium Removal
from Drinking Water," U.S. Environmental Protection Agency,
Small Water Systems Technology Seminar.
Hi82 Hinckley, W.W., 1982, Experimental Water Treatment for a
Drilled Well with the World's Highest Known Radon 222 Levels,
Maine Department of Human Services.
Ka77 Kaufmann, R.F. and Bliss, J.D., 1977, Effects of Phosphate
Mineralization and the Phosphate Industry on Radium 226 in
Groundwater of Central Florida, Office Rad. Progr., U.S.
Environmental Protection Agency, Las Vegas, Nevada.
Ku54 Kuroda, P.K., Damon, P.E., and Hyde, H.I., 1954, Radio-
activity of the Spring Waters of Hot Springs National Park
and Vicinity in Arkansas, Amer. Jour. Sci., 252.
La83 Lassovszky, P., Hathaway, S., Treatment Technologies to
Remove Radionuclides from Drinking Water. Preconference
paper for National Workshop on Radioactivity in Drinking
Water, Easton, MD May 24-26, 1983.
he82 Bondietti, E.A., White, S.K., and Lee, S.Y., 1982, Methods
of Removing Uranium from Drinking Water: II Present Muni-
cipal Treatment and Potential Treatment Methods. EPA-570/
9-82-003, ORNL. EIS-194.
Lo81 Lowry, J.D., and Brandon, J.E., 1981, Removal of Radon from
Grounwater Supplies Using Granular Activated Carbon or
Diffused Aeration.
Mc60 McCabe, L.J., Symons, J.M., Lee, R.D.L., and Robeck G.G.,
1970, Survey of Community Waters Supply Systems, Jour. AWWA
62, 670.
Na76 National Interim Primary Drinking Water Regulations, Part
II Radionuclides, 1976, Promulgation of Regulations on
Radionuclides, Federal Register, 41 No. 133, 28402.
Na77 National Academy of Science, 1977, Drinking Water and
Health, Vol 1.
-------
Oc76 O'Connel, M.F., and Kaufmann, R.F., 1976, Radioactivity
Associated with Geothermal Waters in the Western United
States, U.S. Environmental Protection Agency Technical
Note ORP/LV-75-8A.
Pa79 Partridge, J.E., Horton, T.R., and Sensintaffar, E.L.,
1979, A Study of Radon 222 Released from Water During
Typical Household Activities. U.S. Environmental Pro-
tection Agency Technical Note ORP/EERF-79-1.
Rc81 A Study of Possible Economical Ways of Removing Radium
from Drinking Water, University of Iowa Cooperative
Agreement No. RC 81057501 with U.S. Environmental Pro-
tection Agency.
Re83 Reid, G.W. and Law, A.L., 1983, Chair Report.
Re83 Reid, G.W. and Law, A.L., 1983, Reliability of Water
Treatment Processes, Committee Report.
Sa76 Sastri, V.S. and Asbrook, A.W., 1976. Reverse Osmosis
Performance of Cellulose Acetate Membranes in the Sepa-
ration of Uranium from Dilute Solutions, Sep. Sci. 11(4):
359-374.
Sc76 Schliekelman, R.J., 1976, Determination of Radium Removal
Efficiencies in Iowa Water Supply Treatment Processes
U.S. Environmental Protection Agency, Technical Note ORP/
TAD-76-1.
Sh83 Shelley, W.J., 1983, Private Communication.
Sm61 Smith, B.M., Grune, W.N., Higgins, F.B., and Terrill, Jr.,
J.G., 1961, Natural Radioactivity in Ground Water Supplies
in Maine and New Hampshire, Jour. AWWA Vol. 53.
5080 Sorg, T.J., Forbes, R.W., and Chambers, D.C., 1980, Removal
of Radium 226 from Drinking Water by Reverse Osmosis in
Sarasota County, Florida, Jour. AWWA. 72, 1980
5081 Sorg, T.J., 1981, Process Selection for Small Drinking
Water Supplies, 23rd Annual Public Water Supply Engineers
Conference, University of Illinois, April 21-23.
Te82 Private Communication, Texas Department of Health, 1982,
Austin, Texas Private Communication
-------
TABLE CAPTIONS
1.
Table
1
Boundary Levels, Treatment, Source, User
2.
Table
2
Summary of Process, Levels of Concentration
of Removal
3.
Table
3
Radium Removal in Ion Exchange Plants
4.
Table
4
Radium Removal in Reverse Osmosis Plants
5.
Table
5
Radium Removal in Lime Softening Plants
6.
Table
6
Iron and Manganese Removals by Iron and
7. Table 7
8. Table 8
9. Table 9
10. Table 10
11. Table 11
12. Table 12
13. Table 13
14. Table 14
15. Table 15
16. Table 16
Manganese Removal Processes
EPA Uranium Removal Field Study
Removal of Uranium from Pond Water by
Ca(0H)2 Treatment
Suspected Uranyl Species and Charge
Characteristics of Iron and Aluminum
Hydroxide Floes at Given pHs of Pond
Water
Percent Uranium Removal by Fe2(804)3, FeS04,
and Al2 (SO4)J Coagulants with Varying pHa
EDR Test/Analytical Means
Summary of Radium Removal Technology
Summary of Uranium Removal Technology
Summary of Radon Removal Technology
Summary of Alternatives to the Disposal and
Handling of Treatment Wastes from Lime
Softening, Ion-Exchange, and Reverse Osmosis
Summary of Performance and Economics of
Diffused Aeration and Granular Activated
Carbon to Remove Radon from Water for House-
hold Use (200 gpd Demand)
-------
Icible 1
Boundary Levels, Treatment., Source, User
Element
Raw (pCi/L)
Finish (pCi/L)
Percent Reduction by Treatment
Radium
200
20
150
10
50
5
90
Uranium
200
100
100
40
50
10
95
Radon
750,000
20,000
100,000
10,000
10,000
5,000
95
Landfill
Land Disposal
Ocean Disposal
5 pCi/gm up to 100 pCi/gm (landfill) (under RCFA)
2#/acre/ (based on Cd)
Trace Level Requirement
-------
Table 2
Summary of Process, Levels of Concentration of Removal
Removal
Process Efficiency
Ion-exchange 95% Ra
Effect of Plant
Size
Ion exchange is generally
used in batch process.
Small plants are more cost
effective. The available
cost data only covers plant
capacity up to 10 MGD.
Problems
Treated water shows an
actual increase in
total dissolved solids.
Sodium concentrations
may be elevated.
Raw water requires pre-
treatment if turbidity
and suspendment if
turbidity and suspended
solids, iron and manga-
nese, or bacterial slimes
are present
Treated water may be
corrosive
Disposal of spent brines
can be a serious problem.
Capital cost is substantial.
99% U Experimental Stage
Anion
Exchange
Reverse Osmosis 95% Ra
This process is most
suitable for automated
plant operation and use
in small plants.
High capital and operating
cost.
Considerable pretreatment
requirements particularly
if raw water contains
suspended solids, organic
material, or dissolved
gases.
Reject stream requires
disposal.
Treated water must be
stabilized.
-------
Table 2
Sunmary of Process, Level of Concentration of Removal (Continued)
Process
Removal
Efficiency
Effect of Plant
Size
Problems
Lime and Lime-
Soda Softening
80—90% Pa Best for large plant
capacity over 10 MTCO
This process is the
Choice for large
treatment plants.
Process is more difficult
to control than ion-
exchange or reverse
osmosis.
Operating costs are high,
particularly chemical costs.
System size limitation.
Requires significant
operation attention.
If hardness is not a
problem in the water the
lime-soda softening pro-
cess is not cost-effective
to remove just radium
Greensand
50% Ra
Water having less than 10
pCi/L radium may be treated
with this technique.
Requires disposal of waste
material.
Mn-Fiber
90% Ra
Testing stage
laboratory scale
Manganese oxides are not
completely hound to the
fiber and up to 50% may he
washed off.
The fiber is difficult, to
prepare and handle.
-------
Table 3
PADiriM PEMOVAL IN TON FXCHAMGF PLANTS
Plant
Ra in
pCi/L
Pa out
pCi/L
%Ra
Removed
Hardness
in
mq/lCam-}
Hardness
out
mg/lCaCo-*
Hardness
*
Removed
Eldon, TA
49
1.9
96
375
]0
97
Estherville, IA
5.7
0.3
95
915
46
95
Grinnell, IA
6.7
0.2
97
385
11
97
Ho]stein, IA
12
0.5
96
920
18
98
Dwight Corr.
Inst., IL
3.26
0.36
89
286
43
85*
Hercher, IL
14.31
1.31
91
401
60
85t
Lynwood, IL
14.69
0.41
97
848
78
91
* Removed
t Hardness and %Pa removals are somewhat low due to breakthrough occurring prior to all
samples being collected.
Source: Rennet, D.L, et at. (1976), Schliekelman, P..7. (1976), Brinck, W.L., et al.
(197R)
-------
Table 4
Radium Removal in Reverse Osmosis Plants
Ra in Ra out % Ra
Plants pCi/L PCi/L Rem.
Greenfield, IA 14.0 0.6 96
Sarasota, FL 22.0 0.8 96
Bay Lakes Estates 3.2 0.14 95.6
King Gate 15.4 2.0 87
Spanish Lakes 10.5 1.2 88.5
Sorrento Shores 4.6 0.21 95.4
Venice 3.4 0.6 82.4
Bay Front 12.1 0.6 95
Source: Song, T.J. (1980), Bennet, D.L. et al. (1976).
-------
Table 5
Radium Removal in Lame Softening Plants
Location
Ra In
pCi/L
Ra Out
pCi/L
% Ra
Removed
PH
Des Moines, Iowa
9.3
2.35
75
10.4
Webster City, Iowa
6.1-7.8
0.3-0.9
85-96
10.05-10.95
Peru, Illinois
5.48-6.49
0.51-1.62
70-92
8.2
Elgin, Illinois
3.51-7.45
0.71-0.80
80-90
10.2
Venice, Florida
8.73
2.19
75
9.7
Enqlevood, Florida
1.69
0.69
59
8.5
Source: Bennet, D.C. et al. (1976), Schliekelman, R.J. (1976), Brinck, W.L.
et al. (1978).
-------
Table 6
Ra-226, Iron and Manganese Removals by Iron Manganese Removal Processes
Ra-226 (pCi/L) Iron (mg/1) Manganese (Mg/1)
City
PH
Raw
Treated
Removal
Paw
Treated
%Removal
Raw
Treated
^Removal
Adair
6.7-6.9
6.9
6.7
3
0.5
0.01
R0
0.01
0.01
-
Eldon
7.R
49
43
12
2.0
0.3
85
0.01
0.01
-
Rstherville
7.7
5.7
5.1
11
2.0
0.67
66
0.24
0.27
-
Rrinnell
7.6
6.7
5.7
15
0.7
0.41
42
0.01
0.01
-
Herscher
7.6-8.3
14.9
6.6
56
0.2
0
_
0.47
0.02
Q£
14.5
6.4
56
0.4
0
-
0.41
0.01
98
14.9
6.9
54
0.1
0
-
0.48
0.01
98
14.3
6.9
52
0.1
0
-
0.39
0
ion
14.0
6.9
51
0.1
0
-
0.45
0
100
13.9
6.8
51
0.1
0
-
0.63
0
100
13.9
7.3
47
0.2
0.1
-
0.44
0.13
70
14.1
6.3
55
0.1
0
-
0.53
0.02
96
14.3
6.5
55
0.1
0
—
0.50
0
100
Holstein
7.4-7.6
13
7
46
1.8
0.09
95
0.15
0.01
93
Stuart
7.6-7.9
16
12
25
0.94
0.03
97
0.01
0.01
-
Source: Pennet, D.C. et al. (1976), Schliekelman, P.J. (1976), Rrinck, W.L. et al. (1978)
-------
Table 7
EPA Uranium Removal Field Study by Anion Exchange
Raw Water*
Uranium
(microgram/liter)
Treated Water
Uranium
(microgram/liter)
Bed Volumest
Treated
Capacity-'"''
gran\/liter
Ft. Lupton, CO
35.
35.
ll,867t
0.112
Brighton, 00
23.
23.
24,181t
0.144
Marshdale, 00
28.
<0.1
21,601
Cave, AZ
64.
63.
16,702f
0.272
Church Rock, NM
52.
0.1
10,830
*Uranium 238/235 ratio not determined. If uranium 238/234 is l/l then 1 microgram/liter
equal 0.67 pCi/L
tBed Volumes treated When effluent first reaches influent uranium level
1 BV = 1.88 gallons
ttGrams uranium ranoved per liter of resin (g/L)
-------
Table 8
Removal of Uranium From Pond Water by Ca(0H)2 Treatment3
Ca(OH)s)
doses (mg/L)
50
100
150
200
250
% U Removed
86
85
87
87
90
Final pH
10.6
11.1
11.3
11.5
11.5
aInitial U Concentration: 83 micrograms/liter.
Source: Bordietti, et al., 1982.
-------
Table 9
Suspected Uranyl Species and Charge Characteristics of Iron and
Aluminum Hydroxide Floes at Given pHs of Pond Water
Adjusted pH
4
6
9
10
Uranyl Species
uo22+
uo2oo§
uo2(oo)^-
(U02)3(QH)§
Charges of Floes
X+
(Fe,Al)(0H)3-x
0
(Fe,Al)(PH)3
y-
(Fe,Al)(0H)3-ty
z-
(Fe,Al)(0H)3+z
Uranium Removal
Low (30)
High (88)
Low (48)
High (87)
Source: Bondietti, et al., 1982
-------
Table 10
Percent Uranium Removal by Fe2(804)3, feS04, and A12(S04>3
Coagulants with Varying pHa
Initial
(%)
pH
(mg/L)
F22(S04)3
FeS04
A12(S04)3
Fe2(S04)3
FeS04
A12(S04)3
4
0.5
7
6
7
4.1
4.2
4.4
5
14
8
9
4.3
4.2
4.2
10
8
11
6
3.8
4.1
4.4
15
13
21
15
4.0
4.2
4.7
20
17
26
21
4.1
4.2
4.8
25
18
33
21
4.0
4.1
4.8
6
0.5
16
14
7
6.2
6.1
6.2
5
43
24
30
6.4
6.0
6.2
10
63
33
51
6.2
6.1
6.1
15
76
42
69
6.2
6.2
6.1
20
84
52
80
6.1
6.2
6.1
25
89
44
88
6.2
6.2
6.2
8
0.5
1
6
0
8.4
8.1
8.0
5
4
7
2
8.2
8.1
7.9
10
17
12
9
7.9
8.1
7.9
15
21
11
17
8.0
8.1
7.9
20
33
15
25
7.9
8.0
7.9
25
43
20
48
7.8
8.0
7.8
10
0.5
1
2
8
10.0
10.1
10.0
5
27
32
71
10.0
10.0
9.9
10
83
57
95
9.9
10.0
9.8
15
86
84
98
10.0
10.0
9.7
20
80
92
98
9.5
9.9
9.7
25
87
93
96
10.0
9.9
9.7
Source: Bondietti, et al., 1982
-------
Table 11
EDR Test/Analytical Means
Feed
Brine
Product
Rejection
Radiometrics
pCi/L
pCi/L
pCi/L
Ra 226
667
2,904
64
90.4%
Th 230
54
415
10.0
81.5%
Gross
735
3,294
149
79.7%
Gross
2,182
4,390
379
82.6%
Source: Personal correspondence with Ionics, Inc.
-------
Table 12
Summary of Radium Removal Technology
123 4 5 67 8 9
Ion Exchange
F
Yes
Yes
81-97%
S,M
B
50-3500 pCi/L
M
Be76, Br78
Sc76, E£>77
2
Na -
Corrosion -
Other Ions Removed +
Lime Softening
F
Yes
Yes
80-90%
M,L
S
1-10 pCi/g
L
-
8
Na -
Corrosion -
Cont. Removal +
Softening +
R.O.
F
Yes
Yes
90+%
S,M,L
L
2:1 3:1
H
So80
3
Corrosion -
TDS +
E.D.R.
F
Yes
Yes
90%
S,M,L
L
3:1
H
-
4
Corrosion -
¦IDS +
Manganese/iron
Removal
F
Yes
Yes
25-50%
S,M/L
L
?
-
EJ>77
4
Iron and Manganese
Removal +
MNC>2 Imp. Fiber
L
New
No
?
90-%
S
S
?
-
Co68
-
-
R. S.C.
L,F
New
No
90+%
S,M, L
S
?
-
Bo 82
-
-
Barium
?
New
-
-
-
-
-
-
7
Adv. Health Effects
-------
Table 13
Sunmary of Uranium Removal Technology
1234 5 678 9
Anion Exchange
L,P
Yes
90+%
S,M
B
-
(Bo82, Ha82)
3
?
Lime Softening
Yes
80-90%
M,L
Lot of
S,L
-
(fio82)
9
Na -
Corrosion -
Cant. Removal +
Softening +
R.O.
L
Yes
90+%
S,M,L
L
-
(Sa76,Bo82)
3
Corrosion -
TDS +
£iD»R»
Mo
Yes
-
S,M,L
L
-
-
3
Corrosion -
TDS +
Coagulation
L,F
Yes
80%
M,L
S,L
-
(Bo82)
8
Complex -
Cont. Removal +
Activated
Alumina
L
Yes
90+%
S,M
L
M
(Ha82)
7
-
-------
Table 14
Summary of Radon Removal Technology
1234 5 6 78 9
G.A.C.
L,P,F
Yes
90+%
S,M
S
-
(Lo82)
2
Radon Daughters-
Rad. Licensing -
Aeration
L,P,F
Yes
90+%
S,M,L
G
-
(Lo82)
2
Radon in Air -
-------
LEGEND For Tables 12, 13, 14
1 - Demonstrated Technology
F - Full Scale
P - Pilot Plant
L - Laboratory
2 - Availability
Yes
No
3 - Efficiency
Percent Removal
4 - Size
S - Small
M - Medium
L - Large
5 - Disposal of Waste
L - Liquid
S - Sludge
B - Brine
A - Solid
6 - Reliability
L - Low
M - Medium
H - High
7 - Cost - Capital, Operating and Maintenance
8 - Complexity - 1 least complex 10 most complex
9 - Benefits and Disadvantages
R.O. - Reverse Osmosis
E.D. - Electrodialysis
R.S.C. - Radium Selective Complexes
G.A.C. - Granular Activated Carbon
-------
Table 15. Summary of Alternatives to the Disposal and Handling of Treatment Wastes
Fran Lime Softening, Ion-Exchange, and Reverse Osmosis
Discharge
-To sanitary sewer
-To local receiving waters
a) streams
b) oceans
-By wet pumping or trucking to local sanitary landfill
Storage
-Permanent Lagooning
-Sanitary Landfill
a) with prior temporary lagooning
b) with prior mechanical dewatering: vacumn filtration, centrifu-
gation, pressure filtration, belt filter pressing, and dual-cell
gravity solids concentrations
-Other natural or man made depressions (all with dewatering before
transportation)
a) strip mine areas
b) borrow pits and quarries
c) others
-Tanks or lagoons
a) for settling and decanting into receiving water
b) for settling and pumping supernatant back to plant
-Evaporation lagoons
-Land spreading
Use
-Direct without drying: farmland and pasturelands
-With prior dewatering
a) farmland and pasture land
b) road stabilization
Disposal
-Direct, recharge to aquifers
-With prior dewatering: salt mines, coal mines, and so forth
-As nuclear waste
-In deep aquifers
-In oil well fields
Source: Environmental Protection Agency, 1977, Manual of Treatment Techniques for
Meeting the Interim Primary Drinking Water Regulations, EPA-600/8-77-005
pp. 65-66, Cincinnati, Ohio
-------
Table 16
Summary of Performance and Economics of Diffused Aeration
and Granular Activated Carbon to Remove Radon frcm Mater
for Household Use (200 gpd Demand)
Influent Rn
pCi/L
FCi/L
Effluent
Rn GAC Aeration
Cost (Estimated)
Capital
Operating
GAC Aeration
GAC Aeration
15.000
30,000
150,000
1350-3300 750
2700-6600 1500
1200 (2) <7500
§431-757 $890
$431-757 $890
$1500 $1000
$19 $60
$19 $60
$40 $80
Source: Lowry, J.D., 1983.
-------
FIGURE CAPTIONS
1. Figure 1 Schematic Defining Adjustments Between the
Occurrence of Radionuclide in Water and User
Safety and Health Requirements.
2. Figure 2 Chart Illustrating the Effectiveness of
Uranium Removal by Ion Exchange (Bo82)
3. Figure 3 Chart Illustrating the Effectiveness of
Uranium Removal by Ion Exchange in a Pilot
Plant (Ha82)
4. Figure 4 Chart Illustrating the Effectiveness of
Uranium Removal by Activated Alumina in a
Pilot Plant (Ha82)
5. Figure 5 Chart Illustrating the Costs to Remove
Radionuclides from Drinking Water by
Various Treatment Techniques. $/1000
gallons (Ci83)
6. Figure 6 Chart Illustrating the Operating and Mainte-
nance Costs Associated with the Removal of
Radionuclides from Drinking Waters by
Various Treatment Techniques. $l/year
(Ci83)
7. Figure 7 Chart Illustrating the Capital Costs
Associated with the Removal of Radionuclides
from Drinking Waters by Various Treatment
Techniques. $/year (Ci83)
-------
USER
SOURCE
COST >
EFFECTIVE
ADJUSTMENTS]
Figure 1.
Schematic Defining Adjustments Between the Occurence of
Radionuclide in Water and User Safety and Health Require-
ments .
-------
URANIUM REMOVAL BY ANION EXCHANGE
RAW WATER URANIUM=83 UG/L
A—A—A—A'
1 11
T
T
T
T
T
T
BED VOLUMES THROUGH COLUMN
Figure 2: Chart Illustrating the Effectiveness of Uranium Removal
by Ion Exchange (Bo82).
-------
PILOT PLANT URANIUM REMOVAL
ION EX. A RUN 1-2
300-;
E
F
250-
F
•
200^
U
R
A
150-
N
I
U
100-
M
u
50-
o
-
/
1
0-
0- » iMH< » »H» » » » 44 » »W« »» » » » »-
T
e
T
T
T
3800
6000 9000 12000
BED VOLUMES
RAW WATER URANIUM=300(ug/l)
1 I ¦
15000
T
18000
Figure 3. Chart Illustrating the Effectiveness of Uranium Removal
by Ion Exchange in a Pilot Plant (Ma82).
-------
PILOT PLANT URANIUM REMOVAL
ACT. ALUMINA BUN 1
F 240
U 180
A 150
I 120
1000
BED VOLUMES
RAW WATER URANIUM=273(ug/l)
Figure 4. Chart Illustrating the effectiveness of Uranium Removal
by Activated Alumina in a Pilot Plant (Ma82).
-------
REvE
CTUSI
AmJE
tRATHQ
>.0 10.0 lOfl
AVERAGE FLOW MGD
ODD
Figure 5. Chart Illustrating the Costs to Remove Radionuclides from
Drinking Water by Various Treatment Techniques. S/gallons
(Ci 8 3)
Source: Ciccone and Associates, 1983.
-------
E OSIV
TENING
ATI
ON
i
cr
>
x
&
CO
ANION
AN
H-
CO
O
O
I
J
ILT
LIME
OF
ING
10-
L'TRATI
IOO
I.D
O.I
AVERAGE FLOW _MGD
Figure 6. Chart Illustrating the Operating and Maintenance Costs
Associated with the Removal of Radionuclides from Drinking
Waters by Various Treatment Techniques. $/year (Ci83)
Source: Ciccone and Associates, 1983.
-------
L ME ;S:gFTE^rfvn?]TT7=7;
DAG./ FlLHiRiftTlONl II 'V
iy i si iinf i 111mi i i 111"Tm
\\s" \ ' InLIME SOFTgNlNG MQDlli'l
yr j I III 7 I I I I IT I r i i i! Ill
CAPACITY MGD
Figure 7. Chart Illustrating the Capital Costs Associated with the Removal
of Radionuclides from Drinking Waters by Various Treatment
Techniques. $/year. (Ci83)
Source: Ciccone and Associates, 1983.
-------
COMMITTEE ON HEALTH EFFECTS OF RADON
Chairman: F.T. Cross
Recorder: Neal Nelson
Committee Members: Naomi H. Harley
Werner Hofmann
-------
HEALTH EFFECTS AND RISKS FROM RAD0N-222 IN DRINKING WATER
F. T. Cross
Biology and Chemistry Department
Pacific Northwest Laboratory
P.O. Box 999
Richland, WA 99352
N. H. Harley
Environmental Medicine
New York University Medical Center
500 First Avenue
New York, NY 10016
W. Hofmann
Abteilung fUr Biophysik
Universitat Salzburg
Erzabt-Klotz-Strasse 11
A-5020 Salzburg, Austria
ABSTRACT
This paper presents an evaluation of the inhalation and ingestion
doses from exposure to radon and radon daughters; an overview of the human
and animal health-effects data; estimations of the cancer risks from radon
and radon daughter exposures; and suggested limits for radon concentrations
in drinking water and indoor air. We suggest that a rounded radon-in-water
concentration limit of 10,000 pCi/L can be supported by health-effects con-
siderations alone, based on the conservative "tolerance dose" concept and
other conservative assumptions regarding lung dose. A practical concentra-
tion limit (or action level) of 20,000 pCi/L has been derived by estima-
tions of exposure distributions in the U.S. and in relation to current EPA
standards for uranium-tailings-contaminated buildings. Research needed for
resolution of the uncertainties in these estimates is suggested. We con-
clude that before a maximum contaminant level (MCL) for radon in water can
be firmly established, the broader issue of setting the MCL for radon in
indoor air must be addressed.
-------
INTRODUCTION
The inert, noble gas radon is found everywhere and sometimes occurs
naturally in elevated concentrations that exceed or are substantial frac-
tions of existing regulatory exposure standards for workers. Radon is
soluble in body fluids and fats and, therefore, presents a potential hazard
to the whole body. Radon in water presents a dual pathway for exposure of
individuals: by ingestion, from direct water consumption, and by inhala-
tion exposure when radon emanates from water. As developed in the follow-
ing sections, the dose to the respiratory system outweighs the dose to
other organ systems; therefore, derived limits on exposure to radon in
drinking water are based mainly on the inhalation risk to lung.
2
-------
DOSIMETRY CALCULATIONS
Bale (1951) and Harley (1953) were the first to point out that the
lung-cancer hazard from inhalation exposure to radon and radon daughters
was not from the radon per se but rather from the alpha-dose delivered
218
through lung deposition of the short-lived daughters of radon [ Po(RaA),
214Pb(RaB), 214Bi(RaC), and 214Po(RaC')]. Two alpha-emitters, 218Po(RaA)
214
and Po(RaC'), ultimately deliver the carcinogenic dose to basal cells
of the bronchial epithelium, the presumed critical tissue for induction of
lung cancer. The complexity of the dose estimates (required to account for
daughter deposition, radioactive buildup and decay, removal by physiologic
clearance processes, and physical dose calculations to specific cells in
bronchial mucosa) has been detailed by many authors and considered by
various national and international organizations (see A164; Jac64, 72, 77,
80; Haq66, 67a, 67b; Pa69; Wal70, 71, 79; Har72, 81, 82; Ne74; Fr77; McP79;
Jam80, 81; Ho82a, 82b; Wi82 and USPHS57, 61; FRC67; JCAE67, 69; ICRP77, 81;
UN72, 77; NI71; NAS72, 80; NCRP84a).
Historically, inhalation exposure is defined in terms of the air con-
centration of radon daughters in working level (WL) units. A working level
is defined as a concentration of short-lived radon daughters (through RaC')
5
totaling 1.3 x 10 MeV of potential alpha energy per liter of air. A work-
ing level month (WLM) is an exposure equivalent to 1 WL for 170 hours.
These definitions avoid the problems of disequilibrium of the daughters and
also that of whether the daughters are attached (or are unattached) to a
carrier aerosol. Attached radon daughters deposit with a few percent
3
-------
probability to the respiratory tract surfaces, whereas unattached radon
daughters deposit in the respiratory tract with nearly 100% probability.
Thus, the mix of attached and unattached radon daughters is an important
consideration in assessing lung dosimetry. Fortunately, the mean unattach-
ment fraction values found in the workplace and in the environment are
reasonably constant and sufficiently similar that they do not cause a large
disparity in the radiological dose assessment of environmental and occupa-
tional exposures to radon daughters. The differences in other parameters
that influence the radon-daughter lung dose, such as differences in
daughter product equilibrium, particle-size distributions, breathing
patterns, bronchial morphometry and physiological clearance processes, tend
to produce somewhat compensatory doses to the basal cells of the bronchial
epithelium. Thus, the pertinent radiological doses remain reasonably
invariant with activity and environment and, to some extent, with age.
Inhalation Dose
The more recent radon-daughter lung-dosimetry models, which are in
substantial agreement with one another, place the bronchial epithelium
exposure-to-dose conversion factor at about 0.5 rad/WLM for uranium miners.
The dose per unit cumulative exposure has also been derived for environ-
mental conditions under short-lived radon-daughter equilibrium factors^
^The equilibrium factor is the ratio of the total potential alpha energy
of the actual short-lived daughter concentrations to the total potential
alpha energy that the daughters would have if they were in equilibrium
with radon.
4
-------
of about 0.7 (Har81); the finding is in close agreement for the adult male
(0.71 rad/WLM), adult female (0.64 rad/WLM), the 10-year-old child
(1.2 rad/WLM) and the 1-year-old infant (0.64 rad/WLM). Hofmann (Ho82a)
has also found that basal-cell doses in the bronchi for infants and chil-
dren are higher than for adults by about a factor of two to three in the
first 10 years of life. Subsequent calculations by Harley and Pasternack
(Har82) for a five-lobed human lung have yielded similar dose-factor
values. The small differences in the bronchial dose for the miners and for
those receiving environmental exposures primarily reflect reduced breathing
rates during environmental exposures, differences in lung morphometry,
differences in particle size, and the increased percentage of unattached
RaA in ordinary atmospheres (^7% environmental vs ^4% in mines). These
conversion factors indicate that a cumulative exposure in the nonmining
environment is somewhat more effective in delivering a radiation dose to
basal cells of the bronchial epithelium than are exposures under working
conditions in a mine.
Although it seems intuitive that the reduction in breathing rate (and,
therefore, radionuclide intake), which occurs under environmental exposure
conditions, should reduce the dose to bronchial epithelium, the deposition
onto the tracheobronchial tree increases due to the lower rate of airflow.
This, combined with other compensatory differences between environmental
and occupational exposure conditions, tends to result in somewhat compara-
ble doses to basal cells for the two exposure situations.
Alternative dosimetry models (e.g.,Jac80; Jam81; Ho82a) are based on
slightly different assumptions regarding particle sizes, radon-daughter
5
-------
unattachment fractions, breathing rates, anatomical structure, deposition
equations and mucociliary transport rates. Slight changes in these
physical and physiological parameters do not significantly alter the
bronchial deposition and dose patterns. Differences among the models are
greatest in the assumptions on basal cell depth. In the Harley-Pasternack
models (Har72, 82), doses are calculated for the shallow-lying basal cells
at a 22-ym constant depth below the epithelial surface, resulting in the
doses being highest in the upper bronchial generations. The later calcula-
tions by Harley and Pasternack (Har82), using the Yeh-Schum morphometry
(Ye80), show a more uniform distribution of dose than the earlier calcula-
tions which were based on the Wei be! dichotomous model (We63). In agree-
ment with Altshuler et al. (A164), the shallow-lying basal cells are con-
sidered by Harley and Pasternack to be the target cells in bronchogenic
carcinoma. The other models utilize the basal-cell-depth distributions of
Gastineau et al. (Gas72), which show decreasing epithelium thickness as the
bronchial-tree branches get smaller. The variable depth assumption results
in a relatively uniform basal-cell dose distribution.
In spite of the different assumptions in the models, the adult
exposure-to-dose rounded, conversion factors lie within a relatively narrow
range of values: 0.4 to 0.6 (Jac80), 0.3 to 1.0 (Jam81), 0.3 to 0.8
(ICRP81), 0.6 (Ho82a) and 0.5 to 0.7 rad/WLM (Har81, 82). Because con-
clusive evidence is lacking concerning the location of the relevant target
cells, a reasonable, mean, environmental-conversion factor of 0.7 rad/WLM
for bronchial epithelium has been adopted for this paper. We estimate that
this value is known to within a factor of about 2.
6
-------
In some models of risk for radon-daughter exposure (e.g., EPA80) there
has been a tendency to artificially lower the cumulative exposure in the
environment, presumably to account for the influence of decreased breathing
rates on mean lung dose under nonworking conditions. In our opinion, this
is neither warranted nor justifiable in view of the general agreement, at
least among U.S. modelers, that the relevant doses accrue to bronchial
epithelium and not to lung as a whole, and in view of the compensating
factors mentioned above. Therefore, whether the inhalation exposure is
environmental or occupational, the WL is given comparable weight in our
subsequent treatment of risk.
The effective-dose-equivalent system of the ICRP (ICRP81) assigns
equal weights to the cell dose in the bronchial and pulmonary epithelium
for inhalation exposures. Although their exposure-to-dose conversion
factors for bronchial epithelium (0.32 to 0.85 rad/WLM) are similar to
other values for miners, the Commission's effective-dose-equivalent system
has not been adopted for this paper. Instead, the critical organ/critical
tissue approach of the NCRP, as reflected in their Report No. 39 (NCRP71),
is utilized in our treatment of risk. For the lung, the tissue of concern
is the bronchial epithelium.
Ingestion Dose
While many investigators have addressed the dose to lung from inhaled
radon and radon daughters, comparatively few have calculated the internal
doses from ingested radon and radon daughters. Suomela and Kahlos (Su72)
estimated the dose-equivalent based on whole-body-counting measurements of
7
-------
human volunteers. Using an alpha quality factor of 10, they derived doses
222 222
to the stomach wall of 240 mrem/yCi Rn and 380 mrem/yCi Rn, for full
and empty stomachs, respectively. These values convert to 480 and
222
760 mrem/yCi Rn, using the currently recommended value of 20 for the
alpha quality factor.
Hursh et al. (Hur65) also identified the stomach wall as the tissue
222
which receives the greatest radiation dose from ingested Rn. They state
that the dose contribution from the total decay of the ingested short-lived
daughters is negligible compared with that from the radon itself. Their
222
calculated doses, converted to current rem units, are 412 mrem/yCi Rn
222
and 440 mrem/yCi Rn, for full and empty stomachs, respectively.
The estimates of von Do'beln and Lindell (Vo64) include whole-body
222
doses. Their converted values are 400 mrem/yCi Rn for stomach and
222
4 mrem/uCi Rn for whole body. The converted whole-body dose equivalent
222
of Andersson and Nilsson (An64) is somewhat higher: 14 mrem/yCi Rn.
Current evaluations of these doses appear to be based on the earlier
work (e.g., EPA77, Ka80, Sul82). The analysis of radon-ingestion models by
Sullivan and Nelson (Sul82) concludes that, in spite of the many papers on
radon ingestion, experimental measurements are few and somewhat contra-
dictory. They further state that our present knowledge of the behavior of
ingested radon is insufficient to support a proposed maximum contaminant
level. The uncertainties identified were: (1) the unknown transit time of
radon through the gastrointestinal tract wall, (2) the identification of
the organs receiving the highest doses, (3) the variability in whole-body
8
-------
radon retention, and (4) the unaccountability for all the radon putatively
ingested—up to 50% could not be accounted for in some of the experiments.
In view of the many uncertainties in the data, it is not surprising
that various investigators have apparently assumed inordinately long reten-
tion times for radon in the stomach or other components of the gastroin-
testinal tract, in an attempt to be conservative in their estimates of the
ingestion doses. Sullivan and Nelson (Sul82) estimated that the ingestion
doses to the gastrointestinal tract may be overestimated by as much as
two-orders-of-magnitude and, therefore, they do not consider present data
adequate for assessing radon-ingestion doses.
Using the previous ingestion dose estimates for illustration purposes,
222
reasonable values range from about 400 to 800 mrem/yCi Rn to the stomach
222
and from 4 to 14 mrem/yCi Rn to the whole body. Reasonable mean values
222 222
might be 600 mrem/yCi Rn and 10 mrem/yCi Rn, respectively, to the
stomach and whole body. These values can be converted to annual dose
equivalents if we assume consumption of 0.5 L of untreated (including
unheated and nonaerated) water per day. [The amount of untreated water
consumed is considered by Suomela and Kahlos to vary between 300 and
1200 ml/day; however, many investigators prefer to use the lower number.
The ICRP (1975) estimates daily tap-water consumption at 150, 100 and
200 ml for adult reference man, woman and 10-yr old child, respectively.]
For a radon concentration in water of 1000 pCi/L, the calculated annual
dose equivalent is about 100 mrem to stomach and 2 mrem to whole body.
9
-------
If we use Kahlos and Asikainen's (Ka80) mean transfer coefficient of
-4
10 for radon in water to radon in house air, 1000 pCi/L in water con-
verts to 0.1 pCi/L in air. Using 0.7 rad/WLM as a reasonable lifetime con-
version coefficient for environmental exposures, an alpha quality factor of
20, an equilibrium factor of 0.5, and continuous exposure, the 0.1 pCi/L
converts to a rounded bronchial epithelium dose equivalent of 400 mrem/yr.
-4
The value of about 10 for the mean radon transfer coefficient has been
confirmed by Hess (Hes82) for houses in Maine with 1 air change per hour.
In the UNSCEAR 1977 report (UN77), the mean radon transfer coefficient was
calculated to be 2 x 10"^ for a ventilation rate of 1 h"1. Because some
houses have lower ventilation rates, it is conservative to assume that the
-4 -4
mean radon transfer coefficient ranges from about 1 x 10 to 2.5 x 10 .
The dose-equivalent to bronchial epithelium for 1000 pCi/L in water,
therefore, ranges from about 400 to 900 mrem/yr. Thus, the estimated dose
to bronchial epithelium from radon in water is substantially higher than
the estimated ingestion doses. The actual amount of untreated water drunk
per day is, therefore, relatively unimportant, because the lung dose,
multiplied by the risk factor for lung, compared to the same products for
stomach and whole body, determine the concentration limits for radon in
drinking water.
10
-------
RADON DAUGHTER EPIDEMIOLOGY STUDIES
Introduction
Data from a variety of occupational and medical exposures, and from
disasters, have clearly implicated ionizing radiation as a human car-
cinogen. While there are many epidemiology studies addressing these types
of exposure, comparatively few studies have investigated the exposure to
natural background radiation sources. Even those few show no significant
increase in lung-cancer death rate from inhalation exposure to normally
occurring levels of radon and radon daughters. Positive correlations exist
22fi
for Ra levels in drinking water and cancer incidence (e.g., Be82), but
an unequivocal association of radium and radon in drinking water and cancer
incidence has yet to be demonstrated.
The most notable example of nonmining radon exposures are thorotrast
patients whose lungs continually receive a constant, low-level alpha expo-
220
sure from Rn and its daughters. A higher risk of lung tumors has not
yet been demonstrated in these patients, who have received estimated doses
to the large bronchi of 300 rad over a 30-year period (Va78, 83). Although
222
these exposures are not strictly comparable to Rn daughter exposures, a
222
Rn-daughter exposure which produces about 300 rad to the bronchial
epithelium ranges between 400 and 600 WLM, depending on whether the
exposures are to environmental radon or radon in the mines. These levels
of exposure have an upper estimated lifetime lung cancer risk of about
-1 -3
10 , using the lung cancer risk factor of 5.6 x 10 for persons exposed
to 1 WLM/yr for life that is developed later in this report.
11
-------
The epidemiological data derived from many types of underground mining
show a relatively consistent relationship between lung-cancer incidence
(which is similar to the death rate from lung cancer) and exposure to radon
daughters. This underlying consistency is considered to be related to the
relatively narrow range of bronchial dose per WLM under varying exposure
conditions.
It is difficult to assess the risk of attributable lung cancer through
human epidemiological studies because the detailed information required is
not always available. In the ideal case, the exposure of each miner, as a
function of time, would be available; and the follow-up period would be
long enough for all of the group to have died from lung cancer or other
causes. From such ideal data, attributable lung cancers could be separated
from those arising spontaneously or from cigarette smoking. The cumulative
exposure, person-years at risk, and the number of attributable lung cancers
would allow the exact calculation of a risk factor.
In reality, the data do not fulfill these requirements: estimates of
exposure were often crude, and follow-up periods are not sufficiently long.
Nevertheless, recognizing the limitations of the data, it is possible to
estimate a mean risk factor which we can accept until improved data and
further studies provide more firmly based estimates of risk.
Human data are now available from several groups of underground
metal-ore miners: those in the U.S., Canadian and Czechoslovakian uranium
mines, Swedish and British iron mines, Swedish lead and zinc mines, and
Newfoundland fluorspar mines. Although other potential carcinogens (such
as diesel smoke, traces of arsenic or nickel, and iron ore) are found in
12
-------
these mines, the lung-cancer response appears to be predictable, based upon
radon-daughter exposure. Some studies have divided the workers into sub-
groups on the basis of estimated radon daughter exposure. Eighteen sub-
groups were selected (Ar79) as most suitable (considering both epidemio-
logical and environmental data) for quantitative assessment of the lower
exposure levels. In addition, data on these mining populations have been
reviewed by other authors and organizations (see NI71; NAS72, 80; Se76;
Jo73; Ax78; Sn73, 74; Ren74; Dev64; Wr77; McP79; UN77; Ev81; Ra81;
NCRP84a).
Discussion and Summary
Present data suggest that an absolute threshold exposure for lung-
cancer induction is highly unlikely. [This is also in keeping with present-
day views, in radiation biology and radiation protection, that radiation-
induced cancer is a stochastic (nonthreshold) process.] Evans (Ev67) and
Stranden (Str80) argue that the lung-cancer mortality data at the lowest
reported exposures are not statistically different from expected and that
at least a "practical" threshold for radon-daughter carcinogenesis may
exist. Archer et al. (Ar79) conclude from their analysis of the 18 sub-
groups that, if a threshold exists, it is less than the range from 20 to
30 WLM. Snihs (Sn73, 74) considers that the lowest underground exposure
resulting in an apparent increase in lung-cancer deaths in Swedish miners
is about 15 WLM, although he states that it is impossible to draw conclu-
sions about the exposure-response relationship below 100 WLM. Hewitt
(He79) concludes from an analysis of Canadian uranium miners that if a
13
-------
threshold exists, it is below 60 WLM. These varied opinions seem to indi-
cate the possibility that environmental exposure to radon daughters (or
very-low-level exposures) may result in such a small lung cancer rate as to
be indistinguishable from the natural, nonradiological induction rate.
The incidence of lung cancer attributable to radon-daughter exposure
that has been observed in the various mining subgroups ranges from about
1.5 to 50 cases per WLM/year/106 persons, with a present-day rounded aver-
age value of 10 x 10~6 per person per year per WLM, for estimated mean
follow-up times ranging from about 14 to 48 years. This average value has
been accepted in the lung-cancer prediction model of Harley and Pasternack
(Har81) as reasonably realistic when their modeled data are also compared
to background (normally occurring) lung-cancer incidence in nonsmokers from
environmental exposure to radon. The NCRP (NCRP84a) has calculated the
mean and standard error of the estimated lung-cancer incidence to date to
be 12 ± 2 per WLM/year/106 persons from a data base of 23 exposure groups.
The 95% confidence interval of the mean, therefore, ranges from 8 to 16 per
WLM/year/106 persons.
In estimating the effect of radon-daughter exposure at environmental
levels (normally, less than about 20 cumulative WLM per lifetime), the
attributable risk at high exposures, derived from the mining data, must
somehow be extrapolated to the low-exposure region. In keeping with pru-
dent, conventional practice, the extrapolation is linear, even though some
studies suggest that exposures may be even more efficient in inducing lung
cancer as the exposure rate approaches background levels (Ar78; NAS80).
This hypothesis is in contrast to the possibility mentioned above that
14
-------
very-low-level exposure to radon daughters does not result in distinguish-
able lung cancer. Recent track structure analyses suggest that lung cancer
incidence is a quadratic or cubic function of dose when allowing for con-
comitant cell killing (Ho83). If so, risk at low doses is very much
smaller than indicated by the linear extrapolation.
While risk may be extrapolated to low exposure levels, the accuracy of
any risk estimate depends not only on the extrapolation model but also on
the accuracy of the measurement, or estimation of exposure. From a statist-
ical viewpoint, therefore, risk estimates for low level exposures would
have a large confidence interval and, thus, relatively poor accuracy.
Influence of Modifying Agents
The effect of cigarette smoke in modifying radiation-induced cancer
probabilities remains uncertain at this time. During periods of relatively
short follow-up (15-25 years), cigarette smoking has been associated with a
markedly increased incidence of lung cancer in miners. During 30- to 60-yr
periods of follow-up after initial exposure, lung-cancer incidence is
reported to be either somewhat greater among nonsmokers than smokers (Ax80)
or about the same (Ra81). This latter evidence is in agreement with the
results of studies of beagle dogs that had comparable radon daughter expo-
sures and were exposed to cigarette smoke (Cr78): dogs that "smoked" had
fewer respiratory-tract tumors than dogs that did not "smoke." The current
evidence suggests that the principal role of cigarette smoking in uranium
miners is to accelerate the appearance of lung cancer induced by radiation,
although even this idea has recently been challenged (Sa82). However, the
15
-------
issue cannot be considered resolved as yet, and the role of smoking at low
occupational or environmental radon-daughter levels is unknown.
Circumstantial evidence appears to rule out ore dust and diesel ex-
haust as important contributors to the observed incidence of lung cancer
among underground uranium miners (FRC67). This evidence is also supported
by data from the animal experiments discussed in the following section.
Finally, the co-influence, if any, of other physical, chemical or bio
logical agents at low occupational or environmental radon-daughter levels
is unknown.
16
-------
ANIMAL STUDIES
Introduction
Animal studies have been conducted for several decades to identify the
nature and levels of uranium mine-air contaminants that were responsible
for producing the lung cancers observed among uranium-mining populations.
Many of the initial studies were concerned with early effects or short-term
pathological changes (Jan26; Re39; Ja40). Exposures were based primarily
on radon gas concentrations, giving little or no information on the radon-
daughter concentrations, which subsequently were shown to contribute the
greatest radiation dose to the lung. The early studies (Hu39; Raj42a, 42b;
Ku59), in which lung tumors were produced, were methodologically or sta-
tistically inadequate to show an unequivocal association of lung tumors
with exposure to radon and/or radon daughters.
Beginning in the 1950s, a growing concern emerged that the increased
incidence of respiratory cancer observed in the European uranium-mining
population would also be found in the U.S. mining population (SS55; Wa64).
Systematic studies were subsequently begun in the U.S. to identify the
agents responsible for the increased incidence of lung cancer in miners and
to develop exposure-response relationships in animals. Investigators at
the University of Rochester began to focus attention on the biological and
physical behavior of radon daughters as well as their contribution to the
radiation dose to the respiratory tract (Ba51; Harr54; Mo55). Shapiro
(Sh54) exposed rats and dogs to several levels of radon alone and in the
presence of radon daughters attached to room-dust aerosols. He also showed
17
-------
that the degree of attachment of radon daughters to carrier dust particles
was a primary factor in influencing the a-radiation dose to the airway epi-
thelium. He demonstrated that this dose was due primarily (>95%) to the
218 214
short-lived radon daughters RaA ( Po) and RaC' ( Po), rather than to
the parent radon.
In 1953, Cohn et al. (Co53) reported the relative levels of radio-
activity found in the nasal passages, in the trachea and major bronchi, and
in the other portions of rat lungs after exposure to radon and/or radon
daughters. The respiratory tracts of animals that inhaled radon plus its
daughters contained 125 times more activity than those of animals that
inhaled radon alone.
Beginning in the mid 1950s, Morken initiated a pioneering series of
experiments (Mo66, 73a, 73b) to evaluate the biological effects of inhaled
radon and radon daughters in mice; later experiments used rats as well as
beagle dogs. The essentially negative biological results of these studies
(due primarily, we believe, to inadequate follow-up times in the experi-
ments) suggested that o-irradiation is inefficient in producing tumors in
the respiratory system. The only apparently permanent late changes occurred
in the alveolar and respiratory bronchiolar regions of the lung for a wide
range of exposure levels and for observation times to 3 years in the dog
and 1 and 2 years, respectively in the rat and mouse. Furthermore, injury
in the bronchial tissue was quickly repaired after irradiation ceased. The
carrier aerosols used in these experiments were more typical of environ-
mental aerosols than of those found in the mines.
In the late 1960s and early 1970s, other studies in France and the
U.S. were initiated, which later proved successful in producing lung tumors
18
-------
from inhaled radon daughters. The French exposed rats either to radon
daughters alone or in combination with stable cerium, uranium-ore dust or
cigarette smoke to produce tumors in the lung (Pe70; Ch74, 80). The later
U.S. studies were designed to systematically determine the pathogenic role
of radon daughters alone, or in various combinations with uranium-ore dust,
diesel-engine exhaust and cigarette smoke. These studies involved lifespan
exposures of beagle dogs and Syrian Golden hamsters and chronic exposures
of rats (Cr78, 82).
A discussion of the biological effects in animals of inhaled radon and
radon daughters is included in the recent ICRP Publication 31 (ICRP80). In
general, the lung cancers in rats were noted to be about half bronchogenic
and half bronchioloalveolar in origin, in contrast to the nearly exclusive
bronchogenic origin of human lung cancers. Extrapulmonary lesions were not
a significant finding in the radon inhalation studies. An even more de-
tailed presentation of the animal studies is presented in NCRP Report
No. 78 (NCRP84a).
Discussion and Summary
The animal studies have provided considerable data confirming the
human epidemiology studies:
(1) In rats, primarily, tumor production per WLM at very high exposures
was lower than at moderate exposures (Cr82; Ch80). The lowest attrib-
utable lung cancer rates per unit exposure were observed in miners
exposed to the highest radon-daughter levels in underground mines.
19
-------
(2) In both the human and animal studies, tumor production appeared to in-
crease with decrease from high exposure rate (Ch81; Cr82); the
influence of exposure rate is unknown at current occupational and
environmental exposure levels.
(3) In a small group of Swedish zinc/lead miners, a lower lifetime inci-
dence of lung cancer was observed in those who smoked and were exposed
to radon daughters than in the nonsmokers. This is tentatively
ascribed to the protective effect of increased mucus production from
smoking (Ax78) or of the thickened mucosa resulting from smoker's
bronchitis. A similar result was observed in dogs (Cr78). In rats,
tobacco smoke was found to be cocarcinogenic with radon daughters when
exposure to the smoke followed completion of exposure to the daughters
(Ch80). This effect was not observed, however, when smoking preceded
the radon-daughter exposure (Ch81). Such disparities may partially
explain discrepancies in interpreting epidemiological data.
(4) Emphysema and fibrosis have been attributed to radon-daughter exposure
in animals— hamsters, rats and dogs (Stu78; Cr78)~and underground
miners. Simultaneous exposure to ore dust or diesel exhaust increased
the incidence of these lesions but did not appear to increase the
number of tumors produced by exposure to radon daughters (Cr78, 82;
Ch81).
(5) For equivalent, cumulative, radon-daughter exposure, the older the
animal at the start of exposure, the shorter the latent period (Ch81).
In humans, the highest risk coefficient calculated, about 50 x 10~^
20
-------
lung cancers per year per WLM, is that for persons first exposed when
over 40 years of age (NAS80).
(6) The predictions of the various dosimetric models appear to be borne
out in the various species. The tumors induced in experiments with
animals are commonly more distal than those in humans. Desrosiers'
(De78) modeling of Syrian Golden hamster lungs showed that peripheral
basal and Clara cells may receive doses greater than, or approximately
equal to, those received by basal cells in the central airways. Human
tumors have appeared almost exclusively in the upper generations of
the bronchial tree. Some absorbed-dose calculations show that basal
cells in human upper airways, at about the level of the segmental
bronchi, receive the highest dose from radon daughters (e.g., A164;
Har72, 81).
(7) Lifetime risk coefficients are similar in both animals and humans.
The coefficients based on rat data appear to range between 1 and
4 x 10~4 per WLM for all tumors (benign and malignant) at cumulative
exposures less than 5000 WLM (Ch81, Cross, unpublished data). At
exposures considerably lower than where lifespan is significantly
shortened (<500 WLM), the lifetime risk coefficient appears to be
about 2 x 10"4 per WLM for malignancies and ranges between 2 and 4 x
-4
10 for all tumors. Data are as yet insufficient to determine a
value for exposures below 100 WLM.
21
-------
LUNG-CANCER RISK-PREDICTION MODEL
The predictive model of Harley and Pasternack (Har81, RPC80) has been
adopted for lung-cancer risk predictions because it allows risk coeffi-
cients to be developed for various age groups and exposure periods. This
absolute-risk model is also used by the NCRP in their report on the
evaluation of occupational and environmental exposures to radon and radon
daughters in the United States (NCRP84a). It is based upon the most recent
estimates of lung-cancer deaths among underground miners and accounts for
the apparent increase in lifetime risk with increasing age at first
exposure and duration of exposure (an effect not possible with a
relative-risk model), as noted in the epidemiological studies of these
miners. Although the model represents a reasonable uranium-miner
lung-cancer response, the validity of extrapolation to environmental levels
is unknown.
The adopted average yearly risk coefficient obtained for all exposure
categories and all age groups (10 x 10~6 lung cancers per year per WLM)
corresponds to a lifetime risk (to age 85 years) of about 1 to 2 x 10"4
per WLM, dependent, of course, on activity, age at first exposure and dura-
tion of exposure. For comparison, the ICRP (ICRP81) has adopted a range
-4
for lifetime risk of 1.5 to 4.5 x 10 per WLM, based primarily on
Czechoslovakian underground-mining data. Evans et al. (Ev81) estimated the
lifetime risk (which, they state, is applicable to the general population)
22
-------
-4
to be 1 x 10 per WLM from U.S. and Czechoslovakia!! uranium-miner epide-
miological data. Twice this value (2 x 10"^ per WLM) was adopted by
Jacobi (Jac77) as the lifetime risk applicable to all types of miners; it
is used by Cliff et al. (CI79) to model lung-cancer incidence from environ-
mental exposure. UNSCEAR (UN77) has reviewed the data in uranium miners in
Canada, the U.S. and Czechoslovakia, in Swedish nonuranium miners and in
iron miners in the United Kingdom. UNSCEAR indicates that the probable
-4
lifetime lung-cancer risk is 2 to 4.5 x 10 per WLM. The BEIR III report
contains lung-cancer data in U.S., Canadian and Czechoslovakian uranium
miners, Newfoundland fluorspar miners and Swedish metal miners (NAS80).
The range of risk for all groups was expressed as 6 to 47 x 10~6 lung
cancers per person per year per WLM, the upper value being for the Czech
miners who began exposures at age 40 or older. If we assume that
lung-cancer expression takes place over a 30-year interval (to account for
the BEIR report's exclusion of the latent period in developing the yearly
rate of risk), the 6 to 47 x 10~® per person per year per WLM reduces to
-4
a range of lifetime risk from about 2 to 14 x 10 per WLM. The high
value is considered, by BEIR, to be the most likely risk estimate, at
exposure concentrations of about 1 WL, for those over 65 years of age at
lung-cancer diagnosis. This value is higher than any of the other reported
values and cannot be reconciled until a closer examination of these miners
has been made regarding age at first exposure, cumulative exposure,
exposure rate, smoking history, etc.
23
-------
Except for the high value of BEIR, the lifetime risk estimates for
lung cancer attributable to radon-daughter exposure (per WLM) appear to be
reasonably consistent, considering the difficulty in estimating this
quantity without complete follow-up, and the various methodological prob-
lems encountered in epidemiological studies.
Other features of this predictive model are that lung cancer does not
occur either before a 5-year latent interval or before age 40, and that
risk is corrected from year of exposure by an empirical exponential,
cellular repair factor (20-year half-time). An appropriate life-table
value is utilized to account for competing risks of death.
Although the basic incidence data from the underground-mining epi-
demiological studies cannot be applied directly to environmental situations
(because patterns of exposure differ), a common factor exists in the risk
per rad for bronchial basal-cell dose. The lifetime lung-cancer risk attri-
butable to a dose of 1 rad per year has been calculated, using the conver-
sion factor of 0.5 rad/WLM estimated for miners (RPC80). For environmental
exposure starting at 1 yr of age, the lifetime risk (for exposure to age
_2
85 years) is calculated to be 1.3 x 10 . Because exposure of a popula-
tion involves persons of various ages, it is sometimes necessary to know
the lifetime risk of radon-daughter-induced lung cancer for a population
with age characteristics typical of the United States. This value, using
the 1975 age distribution for the U.S. (WH078), is calculated to be 8.0 x
10 per rad per year exposure. These risk coefficients may be used for
estimating lung-cancer risk from any source of radon-daughter exposure.
Risk from a bronchial dose in rad per year to basal cells is not the
most useful way to evaluate environmental exposures. Two, more useful,
-------
lifetime risk coefficients can be derived that relate risk to environmental
exposure in units of WLM per year, and to an annual exposure to a radon
222
concentration of 1 pCi Rn/L. These coefficients use the previously
indicated, average environmental exposure-to-dose conversion factors for
the adult male, female, 10-year-old child, and infant, as well as assump-
tions on the radon-daughter equilibrium factor. The derivation of these
risk coefficients for environmental exposures is simplified considerably
and contains very little error, if we accept the environmental exposure-
to-dose conversion factor of 0.7 rad/WLM (which applies to adult males) for
all people. The lifetime risk estimate, which includes the effect of the
higher dose-conversion factor in childhood, is within 10% of this value
(see Tables 2 and 3 in Har81). This conversion to WLM units places at
_3
9.1 x 10 the lifetime risk coefficient for beginning exposure at infancy;
for populations with ages characteristic of U.S. inhabitants in 1975, it is
_3
5.6 x 10 per WLM per year for lifetime risk and lifetime environmental
exposure.
For the case of exposure measured as radon concentration over time,
the average annual bronchial dose to an adult male from radon daughters
222
associated with exposure to 1 pCi Rn/L (assuming he is active 16 hours
per day and rests 8 hours per day) is 0.27 rad/year (RPC80). Thus, the
lifetime risk for annual exposures to 1 pCi/L is calculated to be 3.6 x
-3
10 for exposure beginning at infancy and (under the same exposure condi-
-3
tions) 2.1 x 10 for populations of mixed age. Table 1 summarizes the
lifetime lung-cancer risk-coefficient data.
25
-------
A uniform risk/year lung-cancer-prediction model has also been devel-
oped by Harley and Pasternack (Har81) but was rejected as being unrealistic
because lifetime lung-tumor risk was found to decrease markedly with increas-
ing age at first exposure. The ratio of the two lung-cancer prediction
models (uniform risk/year model * decreasing-risk model) decreases with
both duration of exposure and age at first exposure. It is estimated from
the data in Table 1 of Harley and Pasternack (Har81) that the estimate of
lifetime risk from radon-daughter exposures would be increased by approxi-
mately two had the uniform risk/year model been employed instead of the
decreasing risk model.
The environmental lifetime risk coefficients in Table 1 are based on
an unattached RaA/Rn ratio of 0.07 and an equilibrium factor of about 0.7.
The risk coefficients can be adjusted for other unattachment fractions and
radon-daughter disequilibrium conditions. The radon-daugher equilibrium
factor is considered to be the more important adjustment for environmental
exposures. Under some conditions of exposure, the equilibrium factor is
very low; the use of the radon-gas risk coefficients would then produce
unnecessary conservatism in the estimated lung-cancer predictions. On the
other hand, the use of the above radon-daughter risk coefficients when
equilibrium factors are low, results in underestimation of the lung-cancer
risk. Table 2 provides data for adjusting the risk coefficients for other
radon-daughter disequilibrium conditions.
Lifetime Risk From Environmental Exposure
Lifetime lung-cancer risk to populations from continuous environmental
26
-------
radon and radon-daughter exposure may be calculated using any of the four
coefficients described above, depending on the units of exposure. For
example, a lifetime risk for lifetime exposure to the average, outdoor
999 -A
Rn concentration of 0.2 pCi/L (St80; Ge80) is 7 x 10 , 0.07%. For
comparison, Evans et al. (Ev81) calculate a value 1/3 lower for equivalent
concentrations. Indoor concentrations of radon are almost always higher
than those outdoors because vertical mixing cannot take place in the former
case. Furthermore, indoor radon levels may be enhanced, in some cases, by
elevated Ra concentrations in building materials and radon in the water
and fuel supplies. The references cited above indicate that the average
value for indoor radon concentrations ranges between 0.6 and 0.8 pCi/L,
excluding basement concentrations. Using these data, a typical average en-
vironmental exposure for single-family dwellings approximates 0.5 pCi/L
(accounting for both indoor and outdoor exposures), which would result in a
_3
lifetime lung-cancer risk of 2 x 10 , 0.2%. Because a significant frac-
tion of the U.S. population resides in multistory buildings (Har81), and
indoor levels in the U.S. relate primarily to proximity to ground beneath
the structure, the true average environmental risk might be expected to lie
between these two values (e.g., about 0.13%).
Published U.S. annual death rates for lung cancer among nonsmokers
have ranged from about 23 x 10"^ (Ca76) to 47 x 10"^ (Hae58), on average,
for males and females. These values can be multiplied by 45 to calculate
the lifetime risk (ages 40-85), which yields 0.1 to 0.2%. The recent
estimates by Enstrom and Godley (En80) and by Garfinkel (Ga80) give an
average, rounded value for the lifetime risk of 0.6%.
27
-------
It is impossible to confirm the accuracy of the lung-cancer-risk pre-
diction model for environmental background exposures to radon daughters.
The model does project, however, that approximately 20 to 100% of the back-
ground (nonsmoking) lung-cancer incidence can be attributed to environ-
mental radon-daughter exposure. The value is probably closer to 20%,
because of the generally accepted uncertainty in the earlier data on non-
smoker lung-cancer rates.
Accuracy of Risk Prediction from Model
Myers and Stewart (My79) have speculated on an underlying lung cancer
incidence in uranium miners that is not exposure-dependent and may repre-
sent the effect of other carcinogens. It is this factor, among others,
that caused Evans et al. (Ev81) to conclude that 10"4 per WLM was the
upper boundary for the lifetime risk for nonminer exposures. They state
that any value greater than this would be incompatible with both British
and U.S. epidemiological evidence. Exposure to environmental cocarcino-
gens may similarly confound the interpretation of population exposures.
Therefore, the risk factors may be similar for all radon and radon-daughter
exposures.
The prediction model utilizes a lifetime risk coefficient somewhat
higher than the Evans et al. value and concludes that a significant per-
centage (20 to 100%) of the lung-cancer incidence in nonsmokers may be due
to background radon exposures. The prediction model is, therefore, not
unreasonable from the standpoint of grossly overestimating the background
incidence. On the other hand, the upper range of the lifetime risk
28
-------
coefficients of ICRP and UNSCEAR, 4.5 x 10~^/WLM, would indicate that the
true values may be underestimated by as much as a factor of three. An
increase in lifetime risk, however, would also proportionally increase the
predicted incidence from background radon exposures.
Another factor to consider is that a large percentage of the miners
studied are still alive; therefore, one might expect that the lifetime risk
per WLM would increase over the values used in this paper. This possibil-
ity has been considered by the NCRP (NCRP84a), which concludes that, at the
present declining rate of appearance of lung cancer, the total risk, at
least for the U.S. mining group, would not double. All factors considered,
therefore, we cannot conclude from the available evidence what the accuracy
of the prediction model is at this time. Because we assume, in our sub-
sequent treatment of risk, continuous exposure from infancy to the radon
derived from water, along with other conservative assumptions, we believe
that the prediction model provides reasonable estimates of lifetime risk
for deriving concentration limits for radon in water.
ICRP organ-specific risk factors have also been employed in subsequent
sections for a comparison with the lung-cancer-prediction model; the life-
time risks predicted by the ICRP are approximately 60% higher than the values
derived from the lung-cancer prediction model. Insofar as the lung dosim-
etry and ICRP risk factor for lung represent reality, it would appear that
the two methodologies for lung-cancer estimation are reasonably comparable.
29
-------
CALCULATION OF CANCER RISKS FROM EXPOSURE TO RADON IN DRINKING WATER
Lung-Cancer Risk
Indoor radon and radon-daughter concentrations, derived from the water
supply, are multiplied by the appropriate lifetime lung-cancer risk co-
efficients for determining the attributable risk of lung cancer from radon
and radon-daughter exposures. For example, the lifetime risk coefficients
-3 -3
of 3.6 x 10 per pCi/L radon concentration and 9.1 x 10 per WLM/yr (con-
sidered appropriate for exposures beginning at infancy) would be multiplied
by the mean radon concentration or the mean radon-daughter exposure rate
and the number of persons exposed, to provide an estimate of the total num-
ber of lung cancers produced. The risk coefficients applying to exposure
beginning at infancy are employed rather than those applying to populations
of mixed ages.
Regarding actual indoor exposures, Bogue (Bo59) has stated that during
the course of a single year, 20-22% of the inhabitants of the United States
move from one house or apartment to another. Not more than 2% of the adult
population will spend an entire lifetime in the same dwelling, and less
than 15% will spend a lifetime in the same county. The tendency to re-
locate is often least frequent among children and the elderly, and most
frequent among those between 17 and 32 years of age. Mobile homes and
apartment units have high rates of turnover, whereas middle- and upper-
class conventional-type homes have the lowest. The average occupancy time
is probably about 5 years for all ages of residents and all types of dwell-
30
-------
ings. Graphically expressed, distribution of residency time versus fre-
quency would probably appear lognormal.
The significance of population mobility is that the additional indi-
vidual lung-cancer risk for residents in houses with high radon concentra-
tions is small if the duration of residency is short. Population mobility
tends to distribute the cancer risk among a greater number of persons—
those who may at some time reside in one of the dwellings with increased
levels of indoor radon.
In view of the risk-coefficient adjustment factors of Table 2 and the
fact that indoor equilbrium factors for radon daughters can be considerably
less than about 0.7 (Ev69), we prefer to use the rounded, lifetime lung-
_3
cancer risk coefficients of 3 x 10 per pCi/L radon concentration and
_ ?
1 x 10 per WLM/yr exposure rate, for lifetime exposure. Thus, using the
-4
range of mean transfer coefficients of 1 to 2.5 x 10 for radon in water
to radon in house air, the lifetime lung-cancer risk factor for continuous
indoor exposure converts to 3 to 7.5 x 10"7 per pCi/L radon concentration
in water.
For comparison, the ICRP (ICRP77) has chosen 2 x 10"^/rem as an occu-
pational risk factor for lung cancer based primarily on external radiation
exposures. Based on the previously derived lung dose-equivalent of 0.4 to
0.9 mrem/ year/pCi/L radon concentration in water, and a 60-year dose-
accumulating interval, the projected lifetime lung-cancer risk factor for
continuous indoor exposure ranges from 5 to 11 x 10~7 per pCi/L radon
concentration in water.
31
-------
Stomach and Whole-Body Cancer Risks
According to the ICRP (ICRP77), there is evidence that radiation is
carcinogenic to the stomach at moderate doses. Although there is as yet no
value for the stomach risk factor, the ICRP considers that it is likely to
be low. They further estimate that no single "other" tissue (which in-
cludes the stomach) has a risk coefficient exceeding 10"^/rem. Based on
the previously calculated annual dose-equivalent to the stomach of about
100 mrem per 1000 pCi/L radon concentration in water, and an assumed 60-year
dose-accumulating interval, the lifetime accumulated dose-equivalent to
_3
stomach is 6 rem per 1000 pCi/L of radon in water, or 6 x 10 rem per
pCi/L. The lifetime risk to stomach, therefore, is estimated not to exceed
Q
6 x 10 per pCi/L radon concentration in water, especially in view of
the previous discussion regarding the possibility that ingestion doses are
grossly overestimated.
The ICRP also estimates that the mortality risk coefficient for all
radiation-induced cancers from uniform, whole-body irradiation is about
10"^/rem. Based on the previously calculated annual dose equivalent to
whole body of about 2 mrem per 1000 pCi/L radon concentration in water, and
an assumed 60-year dose-accumulating interval, the lifetime accumulated
dose-equivalent to whole body is 0.12 rem per 1000 pCi/L of radon in water,
or 1.2 x 10~4 rem per pCi/L. The rounded lifetime risk to whole-body,
-8
therefore, is estimated to be 1 x 10 per pCi/L radon concentration in
water.
32
-------
Summary of Calculated Cancer Risks from Radon in Drinking Water
Table 3 presents a summary of the estimated lifetime and annual cancer-
death risks associated with drinking water containing 1 pCi/L radon con-
centration.
The total lifetime cancer-death risk is on the order of 4 to 8 x 10"7
per pCi/L radon in drinking water; the annual risk is on the order of 8 to
_q
18 x 10 per pCi/L of radon in drinking water. These estimated risk
coefficients are considered conservative, as they assume continuous expo-
sure from infancy to the radon in, and from water. In fact, it is unusual
for a person to occupy a structure 100% of the time, or always during the
time of water usage.
For comparison, the ICRP (ICRP77) has chosen an acceptable level of
-6 -5
risk for stochastic phenomena in the range of 10 to 10 per year to
any individual member of the public. These risks (based primarily on
external radiation exposure data) imply an acceptable concentration of
radon in drinking water ranging from about 60 to 1300 pCi/L. It is readily
apparent that these concentrations are impractical (and therefore will not
be discussed further in this paper) as very many water supplies exceed
these levels (Du76; UN77, 82; Coh79). The conclusions reached by Duncan et
al. (Du76), based on limited data in the U.S. and Great Britain, are that
radon concentrations in water range from 0 to 30,000 pCi/L, with 25% of the
locations exceeding 2000 pCi/L and 5% exceeding 10,000 pCi/L. In granitic
areas, the concentrations may be an order of magnitude higher. The radon
concentrations reported in the later United Nations report (UN82) range
from practically zero to values up to about 2,700,000 pCi/L in some waters.
33
-------
Table 4 presents total lifetime cancer risks associated with various
radon concentrations in water.
34
-------
LIMITS ON EXPOSURE
It appears unlikely that we can develop reasonable limits on exposure
to natural levels of radon and radon daughters by rigidly following the
suggested limits on population dose or the suggested limits on population
risk proposed by the various regulatory and standards-setting agencies for
manmade sources of radiation. Furthermore, the variations in the con-
centrations of natural radioactivity, and the extensive areas of elevated
levels of natural radioactivity, necessitate that practical limits include
the concentrations normally found in the majority of houses.
A simple first approach toward limiting exposure to radon from water
sources would be to use the historic, conservative "tolerance doses" of
1/10 to 1/100—the exposures known to affect health. (For a discussion of
this concept, see Cr74). The reduction factors are presumably dependent on
the percent incidence or severity of effect. If we assume that indoor
radon and radon-daughter exposures have an associated health effect, and
further assume that average indoor radon levels from all sources, including
drinking water, are on the order of 1 pCi/L (NAS81), the "tolerance" inhala-
tion exposures, per se, are 0.01 to 0.1 pCi/L. (Note: we do not propose
to limit total indoor radon concentrations to these levels, the calculation
is simply for illustration.) Based on the mean transfer coefficients of
-4
1 to 2.5 x 10 for radon in air from radon in water, the derived concen-
tration limit for radon in water ranges from 40 to 1000 pCi/L [(0.01 to 0.1
pCi/L) t (1 to 2.5 x 10"^)], if all of the indoor radon were to come from
the water supply. For houses containing radon in the water supply, 5 to
35
-------
\2% of indoor radon, on average, has been estimated to come from the water
(Ka80; UN81). In our calculations, however, we generally assume that all
indoor radon comes from the water supply. Because the health hazard is
expected to be low for average indoor radon exposures, an upper rounded
value of 1000 pCi/L radon in water is the more reasonable choice.
An even simpler approach would be to assume that the risk associated
with average indoor concentrations of radon and radon daughters is accept-
able to the U.S. population. The possibility that this risk may be less
than the linear hypothesis predicts, or may even be zero, lends some sup-
port to this assumption. Under this assumption, the derived concentration
limit for radon in water ranges from 4000 to 10,000 pCi/L [1.0 pCi/L *
(1 to 2.5 x 10"4)], if all the indoor radon were to come from the water
supply.
Another example of the "tolerance dose" approach is to assume that
lifetime exposures above about 100 WLM have a statistically significant
excess of lung cancer and that exposures below this level have only a low
probability for causing deleterious health effects. This choice is based
on the epidemiological review by the NCRP (NCRP84a) which also concludes
that none of the studies, thus far, have produced data showing a statistic-
ally significant excess of lung cancer in the lowest exposure category
(<60 WLM). Applying the tolerance factor of 1/10, the lifetime exposure
limit is 10 WLM, a value comparable to average lifetime exposures to en-
vironmental radon. The derived radon concentration in air for an average
indoor radon-daughter equilibrium factor of 0.5 and an exposure period of
60 year, is 0.65 pCi/L [10 WLM x 170 WLhr/WLM x 1/8760 yr/hr x 1/60 yr'1 x
36
-------
1/0.5 x 100 pCi/L/WL]. On the assumption that all of the indoor radon
comes from the water supply, the derived concentration limit for radon in
water ranges from 2600 to 6500 pCi/L.
In summary, these three approaches suggest that the upper limit on
radon in water, based on the conservative "tolerance dose" approach;
continuous inhalation exposure; and the assumption that 100% of the radon
in indoor air comes from the water supply, is a rounded value between 1000
and 10,000 pCi/L.
An alternative, more fruitful approach would be to base risk on the
distribution of population exposures to natural radioactivity (NCRP84b);
basing risk on fractions of the dose equivalents allowed for radiation
workers is tenuous at best, because we really do not know the true risk
associated with occupational dose equivalents. If we know the average
exposure, and the range that a population receives, a practical upper limit
on exposure could be established on a cost-risk-benefit, remedial-action
basis. It is unlikely that this approach could be rigorously pursued at
this time, however, due to the paucity of data regarding population expos-
ure to radon and radon daughters in the United States.
The distribution approach has been taken in Canada, confirming an
earlier indoor limit of 0.02 WL (1 WLM/yr) for uranium mining communities
(Ea82). If we apply this radon-daughter concentration to all communities
in the U.S. and assume that the average, indoor, radon-daughter equilibrium
factor is 0.5, a derived upper limit on indoor radon air-concentrations is
4 pCi/L [0.02 WL x 100 pCi/L/WL x 1/0.5]. The derived upper concentration
limit on radon in water (assuming all radon comes from the water) would
37
-------
range from 16,000 to 40,000 pCi/L. The actual value would depend on the
amount of radon coming from soil gas and other sources. It is expected
that the average risk for all U.S. residents would be lower than the
maximum risk associated with the 4 pCi/L limit on indoor air concentration.
While the actual U.S. population exposure distribution is unknown, the
estimate of the mean indoor exposure is 0.2 WLM/yr for first-floor expo-
sures in houses (Har81). The average annual exposure for all members of
the U.S. population may be no greater than about 0.13 WLM/yr, considering
that a significant fraction of the population resides in multistory build-
ings (Har81). Tables 5 and 6 show estimated distributions of the lifetime
risk for average annual exposures of 0.2 and 0.13 WLM, respectively, assum-
ing lognormal exposure distributions and a geometric standard deviation
(GSD) of 2.5. A GSD of about 2 is in keeping with the values measured in
New York and New Jersey residences (Ge80); however, the measurements by
McGregor et al. (McG80) in Canadian houses, and by Prichard et al. in Maine
and Texas houses (Pr81), indicate that mean GSD values are closer to 2.5.
Tables 5 and 6 indicate that the majority of the risk occurs in the
lower exposure categories. Thus, any limits on exposure would have to take
this fact into consideration. The lower of the two distributions
(0.13 WLM/yr, Table 6) indicates that only a very small percentage (<0.03%)
of the population has an exposure to radon daughters exceeding 2 WLM/yr,
whereas approximately 0.35% is exposed to levels exceeding 1 WLM/yr. For
populations exposed to an estimated average of 0.2 WLM/yr radon daughters
(Table 5), approximately 0.14% receive exposures exceeding 2 WLM/yr, while
1.3% exceed 1 WLM/yr. In either case, an exposure limit of 2 WLM/yr would
38
-------
require relatively little remedial action, whereas an exposure limit of 1
WLM/yr would necessitate remedial action on about 10 times more houses.
The tables also indicate that substantial breaks in the lung-cancer
risk occur at the 1-WLM/yr and 2-WLM/yr exposures. This suggests, in the
absence of cost-risk-benefit analyses, that either exposure limit might be
reasonably selected on a practical basis. Limits less than 1 WLM/yr appear
to require remedial action in an impractical number of houses.
If we again choose the 1-WLM/yr exposure limit as an illustration, the
derived, allowable, radon-air concentration (assuming an equilibrium factor
of 0.5 and continuous exposure) becomes 4 pCi/L, a value approximately four
times higher than the estimated average indoor level in U.S. houses (NAS81).
This allows, on average, approximately 3 pCi/L to come from water. The
average, derived, radon level in water sources, therefore, ranges from
about 10,000 to 30,000 pCi/L [3 pCi/L * (1 to 2.5 x 10~4)]. Had 2 WLM/yr
been picked as the indoor annual limit on exposure, the derived, allowable
radon concentration in water would have ranged between about 20,000 and
60,000 pCi/L. For those houses where all of the indoor radon comes from
the water supply, the derived upper concentration limits on radon in water
range from 16,000 to 40,000 pCi/L for the 1-WLM/yr inhalation exposure
limit and 32,000 to 80,000 pCi/L for the 2-WLM/yr inhalation exposure
limit.
Finally, an equivalent-risk approach could be used to derive the con-
centration limit for radon in water. Based on the risk factors in Table 3,
222
20,000 pCi/L Rn in water has an associated lifetime, total, cancer-
_o _7
death risk ranging from 8 to 16 x 10 [20,000 pCi/L x (4 to 8 x 10
39
-------
cancer deaths/pCi/L)], or approximately 1 to 1.5%. The associated
lifetime, lung-cancer-death risk ranges from 0.6 to 1.5%. These lifetime
risks are comparable, by our calculation, to the lifetime lung cancer risk
presently allowed by the EPA in their standards issued for cleanup of
uranium tailings (FR83). Under EPA's standard for buildings, the objective
is to achieve an indoor radon-daughter concentration of 0.02 WL. Tailings
are to be removed from premises where the levels exceed 0.03 WL, but lower-
cost ventilation and air cleaning methods may be employed instead. The
allowable 0.02 to 0.03 WL concentrations have an associated lifetime, lung-
cancer-death risk of 1 to 1.5%, assuming continuous exposure and a risk-
coefficient of 10 per WLM/yr. For comparison, the EPA values for 0.02 WL
lifetime environmental exposures are 1 and 2.3%, respectively, based on
their absolute and relative risk models (RPC80).
Based on these considerations of the estimated distribution of radon
exposures in the U.S., a derived practical limit on radon concentrations in
water is not less than 10,000 pCi/L. A 20,000-pCi/L value is reasonable
and conservative from the standpoints of limiting cost of remedial action
to a more manageable number of houses; the exposure is considered to be
continuous from infancy; the value is based on an air-water transfer
coefficient high enough to accommodate reasonable energy-conservation
measures; and it is based on an assumed, average equilibrium factor of 0.5
for daughters derived from the radon released from water. The derived
water concentration would be higher for equilibrium factors <0.5. In
houses where the total radon-daughter equilibrium factor is substantially
lower than about 0.5, it is almost certain that the radon concentration
40
-------
limit could be higher. Until more data are available regarding the actual
exposure distribution in the U.S., the 20,000 pCi/L choice is considered to
be the best balanced estimate, in the absence of cost-risk-benefit
analyses, of the derived radon-concentration limit for water.
We do not wish to emphasize the radon in water concentration per se,
but rather the primary inhalation exposure limit in WLM/yr. Based on
estimated, natural exposure distributions in dwellings, we suggest that a
limit on exposure less than 1 WLM/yr (including background exposures) might
not be reasonable from a cost-benefit perspective. Some might argue this
to be the case even at 2 WLM/yr. Whatever choice is made for exposures
applicable to all buildings (not just tailings-contaminated buildings), we
believe that the primary emphasis regarding remedial action should be
placed on an annual limit on total inhalation exposure (WLM/yr) and not on
a derived WL-concentration of radon daughters or a derived air concentra-
tion of radon gas. Keeping the emphasis on the primary standard would
allow for variability in occupancy time, ventilation rate, and other fac-
tors bearing on the exposure of individuals occupying a structure. The
derived limit on radon in water (20,000 pCi/L) should, therefore, be looked
on as an action level, above which consideration would be given to char-
acterizing the indoor radon-daughter exposures. Experience may eventually
allow adjustment of this derived concentration to a much higher level.
Finally, it should be pointed out that, based on the previous estim-
ates of dose to stomach and whole body, a 20,000-pCi/L radon concentration
in water would produce estimated annual dose-equivalents to stomach and
whole body of 2 rem and 0.04 rem, respectively. While the current NCRP
41
-------
population limits on annual whole-body dose (0.17 rem/yr) are not exceeded
at this concentration (and would not be exceeded at about 4 times this con-
centration), the dose to the stomach is higher than normally accepted for
population exposures. Four caveats are in order when discussing these
doses. First, the NCRP is restudying the whole issue of population limits
regarding exposures to manmade sources; thus, our comparison base may
change. Second, exposures very often exceed the population limits, which
are based on a factor of 10 to 30 reduction of worker exposures. One might
convincingly argue that the population limits on exposure are unduly con-
servative when compared with certain natural-background exposures. Third,
the radiation risk to stomach is significantly lower than that to lung.
This fact is based on the lack of supportive evidence of stomach cancer in
workers exposed to high levels of radon and radon daughters (such as under-
ground miners) and the additional negative evidence inthe animal experi-
ments. Fourth, the evaluation of ingestion doses by Sullivan and Nelson,
discussed above, indicates that the stomach-dose estimates used in this
paper are unduly conservative. It appears at this time, therefore, that a
dose limitation approach based on the stomach dose would be unrealistic for
deriving allowable radon concentrations in water.
In summary, the derived upper radon concentration in water (assuming
all indoor radon comes from the water supply) ranges from 1000 to
10,000 pCi/L, based on continuous exposure and the conservative "tolerance
dose" approach; 16,000 to 40,000 pCi/L, based on continuous exposure and
the exposure-distribution approach, and the 1-WLM/yr inhalation exposure
42
-------
limit; and 32,000 to 80,000 pCi/L based on the exposure-distribution
approach and the 2-WLM/yr inhalation exposure limit.
We suggest by our analysis that a rounded concentration value of
10,000 pCi/L can be supported by health-effects considerations alone, based
on the "tolerance-dose" concept and other conservative assumptions regard-
ing dose. We further suggest that a practical value of not less than
10,000 pCi/L can be supported by the estimated distribution of environ-
mental radon-daughter exposures in the U.S. The suggested 20,000-pCi/L
concentration limit (or action level) has an estimated lifetime cancer-
death risk comparable, by our calculation, to current EPA standards for
uranium-tailings-contaminated buildings.
Research needed for resolution of the uncertainties in the values dis-
cussed is primarily related to improving the estimations of the inhalation,
rather than the ingestion, exposures and doses. Major examples are:
improved exposure-distribution data for the U.S. population; realistic
values for household occupancy patterns; more accurate data on radon-
daughter equilibrium factors; more accurate data on the ratio of water-
contributed radon to all other sources of indoor radon; more accurate data
on the water-to-air transfer coefficient; and more accurate cancer-risk
factors.
Finally, we would like to emphasize that, before a maximum contaminant
level (MCL) for radon in water can be firmly established, the broader issue
of the MCL for radon in indoor air must be addressed.
43
-------
ACKNOWLEDGMENTS
The authors wish to acknowledge the many helpful suggestions for
improving the text by C. R. Cothern and N. S. Nelson of the U.S. EPA;
D. Crawford-Brown of the University of North Carolina; C. T. Hess of the
University of Maine; and R. 6. McGregor of the Radiation Protection Bureau,
Ontario, Canada.
This work was supported by the U.S. Department of Energy under Contract
DE-AC06-76RL0 1830 and the U.S. Environmental Protection Agency.
44
-------
REFERENCES
A164 Altshuler B., Nelson N. and Kuschner M., 1964, "Estimation of Lung
Tissue Dose from the Inhalation of Radon and Daughters," Health Phys. 10,
1137.
An64 Andersson I. 0. and Nilsson I., 1964, "Exposure Following Ingestion
of Water Containing Radon-222," p. 317. In Assessment of Radioactivity
in Man, Proceedings of the symposium held at Heidelberg, May 11-16, 1964
(Vienna: IAEA).
Ar78 Archer V. E., 1978, "Summary of Data on Uranium Miners," p. 23. In
Workshop on Dosimetry for Radon and Radon Daughters, Oak Ridge National
Laboratory, April 12-13, 1077. ORNL-53481 (Springfield, VA: National
Technical Information Service).
Ar79 Archer V. E., Radford E. P. and Axelson 0., 1979, "Factors in
Exposure-Response Relationships of Radon Daughter Injury," p. 324. In
Conference/Workshop on Lung Cancer Epidemiology and Industrial Appli-
cations of Sputum Cytology (Golden: Colorado School of Mines Press).
Ax78 Axelson 0. and Sundell L., 1978, "Mining, Lung Cancer and Smoking,"
Scand. J. Work Environ. Health 4, 46.
45
-------
Ax80 Axelson 0. and Edding C. "Health Hazards from Radon Daughters in
Sweden." In Health Implications of New Energy Technologies, Proceedings
of an environmental health conference, April 4-7, 1979 (Washington: in
press).
Ba51 Bale W. F., 1951, Hazards Associated with Radon and Thoron. Memo
dated March 14, 1951, Division of Biology and Medicine, (Washington, DC:
Atomic Energy Commission). (Also found in Health Phys. 38, 1061.)
Be82 Bean J. A., Isacson P., Hahne R. M. A. and Kohler J., 1982, "Drinking
Water and Cancer Incidence in Iowa," Am. J. Epidemiol. 116, 924.
Bo59 Bogue D. J., 1959, The Population of the United States, p. 375.
(Glencoe: Free Press of Glencoe).
Ca76 Ca-A Cancer J. Clin., 1976, 18-19.
Ch74 Chameaud J., Perraud R., Lafuma J., Masse R. and Pradel J., 1974,
"Lesions and Lung Cancers Induced in Rats by Inhaled Radon-222 at Various
Equilibriums with Radon Daughters," p. 411. In Experimental Lung Cancer.
Carcinogenesis and Bioassays (edited by E. Karbe and J. F. Park), (New
York: Springer-Verlag).
46
-------
Ch80 Chameaud J., Perraud R., Chretien J., Masse R. and Lafuma J., 1980,
"Combined Effects of Inhalation of Radon Daughter Products and Tobacco
Smoke," p. 551. In Pulmonary Toxicology of Respirable Particles
(edited by C. L. Sanders, F. T. Cross, G. E. Dagle and J. A. Mahaffey),
C0NF-791002. Springfield, VA: National Technical Information Service).
Ch81 Chameaud J., Perraud R., Masse R. and Lafuma J., 1981. "Contribution
of Animal Experimentation to the Interpretation of Human Epidemiological
Data." In Proceedings of International Conference Radiation Hazards in
Mining: Control, Measurement and Medical Aspects, (edited by M. Gomez)
(Kingsport: Kingsport Press, Inc.).
CI79 Cliff K. D., Davies B. L. and Reissland J. A., 1979, "Little Danger
from Radon," Nature 279, 12.
Co53 Cohn S. H., Skow R. K., and Gong J. K., 1953, "Radon Inhalation
Studies in Rats," Arch. Ind. Hyg. Occup. Med. ]_t 508.
Coh79 Cohen B. L., 1979, "Radon: Characteristics, Natural Occurrence,
Technological Enhancement, and Health Effects," Prog. Nucl. Energy 4,
1-24.
Cr74 Cross, F. T., Bloomster C. H., Hendrickson P. L., Nelson I. C.,
Hooper B. L., Merrill J. A. and Stuart B. 0., 1974, Evaluation of Methods
for Setting Occupational Health Standards for Uranium Mines, Research
47
-------
Report prepared for the National Institute for Occupational Safety and
Health (Richland, WA: Battelle, Pacific Northwest Laboratories).
Cr78 Cross F. T., Palmer R. F., Filipy R. E., Busch R. H. and Stuart B. 0.,
1978, Study of the Combined Effects of Smoking and Inhalation of Uranium
Ore Dust, Radon Daughters and Diesel Oil Exhaust Fumes in Hamsters and
Dogs, Battelle, Pacific Northwest Laboratories Final Report PNL-2744
(Springfield, VA: National Technical Information Service).
Cr82 Cross F. T., Palmer R. F., Busch R. H. and Buschbom R. L., 1982,
"Influence of Radon Daughter Exposure Rate and Uranium Ore Dust Concen-
tration on Occurrence of Lung Tumors," p. 189. In Proceedings of Special-
ist Meeting on Assessment of Radon and Daughter Exposure and Related Bio-
logical Effects, (edited by 6. F. Clemente, A. V. Nero, F. Steinhausler
and M. E. Wrenn), (Salt Lake City: R. D. Press).
De78 Desrosiers A. E., Kennedy A. and Little J. B., 1978. " Rn Daughter
Dosimetry in the Syrian Golden Hamster Lung," Health Phys. 35, 607.
Dev64 deVilliers A. J. and Windish J. P., 1964, "Lung Cancer in a Fluor-
spar Mining Community: I. Radiation, Dust and Mortality Experience."
Br. J. Ind. Med. 21, 94.
Du76 Duncan D. L., Gesell T. F. and Johnson R. H., 1976, "Radon-222 in
Potable Water," p. 340. In Proceedings of the Tenth Midyear Health
48
-------
Physics Society Topical Symposium on National Radioactivity in Man's
Environment, C0NF-761031. (Troy: Rensselaer Polytechnic Institute
Press).
Ea82 Eaton R. S., 1982, "Radon and Radon Daughters in Public, Private and
Commercial Buildings Associated with Uranium Mining and Processing in
Canada." In Second Special Symposium on Natural Radiation in the
Environment (New Delhi: Wiley Eastern Ltd.).
En80 Enstrom J. E. and Godley F. H., 1980, "Cancer Mortality Among a
Representative Sample of Non-Smokers in the United States During
1966-68," J. Natl. Cancer Inst. 65, 1175.
EPA77 EPA Office of Radiation Programs, 1977, Radiological Quality of the
Environment in the United States, Environmental Protection Agency Report
EPA5201-77-009, PB-274299 (Washington: U.S. Environmental Protection
Agency).
EPA80 EPA Office of Radiation Programs, 1980, Draft Environmental Impact
Statement (DEIS) for Remedial Action Standards for Inactive Uranium
Processing Sites, Environmental Protection Agency Report EPA520/4-80-011
(Washington: U.S. Environmental Protection Agency).
Ev67 Evans R. D., 1967, "On the Carcinogenicity of Inhaled Radon Decay
Products in Man (CORD)," p. 1188. Report submitted before the Joint
49
-------
Committee on Atomic Energy, Subcommittee on Research, Development and
Radiation, July 1967: Hearing on Radiation Exposure of Uranium
Miners, Part 2 (Washington: U.S. Government Printing Office).
Ev69 Evans R. D., 1969, "Engineers Guide to the Elementary Behavior of
Radon Daughters," Health Phys. 17, 229.
Ev81 Evans R. D., Harley J. H., Jacobi W., McLean A. S., Mills W. A. and
Stewart C. G., 1981, "Estimate of Risk from Environmental Exposure to
Radon-222 and its Decay Products," Nature 290, 98.
Fr77 Fry R. M., 1977, "Radon and its Hazards," p. 13. In Proceedings of
NEA Specialist Meeting, Personal Dosimetry and Area Monitoring Suitable
for Radon and Daughter Products (Paris: Nuclear Energy Agency, OECD).
FR83 Fed. Reg., 1983, "Standards for Remedial Actions at Inactive Uranium
Processing Sites," Vol. 48, No. 3, 590.
FRC67 Federal Radiation Council Staff Report No. 8 (Revised), 1967,
Guidance for the Control of Radiation Hazards in Uranium Mining
(Washington: U.S. Government Printing Office).
Ga80 Garfinkel L., 1980, "Cancer Mortality in Non-Smokers: Prospective
Study by the American Cancer Society," J. Natl. Cancer Inst. 65, 1169.
50
-------
Gas72 Gastineau R. M., Walsh P. J. and Underwood N., 1972, "Thickness of
Bronchial Epithelium with Relation to Exposure to Radon," Health Phys.
23, 857.
Ge80 George A. C. and Breslin A. J., 1980, "The Distribution of Ambient
Radon and Radon Daughters in Residential Buildings in the New Jersey-New
York Area." In: The Natural Radiation Environment III (Houston, Texas).
(Springfield, VA: National Technical Information Service).
Hae58 Haenszel W., Shimkin M. B. and Mantel N., 1958, "A Retrospective
Study of Lung Cancer in Women," J. Natl. Cancer Inst. 21, 825.
Haq66 Haque A. K. M. M., 1966, "Energy Expended by Alpha Particles in Lung
Tissue. Br. J. Appl. Phys. 17, 905.
Haq67a Haque A. K. M. M., 1967, "Energy Expended by Alpha Particles in
Lung Tissue. II. A Computer Method of Calculation," Br. J. Appl. Phys.
18, 657.
Haq67b Haque A. K. M. M. and Collinson A. J. L., 1967, "Radiation Dose to
the Respiratory System due to Radon and its Daughter Products," Health
Phys. 13, 431.
Har72 Harley N. H. and Pasternack B. S., 1972, "Alpha Absorption Measure-
ments Applied to Lung Dose from Radon Daughters," Health Phys. 23, 771.
51
-------
Har81 Harley N. H. and Pasternack B. S., 1981, "A Model for Predicting
Lung Cancer Risks Induced by Environmental Levels of Radon Daughters,"
Health Phys. 40, 307.
Har82 Harley N. H. and Pasternack B. S., 1982, "Environmental Radon
Daughter Alpha Dose Factors In A Five-Lobed Human Lung," Health Phys. 42,
789.
Harl53 Harley J. H., 1953, "Sampling and Measurement of Airborne Daughter
Products of Radon," Nucleonics 11, 12.
Harr54 Harris S. J., 1954, "Radon Levels in Mines in New York State,"
Arch. Ind. Hyg. Occup. Med. 10, 54-60.
He79 Hewitt D., 1979, "Biostatistical Studies on Canadian Uranium Miners,"
pp. 264, 398. In Conference/Workshop on Lung Cancer Epidemiology and
Industrial Applications of Sputum Cytology (Golden: Colorado School of Mines
Press).
Hes82 Hess, C. T., Weiffenbach C. V. and Norton S. A., 1982, "Variations
of Airborne and Waterborne Rn-222 in Houses in Maine," Environ. Inter. 8,
59.
52
-------
Ho82a Hofmann W., 1982, "Cellular Lung Dosimetry for Inhaled Radon Decay
Products as a Base for Radiation-Induced Lung Cancer Risk Assessment I.
Calculation of Mean Cellular Doses," Radiat. Environ. Biophys. 20, 95.
Ho82b Hofmann W., 1982, "Cellular Lung Dosimetry for Inhaled Radon Decay
Products as a Base for Radiation-Induced Lung Cancer Risk Assessment II.
Microdosimetric Calculations," Radiat. Environ. Biophys. 20, 113.
Ho83 Hofmann W. and Katz R., 1983, "Track Structure Analysis of Radiation
Carcinogenesis at Low Doses," p. 565. In Radiation Protection, (edited
by J. Booz and H. G. Ebert), CEC-Report EUR8395.
Hu39 Huech W., 1939, "Kurzer Bericht uber Ergebnisse anatomischer Unter-
suchungen in Schneeberg," Z. Krebsforschung 49, 312.
Hur65 Hursh J. B., Morken D. A., Davis T. P. and Lovaas A., 1965, "The
Fate of Radon Ingested by Man," Health Phys. 11, 465.
ICRP75 International Commission on Radiological Protection, 1975, Report
of the Task Group on Reference Man, ICRP Publication 23 (New York:
Pergamon Press).
ICRP77 International Commission on Radiological Protection, 1977,
Radiation Protection in Uranium and Other Mines, ICRP Publication 24 (New
York: Pergamon Press).
53
-------
ICRP80 International Commission on Radiological Protection, 1980,
Biological Effects of Inhaled Radionuclides, ICRP Publication 31 (New
York: Pergamon Press).
ICRP81 International Commission on Radiological Protection, 1981, Limits
for Inhalation of Radon Daughters by Workers, ICRP Publication 32 (New
York: Pergamon Press).
Ja40 Jackson M. L., 1940, The Biological Effects of Inhaled Radon,
Master's Thesis, Massachusetts Institute of Technology, Cambridge, MA.
Jac64 Jacobi W., 1964, "The Dose to the Human Respiratory Tract by
??? ?70
Inhalation of Short-Lived Rn and Rn-Decay Products," Health
Phys. 10, 1163.
Jac72 Jacobi W., 1972, "Relations Between the Inhaled Potential a-Energy
222 220
of Rn and Rn-Daughters and the Absorbed a-Energy in the
Bronchial and Pulmonary Region," Health Phys. 23, 3.
Jac77 Jacobi W., 1977, "Interpretation of Measurements in Uranium Mines:
Dose Evaluation and Biomedical Aspects," p. 33. In Proceedings of NEA
Specialist Meeting, Personal Dosimetry and Area Monitoring Suitable for
Radon and Daughter Products (Paris: Nuclear Energy Agency, OECD).
54
-------
Jac80 Jacobi W. and Eisfeld K., 1980, Dose to Tissues and Effective Dose
Equivalent by Inhalation of Radon-222, Radon-220 and Their Short-Lived
Daughters, Gesellschaft fur Strahlen-und Umweltforschung MBH, Report
GSF S-626 (Munich-Neuherberg: Institut fur Strahlenschutz).
Jam80 James A. C., Greenhalgh J. R. and Birchall A., 1980, "A Dosimetric
Model for Tissues of the Human Respiratory Tract at Risk from Inhaled
Radon and Thoron Daughters," p. 1045. In Radiation Protection. A System-
atic Approach to Safety, Proceedings, 5th International Congress of IRPA,
Jerusalem, March 1980, Vol. 2 (Oxford: Pergamon Press).
Jam81 James A. C., Jacobi W. and Steinhausler F., 1981, "Respiratory Tract
Dosimetry of Radon and Thoron Daughters: The State-of-the-Art and Impli-
cations for Epidemiology and Radiobiology," p. 42. In Proceedings of
International Conference Radiation Hazards in Mining: Control, Measure-
ment and Medical Aspects, (edited by M. Gomez) (Kingsport: Kingsport
Press, Inc.).
Jan26 Jansen H. and Schultzer P., 1926, "Experimental Investigations into
Internal Radium Emanation Therapy. I. Emanatorium Experiments with
Rats," Acta Radiol. j5, 631.
JCAE67 Joint Committee on Atomic Energy, Subcommittee on Research Develop-
ment and Radiation, Congress of the United States, Ninetieth Congress,
May-August 1967: Hearings on Radiation Exposure of Uranium Miners
(Washington: U.S. Government Printing Office).
55
-------
JCAE69 Joint Committee on Atomic Energy, Subcommittee on Research Develop-
ment and Radiation, Congress of the United States, Ninety-First Congress,
March 17 and 18, 1969: Hearings on Radiation Exposure of Uranium Miners
(Washington: U.S. Government Printing Office).
Jo73 Jorgensen H. S., 1973, "A Study of Mortality from Lung Cancer Among
Miners in Kiruna, 1950-1970," Work, Environ., Health 10, 126.
Ka80 Kahlos H. and Asikainen M., 1980, Internal Radiation Doses from
Radioactivity of Drinking Water in Finland," Health Phys. 39, 108.
Ku59 Kushneva V. S., 1959, "On the Problem of the Long-Term Effects of
Combined Injury to Animals of Silicon Dioxide and Radon," p. 22. AEC
TR-4473.
Mc79 McCullough R., Stocker H., and Makepeace C. E., 1979, "Pilot Study on
Radon Daughter Exposure in Canada," p. 183. In Conference/Workshop on
Lung Cancer Epidemiology and Industrial Application of Sputum Cytology
(Golden: Colorado School of Mines Press).
McG80 McGregor R. G., Vasudev P., Letourneau E. G., McCullough R. S.,
Prantl F. A. and Taniguchi H., 1980, "Background Concentrations of Radon
and Radon Daughters in Canadian Homes," Health Phys. 39, 285.
56
-------
McP79 McPherson R. B., 1979, Environmental Radon and Radon Daughter
Dosimetry in the Respiratory Tract, Pacific Northwest Laboratory Report
PNL-2898 (Springfield, VA: National Technical Information Service).
Mo55 Morken D. A., 1955, "Acute Toxicity of Radon," AMA Arch. Ind. Health
12, 435.
Mo66 Morken D. A. and Scott J. K., 1966, Effects on Mice of Continual
Exposure to Radon and Its Decay Products on Dust. University of
Rochester Atomic Energy Project Report UR-669 (Springfield, VA: National
Technical Information Service).
Mo73a Morken D. A., 1973a, "The Biological Effects of the Radioactive
Noble Gases," p. 469. In Noble Gases, edited by R. E. Stanley and A. A.
Moghissi, National Environmental Research Center Report C0NF-730915
(Springfield, VA: National Technical Information Service).
Mo73b Morken D. A., 1973b, "The Biological Effects of Radon on the Lung,"
p. 501. In Noble Gases, edited by R. E. Stanley and A. A. Moghissi,
National Environmental Research Center Report C0NF-730915 (Springfield,
VA: National Technical Information Service).
My79 Myers D. K. and Stewart C. G., 1979, "Some Health Aspects of Canadian
Uranium Mining," p. 368. In Conference/Workshop on Lung Cancer Epidemio-
logy and Industrial Application of Sputum Cytology (Golden: Colorado
57
-------
School of Mines Press). (Also in Chalk River Laboratory Report
AECL5970.)
NAS72 National Academy of Sciences, 1972, The Effects on Populations of
Exposure to Low Levels of Ionizing Radiation (BEIR Report) (Washington:
National Academy Press, National Academy of Sciences.
NAS80 National Academy of Sciences, 1980, The Effects on Populations of
Exposure to Low Levels of Ionizing Radiation (BEIR-III Report)
(Washington: National Academy Press, National Academy of Sciences).
NAS81 National Academy of Sciences, 1981, Indoor Pollutants (Washington:
National Academy Press, National Academy of Sciences).
NCRP71 National Council on Radiation Protection and Measurements, 1971,
Basic Radiation Protection Criteria, NCRP Report No. 39 (Bethesda:
National Council on Radiation Protection and Measurements).
NCRP84a National Council on Radiation Protection and Measurements, 1984a,
Evaluation of Occupational and Environmental Exposures to Radon and Radon
Daughters in the United States, NCRP Report No. 78 (Bethesda: National
Council on Radiation Protection and Measurements).
NCRP84b National Council on Radiation Protection and Measurements, 1984b,
Exposure From the Uranium Series with Emphasis on Radon and Its
58
-------
Daughters, NCRP Report No. 77 (Bethesda: National Council on Radiation
Protection and Measurements).
Ne74 Nelson I. C. and Parker H. M., 1974, A Further Appraisal of Dosimetry
Related to Uranium Mining Health Hazards, U.S. Public Health Service
Report CPE 69-1131 (Cincinnati: National Institute of Occupational
Safety and Health).
NI71 National Institute for Occupational Safety and Health/National
Institute of Environmental Health Sciences, Joint Monograph No. 1, 1971,
Radon Daughter Exposure and Respiratory Cancer - Quantitative and
Temporal Aspects (Springfield, VA: National Technical Information
Service).
Pa69 Parker H. M., 1969, "The Dilemma of Lung Dosimetry," Health Phys. 16,
553.
Pe70 Perraud R., Chameaud J., Masse R. and Lafuma J., 1970, "Cancers pul-
monaires experimentaux chez le rat apres inhalation de radon associe a
des poussieres non-radioactives," Compt. Rend. Ser. D. 270, 2594.
Pr81 Prichard H. M., Hess C. T., Nyberg P., Weiffenbach C. and Gessell T.,
1981, "Integrated and Grab Sample Radon Data from Dwellings in Maine and
Texas," In Proceedings of International Conference on Indoor Air Pollu-
tion, October 1981 (Amherst, MA); (in press).
59
-------
Ra81 Radford E. P., 1981, "Radon Daughters in the Induction of Lung Cancer
in Underground Miners," p. 151. In Banbury Report 9, Quantification of
Occupational Cancer (edited by R. Peto and M. Schneiderman), (Cold
Spring Harbor: Cold Spring Harbor Laboratory).
Raj42a Rajewsky B., Schraub A. and Schraub E., 1942a, "Uber die toxische
Dosis bei Einatmung von Ra-Emanation," Naturwissenschaften 30, 489.
Raj42b Rajewsky B., Schraub A. and Schraub E., 1942b, "Zur Frage der
Toleranz-Dosis bei der Einatmung von Ra-Em," Naturwissenschaften 30,
733.
Re39 Read J. and Mottram J. C., 1939, "The 'Tolerance Concentration' of
Radon in the Atmosphere," Br. J. Radiol. 12, 54.
Ren74 Renard K. B., 1974, "Respiratory Cancer Mortality in an Iron Mine in
Northern Sweden," Ambio. 3, 67.
RPC80 Radiation Policy Council, 1980, Report of the Task Force on Radon in
Structures (United States Radiation Policy Council Report RPC-80-002,
U.S. Radiation Policy Council, Washington, DC).
Sa82 Saccomanno G., 1982, "The Contribution of Uranium Miners to Lung
Cancer Histogenesis," Recent Results Cancer Res. 82, 43.
60
-------
Se76 Sevc J., Kunz E. and Placek V., 1976, "Lung Cancer in Uranium Miners
and Long Term Exposure to Radon Daughter Products," Health Phys. 30, 433.
Sh54 Shapiro J., 1954, An Evaluation of the Pulmonary Radiation Dosage
from Radon and Its Daughter Products, University of Rochester Atomic
Energy Project Report UR-298 (Rochester: University of Rochester).
Sn73 Snihs J. 0., 1973, "The Significance of Radon and Its Progeny as
Natural Radiation Sources in Sweden," p. 115. In Noble Gases, edited by
R. E. Stanley and A. A. Moghissi, National Environmental Research Center
Report C0NF-730915 (Springfield, VA: National Technical Information
Service).
Sn74 Snihs J. 0., 1974, "The Approach to Radon Problems in Non-Uranium
Mines in Sweden," p. 900. In Proceedings of the Third International
Congress of the International Radiation Protection Association, U.S.
Atomic Energy Commission Report C0NF-730907 (Springfield, VA: National
Technical Information Service).
SS55 Seven State Uranium Mining Conference on Health Hazards, 1955, Hotel
Utah, Salt Lake City, Utah.
St80 Steinhausler F., Hofmann, W., Pohl E. and Pohl-Ruling J., 1980,
"Local and Temporal Distribution Pattern of Radon and Daughters in an
Urban Environment and Determination of Organ Dose Frequency Distributions
61
-------
with Demoscopical Methods," (The National Radiation Environment III,
Houston, TX, April 1978), (Springfield: National Technical Information
Service).
Str80 Stranden E., 1980, "Radon in Dwellings and Lung Cancer. A Discus-
sion," Health Phys. 38, 301.
Stu78 Stuart B. 0., Palmer R. F., Filipy R. E. and Gaven J., 1978, "In-
haled Radon Daughters and Uranium Ore Dust in Rodents," Pacific Northwest
Laboratory Annual Report for 1977 to the DOE Assistant Secretary for
Environment, PNL-2500 PT1 (Springfield, VA: Nationa-1 Technical Informa-
tion Service).
Su72. Suomela M. and Kahlos H., 1972, "Studies on the Elimination Rate and
222
the Radiation Exposure Following Ingestion of Rn Rich Water," Health
Phys. 23, 641.
Sul82 Sullivan R. E. and Nelson N. S., 1982, Ingested Radon Models, memo
dated April 1, 1982, Criteria and Standards Division (Washington, DC:
United States Environmental Protection Agency).
UN72 United Nations Scientific Committee on the Effects of Atomic Radia-
tion, 1972, Ionizing Radiation: Levels and.Effects (New York: United
Nations).
62
-------
UN77 United Nations Scientific Committee on the Effects of Atomic Radia-
tion, 1977, Sources and Effects of Ionizing Radiation (New York: United
Nations).
UN81 United Nations Scientific Committee on the Effects of Atomic
Radiation, 1981, Radon and Thoron and Their Decay Products (Thirtieth
Session of UNSCEAR, Vienna, 6-10 July 1981, prepared in the Secretariat)
(New York: United Nations).
UN82 United Nations Scientific Committee on the Effects of Atomic
Radiation, 1982, Ionizing Radiation: Sources and Biological Effects
(New York: United Nations).
USPHS57 United States Public Health Service, Pub. 494,1957, Control of
Radon and Daughters in Uranium Mines and Calculations on Biologi-eal
Effects (Washington: U.S. Government Printing Office).
USPHS61 United States Public Health Service, Pub. 843,1961, Governor's
Conference on Health Hazards in Uranium Mines, A Summary Report
(Washington: U.S. Government Printing Office).
Va78 van Kaick G., Lorenz D., Muth H. and Kaul A., 1978, "Malignancies in
German Thorotrast Patients and Estimated Tissue Dose," Health Phys. 35,
127.
63
-------
Va83 van Kaick 6., Muth H., Kaul A., Immich H., Liebermann D., Lorenz D.,
Lorenz W. J., Luhrs, H., Scheer K. E., Wagner G., Wegener K. and Wesch
H., 1983, "Recent Results of the German Thorotrast Study - Epidemio-
logical Results and Dose Effect Relationships in Thorotrast Patients,"
Health Phys. 44, Suppl. 1, 299.
Vo64 Von Dobeln W. and Li ridel 1 B., 1964, "Some Aspects of Radon Contamina-
tion Following Ingestion," Ark. Fys. 27, 531.
Wa64 Wagoner J. K., Archer V. E., Carroll B. E. and Holaday D. A., 1964,
"Cancer Mortality Patterns Among U.S. Uranium Miners and Millers, 1950
through 1962," J. Natl. Cancer Inst. 32, 787.
Wal70 Walsh P. J., 1970, "Radiation Dose to the Respiratory Tract of
Uranium Miners," Environ. Res. 14.
Wal71 Walsh P. J., 1971, "Relationship of Experimental to Empirical
Findings and Theoretical Dose Calculations," Final Report of Subgroup IB,
Interagency Uranium Mining Radiation Review Group (Rockville: EPA).
Wal79 Walsh P. J., 1979, "Dose Conversion Factors for Radon Daughters,"
Health Phys. 36, 601.
We63 Weibel E. R., 1963, Morphometry of the Human Lung (New York:
Academic Press).
64
-------
WH078 World Health Organization, 1978, World Health Statistics Annual
(Geneva: World Health Organization).
Wi82 Wise K. N., 1982, "Dose Conversion Factors for Radon Daughters in
Underground and Open-Cut Mine Atmospheres," Health Phys. 43, 53.
Wr77 Wright E. S. and Couves C. M., 1977, "Radiation-Induced Carcinoma of
the Lung - The St. Lawrence Tragedy," J. Thorac. Cardiovasc. Surg. 74,
495.
Ye80 Yeh H. C. and Schum M., 1980, "Models of Human Lung Airways and Their
Application to Inhaled Particle Deposition," Bull. Math. Biol. 42, 461.
65
-------
TABLE CAPTIONS
Table 1. Lifetime Lung-Cancer Risk Coefficients for Lifetime
Environmental Exposures
Table 2. Risk-Coefficient Adjustment Factors vs
Radon-Daughter Disequilibrium
Table 3. Summary of Cancer Risks per pCi/L Radon
Concentration in Drinking Water
Table 4. Lifetime Cancer-Risk Versus Radon Concentration
in Water
Table 5. Calculated Distribution of Lung-Cancer Risk for Populations
Exposed to an Average of 0.2 WLM/yr Radon Daughters (GSD, 2.5)
Table 6. Calculated Distribution of Lung-Cancer Risk for Populations
Exposed to an Average of 0.13 WLM/yr Radon Daughters (GSD, 2.5)
66
-------
Table 1. Lifetime Lung-Cancer Risk Coefficients for Lifetime
Environmental Exposures^
Age at First Exposure Units
Exposure Per WLM/yr Per pCi/L~
1 yr 9.1 x 10*3 3.6 x 10*3
Mixed(b) 5.6 x 10"3 2.1 x 10"3
(^Radon daughter equilibrium factor = 0.71,
unattached RaA/Rn ratio = 0.07.
* Pertains to populations with ages char-
acteristic of U.S. inhabitants in 1975.
67
-------
Table 2. Risk-Coefficient Adjustment Factors vs
Radon-Daughter Pisequi1ibrium
Radon and Radon
Daughter Ratios
1/0.9/0.7/0.7
1/0.9/0.6/0.4
1/0.6/0.3/0.2
1/0.256/0.098/0.084
Equilibrium
Factor^
0.71
0.55
0.29
0.11
Adjustment
Factor for
Radon-Daughter
Risk Coefficient
1.00
1.05
1.30
2.21
Adjustment Factor
for Radon-Gas
Risk Coefficient
1.00
0.81
0.53
0.34
(a
;The equilibrium factor is the ratio of the total potential alpha energy
of the actual, short-lived daughter concentrations to the total potential
alpha energy that the daughters would have if they were in equilibrium
with radon.
68
-------
Table 3. Summary of Cancer Risks per pCi/L Radon
Concentration in Drinking Water
Organ Lifetime Risk Annual Risk^
Lung^
Stomach
Whole Body
3 to 7.5 x lO"7
<6 x 10'8
1 x 10"8
7 to 17 x 10"9
<1 x 10"9
2 x 10"10
^Lung cancer risks are derived from the lung-
cancer-prediction model.
^Assuming 45 years at risk for lung cancer and
60 years at risk for stomach and whole-body
cancer deaths.
69
-------
Table 4. Lifetime Cancer-Risk Versus Radon Concentration
in Water
Radon Concentration
(pCi/L)
5,000
10,000
15,000
20,000
25,000
Lifetime Risk
(*)
0.2 to 0.4^
0.4 to 0.8
0.6 to 1.2
0.8 to 1.6
1 to 2^
^Equivalent to estimated lifetime risk from
estimated average indoor concentrations
(1 pCi/L) in houses.
^Equivalent to estimated lifetime risk of
occupational standard of 4 WLM/yr for 30 yr.
70
-------
Table 5. Calculated Distribution of Lung-Cancer Risk for Populations
Exposed to an Average of 0.2 MLM/yr Radon Daughters (GSD, 2.5)
Fractional
l of LifetT, Fractional Annual °eaths
MLM/yr Range Population Rislc ' Annual Risk 2.2 x 10 Persons^ '
0 - 0.2 68 6.0 x 10"4 1.3 x 10"5 2919
0.2 - 0.5 25 7.9 x 10"4 1.8 x 10"5 3865
0.5 - 1.0 5.7 4.0 x 10~4 9.0 x 10~6 1970
1.0 - 2.0 1.2 1.7 x 10"4 3.8 x 10"6 830
2.0 - 4.0 0.14 4.0 x 10'5 8.8 x 10'7 194
TOTAL 100 2.0 x 10~3 4.4 x 10"5 9778
^The attributable risk from radon (radon-daughter) exposures can be
compared with the present lifetime risk of lung cancer [about 4 x 10"
in the U.S., according to Evans et al. (Ev81)], which is largely
attributable to cigarette smoking.
^Ten percent of the risk occurs above 1 WLM/yr; 2% occurs above 2 WLM/yr.
71
-------
Table 6. Calculated Distribution of Lung-Cancer Risk for Populations
Exposed to an Average of 0.13 MLM/yr Radon Daughters (GSD, 2.5)
WLM/yr Range
% of
Population
Fractional
Lifetime
Risk*a)
Fractional
Annual Risk
Annual Deaths per
2.2 x 108 Persons^
0 - 0.13
68
3.3 x 10"4
7.4 x 10~6
1630
0.13 - 0.5
29.3
7.5 x 10"4
1.7 x 10"5
3652
0.5 - 1.0
2.35
1.7 x 10"4
3.7 x 10"6
812
1.0 - 2.0
0.322
4.6 x 10"5
1.0 x 10~6
223
2.0 - 4.0
0.028
7.9 x 10"6
1.8 x 10"7
39
TOTAL
100
1.3 x 10"3
2.9 x 10~5
6356
^ ^The attributable risk from radon (radon-daughter) exposures can be «
compared with the present lifetime risk of lung cancer [about 4 x 10"
in the U.S., according to Evans et al. (Ev82)], which is largely
attributable to cigarette smoking.
^Four percent of the risk occurs above 1 WLM/yr; 1% occurs above 2 WLM/yr.
-------
COMMITTEE ON OCCURANCE
Chairman: Charles T. Hess
Recorder: William A. Coniglio
Committee Members: Thomas R. Horton
Jacqueline Michel
Howard M. Pritchard
-------
THE OCCURRENCE OF RADIOACTIVITY IN PUBLIC WATER SUPPLIES
IN THE UNITED STATES
C. T. Hess
University of Maine
Department of Physics, Orono, Maine 04469
J. Michel
Research Planning Institute, Inc.
925 Gervais, Columbia, South Carolina 29201
T. R. Horton
Eastern Environmental Radiation Facility
United States Environmental Protection Agency
P.O. Box 3009, Montgomery, Alabama 36193
H. M. Prichard
University of Texas School of Public Health
P.O. Box 20186, Houston, Texas 77025
W. A. Coniglio
Office of Drinking Water
United States Environmental Protection Agency
401 Street, S.W., Washington, D.C. 20460
ABSTRACT
Examination of the collected data for radionuclide concen-
tration measurements in public water supplies in the United
States show more than 51,000 measurements for gross alpha
particle activity iand/or Radium, 89,900 measurements for uranium,
and 9,000 measurements for radon. These measurements were made
as part of national and state surveys of radionuclide con-
centrations in utility water supplies for radium and radon; and
the NURE survey for uranium which included non-utility water
supplies.
-------
Surface water has low values for radium and radon but levels
comparable to ground water for uranium. Separate isotope
measurements were not taken for much of the radium and uranium
data. Because 226Ra to 228Ra ratios and 238|j to 234j ratios are
not fixed in water, further measurements are needed to establish
the specific isotopic concentrations by region. Analysis of the
state average values in geological provinces show the highest
provincial areas for radium are the Upper Coastal Plain, glaciated
Central Platform, and Colorado plateau. For uranium, the highest
areas are Colorado plateau, West Central Platform, and Rocky
Mountains. For radon, the highest provinces are New England and
the Appalachian Highlands - Piedmont. Regional hydrogeological
and geochemical models are suggested for guiding the formulation
of regional standards and monitoring strategies. Utility
supplies serving small populations have the highest concentration
for each radionuclide and have the lowest fraction of samples
measured, which shows a need for further measurements of these
small population water supplies. Risk estimates for the average
concentration of radium in utility ground water give about 2700
fatal cancers per 70.7 year lifetime in the United States. Risk
estimates for the average concentration of uranium in utility
surface and ground water give about 600 fatal cancers per 70.7
year lifetime in the United States. Using 1 pCi/1 in air for
10,000 pCi/1 in water, the .radon in utility water risk estimate
is for 4,400-22,000 ,fatal cancers per 70.7 year lifetime in the
United States.
-------
PREFACE
The sections of this paper are arranged in the order of
introduction, geochemistry and occurrence. A central theme of
all sections of the report is that the geological setting
strongly controls the occurrence of natural radionuclides in
drinking water. The observed concentrations of U, Ra, and Rn in
ground and surface water can be related to the rock types and the
amount and distribution of U and Th in the materials which
constitute the aquifer and surficial deposits. The United States
can be divided into eleven geological provinces, each of which is
characterized by dominant types of rocks or deposits as well as
ground water flow systems, discussed in Table 1 and shown in
Figure 1. (Be81, Sc62). These provinces are discussed in all
sections of the report and provide a framework for understanding
the variations in the distribution and activities of natural
radionuclides in water. In fact, one hypothesis is that certain
provinces or sub-provinces can be characterized as producing
ground water with specific radionuclide problems, or conversely,
without specific radionuclide problems. If this hypothesis can
be verified, it has important applications to the development of
regional guidelines for monitoring requirements in the revised
regulations (La83).
-------
OCCURRENCE OF RADIUM ISOTOPES IN PUBLIC DRINKING WATER
INTRODUCTION
Radium has two natural isotopes which are of concern in
public water supplies. Radium-226 (Ra-226) is generated through
decay of uranium-238 (U-238) and is an alpha emitter with a t^/2
= 1,622 years. This is the isotope which is commonly referred to
as radium and has been measured in many water supplies. The
other isotope, Ra-228, is generated directly by thorium-232
(Th-232) decay and is a shorter-lived, weak-beta emitter (t^/2 =
5.7 years). There is a third isotope of radium which is of
possible concern, Ra-224 with a ti/2 = 3.64 d. Its occurrence
is not well knownf only a few data are available from samples at
the well head. The U.S. Environmental Protection Agency (EPA)
established interim regulations in 1976 tor maximum levels of
radioactivity in drinking water as follows«
Maximum contaminant levels (MCL) of combined Ra-226 and
Ra-228 - 5 picocuries per liter (pCi/l)i gross alpha-
particle activity - 15 pCi/1 excluding radon and uranium.
(Ep76a).
These MCLs were set under the authority of the Safe Drinking
Water Act to protect health, taking treatment costs in con-
sideration. In an effort to minimize the costs of analysis and
monitoring, EPA established a series of screening steps to test
for compliance with the interim regulations. These criteria
stated that when the average gross alpha-particle activity of
four quarterly samples or composites exceeds 5 pCi/1, the same or
-------
-2-
equivalent sample shall be analyzed for ka-226. If the activity
of Ra-226 exceeds 3.0 pCi/1, the sample shall be analyzed for
Ra-228. Inherent in these regulations were the assumptions that
Ra-226 was to be the dominant radioactive contaminant in drinking
water and the Ra-228/Ra-226 activity ratio was less than 1.0.
The regulations required all systems supplying 25 or more people
to be monitored every four years.
Since the interim regulations were established, much more
information on the occurrence of radium isotopes is now available
from state compliance data and from detailed studies on the
correlation and interrelationships of Ra-228 and Ra-226 in ground
water with specific geological provinces ((Mi80), (As81), (Ki82),
(Mi82), and (Kris82)). In light of these new data, the key issues
to be considered for revision of the regulations are«
1) Prioritization of specific areas for monitoring
for Ra-228 and Ra-226j
2) Reduction in the interval frequency or complete
omission for specific areas for repeat monitoring} and
3) Decoupling Ra-228 analysis from Ra-226, with criteria
for when Ra-228 is to be measured.
The purpose of this paper is to concisely review the
existing information on the geochemistry and occurrence of Ra-228
and Ra-226, and to provide guidelines for regulatory revision.
-------
-3-
GEOCHEMISTRY OF RADIUM ISOTOPES
The distribution of Ra-228 and Ra-226 in water is a function
of the thorium and uranium content of the aquifer, the geochemi-
cal setting of the aquifer solids, and the tx/2 of each isotope.
There are specific geological and chemical processes that control
the thorium and uranium content in aquifers, which are discussed
in detail by Olson and Overstreet (0164), Cherdynstev (Ch71), and
Gableman (Ga77). In fact, thorium and uranium have very similar
behavior, with one important exception which is most responsible
for their eventual separation. Thorium has one oxidation state
and is immobile at low temperatures. Therefore, thorium distri-
bution is controlled by primary geochemical processes (such as
magmatic crystallization) or secondary physical processes (such
as sedimentary enrichment in placer deposits). Uranium has two
oxidation states and the +6 state (uranyl) can form highly
soluble complexes which can be transported long distances by oxi-
dizing ground water before being removed by adsorption or reduc-
tion to the +4 state. The estimated average crustal Th/U
activity ratio is 1.2-1.5 so that, in the absence of enrichment
or depletion processes, Ra-228 activity should be higher than
Ra-226. However, the tendency for uranium enrichment under cer-
tain geochemical conditions results in regions of higher Ra-226,
thus EPA's decision to emphasize Ra-226 in the interim regula-
tions.
-------
-4-
Radium enters ground water by dissolution of aquifer solids;
by direct recoil across the liquid-solid boundary during its for-
mation by radioactive decay of its parenf. in the solid (both iso-
topes have thorium as the immediate parent), and by desorption. The
mechanism of alpha recoil is an important factor in the higher
solubility of daughter isotopes compared with their parents.
U-234/U-238 activity ratios in ground water are generally greater
than 1.0 and can be as high as 28 (Gi82). Ra-224/Ra-228 activity
ratios in South Carolina ground water range from 1.2 to 2.0 (W.
S. Moore, unpubl. data) and in Connecticut from 0.8 to 1.7
(Kris82). However, when the daughter/parent pair consists of dif-
ferent elements, geochemical factors become important controls of
their relative solubility. An extreme example is Rn-222, the
immediate daughter of Ra-226| Rn-222/Ra-226 activity ratios in
water can be as high as 10^. Because of alpha recoil and the
different solubilities of the thorium and uranium series isoto-
pes, extensive disequilibrium occurs in ground water.
Recent studies have suggested that radium is rapidly absorbed
from ground water. King et al. (Ki82) proposed that the distance
of Ra transport in ground water was less than that of Rn-222
(with a ti/2 = 3.8 days) due to continual adsorption of radium
onto the aquifer solids. Krishnaswami et al. (Kris82) calculated
adsorption and desorption rate constants for radium in Connecticut
aquifers and proposed that radium removal rates are rapid, as short
as a few minutes. Equilibrium between adsorption and desorption
-------
-5-
is also quickly established, but Krishnaswami et al. concluded
that the partition coefficient strongly favors the solid phase,
and almost all radium introduced into the ground water studied
resides on particle surfaces in the adsorbed state. However, the
extent of sorption is controlled by the geochemical reactivity of
the aquifer material. King et al. (Ki82) notes that the average
Ra-228 and Ra-226 activity in the crystalline aquifers of South
Carolina was lower than for the Coastal Plain sediments, even
though the thorium and uranium content of the rock aquifers was
higher. Furthermore, the Rn-222 activity in the crystalline
aquifers was ten times greater than the aquifers sampled in the
Coastal Plain. King et al. concluded that the affinity of radium
for adsorption sites in the fresh rock surfaces which have higher
cation-exchange capacities was greater than for the sand and gra-
vel deposits composed of refractory minerals such as quartz.
Thus, radium in ground water does not accumulate with ground
water transport in aquifers* it stays very close to the area in
which it is produced.
The insolubility of radium and thorium can be inferred from
studies of potential contamination of ground water due to seepage
from uranium tailings ponds in New Mexico reported by Kaufmann et
al. (Ka76). At one such pond, they estimated that nearly 3 x 10^
L of seepage entered the shallow aquifer over a 20-year period.
The wastes in this pond contained approximately 200 pCi/1 of
Ra-226 and 166,000 pCi/1 of Th-230. Thus, nearly a Ci of Ra-226
and 500 Ci of Th-230 were available to leach with the shallow
-------
-6-
ground water; yet, in 1975, monitoring wells located 1 km down-
gradient from the pond showed no evidence of contamination.
Through an understanding of the physical and chemical pro-
cesses which control radium distribution, we can now begin to
interpret the new data base from state compliance reports, and to
develop predictive models for radium occurrence on which new regula-
tions should be structured. These proposed models would charac-
terize certain geological settings or aquifer types as producing
ground water with high or low radium content. EPA has begun to
develop a predictive model for the occurrence of Ra-228, with a
pilot study completed for two geological provinces, the Atlantic
and Gulf Coastal Plain sedimentary aquifers and the Piedmont rock
aquifer of the eastern United States (Mi82). Information on
areas of high radium occurrence is necessary to provide guidance to
states for additional monitoring. From a regulatory point of
view, areas of low radium activity are very important, in that they
could have a different monitoring priority and schedule. A pre-
dictive model for Ra-228 would also be valuable because so few
samples were measured under the present analytical scheme.
OCCURRENCE OF Ra-226 AND Ra-228 IN DRINKING WATER
All but six states (Illinois, Nebraska, Colorado, Utah,
Montana, and Oregon) have reported known MCL violations for radium
as required by the interim regulations. There are approximately
-------
-7-
200 reported public water suppliers with Ra-226 activities in
excess of 5 pCi/1 after normal treatment (Coj in prep.). The
following sections discuss these results and other studies by
water types, geological setting, and isotope.
Surface Water
The radium content of surface water is usually very low. Ra-226
generally ranges between 0.1 and 0.5 pCi/1 and the Ra-228/Ra-226
activity ratio is generally greater than unity (Mo69)j (E183).
Also, standard water treatment methods are known to remove radium
(Ep76b). To the best of our knowledge, no surface water viola-
tions for radium have been reported by the states. Thus, surface
water systems should be separately evaluated; perhaps they could
be released from monitoring requirements for radium once the source
stream was documented as having low natural radioactivity.
Ground Water
Out of the nearly 60,000 public water supplies in the United
States, about 80 percent use ground-water sources. Over 90 per-
cent of the ground-water supplies serve less than 3,300 people
and are classified as small or very small. In general, radium in
drinking water is a small-system problem. Figure 2 is a com-
pilation of the areas and specific sites which have high radium in
ground water from both state compliance data and published studies.
-------
-8-
The available state compliance data for radium comes almost
exclusively from samples which first showed a gross alpha-
particle activity of > 5 pCi/1. Iowa used a screen of 2 pCi/1
for gross alpha-particle activity. In some areas, states would
analyze additional samples in an area where high radioactivity
was found during the initial sampling. Ra-228 data were provided
for about one-half of the 200 Ra-226 values reported. State-wide
summaries of Ra-226 and Ra-228 data have been published for
Georgia (C183), South Carolina (Ki82), Iowa (Krie82), and Illinois
(Ro77)j Lucas (Lu82) reported results for over 90 percent of the
communities in Illinois, Iowa, Missouri, and Wisconsin.
There have been several studies on the temporal variability
of the activity of radium isotopes in ground water systems. Kriege
and Hahne (Krie82) reported that the mean value for the average
percent deviations of 141 samples over 18 years in Iowa was 21
percent with a relative standard deviation of 15 percent. Michel
and Moore (Mi80) found a maximum variation of 19 percent over 2
years in individual wells. Therefore, in single-well systems,
one sample should be representative of the average annual
activity* also the present requirement for monitoring at 4-year
intervals would not be necessary unless changes to the system
have been made. Systems with multiple wells have the potential
problem of continuously variable radium based on the relative
contribution of each well when sampled.
-------
-9-
From the data reported by the states the mean total radium
activity for supplies exceeding the MCL was almost 10 pCi/1.
Ra-226 activity was generally greater than Ra-228 activity, but
these data were initially biased toward high Ra-226. King et al.
(Ki82) found that the average Ra-228/Ra-226 activity ratio was
1.2 for over 180 samples throughout South Carolina. Of these,
ten samples had Ra-226 less than 3 pCi/1, but total radium
greater than 5 pCi/1. King et al. (Ki82) estimated that perhaps
40-50 percent of the total radium violations for the Piedmont and
Coastal Plain provinces were missed using the prescribed
screening procedure which couples Ra-228 analysis to Ra-226.
Kriege and Hahne (Krie82) reported additional sampling which
identified eight violations for total radium although the Ra-226
was less than 3 pCi/1.
From the available data, there are two specific geological
regions where over 75 percent of the known, radium violations occuri
1) The Piedmont and Coastal Plain provinces in New Jersey,
North Carolina, South Carolina, and Georgia} and
2) A north-central region, consisting of parts of
Minnesota, Iowa, Illinois, Missouri, and Wisconsin.
The rest of the violations are generally scattered clusters,
notably along the Arizona-New Mexico borcier, Texas, Mississippi,
Florida, and Massachusetts (Fig. 2). All of these scattered
violations had high Ra-226 activities, as would be expected from
the screening methods used to detect them. Ra-228 activities in
these systems were very low. We believe that the current analy-
-------
-10-
tical protocol has detected a large percentage of the systems
with high Ra-226. Cothern and Lappenbusch (Co84) have used the
compliance data for Ra-226 to estimate that approximately 500
systems will be determined to exceed the MCL of 5 pCi/1.
Improvement on this estimate is difficult with the existing
data base, which is comprised mostly of reported MCL violations
for Ra-226. Statistical analysis of these data is not possible
because they were not randomly sampled. In this respect, states
should be requested to submit all radium results to facilitate
further analysis. However, some calculations can be made to
corroborate the previous estimates of MCL violations. Data
from South Carolina (Ki82) showed that approximately 3.0 percent
of the ground water supplies exceeded the 5.0 pCi/1 limit for
total radium. (Note that the prescribed screening procedures
detected only one-half of these violations). Applying that per-
centage to North Carolina and Georgia, both of which have similar
hydrogeology, provides an estimate of 150 violations for all
three states. In Iowa, approximately 10 percent of the 605
supplies sampled to date, using a lower screening criteria,
exceed the MCL. Again, applying this percentage to all the
ground water systems of Iowg and half of Illinois, Missouri and
Wisconsin yields 120,violations for Iowa. 75 violations for
Illinois and 50 for. Missouri, and 60 for Wisconsin. We can esti-
mate violations for the states that have not reported as followsi
10 each for Utah, Colorado, and Nebraska, and zero for Oregon and
-------
-11-
Montana. There are 71 violations reported in all other states.
The total of these known and estimated violations is 556.
Assuming 10-25 percent of the actual violations are missed during
the prescribed screening procedure (actual data for Iowa, 8 out
of 60 or 13 percent; for South Carolina, 8 out of 30 or 26
percent), the number of violations ranges between 600 and 700.
Estimates of population exposure nationwide can only be
broadly made without additional information on populations served
by the MCL violations, as well as on the results of all analyses.
Lucas (Lu82) estimated that 91 communities in Illinois, Iowa,
Missouri, and Wisconsin with a population of 599,000 consume
water with Ra-226 greater than 5.0 pCi/1.
Under the present screening methods, however, there was con-
cern that Ra-228 violations were not being detected. Thus, EPA
recently funded a study to determine if a predictive model for
the occurrence of Ra-228 could be developed. The Piedmont and
Coastal Plain aquifers were selected as a pilot study area for
development of a model because the radiochemistry of these pro-
vinces had been extensively studied, these were areas of known
high Ra-228 activities, and nearly 300 values for Ra-228 were
available.
The model comprises a multilevel classification of aquifer
characteristics for each Ra-228 datum. The nature of the data
-------
-12-
precluded use of such analytical tools as regression analysis to
establish a quantitative relationship between, for example, thorium
content and Ra-228. As a result, the underlying model structure
was evaluated to assess the existence of differences not only
between the major aquifer types but also between lower-level
classification within a major aquifer type. A detailed descrip-
tion of the model development, parameters, methodology, and
results can be found in a report by Michel and Pollman (Mi82).
Only a summary is discussed below. Also, although the model was
developed specifically for Ra-228, values for Ra-226 were
available and similar statistics were calculated. There is much
to be learned from these differences in the results for these two
isotopes.
Table 2 summarizes the means and ranges for those aquifer
types which had significantly different Ra-228 distributions.
Note the striking differences in the means, although the ranges
are similar in some cases. Arkosic (immature, feldspar-rich)
sand aquifers had mean values for both radium isotopes up to an order
of magnitude greater than quartzose sands. Limestones and meta-
morphic rock aquifers in the study area had very low activities
of both radium isotopes. Table 3 shows a ranking of all structural
levels used in the pilot study, with symbols indicating groups of
similarity of Ra-228 distribution (with Ra-226 means are also given
without a ranking). Classes identified by the same letter code
in the grouping column represent subsets of a group that are
statistically indistinguishable from other classes of the same
-------
-13-
group. However, groupings that overlap indicate that the par-
ticular individual groups are not unique. For example, group A
as a class is not statistically different from groups B, C, and
Dj groups E, F, and G, however, represent groups of classes with
significantly lower Ra-228 activities than group A. Although
some groups were not significantly different, the ranking
followed the anticipated trends within major types. For example,
the ranking of igneous rocks, with syenite>granite>diorite,
follows thorium abundance in these rock types. Arkosic sand aquifers
are ranked in order of thorium content of the source rock, from high to
low. This pilot study showed that specific aquifer types and
geochemical conditions can be characterized as producing ground
water with high or low Ra-228 activities. Its application can be
demonstrated for the aquifers of the Piedmont and Atlantic and
Gulf Coastal Plain provinces; high Ra-228 was likely to occur in
aquifers composed of (1) acidic, igneous rocks and (2) arkosic
sands with sources having high-to-medium thorium content.
The results from the model were used to map specific areas
(aquifer types), from New Jersey to Alabama, that would be likely
to produce ground water with high Ra-228. In fact, in the
Piedmont province, granitic rock aquifers younger than 350
million years were shown to produce high-radium ground water.
Older rocks had undergone metamorphism which has tended to
recrystallize thorium and uranium into resistate minerals in
which radium is more tightly bound. The arkosic sand aquifers
-------
-14-
were restricted to the upper Coastal Plain from Virginia to
Georgia. These aquifers are composed of sediments eroded from
the nearby Piedmont rocks and are mineralogically immature. They
contain higher amounts of thorium- and uranium-bearing minerals
than the middle and lower Coastal Plain sediments which were
deposited farther from the source rocks. Thus, the trend in the
Coastal Plain aquifers for Ra-228, whose parent is not subject to
secondary transport processes, is one of decreasing activities
with distance from the Piedmont source. Only 1 out of 50 samples
from the quartzose sands of the middle and lower Coastal Plains
aquifers was greater than 3 pCi/1 for Ra-228. In contrast,
Ra-226 in the middle Coastal Plain aquifer is highly variable,
with values from 0 to 196 pCi/1, due to the ability of its parent
to migrate in ground water and undergo secondary enrichment.
Knowledge of the conditions where very low radioactivity
will occur is also very important. In the area studied, low
Ra-228 occurred in aquifers of (1) metamorphic rocks, (23 quart-
zose sands, and (3) limestone. Thus, the lower Coastal Plain,
composed of extensive limestones and deep quartzose sand
aquifers, is notable for its total lack of Ra-228 greater than
1.0 pCi/1 in ground water. Ra-226 will be more variable because
of the high solubility of uranium complexes in the carbonate
system, but it is generally detected by the gross-alpha particle
activity screen. Nevertheless, there have been only three Ra-226
violations reported for the entire lower Coastal Plain province,
from New York to Texas. These violations were all from one
-------
-15-
region, in Florida. The second area is the North-Central region.
Much of the ground water comes from deep aquifers, frequently
having Ra-226 activities of 5-25 pCi/lj Ra-228 can be as high as
32 pCi/1 (Lu82). There is no apparent correlation between Ra-228
and Ra-226 and no specific trends in their distribution by
aquifer, depth in the aquifer, or areal extent. Interpretation
of the radium distribution in this area is complicated by complex
hydrogeology and multiply screened wells in different aquifers.
There is evidence of significant uranium migration, both during
geological time and the present, which provides a mechanism for
high Ra-226 as well as resulting in a complex distribution and
disequilibrium of uranium series isotopes (Gi82) j(Li02 ).
Possible sources for high Ra-228 have not been identified, but
the Ra-228 distribution may be able to be explained by analysis
of the sources, depositional setting, and diagenesis of the sedi-
mentary rock aquifers.
Limited work has been done on radium occurrence in the other
geological provinces. Ra-226 has been found to be high in areas
of uranium mineralization, such as in Texas and the Colorado
Plateau in Arizona and New Mexico (Fig. 2) and violations are
expected in Utah and Colorado when these states report. Thorium
enrichment zones, such as veins and placer deposits, are expected
to produce only scattered, local Ra-228 problems, due to its
limited transport in ground water. These areas would be extre-
mely difficult to locate under the present regulations. However
aquifers with much lower but disseminated thorium and uranium
-------
-16-
(such as granites, tuffaceous rocks, and immature sandstones) are
more likely to have higher background radioactivity and wider
occurrences of both Ra-228 and Ra-226 in ground water.
Occurrence of Ra-224
Data on Ra-224 activities in ground water are scarce.
However, it appears that the activity of Ra-224 is equal to or as
much as twice the Ra-228 activity and therefore could be as high
as 30-40 pCi/1. This Ra-224 activity is unsupported} activities
of its parent, Th-228, are usually less than 0.01 pCi/1. Thus it
enters ground water by alpha recoil during decay of Th-228
adsorbed on the surface of aquifer solids. The radiotoxicity of
Ra-224 and its daughters is small because of their extremely
short half-lives.
RADIUM CONCLUSIONS
From the state compliance data and other studies, much more
is now known about the occurrence of radium isotopes in public
drinking water supplies, and this information should be incor-
porated into the revised regulations.
1) Surface water has very low radium activities! the moni-
toring interval after initial validation should be
significantly lengthened, or perhaps omitted, for
surface-water systems.
-------
-17-
2) Fewer samples may be needed to determine the average
annual activities, particularly for single-well
systems; the monitoring interval could be lengthened
for unmodified systems.
3) Monitoring requirements for Ra-228 should be decoupled
from Ra-226. Instead, separate guidelines for the
occurrence of radium isotopes are needed.
4) Regional Ra-226 problems are fairly well known. The
few additional occurrences of high Ra-226 activity will
be difficult to find without analysis of every system.
As important, however, are those areas which have low
Ra-226. All values should be compiled regionally or
nationally, to document the Ra-226 distribution for
each geological province, with the goal to classify
areas with a high degree of certainty as producing low
Ra-226 ground water. The revised regulations should
include separate, less stringent and less costly moni-
toring requirements for such regions. This approach
would shift monitoring efforts toward known or uncer-
tain areas of high Ra-226 and provide more data on the
actual distribution of high Ra-226 activities, which
will allow for a better risk assessment.
-------
-18-
5) The occurrence of Ra-228 is not well known. It has
been shown that, using the present screening proce-
dures, 10 to 50 percent of the violations for total
radium are being missed. More extensive measurements
of Ra-228 would be difficult because of the problems
with the approved analytical method. An alternative
approach would be to develop a conceptual, predictive
model for Ra-228 occurrence, based on geochemical prin-
ciples, to identify specific types of aquifers which
are likely to have Ra-228 problems. Once verified,
this model should be the basis for developing regional
guidelines for monitoring in areas more likely to have
high Ra-228. This same approach can be used for
refining and interpreting occurrence data for uranium
and radon as part of the regulatory process of deve-
loping standards for these isotopes.
6) Finally, because aqueous radiochemistry is a complex,
technical field, EPA should pruvide State water-supply
personnel with background and explanatory guides in
laymen's language, which will assist them in
understanding radiological problems and in the imple-
mentation of the regulations.
-------
-19-
OCCURRENCES OF URANIUM ISOTOPES IN PUBLIC DRINKING WATER
INTRODUCTION
Uranium has three natural isotopes with long half-lives
(11/2) that permit transport into potable, water supplies. These
isotopes are uranium-238 (99.27% natural abundance), ti/2 =
4.7 x 109 years, uranium-235 (0.72SI5 natural abundance) ti/2 =
7.04 x 108 years, and uranium-234 (.00635 natural abundance)
tl/2 = 2.54 x 105 years. All of these isotopes emit alpha
radiation and produce a long decay series of daughters. The
group of uranium isotopes are found in the Earth's crust with an
abundance of 4 x 10-435 (Hu73) and are found in rocks and minerals
such as granite, metamorphic rocks, lignites, monazite sand, and
phosphate deposits as well as in uranium minerals such as urani-
nite, carnotite and pitchblend (Ca80). It is a trace element in
coal, peat, and asphalt and is present in some phosphate fer-
tilizers at a level of about 100 microgrp.ms/g or 67 pCi/g.
Despite its widespread abundance it has not been shown to be an
essential element for man (Hu73). There is no standard for ura-
nium in water supplies as a radioactive element since, until
recently, it has been considered by the NRC to be a toxic heavy
metal with the standard for ingestion relating to its chemical
toxicity (3 x 10* pCi/1), (10CFR 20,601, Appendix B). However,
some concentration measurements in potable water have been done
in association with gross alpha measurements for the radium
drinking water standard. The uranium activity measured was to be
-------
-20-
subtracted from the gross alpha activity measurement to show
compliance with the gross alpha standard (Dr81). A recent analy-
sis of uranium in water supplies was conducted by Oak Ridge
National Laboratory (Dr81) using 89,944 measurements of uranium
surface, ground, and domestic waters primarily obtained from the
National Uranium Resource Evaluation (NURE) program. The results
of this study are reviewed in this report.
GEOCHEMISTRY OF URANIUM
Although there are geological processes which enrich uranium
in certain rock formations, it occurs as a common trace element
in most rock types. Because of the insolubility of U4+, uranium
must be oxidized in order to be transported in ground water. The
greater solubility of U6+ is due in part to its tendency to form
uranyl di-and tri-carbonate anions. Thus, uranium solubility is
a function of not only the redox potential of water but also of
the pH and the partial pressure of CO2 in the system. In com-
parison to radium, the stability of the uranyl carbonate
complexes and their long half-life allow for uranium to be
transported long distances under oxidizing conditions. Uranium
is removed from solution by sorption or reducinig barriers, a
process which has been well described in the sandstone-type ura-
nium deposits in the western United States (Ga77).
There have been many studies of the isotopic composition of
uranium in natural waters which have shown that most contain more
-------
-21-
activity from U-234 than from U-238. The U-234/U-238 activity
ratio can be as high as 28, but usually ranges between 1 and 3
(Ch71, Gi82). The higher activity of U-234 in water is due to
its selective mobilization by alpha recoil. The natural abundan-
ces of isotopes and the half-lives gives 0.33, 0.015 and 0.33
pCi/yg of natural uranium for U-238, U-2^5 and U-234, respec-
tively, or 0.68 pCi/ug total. Thus, isotopic enrichment can
cause changes in the specific activity of the total sample of
uranium. Total depletion of U-234 from the sample and replace-
ment by an equal activity of U-238 will result in no net change
of total activity* however, the total mass of uranium would
almost double. The human dosimetry will also be changed since
the alpha energies are not the same. Methods which depend on the
mass of uranium will not predict the correct activity for samples
with variable uranium isotope enrichment.
OCCURRENCE OF URANIUM IN GROUND AND SURFACE WATER
Uranium concentration in water depends on factors such as
the uranium concentration in host aquifer rock, presence of oxy-
gen and complexing agents, chemicals, in the aquifer, chemical
reactions with ions in solution and the nature of the contact
between the uranium minerals and the water. These factors vary
(Sc62) with regions of the United States due to rainfall,
geology, and ground water flow patterns, and to anthropogenic
factors such as use rate of ground water and surface water.
Thus, one would expect large variations of uranium content from
-------
-22-
state-to-state. The geological provinces of the coterminous
United States derived from generalizations of rock types and
hydrological flow systems are shown in Figure 1 and Table 1.
These zones can be compared to the uranium concentrations
averaged by state from the (NURE) measurements shown in Figures
3, 4, 5. These averages are surprisingly similar, showing dif-
ferences of only factors of 4 higher concentrations in ground
water than in surface water, with many states identical for
ground and surface water. The average values for each province
is given in Table 4 with average values ranging from .02 in pro-
vince 2, the Appalachian Mountains, up to 2.3 in province 8, the
Colorado Plateau. By grouping low, medium, and high averages,
one sees the four major zones of similar concentration> Zone 1,
the Appalachian Mountains and New England} Zone 2, Appalachian
and Interior Plateau, and Coastal Plain$ Zone 3, the Glaciated
Central Platform, Western Central Platform, Rocky Mountain
System, and Colorado Plateauj Zone 4, the Basin and Range and the
Columbia Plateau and Pacific Mountain System.
The provinces chosen by Beddinger (Be81) may be compared
with those chosen in 1962 by Scott and Barker (Sc62). These pro-
vinces are shown in Figure 1 and compared by region in Table 1.
Concentrations of uranium are given in Table 4 for provincial
schemes. The data of Scott and Barker comprise 561 samples
collected in the coterminous United States from 1954 to 1958 and
are expressed in yg/L (thus they represent U-238 only).
-------
-23-
Table 5 shows population versus uranium concentration for
drinking water sources with more than 10,000 people (Dr81). The
levels are given up to greater than 100,000 population. Due to
limitations in the source information (Dr81), no information is
available for cities of populations less than 10,000, showing a
need for more information on small systems.
RELATIVE SOURCE CONTRIBUTION OF URANIUM
The dietary intake of uranium in United States food is
variously reported from 0.87 to 0.94 picocuries per day (Ha73) to
0.2.to 0.9 pCi/day (UN77) with an average of 0.4 pCi/day. The
comparison with drinking water of average concentration of 2
pCi/L and 2 liter per day consumption gives 4 pCi/day of water
derived uranium which is five to ten times greater than the food
derived uranium. Air contributions of uranium are much smaller
than the food and water contributions.
URANIUM CONCLUSIONS
1. The data shows that elevated levels of uranium found are
found in surface water as well as in ground water.
2. Highest average values of uranium concentration are found in
decreasing order in the following provincest Colorado Plateau,
Western Central Plateau, Rocky Mountain System, Basin and Range
and Pacific Mountain System. The highest state is South Dakota.
Modeling these variations would be very helpful for regional
standards.
-------
-24-
3. High uranium concentrations in the East are widely separated,
most values in the East are low.
4. Isotopic estimates are needed for 238y, 234y since they are
found in disequilibrium in water, regional variations should be
modeled.
5. More analyses are needed for low population systems.
OCCURRENCE OF RADON ISOTOPES IN PUBLIC DRINKING WATER
INTRODUCTION
There are two isotopes of radon, with half-lives long enough
to be considered as drinking water radionuclides. The first is
Rn-222 which is the daughter of Ra-226, called "radon", and has a
half-life of 3.84 days. The second Rn-220 which is the daughter
of Ra-224, was historically called "thoron", and has a half-life
of 56 sec. The time delay from production to consumption of
water of a few hours to a few days for water allows many decay
half-lives for the Rn-220, and it is not observed in water
supplies. Rn-222 henceforth simply radon, is transported by the
water, can lead to public exposures by being ingested and
exposing the digestive system and by becoming airborne and
exposing the lungs. When water is used for cleaning, dish-
washing, bathing, or clothes washing, radon escapes from the
water into building air where it decays into alpha emitting
-------
-25-
daughters. The resulting radon daughters are charged and will
frequently attach to aerosol particles in the air. The dust,
cigarette smoke, or aerosol particles will then be inhaled and
may become attached to the interior of the lung, bringing the
alpha particle emitting radon daughters into close association
with the cell lining of the respiratory system (Ar75).
There is no federal standard for radon in water, although
studies on ingestion doses (Hu65) (As79) and inhalation doses
(Pr81), (He83a), and (He79) have been done. There have been
standards for radon in mine air, and for radon from soil gas in
buildings placed on mine tailings in the United States and Canada
(Us79, At77). Sweden has standards for radon from soil gas in
areas of alum shale and granites (Ak81). Some of these standards
are given in Table 6. In the past, radon and its daughters have
been excluded from the drinking water standards, and considered
only to be an interference in the radium measurements.
Information about levels of radon have been obtained in
state studies and by a federal study done by the U.S.E.P.A.
Environmental Radiation Facility in Montgomery, Alabama and
by the University of Texas at Houston.
-------
-26-
RADON - RADIUM GEOCHEMISTRY
Radon is a water soluble inert gas and its occurrence is
controlled by physical variables such as pressure, temperature,
emissivity of radon from rocks, as well as by time, and by the
geochemistry of its parent Ra-226. High activity of radon is
associated with granitic rocks (St81), uranium minerals (Ta78),
such as uraninite, carnotite and with tailings from phosphate
fertilizer processing (Us79) and uranium mines.
TRANSPORT OF RADON IN WATER FROM ROCKS
As discussed in the geochemistry section, the occurrence of
radon in water is controlled by chemical concentration of radium
in the host soil on rock and by emissivity of radon into the
water. The physical condition of the rock matrix appears to play
a greater role in radon production than does the concentration of
parent radium. Several investigators (An72), (Ra83), (Ta64), and
Ta80) examine the mechanisms influencing the release of radon
from rock grains and the transport of radon through an aquifer.
Experimental and theoretical considerations indicate that dif-
fusion along microcrystalline imperfections dominates the release
of radon into the surrounding interstitial waters. The movement
of radon in water is governed by water transport rather than dif-
fusion in most cases, i.e., cases in whirh the percolation velo-
city is greater than 10-5 cm/sec.
-------
-27-
The variation of radon concentration with rock type for well
water in Maine is illustrated in Table 7. This variation of a
factor of 20 illustrates variation in rock types in Maine.
Public utility water measured at the wells, which is also pre-
sented in the Maine list, is a factor of 5 lower than the state
average. This is due to the higher capacity of the gravel
aquifers used for utility water supplies. These gravel aquifers
allow more water to mix with the same amount of radon, leading to
lower concentrations of radon.
OCCURRENCE OF RADON IN PUBLIC WATER SUPPLIES
Concentrations of radon in various water sources conform to
the log normal distribution. Table 8 shows the results of a
blind sampling of public water supplies in the central United
States. A reanalysis of other published data (Table 9) shows a
similar trend. Some sources appear to be samples from a single
log normal distribution, others from two, or perhaps three distri
butions, as indicated by the sharp breaks or bends in the plots.
For this reason geometrical averages are used for the samples
shown below.
The occurrence of radon in public ground water supplies in
the United States is shown in Figure 7 (taken at the tap) and
Table 9 (mainly utility samples). Radon activities are thousands
of times higher than uranium or radium, probably due to absorb-
-------
-28-
tion of the radium and uranium by the host rock (Ki82). The
results of these geometric means show highest values in mountain
states especially in the Appalachians with the highest states
Rhode Island, Maine, New Hampshire, Vermont, Massachusetts,
Pennsylvania, and Virginia. California is the highest western
state. The high radon values associate with the granitic areas
in the Appalachian Highlands Piedmont Provinces (Figure 1).
Midwestern and coastal plane values are lower and mountain states
in the west are higher. Figure 8 shows the private ground water
supplies (mainly individual samples) for the United States.
These results of geometrical averages show the private supplies
are higher by a factor of 3-20 times the public ground water
samples. This factor results from the use of low capacity wells
for private supplies while public supplies use high capacity sand
or gravel wells. The higher states in the private well list are
Rhode Island, Florida, Maine, South Dakota, Montana and Georgia.
Larger numbers of samples would be desirable to strengthen these
conclusions. Public surface supplies have radon concentrations
less than 100 pCi/1. Table 10 shows a breakdown of the radon
concentrations in water by state and by population of the town
using the well. Highest radon concentrations are found in the
less than 100 category (see Maine and US for examples). The
United States geometric population average is 187 pCi/1.
-------
-29-
MAJOR SOURCES OF INDOOR RADON
Radon produced from radium in the surficial soil and rock
(Ak81) is released into houses from water, soil gas, fuel gas and
construction materials and outdoor air. Both water and soil gas
can be transported into buildings through cracks, drain holes, as
well as water and fuel gas supply pipes (Sc82). The radon in the
ground water is released as it is mixed with air in such indoor
uses as cleaning, bathing, dish and clothes washing and toilet
flushing (He83a) (Pa79). Soil gas will mix into building air and
then diffuse throughout the house. Radon from fuel gas enters
building air from unvented heaters or stoves. Thus, the radon
concentration in air will depend on the sum of all radon sources
(Ge78), the volume of the building, and the ventilation rates of
the building (F180) (Ne81). The average value for radon in house
air due to all these sources has been estimated at 0.3 - 2.2
pCi/1 in normal regions, 1.1 - 1.67 in anomalous regions of the
U.S. (Br81).
The Soil
Radon diffuses from the soil through cracks in foundations,
unventilated crawl spaces, basement drains, and other pathways
into the living space. Direct outgassing from the soil is the
dominant source of indoor radon in most cases contributing .03 -
1.5 pCi/1 in normal regions, 0.3 - 15 pCI/1 in anomalous regions.
-------
-30-
If most radon enters structures through the basement or foun-
dation, radon concentrations would tend to decline markedly with
story above ground, as shown in Table 11. The limitations
of building materials and ground waters as sources of radon (see
below) coupled with the tendency toward single distributions in
given areas combine to suggest that the soil is generally the
largest source.
Ground Water
Ground waters containing radon can add substantially to the
amount of radon in the air of a dwelling. Much of the dissolved
radon can escape when water is used for various domestic purposes
inside a dwelling. The amount of radon in indoor air due to the
use of water depends a great deal on architectural and life-style
related variables. The most sensitive dwellings will be small,
relatively tight structures in which large amounts of water are
routinely used in household appliances. A model for the average
increment to the indoor atmosphere can be expressed as«
C ^
a " MV * ei wi
where Ca and Cw are the concentrations (pCi/1) of radon in the
air and water, respectively, R is the air change rate (hr-1), V
is the dwelling volume, Wi represents the average amount of water
(1) used daily in the ith domestic application, and ei denotes
the transfer efficiency, or the fraction of radon released to the
air for the ith application.
-------
-31-
A number of investigators have made semi-empirical deter-
minations of Ca/Cwj most are on the general order of 10"^ pCi/1
in the air per pCi/1 in the water. Table 12 shows a number of
such estimates and the underlying assumptions used in their deri-
vation. Using Table 12 the range of values for radon in air from
water in the U.S. ranges from 0.2 pCi/1 average for Rhode Island
to as low as 1.2 x 10-3 pCi/1 average for Tennessee.
The relative importance of water as a source of indoor radon
will of course depend upon the amount of radon in the water and
the magnitude of other sources. Upper limit calculations on one
recent data set (Table 12)(Pr83a) show water as the source of up
to 35% of the net (indoor minus outdoor) radon observed in a set
of 81 bedrooms in single family houses in the State of Maine.
The highest radon levels were seen in the basement, which
suggests that the soil is a major source of indoor radon. The
long-term average levels noted in the bathroom were also higher
than those noted in the other living areas, however. Both the
integrated indoor air radon concentrations and the concentrations
of radon in water were distributed log-normally, and a ratio of 4
pCi/1 (bedroom air) per 105 pCi/1 (water) was found by
regression. The application of this factor to the geometric mean
of the water distribution led to a predicted increment of 0.23
pCi/1 in the air versus a net bedroom concentration of 0.66
pCi/1. This is an upper limit estimate, based on the assumption
that the magnitudes of other sources of radon are independent of
-------
-32-
the amount of radon in the water. In fact, the concentration of
radon in the basements was found to be correlated with the con-
centrations of radon in water at the p = 0.01 level.
Construction Materials
A number of recent publications have addressed the problem
of elevated indoor radon concentrations prising from the use of
building materials containing radium-226 (Br81). A portion of
the radon arising from construction materials is able to diffuse
into the living area, where the ultimate concentration increment
will depend on the volume of the dwelling and the ventilation
rate ranging from .003 pCi/1 to 0.3 pCi/1 in the U.S. (Br81).
Perhaps the best publicized case involving construction materials
occurred in Grand Junction, Colorado, where uranium mill tailings
were once frequently used as fill materials around foundations.
Over 5000 buildings were associated with tailings material to
some extent, and 3000 of those buildings were built on top of a
layer of tailings. When radon levels in some of the dwellings
were found to be markedly elevated a general survey was con-
ducted, and eventually a remedial action program was implemented
for those structures exceeding national standards. (See Table
6). A similar situation arose in the phosphate mining area of
Florida. Dwellings were built on mining lands reclaimed at least
in part with phosphate rock residues. Radon concentrations well
over 10 pCi/1 have been noted in buildings containing materials
-------
-33-
such as gypsum wall board derived from phosphate residues or cin-
derblocks made of fly ash or blast furnace slag.(US79).
Fuel Gas
Radioactivity in natural gas was noted as early as 1904
(Sal8), but it was not until the 1970's that the potential
contribution to indoor radon concentrations was seriously
investigated. Interest in the health implications of naturally
occurring radon in gas deposits was stimulated by investigations
of the practicality of stimulating natural gas yields by deto-
nating a nuclear warhead in the appropriate rock formation
(Bu66). Estimates of dose increments caused by the combustion of
natural gas in the home were made by Barton (Ba73), Johnson
(0o73), and Gesell (Ge74). Similar calculations were made for
liquified petroleum gas (LPG), which, because of boiling point
considerations, contains a higher concentration of radon than the
natural gas from which it is made.
The increment to the indoor environment depends on the
amount of gas of LPG burned in unvented ranges or heaters, the
size and infiltration characteristics of the dwelling, and of
course, the concentration of radon in the fuel. The contribution
from this source is usually quite small (0.15 pCi/1) due to the
low use of unvented heaters (Ge77).
-------
-34-
OCCURRENCE OF RADON IN INDOOR AIR
There are a number of situations in which indoor radon
levels are especially elevated. These situations occur when the
structure contains a stronger than usual source of radon, or when
the structure has especially low ventilation and infiltration
rates, or both. Rising heating and air conditioning costs in the
last several years have encouraged people to reduce air infiltra-
tion rates.
A number of trends can be discerned in the recent litera-
ture. Within a given locale, indoor radon concentrations tend to
be distributed log-normally, and sample means vary markedly from
area to area (Ge83). It is becoming increasingly apparent that
local geological factors play a major, if not dominant role in
determining the distribution of indoor radon concentrations in a
given area.
Concentrations of Radon In The Indoor Environment
One of the most extensive studies of radon in dwellings on
record is a recent survey of 12,000 Swedish homes (Hi81). All
the dwellings involved were ones in which elevated radon levels
were expected. The results,as shown in Table 13 are reported in
working levels, and the presumed associated radon concentrations,
based on an equilibrium ratio of 0.5, are added in parentheses.
-------
-35-
The measurements summarized above are not meant to be repre-
sentative and are probably considerably higher than the true area
average. Nevertheless, the number of dwellings involved and the
high values observed combine to demonstrate the extent of the
radon problem in certain areas. Nearly half of the 12,000
dwellings were found to exceed current U.S. occupational standard
for uranium mines (adjusted for constant occupancy vs. a 40 hour
week).
Another extensive survey involved nearly 10,000 randomly
selected houses in 14 Canadian cities (Mc80). Single grab
samples were obtained from basements or the ground floor during
the summer months. In general, data from a given city were better
fit by the log-normal distribution than by the normal distribu-
tion. The geometric means ranged from 0.14 pCi/1 in Vancouver to
0.88 pCi/1 in St. Lawrence, Newfoundland. The geometric standard
deviations ranged from 2.78 to 6.77.
Table 11 summarizes a number of surveys conducted in the
United States in areas not known or suspected to involve anoma-
lies due to mill tailings or unusual mineralization. The data
presented are either the average of a number of grab samples
taken within a single dwelling, or were developed by long-term
measuring devices. The equilibrium fraction (f) of radon
daughters is given where available.
-------
-36-
RADON CONCLUSIONS
1. Radon concentrations in water are highest in ground
water especially in granite areas. Radon concentrations in sur-
face water are very low.
2. Higher concentration occurs in small systems. Domestic
supplies have higher concentrations of radon in water than public
wells. Utility systems are lower by a factor of 10 than private
wells.
3. The highest average radon concentrations in water are
found in the provinces in decreasing orderi New England,
Appalachian Highlands-Piedmont, Pacific Mountain System, Rocky
Mountain System.
4. Ventilation affects radon concentration in air with an
approximate value of 1. x 10-4 for the ratio of radon con-
centration in air to radon concentration in water for a house
with one air change per hour. Soil gas radon contributes a
sizeable portion of the total radon in air.
5. Additional measurements of radon in systems serving less
than a thousand users are required in order to better quantify
exposures to the group that potentially represents the highest
population dose. Because the number of such systems is quite
large (~37,000), these measurements should be obtained from a
representative sampling program guided by geological models.
-------
-37-
CALCULATION OF WATER UTILITY RISKS FROM
RADIONUCLIDES IN WATER BASED ON NATIONAL OCCURRENCE DATA
Each nuclide present in drinking water will present a risk
to the utility users of the water which is related linearly to
the occurrence, concentration, population exposed, and the indi-
vidual risk rate (Ma83). The case of radium in drinking water
data allows a direct calculation. The average concentration of
radium in drinking water from utilities is 1.6 pCi/1 (Ho83), and
the population consuming this water is 70 million people which is
the half of the U.S. population which uses ground water provided
by utilities. Since the individual risk rate is hypothesized to
be by linear dose response 24.5 x 10-6 excess cancers/lifetime
person pCi/1 (using Ma83), we can calculate
1.6 pCi/1 x 24.4 x 10-6 excess cancers 1 x 70 x 106 people
lifetime person pCi/1
~ 2688 people
lifetime in U.S.
Even the elimination of bone sarcomas at low concentration will
leave the sinus carcinomas which are half of this number (Ma83).
An additional 30 x 106 people are exposed at less than .5 pCi/1
surface water provided by utilities. The distribution of
occurrence of radium concentrations permits estimates of the
number of fatal cancers averted when the standard is placed at a
particular concentration. Since the standard is at 5 pCi/1, it
seems reasonable to estimate the fraction of cancers averted by
the standard. The average for the supplies of greater than 5
pCi/1 is 8 pCi/1. This concentration is multiplied by an esti-
mate of the population which uses those supplies obtained from
-------
-38-
rounding the average ground water utility population from the
U.S.E.P.A. Summary.
8 pCi/1 x 1000 persons x 500 x 24.0 x 10-5 excess cancers
supply 1 pCi/1 person lifetime
~ 96 excess cancers
lifetime
Population risk can also be done for uranium in water by
using a similar calculation. For uranium the population weighted
average radioactivity concentration is 0.8 pCi/1 for the whole U.S.
population (Dr81). This permits this result for the risk using
individual risk factors from (Ma83), (Wr83) and (Co83).
[0.8 pCi/l] riOO mrem/yrJ x 220 x 106 people exposed
10 pCi/1
x 34 x 10-6
100 mrem/yr
= 598 persons
lifetime in U.S.
Since there is no drinking water standard for uranium, at this
time, we must use the fraction of people at each occurrence level
to estimate the number of cases avoided by a standard for water
utilities.
Using values from Drury et. al. a water utility users risk
estimate table can be formed (Dr81) (See Table 15.)
-------
-39-
The population risk for radon in utility water supplies can
be calculated using radon concentration in utility water,
exchange rate from water into air, cancer rate per working level
month per million population and the exposed population in
millions. From the work done by the committee on the Health
Effects of Radon in drinking water (Cr83), we can obtain the
individual risk factor which includes the exchange rate for radon
from water into air and the cancer rate pr working level month
per million population exposed to air for a lifetime of 20 years.
This factor is 3 x 10-7 iUng cancers per pCi/1 radon in water.
This factor is increased to 4 x 10-7 cancers when stomach and
whole body cancers are included (Cr83). The radon in water con-
centration data for the whole United States are geologically
controlled and are generally a mixture of low values around
100-200 pCi/1 and high values of 10,000-1,000,000 pCi/1. This
extreme range of values leads to arithmetic averages which are
strongly influenced by the highest few points. The geometric
mean of these data will average the numbers with less weight for
the high values. This geometric mean will be lower than the
arithmetic mean. We have decided to calculate the risk with both
of these means.
Using data from the geometrical and arithmetic means of
radon concentrations for utilities of different sizes, and the
geometrical and arithmetic risk factor, we can calculate the
-------
-AO-
population at risk for the utilities serving less than 100 popu-
lation, 100-1000, population, 1000-5000 population, 5000-10,000
population, 10,000-100,000 population, and 100,000 and above
population. Lifetime risk is shown for both geometric averages
and arithmetic averages in the right column of Table 16.
OTHER NATURAL RADIOISOTOPES OF POSSIBLE CONCERN
The natural radioisotopes discussed above (U, Ra-228,
Ra-226, and Rn-222) are of greatest concern because of their long
half-lives and the health risks associated with the activities
that can be present in public drinking water. There are two
classes of other natural radioisotopes that may be of possible
concerm (1) relatively long-lived isotopes whose activity is
derived from the aquifer, termed unsupported (Th-232, Th-230,
Pb-210, Po-210) and (2) very short-lived isotopes which "grow in"
once the ground water is pumped from the aquifer and thus are
supported (primarily Rn-222 daughters). However, very few data
are available on the activities of these isotopes in drinking
water, primarily due to their low solubility and/or the dif-
ficulty of measuring isotopes of short half-lives.
-------
-41-
Table 17 shows the typical ranges of activities of the
longer-lived radionuclides in ground water. Activities in sur-
face waters will be extremely low (except perhaps in hot springs)
due to rapid sorption onto suspended and bottom sediments. The
highest known activity of Th-232 and Th-230 in U.S. drinking
water is from a well in California that also contained large
amounts of dissolved organic matter which probably complexed with
thorium. Most other values are below 0.1 pCi/1. Th-230 would be
expected to be slightly higher than Th-232 due to generation by
U-234 in solution and by alpha recoil. Likewise, detectable
activities of Pb-210 and Po-210 would be expected because of the
relatively large amount of Rn-222 present in many ground waters.
The insolubility of these isotopes and their short-lived precur-
sors in the aquifer is demonstrated by the fact that more than
99.9 percent of the activity generated by Rn-222 decay in ground
water is removed within the two hours necessary for equilibrium
to be established between Rn-222 and Pb-210. The only known ano-
nymously high Po-210 value in drinking water is the surprisingly
large activity in Louisiana for which the source has not been
determined. In general, these longer-lived isotopes are not
expected to occur in activities greater than 1.0 pCi/1.
The second class of radioisotopes of possible concern are
Rn-222 daughters which reach equilibrium with Rn-222 within two
hours. In untreated ground water systems, removal by adsorption
-------
-42-
would not be as rapid as in the aquifer due to the small surface
area of the distribution system. Exposure from consumption of
these supported, extremely short-lived daughters should be evalu-
ated. Rn-220 and its daughters do not pose a similar problem
primarily because its 54.5-second half-life is too short to allow
diffusion out of the aquifer materials and the initial activities
are much lower.
-------
-43-
REFERENCES
Akerblom, G. V., and Wilson, V., 1981, "Radon Gas - A
Radiation Hazard from Radioactive Bedrock and Building
Materials'*, Bull. Int. Assoc. Engin. Geology 23, 51.
Andrews, J. N. and Wood, D. F., 1972, "Mechanism of Radon
Release in Rock Matrices and Entry into Groundwaters",
Institution of Mining and Metallurgy Bulletin
81.B1980B209.
Archer V. E., Radtard, E. P., and Axelson, 0., 1979, "Radon
Daughter Cancer in Man« Factors in Exposure Response
Reslationships" Proc. Workshop on Lung Cancer
Epidemiology and Industrial Applications to Sputom
Cytology, Col. School of Mines Press, Golden, CO,
p 324.
Asikainen, M., and Kahlos, H., 1979, "Anomalously High
Concentrations of Uranium, Radium and Radon in Water
from Drilled Wells in the Helsinki Region", Geochimica
et Cosmochimica Acta, 43, 1681.
Asikainen, M., 1981, "State of Disequilibrium Between U-238,
U-234, Ra-226, and Rn-222 in Ground Water from
Bedrock", Geochim. Cosmochim. Acta, 45, pp. 201-201.
Atomic Energy Control Board, 1977, "Criteria for
Radioactive Clean-up in Canada", Information Bulletin
77-2, Ottawa, Ontario.
81 Beddinger, M. S. and Sargent, K. A., 1981, "Identification
of Suitable Geohydrologic Environments for the
Disposal of High-Level Radioactive waste."
Proceeding of a Workshop on Siting of Radioactive
Waste Repositories in Geological Formations,
pp. 151-187 Nuclear Energy Agency, Paris.
Breslin, A. J., and George, A. C., 1979, "Radon Sources,
Distribution, and Exposures in Residential Buildings",
Trans Am. Nucl. Soc., 33, 145.
Bruno, R. C., 1981, "Sources of Inioor Radon in Houses",
presented at International Symposium on Indoor Air
Pollution, Health, and Energy Conservation, Oct. 13-16,
1981, Amherst, Massachusetts} U.S. Environmental
Protection Agency, Washington, D.C.
Bruno, R. C., 1983, "Sources of Indoor Radon in Housesi A
Review", Journal of the Air Pollution Control
Association 33, 102-109.
-------
-44-
REFERENCES CONTINUED
Bu66 Bunce, L. A. and Sattler, F. W., 1966, Radiological Health
Data and Health Data and Reports 2 441.
Ca80 Section on Uranium from Canadian Radiation Protection
Bureau Guidelines for Canadian Drinking Water
Quality 1978, 1980 Supports Documentation Health
and Welfare, Ottawa, Canada.
Ch71 Cherdyntsev, V. V., 1971, Uranium-234i 234pp. Keter Press,
Jerusalem.
C183 Cline, W., Adamovitz, S., Blackman, C., and Kahn, B., 1983,
"Radium and Uranium Concentrations in Georgia
Community Water Supplies", Health Physics, 43(7).
Co83a Cothern, C., and Lappenbusch, W. L., 1983, "Occurrence of
Uranium in Drinking Water in the United States",
Health Physics (in press).
Co83b Cothern, C. R., Lappenbusch, W. L., and Cotruvo, J. A.,
1983, "Health Effects Guidance for Uranium in Drinking
Water", Health Physics. 44, 377-384.
Co83c Cothern, C. R. and Lappenbusch, w. L., 1983, "Radium-226
and Gross Alpha Particle Activity Compliance Data for
Drinking Water Supplies in the United States",
Internal Report Office of Drinking Water WH550,
Washington, D.C. 20460.
Cr83 Cross, F. T., Harley, N. H., Hofmann, W., Nelson, N., 1983
"Health Effects of Radon in Drinking Water", Meeting of
the National Workshop for Radioactivity in Drinking
Water Committee, Easton, Maryland.
Dr81 Drury, J. S., Reynolds, S., Owen, P. T., Ross, R. H., and
Ensminger, J. T., 1981, "Uranium in U.S. Surface
Ground and Domestic Waters", EPA-370/9-81-001.
Report by Health and Environmental Studies Center
Complex, Oak Ridge National Laboratory, Oak Ridge,
Tennessee 37830.
Dy76 Dyck, w., Chatergee, A. K., Gammell, D. E., and Murricane,
K., 1976, "Well Water Trace Element Reconnaissance,
Eastern Maritimes, Canada", J. Geochem. Explor. 6, 139.
Ei80 Eichholz, G. G., Matheny M. D. and Kahn, B., 1980, "Control
of Radon Emanation from Building Materials by Surface
Coating", Health Phys. 39, 301.
-------
-45-
REFERENCES CONTINUED
E183 Elsinaer, R. J., King, P. T., and Moore, W. S., 1983,
Ra-224 in Natural Waters Measured by Gamma-Ray
Spectrometry", Anal. Chim. Acta, in Press.
Ep76a Environmental Protection Agency, 1976a, National Interim
Primary Drinking Water Regulations t EPA-570/9-76-093,
159 pp., Office of Water Supply, Washington, D.C. 20460.
Ep76b Environmental Protection Agency, 1976b, Determination of
Radium Removal Efficiencies in Illinois Water Supply
Treatment Processes, Office aF Radiation Programs,
Washington, D.C. 20460.
Ev81 Evans, R. D., Harley, J. H., Jacobi, W., McLean, A. S.,
Mills, W. A. and Stewart, C. G., 1981, "Estimate of
Risk from Environmental Exposure to Radon-222 and its
Decay Products*", Nature, 290, 98.
Fe79 Femleck, 3. K., and Cadigan, R., 1979, "Radium and
Uranium Concentrations and Associated
Hydrogeochemistry in Ground Water in Southwestern
Pueblo County, Colorado**, pp. 34, U.S. Geological Survey
Report No. 79-974.
Fi75 Filliben, J. J., 1975, MThe Probability Plot Correlation
Coefficient Test for Normality**, Technometrics 17, 111.
Fi79 Findlay, W.O., 1979, "Application of Radon Standards to
New and Existing Housing in Elliot Lake, Ontario", in
"Radon in Buildings", Colle, ed., National Bureau of
Standards Special Publication 581, U.S. Government
Printing Office, U.S. Department of Commerce,
Washington, D.C., 1980} Dilworth, Secord, Meagher and
Associates Limited, Elliot Lake, Ontario.
F180 Fleisher, R. L., Mogro-Campero, A., and Turner, L. G.,
1980, "Radon Levels in Homes in the Northeastern
United States* Energy-Efficient Homes", General
Electric Corporate Research and Development Technical
Information Series, Report No. 80CRD288.
General Electric Research and Development Center,
Schenectady, NY 12301.
Ga77 Gabelman, J. w., 1977, "Migration of Uranium and Thorium -
Exploration Significance", Amer. Assoc. Petrol.
Geol., Studies on Geology No. 3, 168 pp., Tulsa, Okla.
-------
-46-
REFERENCES CONTINUED
Ge74 Gesell, T. F., 1974, Estimation of the dose equivalent to
the U.S. Population from radon in liquified petroleum
gas. Proceedings of the 8th midyear topical symposium
of Health Physics Society USAEC Report *cont. 741018
National Technical Information Service, U.S. Department
of Commerce, Springfield, VA 22151, p. 347-354.
Ge75 Gesell, T. F. and Prichard, H. M., 1975. "The
Technologically Enhanced Natural Radiation
Environment". Health Physics 28, pp. 361-366.
Ge78 Gesell, T. F. and Prichard, H. M., 1978. "The Contribution
of Radon in Tap Water to Indoor Radon Concentration" in
Natural Radiation Environment III, pp. 1347-1363. U. S.
Department of Energy Special Symposium series, 51,
CONF 780422.
Ge80 Gesell, T. F. and Prichard, H. M., 1980. "The Contribution
of Radon in Tap Water to Indoor Radon Concentrations",
in Natural Radiation Environment III, Vol. 2, pp. 1347-1363,
Technical Information Center/U.S. Dept. of Energy.
National Technical Information Service, U.S. Department
of Commerce, Springfield, VA 22151.
Ge83 Gesell, T. F., 1983. "Background Atmospheric 222rp
Concentration Outdoors and Indoors« A Review", Health
Physics, 45, pp. 289-302.
Gi82 Gilkeson, R. H. and Coward, J. B., 1982, "A Preliminary
Report on 238u Series Disequilibrium in Ground Water
of the Cambrian-Ordovician Aquifer System of
Northeastern Illinois", .in E. C. Perry and C. W.
Montgomery (eds.), pp. 109-118, Isotope Studies of
Hydrologic Processes, Northern Illinois Univ. Press,
DeKalb, 111.
He79 Hess, C. T., Norton, S. A., Brutsaert, W. F., Casparius, R.
E., Coombs, E. G. and Hess, A. L., 1979, "Radon-222 in
Potable Water Supplies in Maine: The Geology,
Hydrology, Physics and Health Effects", Land and Water
Resources Center, University of Maine at Orono.
He83a Hess, C. T., Weiffenbach, C. V., and Norton, S.A., 1983,
Environmental Radon and Cancer Correlations in Maine",
accepted for publication in Health Phys., 45, 339-348.
-------
-47-
REFERENCES CONTINUED
He83b Hess, C. T-, Weiffenbach, C. V., and Norton, S. A., 1983,
"Variations of Airborne and Waterborne Radon-222 in
Houses in Maine, U.S.A.11, accepted for publication in
Environment International presented to International
Symposium on Indoor Air Polution, Health and Energy
Conservation, Oct. 13-16, 1981, Amherst, Massachusetts.
Hi81 Hildingson, 0., 1981, "Measurements of Radon Daughters in
5,600 Swedish Homes"» presented at International
Symposium on Indoor Air Polution, Health, and Energy
Conservation, Oct. 13-16, 1981, Vol. 8, 1551-6, 59-66
Amherst, Massachusetts} Swedish National Testing
Institute, Boras, Sweden.
Ho83 Horton, T. R. (1983) "Methods and Results of E.P.A.'s Study
of Radon in Drinking Water", 15th Annual National
Conference on Radiation Control, Program Directors,
Reno, Navada, May 16, 1983. Also U.S.E.P.A. Office of
Radiation Programs EPA 520/5-83-027.
Hu65 Hursh, J. B., Morken, D. A., Davis, T. P., and Lovaas, A.,
1965, "The Fate of Radon Ingested by Man", Health
Physics, 11, 465.
Hu73 Hursh, J. B., and Spoor, N. S., 1973, "Uranium, Plutonium,
Transplutonium Elements", Edited by H.C. Hodge, J.S.
Stannard, J.V. Hursh, Springer-Verlag, New York.
ICRP79 International Commission on Radiological Protection
1979 "Limits for Intakes of Radionuclides by Workers",
Annals of the ICRP ICRP Publication 30 Vol. 2 No. 3/4
(Oxford: Pergamon Press).
Jo73 Johnson, R.H., Bernhard D., Nelson, V. S., Calley, H.W.
Jr., Assessment of Radiological Health Effects from
Radon in Natural Gas, 1973, Proceedings of Nobel Gas
Symposium, Las Vegas, Nev., Sept.
Ka73 Kahlos, H., and Asikainen, M., 1973, "Natural
Radioactivity of Ground Water in the Helsinki Area",
Report SFL - A19, Institute of Radiation (Physics,
(Sateilyfysiikan Laitos-) Helsinki, Finland.
Ka80 Kahlos, H., and Asikainen, M., 1980, "Internal Radiation
Doses from' Radioactivity of Drinking Water in Finland",
Health Physics 39. 108-111.
Ka76 Kaufmann, R. F. and Eadie, G. G. and Rusell, C. R., 1976,
"Effects of Uranium Mining and Milling on Ground Water
in the Grants Mineral Belt, New Mexico, Ground Water,
14 (5).
-------
-48-
REFERENCES CONTINUED
Ka77 Kaufmann, R. F. and Bliss, J. D., 1977, "Effects of
phosphate mineralization and the phosphate industry on
Ra-226 in ground water of central Florida", U.S. EPA,
Office of Rad. Programs, Rept. EPA/520-6-77-010, 115
pp. Las Vegas, Nev.
Ki82 King, P. T., Michel, J., and Moore, W. S., 1982,
"Groundwater Geochemistry of Ra-228, Ra-226, and
Rn-222", Geochim. Cosmochim. Acta, 46, pp. 1173-1182.
Ko79 Kobal, I., Kristan, J., Ancik, M., Jerancic, S. and
Skofljanec, M., 1979, "Radioactivity of Thermal and
Mineral Springs in Slovenia", Health Physics, 37,
239-242.
Krie82 Krieg, C. R. and Hahne, R. M. A., 1982, "Ra-226 and Ra-228
in Iowa Drinking Water", Health Physics, 43(4), pp.
543-559.
Kris82 Krishnaswami, S., Graunstein, W. C., Turekian, K. K., and
Dowd, J. F., 1982, "Radium, Thorium, and Radioactive
Lead Isotopes in Groundwater: Application to the in
situ Determination of Adsorption-Desorption Rate
Constants and Retardation Factors", J. Water. Res.
Research, in press.
Ku54 Kuroda, P. K., Damon, P. E., and Hyde, H. I., 1954,
"Radioactivity of the Spring Waters of Hot Springs
National Park and Vicinity in Arkansas", Am. J. Science
252. 76-86.
La83 Lappenbusch, W. L. (1983) Meeting of the National Workshop
of Radioactivity in Drinking Water Committee, Easton,
Maryland.
Le79 Lee, R. 0., Watson, J. E. Jr., and Fong, S. W., 1979, "An
Assessment of Radium in Selected North Carolina
Drinking Water Supplies", Health Physics, 37,
pp.777-779.
Li02 Lively, R. S. and Morey, G. B., 1902, "Hydrogeochemical
Distribution of Uranium and Radon in East-Central
Minnesota", .in E. C. Perry and C. W. Montgomery
(eds.), Isotope Studies of Hydrologic Process, pp.
91-108, Northern Illinois Univ. Press. DeKalb, 111.
Lu82 Lucas, H. F., Jr., 1982, "Ra-226 and Ra-228 in Drinking
Water", (abs. ) 27th Annual Mtg., Health Physics Soc.,
27 June - 1 July, p. .2, Las Vegas.
-------
-49-
REFERENCES CONTINUED
Ma83 Mays, C. W., Rowland, R. E. and Stehney, A. F. (1983)
"Cancer Risk from Lifetime Intake of Radium Isotopes",
Meeting of the National Workshop for Radioactivity in
Drinking Water Committee, Easton, Maryland.
Mc80 McGregor, R. G., Vasudev, P., Letourneau, E.G., McCullough,
R. S., Prantl, F. A., and Taniguchi, H., 1980.
"Background Concentration of Radon and Radon Daughters
in Canadian Homes", Health Physics, 39, 285-289
Mi80 Michel, J. and Moore, W. S., 1980, "Ra-228 and Ra-226
Content of Groundwater in Fall Line Aquifers", Health
Physicst 38, pp. 663-671.
Mi82 Michel, J. and Pollman, C., 1982, "A Model for the
Occurrence of Ra-228 in Ground Water", EPA, Office of
Drinking Water, Environmental Science and Engineering
NO. 81-227-270.
Mi58 Miholic, S., 1958, "Radioactive Waters from Sediments",
Geochemica et Cosmochemica Acta, 14, pp. 223-233.
Mo69 Moore, W. S., 1969, "Oceanic Concentrations of Radium-228",
Earth Planet. Sci. Lett., j6, pp. 437-446.
Mo81 Moschandreas, D. J., and Rector, H.E., 1981, "Indoor Radon
Concentrations", presented at International Symposium
on Indoor Air Pollution, Health, and Energy
Conservation, Oct. 13-16, 1981, Amherst, Massachusetts,
IIT Research Institute, Chicago, Illinois.
Mu82 Mullin, A., 1982, "Abnormally High Alpha Activity in a
Louisiana Drinking Water Supply", (abs.) 27th Annual
Meeting, Health Physics Society, 27 June - 1 July, p.
3, Las Vegas.
Ne81 Nero, A. V., and Nazaroff, W. W., 1981, "Distribution of
Indoor Radon Concentrations and Source Magnitudesi
Measurements and Policy Implications", presented at
International Symposium on Indoor Air Pollution,
Health, and Conservation, Oct. 13-16, 1981, Amherst,
Massachusetts; Lawrence Berkeley Laboratory, University
of California, Berkeley, California.
Na80 National Academy of Sciences, National Research Council,
1980, "The Effects on Populations of Exposure to Low
Levels of Ionizing Radiation", Report of the Advisory
Committee on Biological Effects of Ionizing Radiation
CBEIRIII), Washington, D.C.
-------
-50-
REFERENCES CONTINUED
0164 Olson, J. C. and Overstreet, W. C., 1964, "Geologic
Distribution and Resources of Thorium", U. S. Geol.
Survey Bull. 1204, 61 pp.
Pa79 Partridge, J.E., Horton, T.R., and Sensintattar, E. L.,
1979. "A Study of Radon-222 Released from Water During
Typical Household Activities" EPA Tech. Note.
ORP/EERF-79-1.
Pr81 Prichard, H. M., and Gesell, T. F., 1981, "An Estimate of
Population Exposures due to Radon in Public Water
Supplies in the Area of Houston, Texas", Health
Physics, 41, 599.
Pr83a Prichard, H. M., Hess, C. T. Nyberg, P., Weiffenbach,
C. V. and Gesell, T. F., 1983. "Associations Between
Grab Sample and Integrated Radon Measurements in
Dwellings in Maine and Texas", Environment
International, Vol. 8, 155 1-6, 83-87.
Pr83 Prichard, H. M. "Radon-222 in Municipal Water Supplies in
the Central United States", accepted for publication in
Health Physics.
Ra83 Rama, Moore, W.S. 1983 "Mechanisms of Transport of U-Th
Series Radioisotopes from Solids into Groundwater",
Geochemica Cosmo Chemica Acta. (in press).
Ro77 Rowland, R. E., Lucas, H. F., Jr., and Stehney, A. F.,
1977, "High Radium Levels in the Water Supplies of
Illinois and Iowa", ki Intl. Symp. on Areas of High
Natural Radioactivity, Pocos de Caldos, Brazil, 16-20
June 1975, Edited by Academia Brazileira de Ciencias,
pp. 65-73. Rio de Janeiro, RJ.
Sal8 Satterly, J. and McLennan, J.C., 1?18, Transactions of
Royal Society of Canada 12 Series 3, 153.
Sa 81 Sachs, H. M., Hernandez, T. L., and Rign, J. W., 1981,
"Regional Geology and Radon Variability in Buildings",
presented at International Symposium on Indoor Air
Pollution, Health, and Energy Conservation, October
13-16, 1981, Amherst, Massachusetts) Center for Energy
and Environmental Studies, Princeton University.
Sa78 Sasser, M. K., and Watson, J. E. Jr., 1978, "An Evaluation of
the Radon Concentration in North Carolina Ground Water
Supplies", Health Phys., 34, 667.
-------
-51-
REFERENCES CONTINUED
Sc82 Scott, A.G., 1982, "Remedial Actions, Active and Passive",
presented to Health Physics Society Summer School in
"Radon, A National Health Problem?", June 21-25,
Las Vegas; Dilworth, Secord, Meagher and Associates
Limited, Toronto, Ontario.
Sc62 Scott, R. C., and Barker, F. B., 1962, "Data on Uranium
and Radium in Ground Water in the United States,
1954-1957". Geological Survey paper, 426, pp. 115.
Sh80 Shannon, S., of Los Alamos Laboratories, and Livingston,
D., of the Department of Energy Grand Junction Co.,
1980, private communication.
Sn70 Snihs, J. 0., 1970, "The Content of Some Natural
Radioactive Elements, Especially Radon-222, In some
Potable Waters in Sweden" (SSI»1970-029), Natl. Inst,
of Radiation Protection, Fack, S-104 01, Stockholm 60,
Sweden.
St79 Strain, C. D. and Watson, J. E., 1979, "An Evaluation of
Ra-226 and Rn-222 Concentrations in Ground and Surface
Waters near a Phosphate Mining and Manufacturing
Facility", Health Physics 37, 779-783.
St81 Stranden, E., 1981, "Radon in Houses Using Stone
Magazines for Heat Accumulation", Health Physics,
41, 29.
Ta80 Taniguchi, H., and Vasudev, P., 1980, "Radon and Radon
Daughters Due to Natural Uranium Occurrences in a
Rural Ontario Community", in Natural Radiation
Environment III, Gesell, ed., p. 1623, Technical
Information Center, U.S. Department of Energy.
Ta64 Tanner, A. B., 1964, "Radon Migration in the Groundi A
Review", in» The Natural Radiation Environment (Edited by
J.A.S., Adams and W. M. Lowder, pp 161-190. cFicago,
University of Chicago Press.
Ta80 Tanner, A. B., 1980, "Radon Migration in the Groundt A
Supplementary Review", in« Natural Radiation Environment
III. Edited by T. F. Gesell and W. M. Lowder pp5-56.
U.S. Department of Energy Special Symposium Series 51,
CONF 780422.
-------
-52-
REFERENCES CONTINUED
Un77 United Nations Scientific Committee on the Effects of
Atomic Radiation, 1977, Report to the General
Assembly, Sources and Effects of Ionizing Radiation
(United Nations, New York).
Us79 U.S. Environmental Protection Agency, 1979, "Indoor
Radiation Exposure Due to Radium-226 in Florida
Phosphate Lands? Radiation Protection
Recommendations and Request for Comment1*, Federal
Register, 44, 38664.
Us80 U.S. Environmental Protection Agency, 1980, "Proposed
Cleanup Standards for Inactive Uranium Processing
Sites; Invitation for Comment", Federal Register 45,
p. 27370.
Wr83 Wrenn, M., Durbin, P.W., Nelson, C., Rundo, J., Still,
E.T., Willis, D., "Metabolism and Dosimetry of High
Net Radionuclides" Meeting of the National Workshop
for Radioactivity in Drinking Water, Easton,
Maryland.
-------
TABLE 1—Summary of potential host rocks, geologic framework, and nature of ground water
flow systems in the Provinces of the coterminous United States (Fig. 1)
PROVINCE
1.
New England
Adirondack
Mountains
2.
Appalachian
Highlands-
Piedmont
3. Appalachian
and Interior
Plateaus
GEOLOGICAL FRAMEWORK
New England—complexly faulted
metamorphic and metasedimentary
rocks intruded by large masses
of granite. Adirondacks—
mountains composed of marble
and schist intruded by granites,
anorthosite, and gabbro.
Appalachian Highlands—mountain
belt of granites and metamorphics
thrust westward over Paleozic
rocks. Piedmont—non-mountianous
belt of highly complex metamorphic
rocks with abudant granitus.
Appalachian and Interior Plateaus
consist of gently dipping, gently
folded, sandstones, shales,
carbonates, and evaporites.
In southern Missouri exposing old
crystalline rocks.
NATURE OF GROUND WATER
FLOW SYSTEMS
Flow and head in bedrock of
low permeability and overlying
glacial aquifers greatly
influenced by local
topography and surface-
water features.
Flow and head in metamorphic and
granite bedrock of low
permeability, largely controlled
by topography and surface water
featuresj folded limestone
locally cavernous and highly per-
meable at shallow depths supporting
large springsi sandstone aquifers
of moderate extent and
permeability| flow systems
generally related to local recharge
in interstream areas and discharge
to surface water features.
Regional flow in low to
moderately permeable sandstones
and carbonatesj carbonates locally
of high permeability at shallow
depth due to fractures and solution
channels support large springs.
-------
TABLE 1
flow
— Summary of potential host
systems in the Provinces of
rocks, geologic
the coterminous
framework, and nature of ground water
United States (Fig. 1)--Continuous
PROVINCE
GEOLOGICAL FRAMEWORK
NATURE OF GROUND WATER
FLOW SYSTEMS
4. Coastal Plain
5. Glaciated
Central
Platform
Seward dipping thickening wedge of
sand, sandstones and shales with some
evaporites and limestones) underlain
by a basement of metamorphic rocks.
Igneous and metamorphic rocks on the
northwest overlain by sandstones,
carbonates, shales, and evaporites;
deep basin deposits in Michigan and
Illinois.
Regional flow in sand and limestone
aquifers with intervening clay con-
fining layers? predominant flow
direction seaward; discharge upward
through confining layers and to
streams.
Regional flow in sandstone and
carbonate aquifers; highly
mineralized water at depth in
basins; glacial aquifers locally
overlie bedrock.
6.
Western
Central
Platform
Horizontal to gently dipping
sandstones; deep sedimentary
basins and structural high.
Capped with sands and gravels
Regional flow in layered sandstone
and carbonate aquifers; thick con-
fining beds of shale; deep basins
contain highly saline water.
Extensive fluvial deposit
aquifers Nebraska south into
Texas and glacial aquifers in
North Dakota and South Dakota
overlie sandstones.
-------
TABLE 1—Summary of potential host rocks, geologic framework, and nature of ground water
flow systems in the Provinces of the coterminous United States (Fig. 1)—Continuous
PROVINCE
7. Rocky Mountain
System
GEOLOGICAL FRAMEWORK
Igneous and metamorphic folded core
rocks of Rocky Mountains and
intermontane basins of shales,
carbonates, evaporites, and
sandstones. Intrusive and
volcanic rocks.
NATURE OF GROUND WATER
FLOW SYSTEMS
Regional flow in layered sandstone
and carbonate aquifers with shale
confining beds in intermontane
basins; local recharge and
discharge controlled by topography
and surface water features in frac-
tured igneous and metamorphic
rocks.
8. Colorado
Plateaus
Flat-lying to gently warped layers
of sandstones, shales, limestones
and evaporites with volcanic rocks
Regional flow in layered sedimen-
tary rocks) chief aquifers are
sandstones and carbonates)
discharge to major streams; highly
saline water at depth in deep
basins and in association with salt
beds.
9. Basin and Range
Elongate blocky mountains of faulted
rock complexes) deep alluvium-
filled intermontane basins; intrusive
igneous stocks and plugs; extrusive
ash-flow tuffs, rhyolites, and
basalts.
Flow within closed basins; inter-
basin flow between closed
topographic basins through per-
meable bedrock; interbasin flow in
alluvial channels between basins
with integrated surface drainage;
deep regional flow systems in
carbonate and volcanic rocks.
-------
TABLE 1—Summary of potential host rocks, geologic framework, and nature of ground water
flow systems in the Provinces of the coterminous United States (Fig. 1)--Continuous
PROVINCE
GEOLOGICAL FRAMEWORK
NATURE OF GROUND WATER
FLOW SYSTEMS
10. Columbia
Plateaus
Regional shallow structural basin of
basaltic lava flows} locally faulted
and foldedf mountain range on the
west consisting of elongate chain of
andesitic volcanic cones.
Basaltic lava flows range from
highly permeable to dense nearly
impermeable creating regional
aquifers with perched aquifers
separated by confining beds. The
ground water principally discharges
to the major streams; local
discharge to a few closed basins.
11. Pacific
Mountain
System
Consists of several complex elements*
Large uplifted and tilted blocks
of granite with inliers of
metasedimentsi folded and faulted
folded and faulted sedimentary
rocksj deep elongate troughs filled
with fluvial sediments.
Regional flow in deep intermontane
sedimentary basins; igneous and
metamorphic rocks of low
permeability support shallow local
flow systems related to topography
and surface drainage.
-------
TABLE 2. Summary of Ra-228 and Ra-226 Distribution in Ground Water by
Aquifer Type for the Atlantic Coastal Plain and Piedmont
Provinces.
Aquifer Type
Number
of Values
Ra-
228
Ra-
226
Geometric
Mean
(pCi/1)
Range
(pCi/1)
Geometric
Mean Range
(pCi/1) (pCi/1)
Igneous Rocks
(acidic)
42
1.39
0.0-22.6
1.80
0.0-15.9
Metamorphic
Rocks
75
0.33
0.0- 3.9
0.37
0.0- 7.4
Sand
Arkose
Quartzose
143
92
50
1.05
2.16
0.27
0.0-17.6
0.0-13.5
0.0-17.6
1.36
2.19
0.55
0.0-25.9
0.0-23.0
0.0-25.9
Limestone
16
0.06
0.0- 0.2
0.12
0.0- 0.3
-------
TABLE 3. Ranking of all Structural Levels used Showing Groupings
of Similarity in Ra-228 Distribution. Ra-226 Values
are given for comparison but are not ranked.
Mean
Grouping* Ra-228 No
CpCi/1)
Mean
Structural Level Ra-226
(pCi/1)
A 3.03
A
A 2.49
A
A
B A 2.14
B A
B A
B AC 1.59
B AC
B D A C 0.85
B D C
B D E C 0.52
DEC
DEC
F D E C 0.39
F D E
F D E 0.34
F D E
F D E 0.28
F D E
F D E
F D E 0.28
F D E
F D E
F D E G 0.24
F D E G
F EG
F EG 0.12
F EG
F G
F G 0.09
F G
G
G 0.06
2 Igneous, acidic, composite Th, 0.75
syenite
46 Sand, unconsolidated, arkosic, 2.03
high-Th source
43 Sand, unconsolidated, arkosic, 2.73
medium-Th source
35 Igneous, acidic, composite Th, 2.31
granite
3 Igneous, acidic, composite Th, 0.99
diorite
31 Sand, unconsolidated, quartzose, 1.72
medium-Th source
37 Metamorphic, high-grade, 0.32
nonspecific Th
10 Metamorphic, low-grade, 1.41
nonspecific Th
3 Sand, unconsolidated, arkosic, 0.32
low-Th source
13 Metamorphic, medium-grade, 0.24
nonspecific Th
15 Metamorphic, h.'gh-grade, 0.24
specific Th, monazite
2 Igneous, acidic, refractory Th, 0.14
monazite
19 Sand, unconsolidated, quartzose, 0.09
low-Th source
16 Chemical precipitates, limestone 0.12
*A>B>C>D>E (p<0.05
-------
TABLE 4
Uranium Concentration in the Provinces
Beddinger Provinces
1. New England—
Adirondack Mts.
2. Appalachian
Highlands Piedmont
3. Appalachian
Interior Plateaus
4. Coastal Plain
5. Glaciated Central
Platform Aquifers
6. Western Central
Platform
7. Rocky Mountain
System
8. Colorado Plateaus
9. Basin and Range
10. Columbia Plateaus
Beddinger's
Uranium concentrations
(pCi/1)
Arithmetic Standard
Mean Deviation
0.46
0.020
0.137
0.108
1.04
2.1
1.99
2.31
2.15
0.52
11. Pacific Mountain System 1.41
0.03
0.013
0.27
0.19
2.17
2.2
1.27
1.4
1.53
0.389
1.81
Scott's Uranium
concentrations
(pCi/1)
Arithmetic
Mean
II .34
III .34
I .14
V .34
VI .71
VI 1.5
VII-VIII 1.15
IX .54
VIII 1.15
X 0.14
-------
TABLE 5
Population Versus Concentration of Uranium Distribution for Drinking Water Sources That Serve More
Than 10,000 People. The numbers in parentheses are the number of cities sampled followed by the
total in the category.
Uranium Concentration (pCi/1)*
POPULATION
0.05
-0.5
0.51
-1.0
1.1
-2
2.1
-3
3.1
-4
4.1
-5
5.1
-6
6.1
-7
7.1
-8
8.1
-9
9.1
-10
10. /O
-11
greater than
11 pCi/1
10,001 -
50,000
(532/2407)
292
75
46
14
17
4
75
4
2
3
50,001 -
75000
(63/224)
36
4
4
3
3
10
1
1
(12 pCi/1)
75,001 -
100,000
(31/101)
15
5
3
1
4
3
greater than
100,000
(174/276)
117
15
10
7
8
3
7
4
2
(30.2 pCi/1)
+assumed secular equilibrium of U-234 and U-238.
-------
TABLE 6
Standards for Radon in Air
United States (Us79) In phosphate mining regions in Florida
4 pCi/1 take remedial action
2 pCi/1 reduce to as low as is
reasonably achievable
Canada (At77) In Uranium mining regions
30 pCi/1 take prompt remedial action
4 pCi/1 take remedial action
2 pCi/1 investigate
Sweden (Ak81)
11 pCi/1 for existing buildings
4 pCi/1 for houses undergoing remodeling
2 pCi/1 for newly constructed houses
-------
TABLE 7
Average Radon Values In Private And Public Water Supplies
Arithmetic Number
Mean pCi/1 Maximum pCi/1 of Samples
Maine (He79,83b) 10,000 1,000,000 2,000
In granite zones 22,000 300,000 136*
In sillimanite
grade zone 13,600 100,000 35*
In chlorite 1,100 2,500 56*
In public utilities 2,000 11,700 64
*Rock grade determined by geologist for each private well.
-------
TABLE 8
The Distribution of Radon in Municipal Water Supplies
in the Central United States
Well Waters Distribution Systems
STATE N G.M. G.S.D. N G.M. G.S.D.
(pCi/1) (pCi/1)
Arkansas
6
135
2.66
20
47
1.95
Indiana
10
151
2.13
23
70
2.20
Iowa
33
175
2.13
31
111
2.45
Louisiana
61
151
1.84
22
93
2.54
Minnesota
28
252
2.08
28
183
2.65
Nebraska
47
262
1.85
21
178
3.68
New Mexico
27
287
2.39
20
220
2.16
Oklahoma
7
117
1.74
6
134
1.17
COMPOSITE
209
197
2.10
174
115
2.75
Texas
278
131
2.70
-------
TABLE 9
Radon in Water Results by State and Source
Results are geometric means in units of pCi/1. Parentheses values are
bers of samples.
STATE
PRIVATE
WELL
PUBLIC WATER
SUPPLY*
PUBLIC GROUNDWATER
SUPPLY
PUBLIC SURFACE
WATER SUPPLY
**
AL
120
(22)
8
(31)
70
(182)
ND
(8)
AR
230
C 2)
1400
( 1)
12
( 22)
ND
(1)
AZ
—
—
250
(124)
ND
(6)
CA
43
C 6)
790
( 2)
470
( 15)
ND
(2)
CO
—
—
230
( 76)
—
DE
—
—
30
( 72)
—
FL
6000
C 34 )
320
( 2)
30
(327)
—
GA
2100
( 2)
44
(32)
67
(225)
43
(2)
IA
—
—
220
( 85)
ND
(2)
ID
—
—
99
(155)
—
IL
—
—
95
(314)
—
IN
—
—
35
(185)
—
KS
—
—
120
( 47)
74
(2)
KY
1500
(10)
ND
(18)
32
(104)
ND
(5)
MA
1000
( 8)
7
( 2)
500
(212)
38
(2)
ME
7000
(24)
990
(71)
—
—
MN
1400
C 1)
600
( 1)
130
(233)
—
MO
ND
( 2)
--
24
(138)
ND
(2)
MS
—
260
( 2)
23
(104)
—
MT
4300
( 8)
—
230
( 71)
ND
(6)
NC
15
(29)
27
( 2)
79
(404)
ND
(4)
ND
—
440
( 2)
35
(133)
—
NH
1400
(18)
9
(12)
940
( 52)
ND
(6)
NJ
—
—
300
( 38)
—
NM
59
(14)
45
( 8)
55
(171)
ND
(18)
NV
—
—
190
( 57)
—
NY
1500
( 4)
34
(20)
52
(292)
ND
( 1)
OH
—
—
79
(165)
—
OK
—
—
93
( 83)
—
OR
450
(18)
—
120
( 69)
ND
( 4)
PA
910
(16)
—
380
(105)
—
RI
6500
(69)
5200
( 6)
2400
(575)
ND
(10)
SC
1100
(28)
—
130
(384)
ND
(14)
SD
4200
( 2)
59
( 2)
210
(155)
—
TN
ND
( 2)
ND
( 2)
12
( 98)
—
UT
—
—
150
(195)
—
VA
560
(42)
—
350
(284)
ND
( 4)
VT
210
(23)
840
( 4)
660
( 71)
13
(16)
WI
730
(40)
'28
( 4)
150
(278)
ND
(12)
WY
—
330
( 32)
ND
( 2)
US
920
(434)
68
(224)
130
(6298)
1
(131)
~May include both ground water and surface
**ND - Not detected above background levels.
-------
TABLE 10
Radon
in Water Results
by State and
Population
All
results are
geometric
means
in units of pCi/1.
Parentheses
values ar
in
numbers
of samples.
Population Ranges
STATE <
100
100-1000
1000
-5000
5000-10,000
> 10,
000
Unknown
AL
83 (10)*
59
(76)
82
46)
68
(44)
170 ( 6)
AR
—
32 ( 6)
6
( 3)
12
( 2)
8 (13)
AZ
—
240 ( 2)
200
(68)
350
22)
340
(30)
160 ( 2)
CA
--
—
-
-
— —
470 (15)
CO
—
130 ( 6)
220
(54)
400
8)
300
( 8)
OE
—
100 ( 4)
39
(44)
11
12)
23
(12)
FL
—
320 ( 2)
290
( 4)
49
78)
24(243)
_ _
GA
57
( 4)
42 (12)
130
(56)
190
27)
52
(45)
39 (81)
IA
—
1120 (13)
230
(36)
150
24)
66
(10)
75 ( 2)
ID
6
C 3)
210 (14)
130
(83)
36
25)
100
(30)
IL
—
71 ( 1)
81
(30)
80
98)
not
185)
«•->
IN
—
—
45
(71)
17
58)
50
(56)
KS
260
( 2)
230 (12)
43
(11)
320
4)
58
( 8)
210(10)
KY
—
20 (10)
42
(76)
36
8)
5
(10)
MA
3300
( 2)
—
380
(47)
540
67)
510
(88)
580( 8)
ME
670
( 3)
1600 (23)
690
(33)
2700
7)
450
( 4)
400( 1)
MN
—
180 (22)
150
(76)
68
43)
140
(90)
330( 2)
MO
ND
C 2)
58 (54)
26
(54)
1100
4)
78
( 6)
ND(18)
MS
—
—
15
(45)
51
26)
23
(33)
MT
740
( 4)
280 ( 6)
270
(49)
160
8)
18
( 4)
— —
NC
11,000
( 6)
250(111)
45
(229)
16
32)
21
(16)
8100(10)
ND
—
13 (10)
39
(112)
97
5)
10
( 6)
NH
1700
( 2)
1200 ( 4)
960
(24)
1000
14)
550
( 8)
— —
NJ
—
—
—
360
6)
1200
(12)
120 (20)
NM
1300
( 2)
620 ( 7)
48
(89)
42
16)
33
(49)
450 ( 8)
NV
—
220 ( 2)
240
(36)
72
9)
530
( 2)
180 ( 8)
NY
—
—
56
(59)
31
85)
71 (
113)
56 (35)
OH
—
300 ( 2)
160
( 4)
56
76)
100
(83)
— —
OK
—
260 ( 1)
65
(33)
190
21)
79
(26)
96 ( 2)
OR
—
210,( 2)
110
(48)
23
6)
320
( 6)
180 ( 7)
PA
1900
C 2)
—
260
(34)
450
32)
440
(37)
910 (16)
RI
1700(
91)
3900(325)
980
(71)
1300
30)
1200
(58)
SC
1400
(32)
870 (30)
92
(229)
74
60)
60
(31)
410 ( 2)
SD
310
(16)
300 (41)
160
(85)
210
8)
200
( 5)
_ —
TN
—
160 ( 4)
19
(48)
6
25)
5
(21)
UT
—
260 ( 2)
140
(104)
200
39)
170
(48)
150 ( 2)
VA
85
( 6)
880 (56)
140
(151)
320
12)
720
(14)
2300 (45)
VT
120
C 1)
—
540
(24)
1000
( 2)
750 (44)
WI
—
—
150
(169)
190
61)
130
(48)
— _
WY
—
880 ( 6)
230
(19)
530
6)
54
( 1)
US
990(175)
620(777)
98(2446)
92(1098)
88(1464)
140(338)
~Number of data points used to calculate geometric mean.
**ND - not detected above background levels.
-------
TABLE 11
Indoor Radon Concentrations.(pCi/1) in "Background" U.S. Dwellings
Number of Geom. Geom Eql. Comments Reference
Site Dwellings Mean S.D. Ratio
New York
New Jersey
Basements
18
1.7
2.0
0.48
Ordinary
Ge80
1st Floor
18
0.83
2.0
0.49
Houses
2nd Floor
9
0.77
1.8
0.45
Central Maine
Ordinary
Pr83a
Houses
Basements
77
2.46
2.4
1st Floor
82
1. AO
2.4
Bedroom (2nd)
81
1.12
2.4
Bathroom
81
1.62
2.4
Outside
67
0.46
2.4
Houston, Texas
Houses,
Pr83a
Apts.
Bedroom
103
0.39
2.5
Bathroom
103
0.58
2.5
Outdoors
81
0.22
2.6
Eastern
Pennsylvania
Sa81
Living Area
36
Summer
1.22
4.6
Winter
4.40
3.7
Basement
36
Summer
3.40
6.4
Winter
5.90
4.0
-------
TABLE 12
Factor Relating Radon in Indoor Air to Radon in Water
(pCi/1 In Air Per 105 pCi/1 in Water)
Factor Reference Conditions
14
(Ka80)
50
10
5
< 1
(Ge80)
(Pr83a)
(Mc80)
Calculated value for typical
Finnish single and double
family structures, based on
experimentally determined radon
releases.
Calculations based on 4 occu-
pants, experimentally deter-
mined radon releases
Volume Air Change Rate
15000 1 0.25 hr-1
34000 1 0.50 hr-1
34000 1 1.00 hr-1
Observation obtained by
regression from 3 months
integrated data from 80 houses
in Maine, causality not
strictly implied.
Air and water grab samples in
13 structures in Halifax, Nova
Scotia
-------
TABLE 13
Radon Levels Measured in Air in Houses
Average Maximum Number
pCi/1 pCi/1 of Houses
Maine (He82b) 3.0 63.0 85
New York
Albany Area (F180) 3.1 26.0 21
New York City Area (Br79) 1.0 5.9 21
California
San Francisco (Ne81) 0.4 1.1 28
Pennsylvania
Eastern (Sa81) 10 36
Maryland (Mo81) 3.7 27.0 56
-------
TABLE 14
Radon Daughters in Swedish Dwellings
Radon Daughters Radon
(Working Levels) (pCi/1, f=0.5)
Number of
Dwellings
Percent of
Dwellings
0.000 - 0.054
0.054 - 0.108
0.108 - 0.270
> 0.27
0-11
11 - 22
11 - 54
> 54
6326
4050
1545
162
52
34
13
1
-------
TABLE 15. Cases Prevented for Uranium Concentrations in Public Water
pCi/1
% Supplies
Number
Supplies
People
Exposed*
Average U >
Cases
> 1 pCi
23.6
10,808
10.8 x 10*
7.5
275
> 5 pCi/1
7.88
3,609
3.6 x 106
17.5
214
> 10 pCi/1
3.46
1,584
l.j x 106
30.3
154
> 20 pCi/1
1.33
609
0.6 x 106
54.8
112
*1000/supply
-------
TABLE 16. Assessment of Water Risks for Radon in Public Groundwater
Population
Number of
Utilities
Sampled
Mean Concentration
Of Radon in Water
(pCi/1)
GEOM - ARITH
U.S. Population
Using This Water
(millions)
Lifetime
Risk
GEOM - ARITH
< 100 88
100 - 1000 377
1000 - 5000 1223
5000 - 10,000 549
10,000 - 100,000 704
> 100,000 32
990
620
98
92
92
52
6500
4100
390
350
290
150
1.03
7.4
14.0
8.4
28.3
14.3
59.13
407
1835
548
309
1018
297
4414
2678
12136
2184
1176
3283
858
22315
Population weighted Average 187. pCi/1 - 944. pCi/1.
GEOM
FACTOR USEDt 0.4 x 1Q-6 deaths
pCi/1 water
-------
TABLE 17. Concentrations of Th, Pb, and Po Isotopes in Ground Water (pCi/1).
Description
Th-232 Th-230 Pb-210 Po-210 Reference
New Mexico
Grants Mineral Belt
(54 wells)
Paquate-Jackpile Area
<0.01
<0.02
0.39
Ka 76
Grants-Bluewater Area
<0.01
0-0.04
0-0.66
Ka 76
United Nuclear Area
0
1
o
•
o
0-0.099
0.3-2.3
Ka 76
Ambrosia Lake
<0.03
0.0.08
0-3.8
Ka 76
Gullup Area
<0.02
0-0.09
0-0.6
Ka 76
Rapides Parish
Louisiana (1 well)
290-607
Mu 82
California well
1.3
1.1
ND
This |
paper
Arizona well
ND*
ND
0.9
This |
paper
Connecticut
glacial drift
0.02
<0.001
Kris
82
glacial drift
0.02
0.004
Kris
82
crystalline rock
0.03
0.06
Kris
82
sandstone
0.12
0.020
Kris
82
sandstone
0.07
0.005
Kris
82
sandstone
0.03
0.004
Kris
82
Leesville, S.C.
sand aquifer
0.01
This
paper
ND = not detected
-------
Figure Captions
1. Figure 1, Map of geological provinces of the United States
according to Beddinger (Be81).
2. Figure 2, Map showing the approximate locations and general
areas of public water supplies which exceed 5 pCi/1 of total
Ra (Ra-229 was reported or combined with Ra-226 for about
one-half of the sites). Large dots represent individual
violations. The dot pattern represents the general area of
a group of violations, with the adjacent number indicating
the number of violations in that group. When the locations
were unknown, just the number of violations was indicated
(modified after Cothern and Lappenbusch (Co83)).
3. Figure 3, Map of population averaged uranium concentration in
pCi/1 for surface water in the United States.
4. Figure 4, Map of population averaged uranium concentration in
pCi/1 for ground water in the United States.
5. Figure 5, Map of population averaged uranium concentration in
pCi/1 for domestic water in the United States.
6. Figure 6, Map of geometric average radon concentration in
pCi/1 for public ground water supplies in the United States.
7. Figure 7, Histogram showing radium concentrations in pCi/1
for public water supplies exceeding 5 pCi/1 in the United
States.
8. Figure 8, Histogram showing uranium concentrations in pCi/1
for domestic water supplies in the United States.
9. Figure 9, Histogram showing radon concentrations in pCi/1
for public ground water supplies in *-.he United States.
-------
Acknowledgements
We would like to acknowledge the help of the United States
Environmental Protection Agency especially Dr. Rick Cothern, and
Dr. William Lappenbusch and Mr. Drury of Oakridge National Lab
for helping us to obtain information about uranium occurrence.
We would like to acknowledge the assistance of Dr. Moore of the
University of South Carolina for previewing the manuscript and
suggestng useful changes. We would also like to acknowledge the
objections of Mr. Thomas Horton, to the inclusion of the risk
section in this paper. We thank the typist, Mrs. Patricia Heal
for her patience in making revisions to the manuscript.
-------
O NEW ENGLAND-ADIRONDACK MOUNTAINS
5 APPALACHIAN HIGHLANDS-PIEDMONT
S APPALACHIAN AND INTERIOR PLATEAUS
CD COASTAL PLAIN
HQ GLACIATED CENTRAL PLATFORM Bi
~ WESTERN CENTRAL PLATFORM E3
E3 ROCKY MOUNTAIN SYSTEM MSI
E3
COLORADO PLATEAUS
BASIN AND RANGE
COLUMBIA PLATEAUS
PACIFIC MOUNTAIN SYSTEM
-------
- ONLY GROSS ALPHA PARTICLE VIOLATIONS
ARE AVAILABLE 144 SITES WITH >15 pCi/ll
^ NO DATA REPORTED
-------
URANIUM CONCENTRATION (pCi/l) SURFACE WATER
MONT.
NDAK.
MINN.
S-DAK.
MICH.
CONN
N.J. 01
75 VW.VA
MO. .075
CALIF.
OKLA.
N.MEX
TEXAS
ALASKA
-------
URANIUM CONCENTRATION (pCi/l) GROUND WATER
N.H.
.035 j ME\
W.035
VT.
.01
N.DAK.
.35
MINN.
MA. .01
.15
NY
.035
WIS
.01
SOAK.
7.5
I 075
CONN. .075
N.J. .01
WVo.
MICH.
) .35
PA.
.01
IOWA
.75
3.5
NEB.
OHIO
3.5
DEL.
MD. .075
.75
COLO.
.75
MO.
KAN.
CALIF.
KY.
.01
3.5
.01
3.5
N.C.
.035
TENN.
.01
OKLA.
ARK.
.075
ALA. \ GA.\035
3.5
MISS
.01
.075
TEXAS
LA.
.15
.075
ALASKA
PQ
HAWAII
o
-------
URANIUM CONCENTRATION (pCi/l) DOMESTIC WATER
R.I. 075
CONN. .035
CALIF.
OKLA.
TEXAS
ALASKA
-------
GEOMETRIC AVERAGE RADON CONCENTRATION IN PUBLIC GROUNDWATER SUPPLIES (pCi/l)
N H. f\
VT 940 I ME.
660
N.DAK.
35
990
MINN.
230
MA.500
NY.
WIS
S.DAK.
130
WYO.
R.I. 2400
CONN.
N.J. 300
MICH.'
150
210
330
PA.
380
IOWA
220
NEB.
180
ILL. IND. I 0H,°
95 I35
1 J ky
OEL. 30
MO.
190
COLO.
VA.
350
MO.
KAN.
CALIF.
230
24
120
470
N.C.
79
TENN
OKLA.
93
ARK.
130
GA
250
ALA.
55
MISS.
67
70
TEXAS
LA.
23
130
90
ALASKA
HAWAII
o
130
-------
30-
25-
co 20-
UJ
co
<
o
U-
° 15-
tr.
LlI
CD
10-
5-
1
\n
/lh
10
15
20
25
I
30
COMBINED RADIUM CONCENTRATION (pCi/l)
-------
CO
Ui
-J
Q.
2
<
CO
u.
o
IT
UI
CD
6000
5000-
4000 -
3000-
2000 -
1000 -
(0
CM
00
*¦
0>
00
ID
ro
N-
ro
O
lO
N-
00
CM
2 ®
0> O
CM Si
00
ro
CM
a>
*
CM
CM
o _•
to o>
Is- - •
m CM £- ro
n
h- —
.0 .05 .20 I 5 20 100 500
.02 .10 .50 2 10 50 200 1000
URANIUM CONCENTRATION
-------
PUBLIC GROUNDWATER SUPPLIES (ALL)
100.00
80.00-
>¦ 60.00-
o
z
UJ
z>
o
jj! 40.00
CM
o
CM
o>
o>
(D
(0
CM
lO
*
to
CM
CM
IO
O
IO
0.10 0.50 1.00 5.00 10.00 15.00 20.00
RN-222 CONCENTRATION (PCi/l) *I03
—»
50.00
—T=
100.00
-------
COMMITTEE ON METABOLISM AND DOSIMETRY
OF HIGH LET RADIONUCLIDES
Chairman: McDonald E. Vrenn
Recorder: Chris Nelson
Committee Members: Blaine Howard
Joyce Lipzstein
John Rundo
Edwin T. Still
David L. Willis
-------
METABOLISM CF INGESTED URANIUM AND RADIUM
McDonald i. Wrenn
Radiobiology Division
University of Utah
Salt Lake City, Utah 84112
Patricia W. Durbin
Lawrence Berkeley Laboratory
Berkeley, Ca. 94720
Blaine Howard
Division of Environmental Health Services
Utah State Health Department
Salt Lake City, Utah 64110
Joyce Lipsztein
Coirissao Kacional de Energia Nuclear
InstitJto de Radio Protecas E Dosimetria
Avenidadas Aaericas, KM11,5
Rio de Janeiro, Brazil 22600
John Rundo
Argonne National Laboratory
Argonne, IL 60439
Edwin T. Still
Environment and Health Management Division
Kerr-McGee Corporation
Oklahoaa City, OK 73125
David L. Willis
Department of General Science
Oregon State University
Corvallis, OR 97331
-------
Abstract
The literature on metabolism of 0 and Ra for man relevant to deriving drinking
water standards has been reviewed and summarized. Ra is well understood, but
significant gaps remain in our knowledge about U metabolism. Limits should be
based on an equilibrium model where a constant relationship between intake and
organ burden is established, using the best and most likely metabolic para-
meters. For the skeleton we conclude that the best estimate of skeletal
22 c
burden expressed in days equivalent intake are 25 days for Ra, 10 days for
Ra, and 0.3 days for Ra. For longlived isotopes of U, we chose 11 days,
with a range between 1 and 35 days. T^e committee believes that intake of
natural 0 in water should be limited by considerations of toxicity to the
kidney, and we believe that the metabolic* model of Spoor and Hursh with a
modified GI absorption (1.4%) should be used to infer kidney content. Our
review and analysis of the world literature lead us to believe the average
human GI absorption of 0 is most likely 1 to 2% and is probably reasonably
independent of age or the mass of U ingested. Using a safety factor of 50 to
150, the committee recommends a limit of D in water of 100 micrograms/liter in
order to limit toxic effects in the kidney. 100 micrograms/liter is equi-
valent to 67 pCi/liter of longlived alpha-emitting natural 0 isotopes. Further
research into the distribution of U in the human body is desirable, especially
at natural levels in kidney and skeleton, the time dependent pharmacokinetics
of U in animals, the GI absorption of U in man from water and food, toxico-
logical and U distribution studies in animals under conditions of chronic oral
U intake, and metabolic model error propagation.
* In most cases throughout the text and appendices, "metabolism" can be used
interchangeably with "biokinetics".
-------
1
Int roduction
The metabolic models foe U and Ra which are described in this section are
required to estimate the risks to human health from ingesting these elements
in drinking water. Chemical toxicity, which is relevant to 0 in its natural,
depleted or slightly enriched state, is addressed, as are the radiotoxicity
and the radiobiological effects of the important alpha-emitting isotopes of
Ra, including ^*Ra, ^®Ra, and ^®Ra.
Although no radiobiological effects from injected or ingested natural U
have been, or probably can be, observed in occupational!-/ exposed populations
or in experimental animals, this paper estimates the kinetics of skeletal D
deposition, so that risk coefficients for bone cancer induction developed by
the subcommittee on risk can be applied JMays84). This procedure utilizes the
average dose to bene, rather than endosteal dose or dese to bone marrow. Dose
226 228
to bone marrow is not used because populations exposed to Ra + Ra and
224
Ra developed no additional leukemias above expectation (Spi83). However,
for these radionuclides, many bone sarcomas were observed, indicating that
dose to red marrow is less effective in inducing leukemia than the dose to
bone surfaces is in inducing bone sarcomas. tM&Cwill be regarded as
the major potential radiobiological effect of ingested alpha-emitting radio-
isotopes of Ra and the presumed radiobiological effect of U, if any.
Finally, best estimates of normal D metabolism will be used, because even
in extreme cases the amounts of 0 or Ra ingested in potable water are not
great enough to chemically or radiobiological^ modify their metabolic beha-
vior (Co83a). With best estimates, known factors of safety can be introduced
at the ends of the assessments.
-------
2
Scope of Review
This paper cites more than 120 references, many of them review articles
which, in turn, summarize large numbers of scientific papers. Periodically
since 1958, and most recently in 1982, the United Kations Scientific Committee
on the Effects of Atomic Radiation (UNSCEAR) has summarized what is known
about the levels of Ra, 0, other naturally-occurring radionuclides and manmade
alpha-emitters in air, food, water, soils and rocks, as well as inhalation and
ingestion of these materials in food and water and the relationship of that
intake to their accumulation in the body (uN58, 62, 66, 69, 72, 77). Earlier,
the Manhattan Project investigated and reported on the potential chemical tox-
icity of 0 in the National Nuclear Energy Series published in 1949 and '.953
(Ta51, Vo49, Vo53). Health and safety data for 0 in the workplace were pub-
lished in 1958 (HA58) and in 1975 (Wr75). The toxicology and metabolism of U
in man and animals, as well as the relevance of that data to established
protection limits, were thoroughly reviewed in 1973 (Hod73a,b; Hu73, Sp73,
Yu73). The metabolism and biological effects of Ra have been examined in a
number of important articles (Ev66, ICRP73, Ro78, Ro83), and constitute a
large part of the first supplement to Health Physics (Vol 44, Suppl. 1,
1983). The metabolism of environmental levels of Ra and U was also reviewed
in a symposium on high natural background areas (Wr77b). The comparative
distribution of D among lung, liver, bone and kidney, as well as local distri-
bution in bone, was described in a followup symposium (Wr82).
-------
3
226
In view of the availability of so much data, (especially for Ra), the
scope of the discussions of Ra and U metabolism in this paper will be limited.
Interim drinking water standards are presented in terms of concentration
(amount per unit volume of water). In that the issue under consideration is
ingestion, such a restricted form of expression of limits is believed to be
valid, even though the underlying assumptions about the amount of daily water
intake must be consistent with the objectives of the risk limitation proce-
dure. Recommendations will be made in a form relevant to isotopes of U and Ra
likely to be present in soluble form in drinking water, and it follows that
water samples should be filtered prior to analysis for compliance.
The criterion we adopt to control radiological risk is to limit the like-
lihood of bone cancer induction in populations, not in individuals. It fol-
lows from the adoption of this criterion that the number of interest is the
average intake in the population on a per capita basis, rather than a maximum
intake for a maximally exposed individual in the population. Therefore, this
paper uses best estimates of average intake rates of water for input to the
metabolic model. The procedure developed should be sufficient to limit per
-4
capita lifetime risk to 10 for bone sarcomas and carcinomas in soft tissues
adjacent to bone (head sinus carcinomas).
There are no data on experimental induction of bone cancer by ingested,
injected or inhaled natural U in soluble form. Bone cancer has been experi-
mentally induced in animals by D, but only higher specific activity U isotopes
or mixtures of U isotopes: 232u or 233u injected as U02(N03)2 into mice
(Fin53); enriched U+® (90% 23*U and 23^u by weight) intratracheally instilled
in rats (Fil78); 232U or 233U inhaled as 002(N03)2 by rats (Bal83). No bone
cancers or leukemias have been reported among the large numbers of rats,
-------
4
rabbits or dogs that were injected with, fed, or inhaled soluble ..or insoluble.
0 compounds (Vo49,53; Fin53; Yu73).
For collective dose and risk, we will consider exposures leading to cum-
ulative bone tumor risk of 1C% of the natural incidence of bone cancer, or an
-4
approximate per capita lifetime risk of 10 in man, while noting that total
"natural" cancer mortality lifetime risk in the U.S. is about 1.7 X 10~\
The International Commission on Radiological Protection (ICRP77a,b) has
suggested that "a risk of cancer in the range of 10"® to 10~5 per year would
be likely to be acceptable to any individual member of the public." Over a
7C-vear lifetime, it corresponds to a lifetime risk in the approximate (be-
cause of uncertainties in latency periods and "plateau" effects) range of 7 X
10~5 to 7 X 10"4.
Choice of a Metabolic Model
Radiation dose from internally deposited radionuclides is rarely obtained
directly, particularly at environmental intake levels, because of the near
impossibility of in situ measurements, particularly of alpha particles.
Instead, other quantities are measured from which doses can be inferred. The
closer the relation between the measured quantity and the dose, the fewer are
the assumptions required, and the more realistic the estimate of dose is
likely to be. Thus, dose is best calculated from knowledge of the amount and
distribution of radionulcides in organs, which in turn may be inferred from
either direct in vivo measurements or metabolic models relating intake to
accumulation, distribution and excretion.
-------
5
A number of mathematical models have been developed which relate body 0
or Ra content to intake (ICRP59, 67, 72, 79). Most of these are not easily
adaptable to conditions in which both the intake and the dose rate are chron-
ic, as in a natural environment. An exception is the model developed by the
United Nations Scientific Committee on the Effects of Atomic Radiation (UN62,
66, 69, 72) which was used to estimate radiation doses from chronic intakes of
naturally occurring radionuclides.
Important considerations for determining body content of a radionuclide
after it is ingested ingestion in water are time and age dependence of the
intake. For purposes of the present analysis, it is assumed that organ con-
centrations remain constant at the equilibrium value of an adult with fixed
daily intake. Of course, this is an oversiraplification, since dietary com-
position and source of water supply can change with age and residence, as can
some of the relevant metabolic parameters. However, there is evidence that
226 228 210
such a simplification is appropriate for Ra, Ra, Pb, and U (Wr77b).
Finally, one needs to consider the interval between conception and birth, when
the accumulation of minerals depends on (a) the mother's intake, (b) the
reservoir of elements in the mother's body, and (c) the ability of the
placenta to discriminate between essential and non-essential (but chemically
similar) elements.
In the equilibrium model, skeletal concentration is a constant and there
is no dependence on age. This is relevant to a substance which may experience
discrimination by the GI tract, but not by the placenta (Wr77b). In this
case, the concentration of an element in the fetal skeleton should reflect
that in the mother, and the ratios of the element to calcium in the newborn
and in the maternal blood should be the same. Elements which satisfy this
-------
6
condition are the alkaline earths, which include the longlived Ra isotopes and
U, which at physiological pS is in stable 6+ state as the divalent cation
uo2+2.
The approaches we will take are to identify empirical relationships
between equilibrium concentrations in the environment and the human body, oc
to adduce sufficient information about metabolism to predict the uptake and
concentration in the body as a function of time. The latter approach will be
restricted to the case of 0 in the kidney. There may be circumstances which
alter the proportionality constant between intake and equilibrium amount in
the body; for example, chemical congeners in food may influence (usually
reduce) transport across the GI tract.
Metabolic Model for Ra
This section examines placental discrimination and variation of Ra con-
centration in bone with age. The normal intake of Ra has been the subject of
several reviews: food is normally the major source of intake (UN62, 66, 69,
226 228
72, 77). Whenever the concentration of either Ra or Ra exceeds several
pCi/1 in water, water may be the dominant source of intake. Generally, air is
not an important contributor. 1*ie C7N5C2AR examined normal and elevated intake
of Ra in air, water, and food, and its content in the human body in the
reports issued in 1958, 1962, 1966, 1972 and 1977. The following discussion
draws heavily on the UNSCEAR summaries. The ICRP report entitled 'Alkaline
Earth Metabolism in Adult Man" (ICRP73) focused on the metabolism of transient
intakes in both the occupational and medical Ra cases. The dosimetry of Ra
and other alpha-emitters in bone was summarized by Spiers (Spi68) in his mono-
graph on internal emitter dosimetry, and by the National Council on Radiation
-------
7
Protection monograph on :Natural Background Radiation in the United States"
(NCR?"? 5).
The limited evidence shows that under conditions of chronic intake the
concentration of longlived naturally occurring alpha emitters per gram of body
ash is nearly invariant throughout life JMayn61, St€4# Ra65, 3o63). Mayneord
226 226
showed this is specifically correct for Ra. Thus, the natural Ra/Ca
ratio in the human skeleton does not change (St64) from 4 months of fetal life
(organogenesis is not complete until 3 months) through old age. Fisenne
reported measurements of Ra in human bone which are consistent with the hypo-
thesis that Ra in the skeleton is reasonably age-independent (Fi~9). Accor-
dingly, the model of an age-independent Ra/Ca ratio in the skeleton appears to
be appropriate for continuous Ra intake (UN66).
The observed ratio in bone and diet is defined as:
g Ra/g Ca in bone
OR (bone/diet) = .
g Ra/g Ca in diet
The OR is considerably less than unity for Ra.
226
A global mean OR of 0.024 or median OR of 0.020 for Ra is suggested by
UNSCEAR in its 1977 survey of the worldwide measurements (UN77). This survey
presents the results of eight studies in six countries in which the mean OR
ranged from 0.013 to 0.039. The highest single value (0.039] was for San
Francisco; the second highest U.S. value (0.024), for New York City. In any
given study, sampling or other errors which result in a high estimate in bone
or a low estimate in diet, or the failure to include significant non-dietary
sources of Ra, would lead to an overestimate of OR. Using the global average
intake of 0.92 pCi/day, which lies between the two U.S. values, and a mean
-------
8
concentration of 5.1 pCi/kg in bone (rtass of mineralized tissue is 5 kg for
*
Reference Man ), leads to a net burden in the skeleton that is equivalent to
28 days intake with a range from 20 to 37 days. Considering the potential
variability of the data from which these numbers are derived, this range seems
small indeed.
However/ in order to make the best estimates of the Ra content of the
human body from dietary information, the Ca intake should be known as well, as
pointed out by Penna-Franca et al. (Pe65).
226
Pot Ra, it is possible to measure the body burden in vivo at low natu-
ral levels (i.e., about 20 pCi) by use of the breath radon technique, provided
that care is taken to have the subject breathe radon-free air for a suffi-
226
ciently long period. Stehney and Lucas (St56) measured body burdens of Ra
2 26
using the breath radon technique; intake rates of Ra were determined for
226
the same population by measuring Ra in feces, which should give a good
indication of total intake. Measurement of Ra excreted in feces avoids the
need to reconstruct representative diets; such reconstruction introduces
errors if foods and water are not included in their proper representative
proportions. Food preparation may also alter the Ra content of the water used
(e.g., in brewing coffee). Four groups were measured as shown in Table i.
The Statevilie samples were from a prison having high ^^bRa in water, and the
s-ufcjects were presumed to have had a constant intake during their incarcer-
226
ation. Since incarceration represented a change, most likely to higher Ra
intake, those subjects were not in equilibrium. The elevated body burdens of
The 5000 g of mineralized bone is water, organic, and bb% ash by
weight; the dry mass is 4400 g (ICRP/4).
-------
9
*5 o a ^ j
^°Ra were not established rapidly, and between 10 and 10 days was required
to approach equilibrium.
Stehney and Lucas' evaljations of "days of intake equivalent body burden"
range from 17 for the Stateville group (11 subjects with a mean tine at State-
ville of 20 years, probably not in complete equilibrium) to 45 days for eight
boys from Iockport, Illinois (ages 15 to 18). With the exception of the latter
group, the results are consistent with a value of 23 days derived from the
alkaline earth model adopted by the ICRP (ICRP73). Further, assuming dietary
intakes of 1 g Ca/day, the results are consistent with an OR(bone/diet) of
0.02.
As is well known from animal studies, although Ra initially deposits on
bone surfaces, especially in areas of rapid bone formation, continual intake
and bone remodeling cause the Ra to become uniformly distributed throughout
bone.
226
In the past, between 10% and 30% of the Ra in the adult human body was
believed to be in soft tissues. The ICR? model of alkaline earth metabolism
suggested that 15% is an appropriate estimate under chronic exposure condi-
tions (ICRP73). Schlenker et al. recently adjusted the ICR? aodel on the
226
basis of data on the Ra content of soft tissues from 17 subjects who
received Ra by injection or ingestion 5 days to 53 years before measurement
(Sc82). They concluded that soft tissue retention peaks at 58% of the whole
body retention 18 days after a single intake and then falls steadily to 33% at
100 days and 6% at 1000 days. Under conditions of chronic intake, 5.5% of the
body's Ra content would be in soft tissue.
In summary, the global average approaches 28 days equivalent intake of
226
Ra in the body (UN77). The ICRP alkaline earth model predicts 24 days, and
-------
10
the data on measured human intake rate and accumulation range from 17 to 45
days (St56). Since the alkaline earth model includes biological knowledge not
226
inferrable from empirical measurements of Ra in ingesta and bone, it is
reasonable to select 25 days intake equivalent as a best estimate of the
equilibrium content of Ra in the skeleton.
2 2 d
The days equivalent intake for Ra and Ra will be inferred using
228 1
parameters adopted by 1CRP for the alkaline earth model. Ra (T^ = 5.77
years) is ubiquitous in soils, rocks, foods, and water, and it is metabol-
226 22 4
ically the same as Ra. Ra accumulation is limited by its short physical
half-life of 3.62 days? it is commonly associated with ^®Ra via ^®Th.
2 2 fl 55i
We can deduce values for the body contents of Ra and Ra in units of
their daily intakes, from their effective retention integrals given in Table
36 of ICRP Publication 20 (1CRP73) or in Table 8 of reference Sc82, leading to
equilibrium values of body contents of 10 and 0.3 times their daily intakes,
respectively (GI absorption taken as 0.2). The Ra will be distributed on
bone surfaces, but will have insufficient time before decay to become well
distributed throughout bone volume. For these parameters, the best estimate
of the equilibrium Ra content in human skeleton is given in Table 2.
Metabolic Model for U Isotopes
The metabolism of D has not been as well studied as that of Ra. Details
about the gastrointestinal absorption of D and its distribution in the body
are reviewed thoroughly in Appendices A and B, the results of which are
summarized below.
The first topics considered are placental discrimination against U and
the age-dependence of U concentration in bone. TSiere is evidence that some 0
-------
11
is accumulated in bone before birth. Sikov and Mahlum (Si68) stjdied
233
placental transfer of 0 ( U injected as citrate) in rats. The concentration
233
of 0 in placenta relative to that in the fetus was 1.7 at 15 to 20 days of
gestation. No measurements for calcium were reported. When Sikov (personal
communication) compared placental transport of a n^ber of nuclides at 19 to
20 days gestation, D showed the lowest placenta-to-fetus ratio (1.7), while
the 1^Cs ratio was 2.0. This animal work suggests that U/Ca ratio in fetal
bone relative to maternal bone may be close to unity.
Masjda et al. (Mas71a,b,c,d) investigated four villages in Japan where
the intake of D varied between 1 and 9 yg/day. TCieir work showed that urinary
excretion of U appeared to be nearly independent of age after puberty. This
provides presumptive evidence that U in the body is in equilibrium with
intake, and the absence of placental discrimination against 0 relative to
calcium would imply that this equilibrium is established early in life, as is
the case for *^®Ra.
The distribution of 0 in the body of Reference Man is summarized here and
given in detail in Appendix B. Animal studies indicate that the amount of
soluble 0 accumulated internally is proportional to intake from inhalation
and/or ingestion.
Natural levels of 0 in human bones collected in the U.S. were measured in
9 studies without regard to geographical location sampled, sample treatment,
or part or number of skeletons sampled? the range of skeletal 0 content
calculated for Reference Kan is 2.3 to 61.6 yg, with median and mean values of
12.9 and 24.9 wg, respectively.
Measurements of 0 in soft tissues are sparse. The natural U concentration
in kidney appears to be only about twice that in liver, and less than that
-------
12
reported for some other soft tissues (i.e., fat, lung, gonads—see Table 3-
2). Kidney contains about 0.3% of the U estimated to be in the body of
Reference Man.
The best available value for the normal U content of Reference Han deter-
mined from measurements is about 38 yg, of which 66% or 24.9 \q is in the
skeleton. The proportions in bone and soft tissue are in reasonably good
agreement with the results of experimental 0 administrations to large animals
(Du75, Ste80), and human subjects (Ber57). If one estimated dietary D intake
at 1.75 ug D/day, then there is 14 days equivalent U intake in skeleton. In
Appendix C, kinetic modeling of U uptake in bone, using the ICR? model with
modified GI uptake and modified parameters for U distribution in bone
compartments, gives 11 days equivalent accumulation of D in skeleton.
Neutron-induced autoradiographs of natural D in bone show that under
equilibrium conditions it is diffusely distributed throughout the bone volume
(Sc73, Wr82). Higher than average 0 concentrations are found on bone surfaces
shortly after a single intake (Ro68; Ste80), but the U gradually assumes a
pattern which is more diffuse. Neutron track autoradiographs of bones from
dogs exposed only to environmental U in the diet show uniform distribution.
Under chronic exposure conditions, 0 in human bone is likewise expected to be
reasonably uniformly distributed throughout the volume.
GI absorption studies of U include single oral administration experiments
of soluble uranyl compounds to rats, dogs, hamsters, and a baboon, continuous
feeding of dry salts to adult rats, and single administrations of soluble
salts to neonate rats and swine. GI absorption is consistently lower in the
rat than in the other animals studied. Por this reason, the data on GI
absorption in the rat should not be used uncritically to infer human
-------
13
absorption. There is some evidence that GI absorption may decrease very
slowly with increasing mass of U fed (see Figure A-1).
There are four sets of data dealing with measurements of 0 in diet and
excreta of man which can be used to infer average GI absorption from food and
water. Tt.ere are six sets of animal experiments which are suitable for the
same purpose, for a total of 10 suitable experimental estimates of GI absorp-
-2 3
tion of soluble U over a range of 0 intakes from 2.5 x 10 to 6.7 x 10 y
g/kg/day. U*.e mean values range from 0.3 to 7.8% of ingested U absorbed,
with a mean of 1.6%. If the single highest value, 7.7%, is excluded, the mean
absorption is 1.1%. Least squares fits of linear, semi-log and log-log func-
tions to GI absorption vs. dosage (pgAg) yielded U absorptions at natural
intake levels in the range of 1.3 to 1.9%. The highest single value of a
absorption, 7.7%, cooes from coupling an older study of urinary D excretion in
man and a U.S. dietary survey and measurement program (fluorescence analysis
of 0 was used for both studies). 3ecause of limitations on sensitivity of the
fluorometric technique for the urinary measurements and the tenuous relation-
ship between the urine 0 levels in a few persons and the D content of their
diet (inferred from a small-sample diet survey), this value is the most
doubtful of those available. For this evaluation, we believe that the value
for U absorption obtained in that way may properly be excluded in inferring GI
absorption in man. The power function fit to the nine other studies gives a
GI absorption of 1.4% at environmental levels of O intake, and may decrease
slightly with increasing 0 intake. A value of 1.4% is adopted as our best
estimate for subsequent analysis.
The multiple compartment U metabolic model of Struxness (Str55) was
improved to include a two-conpartment skeleton 0»r77a, Wr78, Rcs80). In
-------
14
addition, an elegant direct solution of a five-corapartment nodel by Lipsztein
et al. (Li61), a blood-organ transfer kinetic model by Skrable et al. (Sk80),
and the non-connecting models of Spoor and Hursh (Sp73) and ICRP (ICRP79),
which have no feedback through blood, have also been developed. Not only the
models, but also the best values of the parameters to insert in the models,
vary with author preference and the kind and quality of the biological data
available to them. All of the models cited, using the biological parameters
preferred by their authors, give results for equilibrium skeletal D
accumulation in the range of 1 to 40 days of intake. Calculations of the
predicted equilibrium concentrations in human kidney and skeleton for several
of the models are presented in Appendix C.
Dosimetry
The average radiation dose rate to bone can be calculated in a relatively
straightforward manner, and Harley (Har74) derived the following formula in
convenient units for environmental work:
D = 18.7 CE,
where D is the dose rate in mrad/yr, C is the radionuclide concentration in
pCi/g, and E is the energy absorbed pe-r disintegration in MeV. Average doses
calculated by this formula, normalized to 1 pCi/g of parent nuclide and
adjusted for appropriate daughter equilibria and radon emanation, are sumita-
rized in Table 3. fte dcse rates per unit concentration of parent nuclide in
bone decrease in the following order: ^®Ra > > ^®0.
-------
Development of Limits for U
The committee believes that limits for natural U in drinking water should
be based on chemical toxicity (which has been observed in man and quantified
in animals), rather than on a hypothetical radiological toxicity in skeletal
tissues (which has not been observed in either man or animals). The naturally
5^ j 57ft
occurring mixture of 0 isotopes, U, tJ and U, has such a long col-
lective rate of radioactive decay (half-life), and a correspondingly low
specific activity, that its biological action is predominantly that of a non-
radioactive elennent. That fact was considered to be compelling by the ICK? in
its 1959 recommendations on permissible dose for internal radiation I.ICRP59).
Of all the radionuclides for which the ICR? recommended body burden or maximum
238
permissible concentration limits, U and its naturally occurring mixture
with U and D were the only nuclides for which chemical toxicity was con-
sidered to be the limiting criterion.
There is as much or more information on the chemical toxicity of D as
there is for any other metal (Tan51, Vo49, Vo53), and it has been the subject
of more recent studies and reviews (3od73a,b; Yu73, Hu73, Du75, Mo82, Wrenn,
Morrow and Hursh, private communication to Union Carbide Corp., 1982-34). The
quantitative relationships between D intake and kidney damage have been meas-
ured in several species over a large range of dosages of many soluble U
compounds administered by various routes for, in some cases, extended periods
of time. Only transient kidney dysfunction has been observed in patients
given homeopathic injections of (Hod 73b) or in U workers (Eas58,
Wr75). Most recently, Moss and McCurdy (Mos82) reported increased 2^ micro-
globulin excretion in urine that could be correlated with 0 in well water at
concentrations ranging from less than 0.5 to greater than 80 ;jg/liter;
-------
:6
however, the influence, if any, of other unspecified constituents of those
waters, and total U intake, has not yet been taken into account, nor has the
nor has the specific effect of U on £2"®^cto9lo5:>ul^n excretion been
quantified.*
Prcximal renal tubule cells are killed by high acute or lower chronic
dosages of soluble U compounds administered by any route to experimental
animals (Yu73, Du75). If the 0 dosage is not great enough to destroy a crit-
ical mass of kidney cells, the animal survives and the lost tubule cells are
replaced. Scwever, the new cells are not structurally identical to these
lost, and presumably they are not functionally identical to them. It also
appears that not all U-damaged tubule cells are replaced and that their neph-
rons are ultimately lost, as is shown by an increase in the proportion of con-
nective tissue and a reduction in weight of chronically O-poisoned kidneys.
Below some critical 0 concentration, bsss law relationships favor the presence
of U in glomerular filtrate as a soluble complex that does not react with the
tubule cells; however, above that critical U concentration kidney damage might
* U concentrations were increased as U concentrations in well water increased
in the urine and hair of 133 persons who draw household water from wells with
U concentrations ranging from less than 0.5 to greater than 80 Ag/l (Mos32) .
The relationships are not altogether clear, e.g., linear regression equations
predict significant U in both hair and urine of this group at zero
concentration in the water. There are neither histories of, nor overt clin-
ical signs of, kidney dysfunction even among those drinking the waters with
the highest U concentration. No subtle changes of kidney function were
revealed by clinical chemistry except an apparently elevated excretion of ^2~
microglobulin (SMG, a low molecular weight protein, considered to be a
sensitive indicator of certain renal tubular disorders). 3MG excretion
increased per unit of creatinine excretion with increasing U concentration in
water, but BMG excretion was substantial, about 80 units 3MG per unit of
creatinine, even among those drinking waters with U concentrations less than
0.5 4
-------
17
be expected to occur gradually. Although the renal reserve is great (only a
fraction of nephrons are functional at any tiire), if U exposure is high for a
long enough time, damage may accumulate to a level that compromises function.
As was noted in the Introduction, bone cancer has been induced in exper-
imental animals by injection or inhalation of soluble compounds of high
specific activity 0 isotopesf oc (Fin54, 3al83), but neither bone
cancer nor marrow dysplasia have been reported among the many animals that
have received compounds containing only the natural mixture of U isotopes.
There are only two modern controlled animal experiments in which cancer was
induced by natural 0, but neither supports a conclusion that bone cancer would
result from ingesting soluble D in drinking water or foods. In one experi-
ment, lung cancer was produced in rats and dogs but not in oor.keys following
inhalation of large amounts of highly insoluble UOj continuously for 2 to 5
years (I*e70, Le73): clearance of the UC>2 from the lungs was slow, and radi-
ation doses of the order of 200 to 500 rad were accumulated in the lungs of
those animals that developed lung tuaors. In the other study, metallic 0
powder dispersed in lanolin was implanted in rats, either once into the marrow
cavity of the femur (5C og U) or in six monthly injections into the pleural
cavity (50 mg U each). A small fraction of the U dissolved, as indicated by
some acute mortality and late renal damage. Sarcomas containing U particles,
reported to be of periosteal or connective tissue origin, developed at or
adjacent to the U injection sites. Bie same or greater incidences of similar
tumors at these same sites were seen in duplicate experiments with metallic
Ni. The authors were unable to decide whether the local tumors induced by
insoluble 0 were caused by its chemical or physical properties (Hue52).
Sarcomatous tumors have also been induced in man by large, long-standing
-------
18
deposits of extravasated, insoluble, finely-divided ThO^. (The primary
232 218
alpha-emissions of Th are con^arafcle in energy to U.) ;Sv6?; Mays~4).
These observations are not relevant to ingestion of U in drinking water.
While the information concerning renal toxicity as a consequence of con-
tinuous intake of very small amounts of D is not as extensive or convincing as
that obtained for higher levels of 0 intake, the committee believes that if
any late effects will be associated with chronic ingestion of soluble 0 in
drinking water, they are more likely to be chemical.
Our consensus metabolic roodel for 0 uses a GI absorption of 1.4%, single
exponential elimination from the kidney with a half-time of 15 days, and a
blcod-to-kidney transfer of 11% to calculate the concentration in kidney as a
function of time. In accordance with SAS guidance (NAS77, p. 8C4), the com-
mittee introduced a safety factor (called uncertainty factor by HAS) of 50 to
ensure that individuals will be unlikely to experience permanent kidney damage
from the ingestion of water containing uranium. The no-toxic-effects concen-
tration limit in the kidney was taken as 1 ug U/g kidney. The committee con-
siders irreversible kidney injury a non-stochastic effect with a threshold.
In the past, 3 ;jg/g has been considered the approximate threshold (Sp"3). As
is described in Appendix C, we concluded daily intake should be limited to 187 y
g 0/day. For 1.7 1/day intake of water (NAS77, p. 11), the limiting concen-
tration in water should be about 110 jg/1. A rounded value of "00 jg/1, is
suggested, which is equivalent to 67 pCi/1 of longlived alpha-emitting D
o ^ q 5^4
isotopes, if the U and U isotopes are in radioactive equilibrium.
-------
19
Combining metabolism, dosimetry and risk estimates
The average per capita risk of bone sarcoma can be related to intake rate
by the following formula, assarting linearity of risk and dose:
(IC)k ,
where: 1^ is the per capita lifetime average risk limit chosen for bone
sarcoma per capita in the population of interest,
I is the per capita average fluid intake in liters/day,
C is the average concentration of Ra or 0 isotopes in water consumed
(pCi/1),
k is the lifetime risk of bone sarcoma induction from 1 pCi/day intake
over a lifetime. (From Ma84).
k » 3 x 1
-------
20
time linear stochastic risk of bone sarcoma induction of 3 x 10"®; for 1 1/day
average water intake the risk estimate is equivalent to 2 x 10~5. We note
that a somewhat lower lifetime (7C-year) bone tumor risk of 1.6 x 10~® is
calculated for intake of natural U in 1.7 1/day of drinking water containing
100 '^g U/l, if we use the risk factors, bone surface cell dosimetry, and
metabolic model and parameters for an equal activity mixture of and
as recently recommended by ICRP (ICRP77a, 79), but substituting a GI absorp-
tion of 1.4% for the ICRP value of 5%.
The formula may be used to construct linear estimates of average risk.
True expectations depend on the shape of the dose response, which if propor-
2
tional to D for example, would be near zero (see KaysS4).
Table 4 lists the concentrations of 233,234,235,236 or 238^ an^
228,226, or 224^ ,n <3rin|
-------
21
Summary
1. The metabolisms of U and Ra have been discussed with reference to
limiting chemical toxicity effects on the kidney for U, and evaluating the
risk of bone sarcoma induction, a stochastic effect, for Ra isotopes. In
addition, estimates of potential risk of inducing bone sarcoma were made for U
(Mays84), even though it is not established whether natural U can induce bone
sarcomas.
2. We use the best estimates of each metabolic parameter to estimate in-
take, GI absorption, distribution, excretion and stochastic risk.
3. We use best estimates of the coefficients along with a safety factor
to derive limits for chemical toxicity for 0 in water.
4. The committee concluded that a simple *odel of bone metabolism would
suffice to evaluate the concentration of D or D isotopes in human bone under
equilibrium conditions.
5. Equilibrium models for ingested nuclides that experience little
placental discrimination are nearly independent of age. The number of days
equivalent intake for the following nuclides in the skeleton at equilibrium
are as follows:
226.
Ra
25 days
228
Ra
10 days
0.3 day
Longlived U
11 days (range 1 to 35 days)
6. A great deal of environmental and human and animal metabolic and
toxicologic data underlies the Ra work. A lesser amount of human and animal
-------
22
data are available for U. The epidemiological studies of 226Ra + 229Ra and
22*Ra in man show that all these nuclides can induce bone sarcomas, that 22®Ra
can induce carcinoma of the soft tissues, no statistically significant in-
crease in leukemias is noted (Spi83). To date, there are no data suggesting
any radiological effects from ingested natural U in either animals or man.
7. The average dose rate to bone was summarized for these nuclides for a
reference concentration of 1 pCi/g of 5000 g of mineralized tissue in
Reference Man skeleton.
8. A recommendation was developed to limit chemical toxicity to the
kidney from the ingestion of natural 0 in drinking water. The appropriate
limit for soluble U in drinking water would be 100 i^U/1, using a metabolic
oodel which estimates a higher kidney uptake of 0 than the new ICR? model
(ICR? 79), limiting the D concentration to 1 C/g of kidney, and applying a
safety factor of 50. T*)is is equivalent to 67 pCi/1 of longlived alpha-
emitting natural U isotopes in their ordinary radioactive equilibrium.
9. The equivalent of 11 days intake was chosen as the equilibrium content
of 0 in the adult human skeleton. "Kie range of measured values is from 2 to
52, indicating that our estimate could be as much as a factor of 5 higher or
lower than the extreme values.
22€ 228
10. The dose rate to bone from Ra was compared to that from Ra.
228
For uniform distribution of the dose and for 90% retention of the Ra
daughters, the dose rate to bone at equal intake rates of the two Ra nuclides
226
is about the same. For practical purposes, it might be concluded that Ra
228 228
and Ra are equally important in drinking water on an activity basis. Ra
in the skeleton equals 10 days equivalent intake, but this is compensated by
the greater average energy per disintegration delivered by its daughter
series.
-------
23
-4
11. A lifetime risk of bone sarcoma induction equal to 10 would be
conferred by lifetime consumption of 1.7 1 of water per day containing 31, 32,
and 174 pCi/1 of 226Ra, 228Ra, and a combination of 234U, 235U, and 238U,
-4
respectively. A lifetime acceptable risk of cancer induction from 0.7 X 10
to 0.7 X 10~3 is suggested by ICRP (ICRP77a,b).
12. The committee concluded that a best estimate of dose and risk per
capita should be made; uncertainties should be propagated; and conservative
coefficients should not be used to evaluate the limiting cases for intake.
The committee did not have the time a.nd resources to analyze the data taking
into account propagation of errors for the whole modeling process. This
should be done.
13. The committee recommends a limit of 100 i>g/l for natural U based on
chemical toxicity to the kidney. This should be sufficiently low to sake the
likelihood of kidney damage to individuals remote. In addition, using a
linear dose response model, the lifetime risk of bone sarcoma induction fron
natural 0 at that toxic limit for continuous intake would be between 10~^ and
10"4.
Recommendations
1.Use best estimates of metabolic parameters, realistic models, and
propagate the uncertainties. The latter may require some research.
2. Further research is needed on GI absorption of U in animals (excluding
rats), and on the toxicity and pharmacokinetics of D under conditions of
chronic oral intake or its equivalent.
-------
24
3. It is important to obtain more and better (fata for natural levels of U
in water, diet, and bar,an bor.e in different geographic areas. The distribu-
tion of U in the human body should be investigated more thoroughly, including
soft tissues as well as the skeleton.
4. GI absorption of U should be inferred by measuring 0 intake and both
fecal and urinary D excretion in man under controlled intake conditions.
5. Surveys of current water (and other fluid) intake are needed to better
understand the average per capita intake of local drinking water.
6. Final limits for U in drinking water should not be set until the
research identified above (at least items 3, 4, and 5) is complete. Hie
research is reasonably short-term in nature and could be completed within a
few years.
226
7. The interim Ra limits in water could be relaxed by a factor of at
least 4, and still provide a very high degree of protection for individuals.
8. For interim guidance for U, 100 \»g 0/1 of water was chosen as a
reasonable value, based on considerations of kidney toxicity, with the
application of a safety factor of 50 to 150.
-------
25
ACKNOWLEDGMENTS
We acknowledge and appreciate the assistance of Diane Pouts for her
typing and re-typing of the manuscript, Ror.a Hoffman for her assistance in
same, Dr. Robert Larsen (Argonne National Laboratory) for his many useful
suggestions on assessment of gastrointestinal absorption .and the relevant
literature, and Christopher Nelson of EPA for his participation in the panel
and his valuable support. Part of the work underlying this review was
supported by DOE Contract DE-AC02-7 6EV00119 and the contracts of other
committee members. The authors also wish to thank Dr. Charles Kays for his
review of the manuscript and Dr. Richard Cothern for his encourageir.ent and
patience.
-------
Appendix A: Gastrointestinal absorption and accumulation of uranium in the body.
Patricia w. Durbin and McDonald E. Wrenn
I. Animal Experiments
A. Single administration of soluble U salts to adults
233
1. nitrate m solution at pH 4 was given to 60-day-old
rats by gavage at a dosage of 0.3 mg 0/kg. Absorption was originally reported to
be < 0.05% (Hait48, Du75). The original data were reanalyzed to take account of
technical errors*. Absorption may have been as much as 0.35%, when recalculated
by summing the D contents of tissues (except GI tract and skin) and also by com-
paring the U contents of skeleton, liver and kidneys with those of rats injected
with D intramuscularly.
233
2. OO2F2 enriched with U was dissolved in water and fed to dogs by
mouth at a dosage of 0.7 mg U/kg (Fis60). Average absorption was reported to be
1.55% with a range for 7 dogs of 0.83 to 2.3%, based on assay of tissues and
excreta. Only the final calculated result was originally reported, but the raw
data were recently located and confirm the published mean value.
3. ^^^02(1*03)2 or in solution at pH 1.5 to 2 was given
232 233
to adult rats by gavage at dosage levels of 2.3 ug U/kg or 4 mg U/kg
(Su8Ca). Absorption at both dosages was 0.06%, determined by summing assayed
tissues and urine to 7 days. The same results are obtained if the skeleton and
liver U contents are compared with those of rats killed 7 days after an intra-
muscular injection of U (Harr81, Ham48, Du75).
233
* Assay of U was accomplished by dissolving ashed tissues, evaporating small
aliquots on metal plates, and alpha counting. In this particular study, correc-
tions for the self-absorption of the alpha particles, which is substantial in the
aliquots of bone.samples used, was inadvertently omitted. The minimum detectable
amount had been assumed to be one-half of the background count regardless of the
length of time of counting, and all samples yielding a net count less than one-
half of the background count were disregarded. Recalculation involved applying
the appropriate self-absorption corrections and using all data.
-------
A-2
233
4. "-"'JOjCNOj) j in solution at pH 1.5 to 2 was given to adult rats by
gavage at dosages of 5.1, 12.6, or 25.3 mg U/kg (Su83). At 7 days the skeletons
contained 0.01, 0.01 and 0.02% of the gavaged D, respectively. Absorption of
0.C44, 0.044 and 0.088%, respectively, can be calculated by comparison with 0-
injected rats (Harr81, Earo48, Du75).
5. in 0.1 M HNOj was administered to adult Syrian ham-
sters by gavage at a dosage of 0.63 mg U/kg. Absorption was calculated to be
0.77% based on a comparison of the U contents of several body parts at 14 days
with those of intravenously injected hamsters (Harr81).
233
6. 0 in a chlorinated solution of 0.01 M NaHCOj was given by gavage
to a fasting adult baboon at a dosage of 0.53 ug 2*30/kg. At the same time, 0.42
236
Ug 3/kg was administered to the same baboon by intravenous injection. Absorp-
233 236
tion was estimated to be 1.2% based on comparison of the U and U contents
of several body parts at 32 days (Lar84).
3. Continuous feeding of U salts to adult animals
1. Rats were fed for up to 2 years on pelleted diets containing 0.05
to 0.5% of 002^*2 or to 2% of (Ma53). If daily food intake was 5
g/100 g weight, daily U intakes ranged from 20 to 200 mg O/kg of uranyl fluoride
and from 125 to 500 rag U/kg of Animals were killed at 2 years, and
the U content of their bones was determined, as shown in Table A-1.
If it is assumed that the ash of the rat skeleton is 4 g/100 g body weight,
that the overall turnover rate of 0 in rat bone is 0.01 day"1, and that the
fraction of absorbed U deposited in bone is 0.2, then intestinal absorption is
estimated as follows;
yg 0/g bone ash x 40 g bone ashAq x 0.01 day^ x 100
Absorption (%) ¦ ^ a/kg/day intake x 1000 yg/ag x 0.2
As 3hown in Table A-1, the range of estimated absorption is from 0.038 to 0.078%,
and it appears to be independent of the concentration of D in the diet over the
range investigated.
-------
A-3
2. Adult male and female Sprague-Dawley rats (approx. body weight, 300
g) and male New Zealand rabbits (approx. body weight, 3.5 kg) were given drinking
water ad lib containing 600 rog/liter of (284 mg U/liter) for 90 days.
The animals were killed, and the U content of bones and kidneys was measured.
4
Intake of U was estimated to be 10 mg/day (3.3 x 10 jjg U/kg) for the rats and 60
4
n>g U/day (1.7 x 10 wg U/kg) for the rabbit9. Absorption was calculated to be
0.035% for the rats and 0.23% for the rabbits, based on measured body content of
U, the estimated daily intake, and the 9.74-day retention integral used for 0 in
xan (ICRP79, Tr83b). The value calculated for GI absorption of 0 in rats agrees
reasonably well with values reported by others for single administration of U
salts by gavage (Eam48, Su80a, Su83) and for continuous feeding of dry U salts
mixed with the diet (Hav49, Ma53). See Table A-2.
3. was fed to rats for 30 days at dietary levels from 0.5 to
12%. The mean lethal dosage at 30 days {^50/30) was a dietary level of 4% (972
mg C/kg/day based on an assumed food intake of 5 g/100 g weight/day). The LD5()/30
for intraperitoneally injected UOjCNO^^ in solution was 0.39 mg 0/kg/day, from
which it could be inferred that 0.39 x 100/972 = 0.04% of the dietary U had been
absorbed (Hav49; Ma49).
C. Single administration of soluble 0 salts to neonates
1. or solutions at pH 1.5 to 2 were given by
232 233
gavage to 2-day-old rats at dosages of 0.12 yg U/kg or 211 jg U/kg (SuSOb).
The skeletons contained 6.23 and 0.82% of the gavaged U at 7 days respectively,
from which absorptions of 6.7 and 1.3%, respectively, were calculated.
2. 233U02(NOj)2 in solution at pH 1.5 to 2 was given by gavage to 1-
day-old miniature swine at a dosage of 1.5 to 2 mg C/kg (Su82). Skeleton and
tissues (except GI tract) were assayed at 12 days and contained 31.3 and 3.2% of
the gavaged 0, respectively. Absorption was estimated to be at least 34.5% from
the body content, but a comparison with injected U would provide a more accurate
result.
-------
A-4
II. Experimental administration of U solutions in man
1. dissolved in water was drunk by an adult volunteer. The
amount administered was 417 mg 0, which foe the Reference Man weight of 70 kg,
was 6.7 mg U/kg (Bu58). An unknown amount of the U was lost by immediate vomit-
ing, and the intestinal tract was cleared rapidly by the onset of diarrhea. In 7
days about 2.5 rog of 0 was recovered in urine. Comparison with urinary excretion
of about 8C% of intravenously injected U in 7 days (5a48jSu73) indicates that
about 3.1 mg or 0.73% of the total ingested D had been absorbed.
2. dissolved in Coca-Cola was drunk by four male hospital
patients (56 to 78 years old) after an overnight fast (Hu73). The amount of 0
administered ranged from 0.08 to 0.17 mg 0/kg. Uranium in urine was measured for
one to two weeks, and mean cumulative excretion was 1.15% ± 1.26 at 7 days.
Absorption, estimated by comparison with the U-injected hospital patients (SC% of
injected D excreted in urine in 7 days, Ba48), ranged from 0.3 to 3.4%, with a
mean of 1.4% ± 1.4.
The results of the experimental studies of D absorption in animals and man
are collected in Table A-2.
III. Estimation of the absorption of U from environmental data
There is little endogenous fecal excretion of U after its injection in ani-
mals or people. Under equilibrium intake conditions, if the amount of U inhaled
is small, daily excretion of 0 in urine approximates the U absorbed from food and
drink. Thus, it should be possible to estimate intestinal absorption of U in man
and animals from the U content of the alimentary intake, urine and feces (3u~3;
Ad74; Wr77b).
A. 0 in foods—dietary surveys
In regions where treated surface waters are used for cooking and drink-
ing, foods appear to be the major source of environmentally acquired U C«*e67;
No70; Ha72). The average concentration of U in rocks is about 4 ppm (A160), and
on the average, U is taken up by plants from the soil to the extent of 7.5 x 10 ^
g 0/g fresh plant material (>3 0/g dry soil) (Tr83a). Industrial fume and
tobacco smoke may contribute small amounts of U; < 0.01 ,jg U/day from breathing
city air and < 0.05 yg U/day from cigarette smoking (two packs) (Ea72; Lu7 0).
-------
A-5
Average dietary intake of U has been calculated for three U.S. cities from
the U content of a variety of foodstuffs purchased in those cities and the
average annual consumptions of each of those foods obtained from a survey of
foods purchased by households (USDA 1955, updated 1967, quoted by Ke67; Hu"3).
The per capita intake for household members in New York City and Chicago, cities
from which urine samples were also obtained, was calculated to be 1.38 >g/day,
with 1.35 wg/day derived from foods and 0.03 ;jg/day derived from 1 liter of water
(We67).
Household food surveys do not take account of the differing intakes of men,
women, and children. 3ased on caloric need (body size, activity, growth), the
daily food intake foe Reference Man (1CRP74) is 1.4 times that of women and 10-
year-old children. All of the calculations in this review are for Reference Man,
and we estimate that the intake of D by Reference Kan in Sew York City and
Chicago is 1.75 ^g/day*.
A per capita dietary intake of 1 yg U/day was estimated for the U.K. in the
same way (Ha72). Prepared foods, condiments, and dry tea and coffee were assayed
for U in addition to foods as purchased. TOe U concentrations of prepared foods
were all greater than those of the same foods as purchased. Table salt was found
to contain 40 ppb of U, and 5 g of salt would add 0.2 i»g U to the daily intake.
Likewise, if U in tea and coffee were leached in preparation, that U would be
added to the daily intake. The D in drinking water was not measured in the U.K.
study, so it is incomplete in that respect. The data suggest that actual intake
of U with foods and beverages is somewhat greater than intake calculated from
surveys of foods purchased.
* If, at the time of the food surveys, the average U.S. household was one
adult male, one adult female and two children, their total caloric need would be
about 3 x 1 unit plus 1.4 unit or 4.4 units of caloric need. The total U intake
of the four-person household with a per capita intake of 1.38 ^
-------
A-6
Ranges of the measured U concentrations in the foods analyzed were also
reported as follows: plant products (fruit, vegetables and grains, both raw and
prepared) 0.3 to 30 ng U/g wet weight; and animal products (meat, fish, poultry,
eggs, milk, both raw and prepared) 0.0005 to 4 ng U/g wet weight (Ha72). Tt.e U
concentrations in the animal products were 1/10 or less of those in the plant
products, which is not surprising considering the demonstrated discrimination by
domestic animals against U in their diets (see Table A-2 and We67) and in
secretion of milk (McC63; La82).
Replicates of daily diets of adults were measured for U in two Japanese
cities, and the median U concentration was 0.042 >g C/g of food ash (So70). The
ash content of the diet was estimated to be about 35 g/day, so the total intake
of U in food by adults was calculated to be 1.5 yg U/day. Drinking water contri-
buted 0.0C9 pg U/liter. When account is taken of the smaller size of adult
Japanese (60 kg; Ta79) compared to U.S. males (Reference Man weighs 70 kg), that
value for U intake is in good agreement with our calculated value of 1.75 ug
U/day for the dietary U intake by Reference Man in the U.S.
B. Excretion of U in urine
Urine samples were obtained from adults (probably all males and probably
all laboratory employees) living in the New York City and Chicago areas, and the
U concentrations were determined (We60; We67). The range of U concentration was
the same for the two sample sets, and the results are combined in Table A-3. The
distribution appears to be log normal, with a geometric mean of 0.097 ug U/liter
of urine. Higher values have been reported froc an industrial laboratory in a U-
processing facility (Wi65) and quoted from an undocumented personal communication
(M.H. Dean, quoted by Hu73). 3owever, the urine U concentrations shown in Table
A-3 were obtained by the same analysts who measured the U content of the U.S.
foodstuffs, and the results of the two sets of measurements should be internally
consistent.
C. Calculation of U absorption
The daily urine volume of Reference Man is 1.4 liters, and for a U
concentration in urine of 0.097 yg U/liter, urinary elimination of U would be 1.4
liter/day x 0.097 ;»g U/liter = 0.135 ug U, which is nearly equal to the amount
-------
A-7
per day at equilibrium. Por a daily intake of 1.75 ug U by Reference Man, GI
absorption would be (0.135 pg U/day) X 100/11.75 ;jg 0 intake/day) « 7.7%.
D. Concurrent analyses of 0 in urine and local diet
Soils, stream and irrigation waters, local vegetables, composited human
diets (food and drink), and human urine were obtained in and around three
Japanese villages near U mines and mills, and from a remotely located control
village. U was analyzed fluorometrically, and the minirrum detectable amount was
about 0.01 uq O. Only the results from Ten-no, the village with the highest
level of environmental D, are considered here, because U was detectable in all
the urine samples from that village. Daily intake of U in foods and water was
about 9.2 yg for residents of Ten-no over 10 years old. Twenty-four-hour urine
specimens were obtained on four occasions from about 25 men and women over the
age of 10, for a total of 116 samples. The range of urinary U was 0.02 to 0.24 u
g/24-hours, and the mean was 0.15 ± 0.043 ng O/person/day. The urinary 0 values
appeared to be normally distributed about the mean and were independent of sex,
age, or time of sampling (Ya69, Mas7la,b,c,d).
An average apparent U absorption, 0.15 ± 0.043 wg U in 24-hr urine/9.2 ug U
intake/day = 1.6% ± 0.5, can be calculated from these data. Data from the
control and the two other U mining villages yielded estimates of U absorption
less than 1.6%; and they are skewed to low values by the large numbers of urine
samples with less than the minimum detectable amount of 0, to which zero values
were applied. The strengths of this body of data are the large number of urine
samples analyzed, the breadth of sampling, and the concurrent analysis of U in
local diets. Its weaknesses are the insensitivity of the analytical method used
for 0 and the poor recoveries of D from food and soil samples.
E. Estimation of D intake from fecal U excretion
2^0
1. Control subjects in a study of the differential metabolism of * Th
A MO
and U and D inhaled in 0 ore dust included three retired U mill workers (4
to 14 years since last employment as U ore crushermen), and three volunteers who
lived in U Billing communities but had no U work history. Two consecutive 24-hr
23 a 2 38
urine and fecal collections were obtained and analyzed for U and U. The
238 5^1
data for U, which are supported by the results for U, are shown in Table
-------
A-8
A-4 (Fish83). Apparent 0 absorption can be calculated assuming that the sum of
daily urinary and fecal U excretion is equal to the alimentary intake. It should
be noted that U absorption will be underestimated if insoluble U compounds are
being inhaled, or if, in the case of the retirees, insoluble U is still being
cleared mechanically from the lungs by way of the GI tract. Conversely, U
absorption will be overestimated if soluble U is being inhaled, or if, in the
case of the retirees, 0 ore dust in the pulmonary tract is still being solubil-
ized and absorbed into the body.
The apparent total intakes of D of these individuals ranged from 11 to 18
yg U/day for the controls and from 5.3 to 71 yg U/day for the retirees. Although
large compared to 0 intakes estimated for city dwellers (We67; Kail), the 0
intakes of these individuals are not unreasonable, because U in potable waters
and locally grown foods tends to be higher in U mining and milling communities.
The mean U absorption calculated for the three controls (0.82%, range 0.6 to 1%)
was not significantly different from that calculated for the retired U workers
(0.94%, range 0.55 to 1.6%) and the two groups can be combined. Ihis body of
data suggests that at an apparent mean daily intake of 24 yg U/day (0.34 yg U/kg
for 70 kg Reference Man), the GI absorption of U is 0.76% (range 0.4 to 1.6%).
2. In the city of Ahmedabad (Gujarat State of India) about 2 million
people consume well water with 0 concentrations up to 22.4 yg 0 per liter. Com-
plete 24-hour urine and fecal specimens were collected from an individual once in
summer and once in winter and analyzed for D content. U absorption can be esti-
mated assuming that it is not lost in sweat. The estimated U absorption for the
summer sample, when about 3 liters of water was consumed daily, was urinary
0/(urinary 0 + fecal 0) = 2.25 yg 0/(2.25 + 56.9) yg 0 ¦ 3.8%. For the winter
sample, when water intake was about 1 liter/day, 0 absorption was 0.18 yg U/(0.18
+ 31.2) yg 0 = 0.57%. The mean of the two estimates is 2.2% U absorbed (So80).
The high U content of the prepared vegetarian diet, about 20 yg 0/kg, is not sur-
prising considering the high 0 content of the local water. For an individual with
a body size similar to Japanese Reference Man (Ta79), the nean daily intake is
0.76 yg U/kg.
-------
A-9
The results of the four estimates of U absorption in man obtained from
measurement of environmental 0 in ingesta and excreta are collected in Table A-5.
IV. Estimation of U absorption by comparison with Ca (Observed Ratio method)
The observed ratio (OR) was originally defined to predict the behavior of a
90
metal (M), such as Sr, with respect to its well-studied essential analogue,
calcium (Co60), as follows:
CR(body/diet) = M(body)/Ca(body) * M(diet)/Ca(diet).
The U in water, plant products, and mammalian tissues is presumed to be in
the stable +6 state as (I>c49; Lat52; A160; Ru66; Ya68). ion can (1)
enter all domains of bone water, e.g., canaliculi, (2) participate in ion
exchange processes at mineralized bone surfaces both by replacing Ca and subse-
quently being replaced by Ca, and (3) when accumulated at growth sites, be buried
by formation of new bone (Ne51j Tan51; Ne58; Ro68; ?r82; Ste80).
When U is administered to animals or to man, most of the D present in the
body after the initial phase of plasma clearance and urinary excretion is in the
skeleton (Yu73; Du75; Ste80), and the major fraction of U acquired by people from
the environment is in the skeleton (see references in Tables B-1 and B-2). The
environmentally acquired D present in the human skeleton appears to be fairly
uniformly distributed in bone mineral; however, there may be some accumulation of
U on bone surfaces in adult life (Ha71). Although the chemistry and kinetics of
Ca and UC>2++ in bone are not exactly alike, is similar enough to Ca to
warrant application of the OR method to obtain an independent estimate of the
absorption of environmentally acquired 0.
The OR(body/diet) is the resultant of several processes in which the foreign
metal is not handled as efficiently as Ca, i.e., there is discrimination against
the foreign metal. The most important discriminations against tf02++ occur in the
kidneys, where Ca is efficiently reabsorbed; in the intestine, where to the de-
gree that metabolic needs are met, Ca is actively absorbed; and in the skeleton,
where U02++ is limited to the hydration shell of bone mineral crystals (Ne58).
The following relationships apply:
OR(body/diet) ® DF(absorption) x DF(excretion),
DF(absorption) = fraction of metal absorbed/fraction of Ca absorbed,
DF (excretion) * fraction of metal retained/fraction of Ca retained.
-------
A-10
Table A-6 contains the data required to calculate OR{body/diet) for C and the
discrimination factors needed to estimate intestinal absorption of D.
For U, OR(body/diet) = (38 x 10"6g/'000 g)/(1.75 * 10~6g/day/1.1 g/day)
» 0.024.
Not surprisingly, that low value OR for U is the same as has been calculated for
Ra (UN77, see text).
DP(excretion to 10 days) ¦ 20/73 ¦ 0.27.
U absorption (%) = OR(body/diet)/OF(excretion to 10 days) x % Ca absorbed
= 0.024/0.27 x 34 => 3.C%.
The intestinal absorption of U calculated from the "best" values for body D using
the OR method is thus 3.%, less than one-half the value of 7.7% calculated from
U.S., urine and diet data.
If the procedure outlined above is reversed, and it is assumed that the U
analyses of U.S. diet and urine are the more reliable, the calculated U body
content is about 100 yg, providing support for the higher reported values for D
in bone.
V. U absorption and intake level
The mean 0 absorption determined in 13 experiments with rats (Table A-2) was
0.078% ± 0.083, and if the least reliable result is omitted (Eam48), the mean for
12 rat studies is 0.055% ± 0.016. The mean U absorption for 6 experimental
studies of U absorption in man and animals other than rats (Table A-2) combined
with the four estimates of 0 absorption obtained from environmental data (Table
A-5) is 1.86% ± 2.16. If the least reliable result, that based on U.S. dietary
survey data (We60, We67) is omitted, the mean of nine experimental results is
1.2% ± 0.57. The difference between the mean U absorption determined for rats
and that for all other animals is statistically significant (t-test, p < 0.01),
regardless of whether the least reliable point in each group is omitted, indi-
cating that the rats constitute a separate population, which should not be used
in predicting U absorption for man.
Linear regression analysis was used to test the two data sets (rats and all
other animals) for dependence of D absorption on D intake level. The regression
equations and correlation coefficients of linear, semilogarithmic and logarithmic
-------
A-1 1
fits are shown in Table A-7. The logarithmic fit, selected for display in Pig.
A-1 because all the data can be shown together, also provides the best fit to
each set of data as indicated by the correlation coefficients. The slopes of the
fitted lines for the two sets of U absorption data are probably not different
from each other. An inverse dependence of the GI absorption of U on 0 intake is
suggested by these data, but is not proven because the scatter of the data does
not permit rejection of zero slope for either regression equation. All of the
available data for man and animals other than rats could be fitted by a single
line. The large displacement from the fitted line of the environmental point
based on U.S. dietary survey data suggests that additional studies are needed to
determine the shape of the U absorption curve in the region of very low intake.
Calcium is absorbed from the intestine by a specific active transport
mechanism. For Ca intakes below that required to meet metabolic needs, frac-
tional absorption is inversely dependent on intake, and thereafter fractional Ca
absorption is nearly constant. GI absorption of the other alkaline earths also
depends on dietary Ca levels, and they are presumed to be absorbed to some degree
via the Ca transport system. The relationships between U absorption and levels
of dietary Ca have not been studied, but the divalency and small ionic radius of
the uranyl ion (Lat52, Sh76) may allow U02*^ to be absorbed via the Ca transport
system.
Some dependence of 0 absorption on intake is anticipated from chemical
considerations. At very low levels of U intake, absorption should be maximal
because of (a) the availability of natural complexing agents, e.g., citrate, to
+2
stabilize UO2 against reduction and precipitation and facilitate absorption,
and (b) the therroodynamically favorable (mass law) condition of a low 0 concen-
tration. Exhaustion of complexing agents in the intestinal contents and forma-
tion of insoluble diuranates and phosphates in the neutral to alkaline intestinal
contents (Lat52) would act to reduce U absorption at high intake levels.
It is obvious that the point giving 7.7% absorption based on dietary survey
data and fluorometric analysis of urine greatly influences the predicted D
absorption at environmental levels. Curve 1 of Figure A-1, which includes that
point and the other experimental data for animals other than rats, predicts 3.4%
of 0 absorbed at an intake of 2 x 10~2 ug U/kg/day. Elimination of that point
from the data set yields a predicted absorption of 2.1% of U at that intake level
-------
A-12
(curve 2). In the absence of the resolution of some of the uncertainties, we
consider that for the purposes of this report, and over the range of U intake of
concern in setting drinking water standards (up to 2 ug/kg/day, 1.4 1/day of
water containing about 35 pCi U/l consumed by Reference Kan), it is reasonable to
use a rounded-off value of the mean that was calculated for nan and animals other
than rats, 0 absorption ¦ 2%. Our best estimate of absorption at 1 pg/kg/day U
intake is 1.4%. None of the available experimental or environmental data support
a fractional U absorption greater than about 5%, even at intakes of the order of
1 to 2 jjg/day for Reference Man. A higher value for U absorption (about 20%)
based on dietary U data from the U.K. (Ha72) and unpublished analyses of U in
urine (M.H. Dean, quoted by Hu73) seems unlikely on physiological grounds,
because it approaches the fractional absorption of Ca and Sr (ICRP7 3).
-------
Appendix B: Distribution of U in the body
Patricia W. Durbin and McDonald E. Wrenn
Distribution of 0 has been studied after administration to animals by
injection, feeding or inhalation and to people by injection (see references in
Hu73, Yu73, Du75, and Ma53, Sto51, 3er57, Li81, Harr81, Ste30, SuSOab, Su82,
Su83, Pr82). Bone is the major site of 0 accumulation in all the mammals
studied, whether the U exposure was a single or prolonged experimental admini-
stration or by intake from the environment. The relationship between the 0
content of the human body and the levels of 0 in the environment (food, water,
and air) is not clearly established {3e64, Bu73, Sp73, No62); however, continuous
0 feeding or inhalation in animals demonstrate that the amount of 0 in bone, and
to a lesser degree in soft tissues, is directly proportional to the U exposure
level. Equilibrium 0 concentrations were rapidly established in all the tissues
of growing animals and in the soft tissues of animals that were skeletally mature
at the start of exposure (Ma53, Sto5l, Le70). Prom the animal experiments, it
can be inferred that nearly constant D concentrations should be found in people
exposed to a constant level of environmental U, that U turnover in soft tissues
is fairly rapid, and that turnover of some fraction of 0 in bone is slow.
The chief difference between the distributions of administered U in people
(3er57) and larger animals (dog, Tan51; monkey, Le70; baboon, Li81) and those in
small animals (mainly rodents? Yu73, Du75) is that for the first few days to
weeks a larger fraction of the U is present in the soft tissues of the people and
larger animals. For that reason, only the U studies in the large animals appear
to be suitable for assessing the validity of the measurements of the distribution
of environmental U in man. Urinary excretion of U is rapid but not highly effi-
cient, and the significant U content of the soft tissues must be taken into
account in the development of a model that accurately describes the metabolism of
environmentally acquired 0.
I. Bone
Several sensitive analytical techniques have been applied in the U.S. and
elsewhere to measure the level of U in the skeletons of persons with no known
occupational U exposure, and the results are collected in Table 3-1. [Note:
-------
3-2
Other reports have appeared of "normal" U in the human skeleton and tissues with
results that are ten or more times higher than those shown in Table B-1. How-
ever, the analytical methods used were not as specific for U nor as sensitive as
these used in the studies tabulated, and more importantly, most of those other
measurements were done in laboratories with high potential for contamination of
environmental samples (Ca75, Bu58, Qu53; see also references in Hu73).
Consequently, those results have not been included.] Nine sets of analyses for U
in human bone specimens from several U.S. locations have been reported. The
number of individual skeletons sampled is not available in all cases, nor can the
associated alimentary and air intakes of 0 be assessed. The reported average
values (recalculated to a common basis of ng U/g bone ash) range from 2.3 to 61.6
g U in the skeleton of Reference Man. The best value to use at the present time
appears to be the arithmetic mean (± S.D.) of the nine studies, 24.9 ± 22 ;g U,
in the 5000 g of bone (2800 g of bone ash) of Reference Man skeleton. That
choice is supported by the mean of the five skeletal U values reported for other
places, 29.2 ± 20.2 yg D in the Reference Man skeleton.
II. Soft Tissues
The amount and distribution of D normally present in human soft tissues is
even more difficult to assess from the available information. There are fewer
reported analyses (see Table B-2); some important tissues have not been analyzed,
(skin, GI tract); some tissues have been analyzed by only one investigator in one
location; and fat and skeletal muscle, which appear from the injection experi-
ments to contain most of the V in soft tissues, have been analyzed only once in
samples obtained outside the O.S. Consequently, it has been necessary to make
some judgments, based on the results from the human U injection cases and from
experiments in large animals.
Lung: The reported U concentrations in normal lung are 8 to 10 times higher
than would be expected from the human and animal injection studies (Ser57,Tar.51,
Ste30, Li81). The unexpectedly high lung 0 content, and in addition the very
high 0 concentrations in the tracheobronchial lymph nodes (T3LN), indicate that
much of the V in the lungs (we shall assume 85%) was acquired by inhalation of
insoluble material (StoSI, Le70).
-------
3-3
Fat; Only one analyst reported on D in human fatty tissue, and then only
for two specimens (whether those were from different individuals was not
stated). That U concentration, 0.6 ng U/g wet tissue, is 1.5 times greater than
the values reported by that and other analysts for spleen and kidneys, and two to
three times greater than the reported normal U concentrations in liver, muscle
and heart. In the human U-injection study, the concentration in fat ranked
seventh or eighth of the eight tissues that contributed roost to the total U in
soft tissues (apart from kidney) (Ber57). The U concentration in fat was not
more than two times that of skeletal muscle. The 0 concentrations in fat of one
baboon 4 days after U injection was low; it was 2.9 times that of muscle, and
less than one-tenth that of liver and spleen (Li81). In the other animal
studies, U concentration in fat was not reported, but bone, urine and other soft
tissues approached 10C% of the administered U, so there was little D left over
that could have been present in fat. On balance, it does not seem likely that
fat contains 50% of the total U normally present in human soft tissues. As an
alternative, we have elected to use the mean U concentration in heart and
skeletal muscle as representative of the 32 kg of soft tissues and fat otherwise
unaccounted for, and we estimate the 0 content of all of those tissues of
Reference Man to be 5.6 yg.
Blood: The mean of the three repotted sets of U assays of human whole
blood, 0.46 ng U/ml, would lead to a calculated total of 2.4 yg of U in the blood
volume of Reference Man. At an intake of 1.75 yg O/day, and an absorption of 2%,
the blood volume would contain all of the U absorbed in 69 days. Considering the
rapid efflux of parenterally administered u from plasma to extracellular fluid,
bone and urine (Tan51, Ber57, Ro68, Li81, Ste80), the values reported for U in
blood seem unduly high, unless a significant fraction of blood uranium is
associated with cells. Analysis of D in separated normal human plasma and red
cells (Lu70) and of the separated blood constituents in C-injected baboons (Li81)
indicates that within a few hours after intake most of the D present in blood is
associated with cells, "fte value assigned to blood in Table B-3 has been
calculated to include the reported approximate partitioning of blood U between
the plasma and cells of the baboon, assuming a venous hematocrit of 0.45.
-------
Total soft tissue U; The value we have selected to represent soft tissue U,
13 yg, therefore excludes the T3LN and 85% of the measured lung 0 content, uses
the mean reported U concentrations in heart and skeletal muscle as representative
of the non-parenchymous soft tissues, and excludes the one-half of the blood
volume that remains in organs obtained in a routine autopsy.
The "best" value for the normal U content of Reference Kan, 38 ;ig 0, of
which 66% is in the skeleton, is in reasonably good agreement with the
distribution of U experimentally administered to animals (Du75, SteSO) and human
subjects (Ber57).
-------
Appendix C: Metabolic Models for Uptake of Uranium in the Human Kidney and
Skeleton
Kidney
For a two-cornpartroent model of kidney, as proposed by ICRP (ICRP79), the
equilibrium concentration at constant daily intake (I) is as follows:
A f,
—— = ¦ f T + f T 1
Ira ni (ln2) ^ 21 1 r22 2J
where
A
CD
S
amount of U in the kidney at equilibrium
m
s
kidney mass (310 g)
fi
at
fractional GI absorption from GI tract to blood
f21
s
fractional transfer from blood to kidney compartment 1
f 22
3m
fractional transfer from blood to kidney compartment 2
T1 (2)
s
half-times in kidney compartments 1 or 2
I
s
daily intake (wg).
Then
C A
CD OD
T~ = Im
is the concentration (C) in ug U per g kidney per yg./day intake. This value is
listed in Table C-1 using the most recent ICRP model for U (1CRP79), and the con-
tinuous intake model used earlier by ICRP (ICRP59) and discussed by Spoor and
Harsh (Sp73). In the table, f^ £S varied from 1 to 1C% to show the effect of GI
-------
C-2
absorption on eventual kidney content. The measurements made after intravenous
administration of U to comatose patients have been summarized (Hu73), and based
on these, a blood-to-kidr.ey transfer factor of 11% is assumed, with a 15 day
half-time in the kidney.
If CL is the limiting concentration in kidney, and S is the additional
safety factor desired, then the limiting daily intake (IL) can be derived as
lCl/S)
XL * "(A JZ) '
CL has been generally accepted (Sp73) as about 3 Wg/g kidney, but more recent
data in dogs (Mo82) suggest that a value of 0.6 wg/g is below the injury thres-
hold in man. We will use 1 wg/g to derive a limit; Ais taken from the last
column in Table C-1. Por the same value of fj, the Spoor and 3ursh model gives
nearly the same equilibrium uptake in kidney as the ICRP model, exceeding that
estimate by only 15%. We will use the simpler Spoor and Hursh model here.
We have used 1.4% as the best available GI absorption estimate for man at
environmental levels of D intake and introduced a safety factor (S) of 50. The
size of the safety factor depends on our certainty about the potential toxicity
of U in man: safety factors have been called uncertainty factors by NAS (NAS77).
We have some information about U uptake, absorption and metabolism in man, and a
wealth of animal toxicology data, even though pieces of information that would
improve the reliability of our estimates are missing.
The committee believes that based on the NAS definition, U should be
assigned an uncertainty (safety) factor between 10 and 100 (NAS77, p. 804). We
have chosen 50 as a factor that should provide a high margin of safety. One
might argue that we have actually introduced a safety factor of 150, because we
-------
have also chosen a more conservative estimate (1 ug/g) for the limiting concen-
tration of U in the kidney than was used in the past (3 yg/g).
We note that ICRP (ICRP64) derived a limit for occupational exposure to D in
water of 2 x 104 pCi/1, equivalent to 6.1 x 104 pg/1 for 750 ml/day intake
4
of water (Hu73, p. 253), or an intake by mouth of 4.6 x 10 ;jg/day. For indivi-
duals in the general population, the ICRP recommendations were equivalent to 1800
g/1 in water (Hu73, p. 263). Inasmuch as the ICRP had inadvertently underesti-
mated GI absorption by about a factor of 10, this would be equivalent to 180 ug/1
for 1% GI absorption.
We obtain A Jm « 1 x 10~* and IL * (1/50J/10.7 x 10"^, or 200 -^/day. For
1.7 1 of water intake/day for a reference individual (NAS77) and a safety factor
of 50, this is equivalent to 120 The committee recommends 100 yg/1
»
(rounded off) as the intake limit for U in drinking water. It is designed to
limit toxic effects in the kidney from chronic U intake. This U concentration
has a radioactivity of 67 pCi/1 of alpha-particles, if the naturally-occurring 0
isotopes are present in their normal relative abundances.
Skeleton
The equilibrium level of 0 in the human skeleton resulting from a constant
and continuous daily intake at 1 unit per day in food and/or water is given by
A^/I in the same formula as for kidney uptake, but with different numerical
values of the coefficients. The general form of the ICRP model for 0 in bone
will be retained here, namely two compartments for bone, one with a short half
time (we choose 300 d instead of the 20 d chosen by ICRP), and the other with a
long half time (5000 d, the same as the value chosen by ICRP).
-------
Use of the coefficients for 0 uptake and fractional transfer from blood
chosen by the ICR? for workers (f1 ¦ 0.05, f21 a 0.2, f22 ¦ 0.23) leads to
equivalent days accumulation in skeleton of - 8.3 days. Changing the half time
for short-term bone to 300 d leads to an estimate of 12.5 days equivalent uptake.
This appears to be on the low side of the various estimates of skeletal accumu-
lation (see Appendix A), even though the value for f^ used in the model is also
higher than our best estimate of GI uptake (1.4%).
One may vary the coefficients in the model within ranges which are reason-
able and consistent with animal experiments and human data (Du75). The animal
data all point to a blood to skeleton transfer from 10 to 2C% (f2-j + ^22'* For
example if we assume f, * 0.01, f21 + f22 = 0.2, blood to short-term bone trans-
fer of 0.18 and to long-term bone, 0.02, the model predicts 1.5 days equivalent
accumulation, which would correspond to levels roughly equivalent to those
reported by Fisenne (Fi80). At the other extreme using f ¦ 0.05, and equal
partitioning between long- and short-term bone, the equation predicts 26.5 days
equivalent intake in bone.
We believe that estimating days equivalent intake from various analyses of
human bone samples, coupled with dietary measurements of a general nature, pro-
duces an estimate of days equivalent intake with a large potential error. If we
use our best estimate of human GI absorption the model predicts a skeletal con-
tent at the lower range of the observed values.
To estimate potential radiation risk to skeleton we will assume an equilib-
rium with 11 days equivalent intake in the skeleton as our best estimate (see
Appendix A for the derivation).
Higher estimates are not excluded. For example, the ICRP metabolic data for
Reference Man predicts 59 'rig U in skeleton from a daily intake of 1.9 wgU/day or
-------
31 days equivalent intake, while the metabolic model ICRP adopted gives 8 days
+ 2
(ICRP79). On grounds of comparative chemistry and skeletal biochemistry, UC>2
+2
should be less tightly held in bone than Ra , and one would predict a shorter
net residence time and smaller value for days equivalent intake for D than for
Ra.
Cothern (Co83b) has obtained 33 using the ICRP model modified for 2C% GI
absorption. The 11-day equivalent accumulation in skeleton (range 2 to 33 days)
cannot be said to differ significantly from the 8-day ICRP prediction, using the
committee*s consensus for the best values of the aetabolic parameters. In order
to determine the reasons for the difference, skeletal U burden data are required
in a population with 0 intakes that have been well-characterized over decades.
226
Data are available for Ra; however, for practical reasons, U data are not
obtainable directly from living man, but must be inferred from bone samples
obtained at autopsy.
-------
R-1
REFERENCES
Ad74 Adams N. and Spoor N.L. , 1974, "Kidney and bone retention functions in
the human metabolism of uranium," Fhys. Med. Biol. 19: 460-471.
A160 Alexander L.T., Hardy E.P. Jr., and Hollister, H.L., I960, "Radio-
isotopes in soils: particularly with reference to strontium in:
Radioisotopes in the 3iosphere, Proc. of a conference, Minneapolis,
Minnesota, October 1959, (R.S. Caldecott and L.A. Snyder, Eds.) pp. 3-
22 (Minneapolis: Univ. of Minnesota Press).
Ar72 Arnold J.S. and Wei C.T., 1972, "Quantitative morphology of vertebral
trabecular bone," in Radiobiology of Plutonium, (B.J. Stover and
W.S.S. Jee, Eds.), pp. 333-354, (Salt Lake City: J.W. Press, Dept. of
Anatomy, University of Utah).
5a48 Bassett S.H., Frenkel A., Cedars N., Van Alstfine H., Waterhouse C.,
and Cusson K., 1948, The excretion of hexavalent uranium following
intravenous administration. II. Studies on human subjects,
University of Rochester report CR-37, (Rochester, NY: University of
Rochester).
Bal8 2
Ballou J.E., 1983, "Toxicity of thorium cycle nuclides," Pacific
Northwest Laboratory Annual Report for 1982, Pacific Northwest
Laboratory, Richland, WA, PNL-4600, pt.1, pp. 83-85.
-------
3-2
3e64 Berdnikova A.V., 1964, "On the uranium content in the external
environment and in human excreta," Vopr. Pi taniya 23: 17-20 [In
Russian, cited by Hu73].
3er57 Bernard S.R., and Struxness B.G., 1957, A study of the distribution
and excretion of uranium in can: An interim report, Oak Ridge National
Laboratory, Oak Ridge, TN, ORNL-23C-4.
Br80 Broadway J.3. and Strong A.B., 1980, Radionuclides in human bone with
estimates of population dose and health impact, U.S. Environmental
Protection Agency, Washington, DC, EPA520/5-8C-006.
Br83 Broadway J.H., and Strong A.B., 1983, "Radionuclides in human bone
samples," Health Phys. 45: 765-768.
Bu58 Butterworth A., 1958, "3aman data on uranium exposure," in Symposium
on Occupational Health Experience and Practices in the Uranium
Industry, Proc. of a conference, Sew York, Sew York, Oct. 1958,
D.S.A.E.C. Health and Safety Laboratory report HASL-58, pp. 41-46.
Ca75 Campbell E.E., Mclnroy J.P. arid Schulte H.P., 1975, "Uranium in the
tissue of occupationally-exposed workers," in Conference on Occu-
pational Health Experience with Uranium, (M.E. Wrenn, Ed.), Arlington,
VA, April 1975, D.S. Energy Research and Development Administration,
Washington, DC, ERDA-93, pp. 323-350.
-------
R-3
C06O Comar C.L. , and Wasserxan R.H., I960, "Radioisotope absorption and
methods of elimination: Differential behavior of substances in zeta-
bolic pathways," in Radioisotopes in the Biosphere, Proc. of a con-
ference, Minneapolis, Minnesota, October '959, (R.S. Caldecott and
L.A. Snyder, eds.), (Minneapolis; Univ. of Minnesota Press), pp. 526-
540.
Co83a Cothern C.R., and I^ppenbj sch, W.L., 1933, "Occurrence of uranium in
drinking water in the O.S.," Health Physics 45; 89-100.
Co83b Cothern C.R., Lappenbusch W.L., and Cotruvo J.A., 1983, "Health
effects guidance for uranium in drinking water," Health Physics 44
(supp.1): 377-384.
Do49 Dounce A.L., 1949, "The mechanism of action of uranium compounds in
the animal body," in The Pharmacology and Toxicology of Uranium
Compounds, (C. Voegtlin and H.C. Hodge, Sds.), National Nuclear Energy
Series Div. IV, Vol. 1., (Sew York: McGraw-Hill Book Co.), pp. 951-
991.
Du75 Durbin P.W. and Wrenn M.E., 1975, "Metabolism and effects of uranium
in animals," in Occupational Health Experience with Uranium, Proc. of
a conference, Arlington, VA, April 1975, (M.E. Wrenn, Ed.), U.S.
Energy Research and Development Administration, Washington, DC, ERDA-
93, pp. 67-129.
-------
R-4
Ed67 Edgington D.N. , 1967, "The estimation of thorium and uranium at the
submicrogram level in bone by neutron activation," Int. J. Appl.
Radiat. Isotopes 18s 11-18.
Ev66 Evans R.D., 1966, "The effect of skeletally deposited alpha-ray
emitters in roan," Brit. J. Radiol. 39: 881-895.
Fil78 Filippova L.G., Nifatov A.P., and Lyubchanskii E.R., 1978, "Some of
the long-term sequelae of giving rats enriched uranium," Radiobio-
logiya 18:400-405. English translation, Jane 1979, D.S. Department of
Energy, Washington, DC, DCE-tr-4/9, pp. 94-100.
Fin53 Fir.kel M.P., 1953, "Relative biological effectiveness of radium and
other alpha emitters in CF No.1 female mice," Proc. Soc. Exptl. Biol.
Med. 83: 494-498.
Fi79 Fisenne I.M., and Keller K.W., 1979, "The world-wide distribution of
radi^a-226 in ashed human bone," in Environmental Measurements
Laboratory Report, EML-356, pp. 1-47 - 1-58.
Fi80 Fisenne I.M., Perry P.M., and Welford G.A., 1990, "Determination of
uranium isotopes in human bone ash," Anal. Chem.: 777-779.
Fi83 Fisenne I.M., Perry, P.M., Chu, N.Y., and Harley, N.H. 1983.
Measured 234,238^ an(j fanout 239,240pu ^-oman bone ash from Nepal
and Australia: Skeletal alpha-dose. Sealth ?hys. 44 (supp. 1): 457-
467.
-------
?is60 Pish B.R., Payne J.A. and Thompson J.L., 1960, "Ingestion of uranium
confounds," in Oak Ridge National Laboratory Sealth Physics Division
Annual Report, Oak Ridge National Laboratory, Oak Ridge, TN, ORNL-
2994, pp. 269-272.
Fish83 Fisher D.R., Jackson P.O., Brodaczyr.ski G.G. and Scherpelz R.I., 1983,
^i 2 ^ ft a
"Levels of 0, 0, and Th in excreta of uranism mill crusher-
men," Health Phys. 45: 617-630.
Ga70 Ganguly A.K., 1970, "Assessment of internal radioactive contamination:
The present trends in the Indian programme. The detection and
assessment of uranium and plutoniura in the whole body and in the
critical organs," (Vienna: International Atomic Energy Agency), Report
118.
Go64a Gong J.K., Arnold J.S., and Cohn S.H., 1964, "The density of organic
and volatile and non-volatile inorganic components of bone," Ar.at.
Rec. 149: 319-324.
Go64b Gong J.K., Arnold J.S., and Cohn S.H., 1964, "Composition of
trabecular and cortical bone," Ar.at. Rec. 149: 319-324.
Ha71
Hamilton E.I., 1971, "The concentration and distribution of uranium in
human skeletal tissues," Calcif. Tiss. Res. 7: 150-162.
-------
R-6
Ha72 Hamilton E.I., '972, "The concentration of uranium in man and his
diet," Health Phys. 22: 149-153.
Ham48 Hamilton J.G., 1948, The metabolic properties of the fission prodjcts
and actinide elements, Rev, Mod. Phys. 20: 718-728.
Har74 Harley J.H., 1974, Concentration of dose factors for internally
deposited natural radionuclides, New York City, USAEC Health and
Safety Laboratory, EASL TM 74-2.
Harr81 Harrison J.D., and Stather J.W., 1981, "The gastrointestinal
absorption of protoactinium, uranium and neptunium in the hamster,"
Radiation Res. 88: 47-55.
HAS58 Health and Safety Laboratory, 1958, Symposium on the Occupational
Health Experience and Practices in the Uranium Industry, (Sew York
City: OSAEC Division of Biology and Medicine, Health and Safety
Laboratory), HASL-58.
Hav49 Haven P.L., and Hodge H.C., 1949, "Toxicity following the parenteral
adminstration of certain soluble uranium compounds," in The Pharmaco-
logy and Toxicology of Uranium Compounds, (C. Voegtlin and H.C. Hodge,
Eds.), National Nuclear Energy Series Div. IV, Vol. 1, (New York:
McGraw-Hill Book Co.), pp. 281-308.
-------
R-7
He64a Heaney R.P., and Skillman T.G., 1964, "Secretion and excretion of
calcium by the human gastrointestinal tract#" J. Lab. Clin. Med. 64:
29-41.
He64b Heaney R.P., Bauer G.C.H., Bronner P., Dymling J., Lafferty F.W.,
Nordin B.E.C. and Rich C., 1964, "A normal reference standard for
radiocalcium turnover and excretion in humans," J. Lab. Clin. Med. 64:
21-28.
Bo63 Boltzman R.B., 1963, "Measurement of the natural contents of RaD
(Pb^®) and RaF (Po^1^) in human bone: Estimates of whole-body
burdens," Health Phvs. 9: 385-400.
Hod73a Hodge H.C., Stannard J.N., and Hursh J.B. (Eds.), 1973, Cranium,
Plutonium, Transplutonic Elements, Handbook of Experimental Pharma-
cology, Vol. 36, (Berlin: Springer-Verlag).
Hod73b Hodge H.C., 1973, "A history of uranium poisoning (1824-1942)," in
Uranium, Plutonium, Transplutonic Elements, Handbook of Experimental
Pharmacology, (H.C. Hodge, J.N. Stannard, and J.B. Hursh, eds.), Vol.
36, (Berlin: Springer-Verlag), pp. 5-68.
Hue52
Hueper W.C., Zuefle J.H., Link A.M., and Johnson M.G., 1952,
"Experimental studies in metal carcinogenesis. II. Experimental
uranium cancers in rats," J. National Cancer Institute 13: 291-301.
-------
R-8
Bu68 Hursh J.B., and Sjomela J., '.968, "Absorption of Pb from the
gastrointestinal tract of roan," Acta Radiologica 7: 1C8-120.
Hu69 Harsh J.B., Neaman W.R., Toribara T. , Wilson H. , and Waterhouse C.,
1969, "Oral ingestion of uranium by nan," Health ?hys. 17: 619-621.
Bu73 Hursh J.B., and Spoor N.L., 1973, "Uranium data on man," in Uranium,
Plutonium, Transplutonic Elements: Handbook of Experimental
Pharmacology, (B.C. Hodge, J.N. Stanr.ard, and J.B. Hursh, eds.),
Volume 36, (Berlin: Springer Verlag), pp. 197-239.
ICRF59 International Commission on Radiological Protection, 1959, Permissible
Dose for Internal Radiation, Publication 2, (Oxford: Pergamon Press).
ICRP64 International Commission on Radiological Protection, 1964,
Recommendations of the International Commission on Radiological
Protection (As Amended 1959 and Revised 1962), Publication 6, (Oxford:,
Pergamon Press).
ICRP67 International Commission on Radiological Protection, 1967, Evaluation
of Radiation Doses to Body Tissues from Internal Contamination Due to
Occupational Exposure, Publication 10, (Oxford: Pergamon Press).
ICRP73
International Commission on Radiological Protection, 1973, Alkaline
Earth Metabolism in Adult Man, Publication 20, (Oxford: Pergamon
Press). Also Health Physics 24: 125-221.
-------
R-9
ICRP74 International Commission on Radiological Protection, 1974, Report of
the Task Group on Reference Han, (W.S. Snyder, Chairman), Publication
23, (Oxford:Perganion Press).
ICRP77a International Commission on Radiological Protection, 1977, Radiation
Protection: Recommendations of the International Commission on
Radiological Protection, ICRP Publication 26. (Oxford: Percamon
Press).
ICRP77b International Commission on Radiological Protection, 1977, Probiems
Involved in Developing an Index of Harm, Publication 27, (Oxford:
Peigamon Press).
ICRP79 International Commission on Radiological Protection, 1979, Part I:
Limits for Intakes of Radionuclides by Workers, Publication 30,
(Oxford: Pergamon Press), Supplement Part 1, pp. 102-104.
La52 Latimer W.M., 1952, Oxidation Potentials, 2nd ed. (New York: Prentice-
Hall) .
La82 Lai N., Sharma Y.P., Sharma P.K., Talvar I.M., Nagpaul K.K., and
Chakarvarti S.K., 1982, "Uranium assay in milk,* Health Phys. 43: 425-
428.
-------
R-10
Lar84 Larsen R.P., Bhattacbarrya M.H., Oldham R.D., Morretti E.S. and Cohen
N., 1984, "Gastrointestinal absorption and retention of plutonium and
uranium in the baboon," Argonne National Laboratory Environmental
Research Division Annual Report, July 1982-June 1983, Aroon.ne National
LAboratory, Argonne, IL, ANL-83-100, pp. 51-60.
Le70 Leach L.J., Maynard E.A., Hodge H.C., Scott J.K., Yuile C.L.,
Sylvester G.E., and Wilson H.B., 1970, "A five-year inhalation study
with natural uranium dioxide (002) dust: I. Retention and biologic
effect in the monkey, dog, and rat," Health Phys. 18: 599-612.
Le73 Leach L.J., Yuile C.L., Hodge H.C., Sylvester G.E., and Wilson H.B.,
1973, "A five-year inhalation study with natural uranium dioxide (00^)
dust. II. Postexposure retention and biologic effects in the .monkey,
dog, and rat," Health Phys. 25: 239-258.
Li81 Lipsztein J.L., 1981, An improved model for uranium metabolism in the
primate, Doctoral Thesis, Environmental Health Sciences, New York
University, New York, NY.
Lu7 0 Lucas H.P. Jr., and Markun F., 1970, "Thorium and uranium in blood,
urine, and cigarettes," Argonne National Laboratory Radiological
Physics Division Annual Report, Argonne National Laboratory, ARgonr.e,
IL, ANL-7760, Part 2, pp. 47-52.
-------
R-11
Mas71a Masuda K., 1971, "Intake and urinary excretion of uranium in nor.-occa-
pationally exposed persons: I. Recovery of uranium in the daily diet
and urine," Japanese J. Hyg. 26: 434-437.
Mas71b Masuda K., 1971, "Intake and urinary excretion of uranium in non-
occupatior.ally exposed persons: II. Uranium in the daily diet,"
Japanese J. Hyg. 26: 438-441.
Mas71c Masuda K., 1971, "Intake and urinary excretion of uranium in non-
occupational ly exposed persons: III. Urinary excretion of uranium,"
Japanese J. Hyg. 26:442-446.
Mas71d Masuda K., 1971, "Intake and urinary excretion of uranium in non-
occupationally exposed persons: IV. Discussions on dietary intake and
urinary excretion of uranium," Japanese J. Hyg. 26: 447-450.
Ma49 Maynard E.A., and Hodge H.C., 1949, "Studies of toxicity of various
uranium compounds when fed to experimental animals," in The
Pharmacology and Toxiciology of Uranium Compounds, (C. Voegtlin and
H.C. Hodge, Eds.), National Nuclear Energy Series Div. IV, Vol. 1,
parts 1 & 2, (New York: McGraw-Hill 3ook Co.), pp. 3C9-376.
Ma53 Maynard E.A., Downs W.L. and Hodge H.C., 1953, "Oral toxicity of
uranium co^>ounds," in The Pharmacology and Toxicology of Uranium
Compounds, (C. Voegtlin and H.C. Hodge, Sds.), National Nuclear Energy
Series Div. VI, Vol. 1, parts 3 6 4, (New York: McGraw-Hill Book Co.),
pp. 1221-1369.
-------
R-12
Kay61 Hayneord M.V. , 1961, "The natural radioactivity of the hurr.an body,"
in Progress in 3iophysics and Biophysical Chemistry, Vol. 11, (J.A.V.
Butler, B. Katz, R.E. 2irkle, Eds.), (New York: Pergamon Press), pp.
2-24.
22*
Mays78 Mays C.W., Ed., 1978, Biological Effects of Ra and Thorotrast,
Proceedings of a symposium, Alta, UT, 21-23 July 1974, 3ealth Physics
Volume 35, No. 1.
Mays84 Kays, C.W., Rowland, R.E., Stehney, A.F. 1984. Cancer risk from the
lifetime intake of radium and uranium isotopes. Health Phys. (this
issue).
KcC63 KcClellan R.O., Casey B.W., and Bustad L.K., 1963, "Transfer of some
radionuclides to milk," Hanford Biology Research Annual Report, USAEC
report HW-76000, pp. 98-108.
McI79 Mclnroy J.P., Campbell E.E., Moss W.D., Tietjen G.L., Eutsler B.C.,
and Bcyd H.A., 1979, "Plutonium in autopsy tissue: A revision and
updating of data reported in IA-4875," Health Phys. 37: 1-136.
-------
R-13
Mo82 Morrow P.E., IJeach L.J. , Smith F.A., Gelein R.M., Scott J.B. , Beiter
H.D., Amato F.J., Picano J.J., Yuile C.L., and Consler T.G., 1982,
"Metabolic fate and evaluation of injury in rats and dogs following
exposure to the hydrolysis products of uranium hexafluoride: Impli-
cations for a bioassay program related to potential releases o£
uranium hexafluoride," D.S. Nuclear Regulatory Commission, Washington,
DC, NUREG/CR-2268.
Mos82 Moss M.A. and McCurdy R.F., 1982, "Clinical study of a population
exposed to excessive levels of uranium in drinking water," Presented
at the 51st Annual Meeting of the Royal College of Physicians and
Surgeons of Canada, held in Quebec City, September 1982. Abstract:
Annals R.C.P.S.C. 15:4 (June 1982).
NAS77 National Academy of Sciences, 1977, Drinking Water and Health.
National Academy of Sciences, Washington, DC.
NCRP75 National Council on Radiation Protection and Measurements, 1975,
Natural Radiation Background in the United States, NCR? Publication
45, Washington, DC.
Ne51 Neuirsan W.F., 1951, "Deposition of uranium in bone," in The
Pharmacology and Toxicology of Uranium Compounds, (C. Voegtlin and
H.C. Hodge, Eds.), National Nuclear Energy Series Div. VI, Vol. 1,
(New York: McGraw-Hill Book Co.), pp. 1911-1991.
-------
R-14
Ne58 Seaman W.P., and Seaman M.W., 1958, The Chemical Dynamics of Bone
Mineral, (Chicago: University of Chicago Press).
No62 Novikov Y.V., and Rezanov I.I., 1962, "The content of uranium in
drinking water of soil origin and its excretion in urine and feces,"
Gig i. Sunit. 27: 103-105 [In Russian, cited by Bu73].
No70 Nozaki T., Ichikawa M., Sasuga T., and Inarida M., 1970, "Neutron
activation analysis of uranium in human bone, drinking water and daily
diet,* J. Radioanalytical Chem. 6; 33-40.
Pe65 Penna Franca E.J., Almeida C., Seeker J., Emmerich M., Roser P.X.,
Kegel G., Hainsberger L., Cullen T.L., Petrow H., Drew R.T. and
Eisenbud M., 1965, "Status of investigations in the Brazilian areas of
high natural radioactivity," Health Phys. 11: 699-712.
Pr82 Priest N.D., Bovells G.R., Green D., and Haines J.W., 1982, "Uranium
in bone: Metabolic and autoradiographic studies," Banian Toxicol. 1:
97-114.
Qu58 Quigley J.A., Heatherton R.C., and Ziegler J.P., 1958, "Studies of
human exposure to uranium," in Symposium on Occupational Health
Experience and Practices in the uranium Industry, D.S.AEC Health and
Safety LAboratory, Hew York, HASL-58, pp. 34-46.
-------
R-15
Ra65 Rajewsky B., Belloch-ZiitLT.erman V, , Lohr E., and Stahlhofen W., 1965,
2 26
Ra in human embryonic tissue, relationship of activity to the stage
226
of pregnancy, measurement of natural Ra occurrence in the human
placenta," Health Phys. 11:161-169.
Ro68 Rowland R.2., and Farnham J.2., 1968, "The deposition of uranium in
bor.e," in Radiological Physics Division Annual Report, Argonne
National Laboratory, Argonne, IL, ANL-7489, pp. 33-36.
Ro78 Rowland R.B., Stehney A.P., and Lucas H.P., 1378, "Dose-response
relationships for female radium dial workers," Radiat. Res. 76: 368-
383.
Ro83 Rowland R.B., Stehney A.P., and Lucas H.P., 1983, "Dose-response
relationships for Ra-induced bone sarcomas," Health Physics 44 (supp.
1): 15-31.
RosSO Rcswell R.L. and Wrenn M.E., 1980, "An improved model for uranium
metabolism in man," in Research in Radiobiology; Annual Report of Work
in Progress in the Internal Irradiation Program, Radiobiology
Division, University of Utah, Salt Lake City, COD-119-256, pp. 115-
128.
Ru66
Russell R.S., 1966, Radioactivity and Human Diet, (Oxford: Fergamon
Press).
-------
R-16
Sc73 Schler.ker R.A., and Oltrtan 3.G.f 1973r "Fission track ajtoradio-
graphs," in Argonne National Laboratory Annual Report, Argonne
National Laboratory, Argonne, IL, ANL-8C60, Part II, pp. 163-168.
Sc81 Schlenker R.A., and 01 titan B.G., 1981, "Uranium concentrations in
hu.uan bone," in Actinides in Man and Animals, (M.S. Wre.nn, Ed.), (Salt
Lake City: RD Press), pp. 473-475.
Sc82 Schler.ker R.A., Keane A.T., and Holtzman R.B., 1982, "The retention of
226
Ra in human soft tissue and bone: Implications for the ICR? 20
alkaline earth model," Health Phys. 42: 671-693.
Sh76 Shannon R.D., 1976, "Revised effective ionic radii and systematic
studies of interatomic distances in halides and chalcogenides," Acta
Crystal. A32: 751-767.
Si68 Sikov M.R., and Mahlum D.D., 1968, "Cross-placental transfer of
selected actinides in the rat," Health Phys. 14:205-208.
Sin83 Singh N.P., Lewis L.L., and Krenn, M.E., 1983, "Concentrations of
uranium isotopes in human tissues from 2 states in the United States,"
in Proceedings of the 7th International Congress of Radiation
Research, (J.J. Broerse, et al., Eds.), (Amsterdam: Martinus Nijhoof
Publishers), Paper E5-13.
-------
R-17
Sk80 Skrable K.W., Chabot G.E., French C.S., Wrenn M.E., Lipsztein J.,
LoSasso T. and Durbin P.W., 1980, "Blood-organ transfer kinetics,"
Health ?hys. 39: 139-209.
So80 Somayajulu B.L.K., Meghani C.A., and Borole D.V., 1980, "Uranium in
drinking waters: Rejection by hoitan body," in Management of
Environment, (B. Patel, Ed.), (New Delhi: Wiley Eastern Liirited), pp.
528-532.
Spi68 Spiers F.W., 1968, "Radioactivity of the human body," in Radioisotopes
in the Hj.T,an 3ody Physical and Biological Aspects, (Sew York:
Academic Press), pp. 257-296.
Spi83 Spiers F.W., Lucas H.F., Rundo J., and Anast G.A., 1983, "Leukaemia
incidence in the U.S. dial workers, Health Physics 44 (supp. 1): 65-
72.
Sp73 Spoor N.L., and Harsh J.3., 1973, "Protection criteria," in Handbook
of Experimental Pharmacology, (Hodge H.C., Stannard J.N., and Harsh
J.3.), Vol. 36, (3erlin: Spring-r-Verlag), pp. 241-270.
St64
Stahlhofen W., 1964, "Measurement of the natural content of Th^®,
226
Ra , and their daughters in the human body," in Assessment of
Radioactivity in Wan, Vol. 2, (Vienna: International Atomic Energy
Agency), pp. 505-519.
-------
R-18
St56 Stehney A.P., and Lucas H.?. Jr., 1956, "Studies on the radium content
of humans arising from the natural radium of their envi rorjsent," in
Proc. First Conf. on Peaceful Uses of Atomic Energy, Vol. 11, (Geneva:
International Atomic Snergy Agency), p. 49.
Ste80 Stevens W., Bruenger P.W., Atherton D.R., Smith J.M. and Taylor G.N.,
1980, "The distribution and retention of hexavalent in the
beagle, Rad. Res. 83: 109-126.
Sto51 Stokinger H.E., Baxter R.C., Dygert H.P., Labelle C.W., Laskin S.,
Pozzani O.C., Roberts E., Rogherroel J.J., Rothstein A., Speigl C.J.,
Spragae G.P. Ill, Wilson H.B., and Yaeger R.C., 1951, in The
Pharmocology and Toxicology of Uranium Compounds, (C. Voegtlin and
H.C. Hodge, Eds.), National Nuclear Energy Series, Div. VI, Vol. 1,
(Sew York: McGraw-Hill Book Co.), pp. 1370-1778.
Str55 Struxness, E.G., Luessenhop, A.J., Bernard, S.R., and Galliroore J.C.,
1955, "The distribution and excretion of hexavalent uranium in man,"
Proceedings of International Conference on the Peaceful Uses of Atomic
Eneigy, Vol. 10, (Geneva: International Atomic Energy Agency) pp. 186-
196.
SuSOa
Sullivan M.P., 1980, "Absorption of actinide elements from the
gastrointestinal tract of rats, guinea pigs, and dogs," Health Phys.
38: 159-171.
-------
R-19
Su80b Sullivan M.F., '930, "Absorption of actinide elements from the
gastrointestinal tract of neonatal animals," Health Phys. 38: 173—'65.
Su82 Sullivan M.P., and Gorham L.S., 1982, "Further studies on the
absorption of actinide elements from the gastrointestinal tract of
neonatal animals," Health ?hys. 43: 5C9—519.
Su83 Sullivan M.F., 1983, "Gut-related studies of radionuclide toxicity,"
in Pacific Northwest Laboratory Annual Report for 1982, Pacific
Northwest Laboratory, Richland, WA, PNL-4600, pt. 1, pp. 95-99.
Sw67 Swarm R.L., Ed., 1967, Distribution, Retention and Late Effects of
Thorium Dioxide, Proceedings of a conference, New York, NY, April
1966, Ann. N.Y. Acad. Sci. 145: 523-858.
Ta79 Tanaka G., Kawamura H., and Nakahara Y., 1979, "Reference Japanese
Man: I. Mass of organs and other characteristics of normal Japanese,"
Health Phys. 36: 333-346.
Tan51 Tannenbaum A., Silverstone H., and Koziol J., 1951, "Tracer studies of
the distribution and excretion of uranium in mice, rats, and dogs," in
Toxicology of Uranium, (A. Tannenbaom, Ed.), National Nuclear Energy
Series, Division VI, Vol. 23, (New York: McGraw-Hill 3ook Co., Inc.),
pp. 128-181.
-------
3-20
, 2 **>6
Tr83a Tracy B.L., Prantl F.A., and Quinn J.M., 1983, "Transfer of Ra,
210
?b, and uranium from soil to garden produce: Assessment of risk,"
Health ?hys. 44: 469.
Tr83b Tracy B.L., Quinn J.M., Gilaian A.?., Villenueve D.C., Secours V.B.,
Valli D.E., and Yagminaf A.P., 1983, "The metabolism of ingested
uranium in ir,ancr.als," in Proceedings of the 7th International Congress
of Radiation Research, (J.J. Broerse et al., eds.), (Amsterdam:
Kartinus Nijhoff Publishers), paper E5-18.
UN62 United Nations Scientific Committee on the Effects of Atomic
Radiation, 1962, (New York: United Nations).
UN66 United Nations Scientific Comittee on the Effects of Atomic
Radiation, 1966, (New York: United Nations)
UN6 9 United Nations Scientific Committee on the Effects of Atomic
Radiation, *969, (New York: United Nations)
UN72 United Nations Scientific Committee on the Effects of Atomic
Radiation, 1972, (New York: United Nations)
5JN77 United Nations Scientific Committee on the Effects of Atomic
Radiation, 1977, (New York: United Nations)
-------
R-21
Vc-49 Voegtlin C., and Hodge H.C. (Eds.), 1949, The Pharmacology and
Toxicology of Uranium Compounds, National Nuclear Energy Series, Div.
IV, Vol. 1, Parts 1 & 2, (New York: McGraw-Hill Book Co.).
Vo53 Voegtlin C. and Hodge 3.C. (5ds.), 1953, The Pharmacology and
Toxicology of Uranium Compounds, National Nuclear Energy Series, Div.
VI, Vol. 1, (New York: McGraw-Hill Book Co.).
We60 Welford G.A., Morse R.S., and Alercio J.S., 1960, 'Urinary uranium
levels in non-exposed individuals," Am. Ind. Byg. Assn. 21: 68-70.
We67 Welford G.A. and Baird R., 1967, "Uranium levels in human diet and
biological materials," Health Phys. 13:1321-1324.
We76 Welford G.A., Baird R., and Pisenne I.M., 1976, "Concentrations of
natural uranium in the human body," in Proc. of the Tenth Midyear
Topical Symposium of the Health Physics Society, Rensselaer
Polytechnic Institute.
Wi65 Wing J.F., 1965, "Background urinary uranium levels in humans," Health
Phys. 11: 731-735.
Wr7 5
Wrenn, M.E., Ed., 1975, Conference on Occupational Health Experience
with Uranium. Energy Research and Development Administration, Wash-
ington, DC, ERDA-93.
-------
R- 22
Wr77a Wrer.n M.S., Losasso T., and Durbin P.W., 1977, "A metabolic model for
uranium metabolism in man," in Radioactivity Studies, Progress Report
to the U.S. Energy Research and Development Administration, Institute
of Environmental Medicine, New York University Medical Center, New
York, NY, COO-3382-16, Section XI.
Wr77b Wrenn M.E., 1977, "Internal dose estimates," in International
Symposium on Areas of High Natural Radioactivity, (Rio de Janeiro:
Acad. Brasileira de Ciencias), pp. 131-157.
Wr78 Wrenn M.E., Lipsztein J., LoSasso T., Durbin P., and Skrable K., 1978,
Further Developments on Improved Metabolic Models for Uranium
Metabolism in Man, Institute of Environmental Medicine, New York
University Medical Center, New York, NY, COO-3382-17, EY-76-S-02-3382,
Section VI.
Wr82 Wrenn M.E. and Singh N.P., 1982, "Comparative distribution of uranium,
thorium and plutonium in human tissues of the general population," in
Natural Radiation Environment, (K.C. Vohra et al., Eds.), (New Delhi:
Wiley Eastern Ltd.), pp. 144-154.
Ya68 Yamamoto T., Masuda K., and Onishi N., 1968, "Studies on envirorjnental
contamination by uranium. I. Environmental survey of uranium in
Kamisaibara Village, Okayama Prefecture," J. Rad. Res. (Japan) 9 (3-
4) : 92-99.
-------
R-23
Yu73 Yuile C.L., 1973, "Aniinal experiments," Handbook of Experimental
Pharcacology, (Hodge H.C., Standard J.N., and Harsh J.B.), Vol. 36,
(Berlin: Springer-Verlag), pp. 165-196.
-------
*y OfL O Ot
TABLE 1. Body Ra contents and daily Ra intakes (inferred from fecal excretion) of adult and
22fi
adolescent human males. Their Ra was accumulated entirely from environmental sources (food, water,
air). Data of Stehney and Lucas (St56).
Mean No. of Excretion per day* Ratio
No. of age fecal Dry feces Radium of body Ra
Group subjects (years) samples (gm) (10~^gm) to fecal Ra
Adult Control 1
Stateville 11
Chicago boys 7
Lockport boys 8
29
44
17
17
3
19
26
24
26.6 ± 0.1
24.1 ± 3.8
24.7 ± 1.1
37.3 ± 3.2
1.7 ± 0.4
12.6 ± 2.4
1.6 ± 0.1
8.2 t 0.6
24
17**
22
45
* Errors shown are the standard deviation of the mean for the group.
** Ratio for Ra retained by 11 subjects during a 19.7-year average period at Stateville.
-------
Table 2. Days equivalent intake of radium
isotopes in bone.
Days Equivalent Distribution
Nuclide Intake In Bone
Ra-226
Ra-228
Ra-224
25
10
0.3
volume
volume
surface
-------
3
TABLE 3
Average skeletal dcse rates (mrad/yr) from constant concen-
trations of naturally-occurring uranium and radium radio-
nuclides, assuming a constant concentration of 1 pCi/g of
parent nuclide and a uniform distribution.
Nuclide mrad/yr
__
2340
O -v Q 2 2 0
* Ra (+ dtrs. adjusted for 90% Rn retention)
22®Ra (+ dtrs. adjusted for 33% retention of 222Rn
210
in bone, excluding ?o)
226Ra (+ dtrs. adjusted for 33% retention of 222Rn
in bone, including ^®Po)
From Wr77b. If U and U are in equilibrium, add the
two doses together. Alpha dose rates at bone/surface inter-
faces will generally be less than half the average dose
rates in mineral bone.
78
89
560
207
240
-------
4
Table 4. Computed limiting concentrations (pCi/1) of longlived U,
226 224
Ra, and Ra in water for given lifetime risks (LR) of bone sarcoma (plus
226
head carcinoma for Ra) based on drinking water intake of 1 1/day
Li fetime
Risk Computed Limiting Concentration (pCi/1)
(Lr) 233, 234, 236, 236 or 238^, 228Ra 226Ra 224Ha
10~4 330 23 24 310
10~5 33 2.3 2.4 31
-------
5
Table A-1. Estimated gastrointestinal absorption of U continuously fed to rats
as dry salts mixed with diet. Data of Msynard et al. (Ma53). Food intake is
assjjned to be 5 g/100 g weight/day.
Compound
Dietary U
(g/100 g)
U intake
(ing O/Vg/day)
0 in femur
(pg D/g ash)
Calculated
absorption (%)
uo2f2
(0.77 0 by
weight)
0.01
0.05
0.25
0.50
3.8
20
96
195
control
5.2
28.4
58.2
0.052
0.059
0. 059
0.06
U02(N03)2 6H20 0.50
(0.47 0 by 2.0
weight)
118
472
22.5
184
0.038
0.078
-------
6
Table A-2. Ex
per imental
determinations of
gastrointestinal absorption
of U.
0 Absorption
0
intake
Species
(%)
(wg U/kg/day)
Comments
Reference
Adult animals
1 hamster
0.77
6.3
X
102
233U02(N03)2, pH 1
Harr81
2 rabbit
0.28
17
X
103
°°2^3^2 *n water
Tr 83b
3 dog
1.55
7
X
102
233U02K2 in water
FisSO
4 baboon
1.2
5
X
10~1
233D02:nD3)2 in 0.1
NaSCOj, p3 8.1
M
LarS4
5a man
0.73
6.7
X
103
002(NOj)2 in water
Bu58
b man
1.4
1.32x
102
D02(H03)2 in cola drink
Hu69
6a rat
0.35
3
X
102
233U02(N03)2, pH 4
Eam43
b rat
0.06
2.3
232U02(N03)2, pH 2
Su8Ca
0.06
4
X
103
233002(N03)2, pH 2
Su8Ga
6c rat
0.044
5.1
X
103
233U02(N03)2, pH 2
Su83
0.044
1.3
X
104
233D02(N03)2, pH 2
Su83
0.088
2.5
X
104
233U02(N03)2, pH 2
Su83
6d rat
0.035
3.3
X
104
U02(H03)2 in water
Tr 83b
6e rat
0.052
2
X
104
U02F2 in diet, 0.05%
Ma 5 3
0.059
9.6
X
104
U02?2 in diet, 0.25%
Ma53
0.06
2
X
105
002F2 *n diet, 0.5C%
Ma 5 3
0.038
1.2
X
105
002(ND3)2 in diet,
0.50%
Ma53
0.078
4.7
X
105
002(N03)2 in diet,
2.0%
Ma53
0.04
9.7
X
105
U02(N03)2 in diet,
^50/30 days
4.0%
Sav49
Neonatal aniica
Is (one to
two
days
old)
7a rat
6.7
0.12
232002{N03)2, pH 2
SuBOb
1.3
2.1
X
102
233002(N03)2, pH 2
SuSOb
8 swine
34.5
2
X
103
233002(N03)2, pH 2
Su82
-------
7
Table A-3. Measurements of U in human urine. Data of Welford
et al. (We60) and Welford and Baird r*e67).
U Concentration Number of Samples
(ug U/liter) Chicago New York Total
< I
3.i
050
1
3
4
0.051
-
0.075
4
4
8
0.076
-
0.10
4
6
10
0.101
-
0.125
0
3
3
0.126
-
0.15
0
4
4
0.151
-
0.175
1
2
3
0.175
-
0.20
1
1
2
0.201
-
0.225
0
1
1
0.226
-
0.25
0
0
0
0.251
-
0.275
0
0
0
0.276
-
0.30
0
2
2
Total of 37 samples: arithmetic mean, 0.C98 ± 0.067 ug
U/liter; modal value ¦ 0.C88 ug U/liter (between 0.076 and
0.101 ug U/liter); geometric mean » 0.C97 ug U/liter.
-------
8
Table A-4. Gastrointestinal absorption of U calculated from urine and fecal data
reported by Pishet et al. (Fish83).
Subject
No. days
Apparent intake8'15
Absorption
(%)c'd
(yg)
23 40
238u
C1
1
18
0.43
0.64
C2
2
12
0.96
0.98
C3
2
11
0.39
0.85
R1
2
24
0.58
0.55
R2
2
5.3
1.11
1.62
R4
2
71
0.40
0.55
Average
24
0.65
0.87
a 258
Apparent intake « urinary U/day + fecal U/day. It is given only for 0 in y
9-
^ The daily intake of 0 is inferred from the specific disintegration rate of
U: 0.742 d?m/yg 0.
c The data were reported as and dprt/day. The ratios of were
close to 1. The absorption calculated for 6 individuals was 0.65% for and
0.87% for the difference is probably related to counting error for the
isotopes in urine. The mean, 0.76%, is taken as the best estimate of GI
absorption in these persons.
Absorption is taken as (urinary U/day)/(apparent intake).
-------
9
Table A-5. Estimated gastrointestinal absorption of U in man, based on
environmental data.
Estimated 0 intake
(|jg U/kg/day)
GI absorption
(%)
Method and references
0.025
7.8
Urine analysis, U.S. dietary
survey (We60, We67)
0.15
1.6
Analysis of urine and local
diet and water (Ya68,
Mas71a,b,c,d)
0.34
0.76
Analysis of urine and feces
of 6 subjects (Fish83)
0.76
feces of one subject (So80)
2.2
2 analyses of urine and
-------
10
Table A-6. Distribution, excretion and retention parameters of 0 and Ca in
Reference Man.
Caa Ub
Body content 1000 g 46 yg
Bone 995 g 25 wg
Soft tissues 5 g 13 vg
Daily intake 1.1 g 1.75 pg
Intestinal absorption (%) 34 ?
Retention in body (% at 10 days after injection) 73 20c
a Ca metabolic and distribution data from He€4a,b; ICRP73,74; Spi68.
k U metabolic data from Tables B-1 and B-2.
0 U retention data from Sa48j Hu73; 3er57.
-------
11
Table A-7.
Linear regression
analysis of
dependence
of U absorption
on U intake.
Animal®
Correlation
(No. of
Scale
Regression equation*5
coefficient
studies)
U absorption
U intake
U absorption (%) =
(c)
rat (13)
1inear
linear
0.C85 -
5.2x10~8 I
-0.174
linear
log
0.16 -
0.02 log I
-0.351
log
log
0.11 I"
0.06
-0.353
all others
linear
linear
2.2 - 1.
3x10~4 I
-0.323
(10)
linear
log
2.6 -
0.6 log I
-0.581
log
log
1.8 I"
0.12
-0.705
(9)
log
log
1.4 I"
0.08
-0.601
a Data from Tables A-2 and A-5.
^ 1 is expressed in yg/day intake/kg of body weight.
-------
Tteble B-1: Uranium in skeleton of persons with no kncwn occupational exposure
Source of senples
Skeletal parts measured
Reported U
ocroentratlfgi
ng u/g ash
U in Reference Man skeleton
Tbtal U (pg)b
Iteferenoe and method
Uiited States
U.S. (?)
faiur
0.004 iq 23fyg ash
4.0
11.2
Hfl 67;
Illinois
variety of bones
0.022 ig U/g ash
22.
61.6
We67; fluorcmetry
Illinois
same sanples
8.6 ng U/g ash
8.6
24.
Vle76; fluorcmetry and
mass spectrometry
New York City
vertebral bodies
0.32 ng U/g wet
2.3
6.4
We76: flmrometry
3 U.S. cities
vertebral ash
0.6 c£in 23Vkg ash
0.81
2.3
Fi80; 23*\j alpha spectr.
Wisoonsin
variety of tones
0.02 ig U/g ash
20.
56.
Li81; 238U alpha spectr.
Cbloradn and
Pennsylvania
vertebral bodies
0.5 pCi 23fykg V*etc
10.6
38.8
Sin83; 23f\j alpha spectr.
New York State
long bone oortex
21 pg ^^tl/g wet^
4.6
12.9
ScBl; NIAR
5 U.S. regions®
vertebral bodies^
199 fCi 23Vkg wet
22 fCi 23SUAg wet
150 fCi 238UAg wet
3.8
10.6
Br80,83; 234'235'238u
alpha spectr.
Other Countries
United Kingdom
variety of bones
2.03 x 10"8g U/g ash9
20.3
56.8
Ha71, 72; delayed n
Japan
ribs
6 ng U/g ash
6.0
16.8
Nd70; delayed n
India
not specified
4.9 ng U/g ash
4.9
13.7
Ga70;
N*pal
vertebral hodles^
11.8 dpPB 23t\j/tog ash
16.0
44.8
Fi83» 23{\) alpha spectr.
Australia
vertebral bodies and ribs
3.65 dpn 23fykg ash
4.9
13.7
n m
-------
^~bnversion factors userl to obtain U concentration as ng U/g ash: Ut anion oanpoeition arri activity— 1 pCi = 2.22 cfcxn; specific activity of =
7.42 x 10^ dpn/q, only data for ^®U were used; abundance of = 0.007196. Factors related to bone composition— Ash fractions of cortical and
cancellous human bone are 0.639 and 0.599, respectively (Gb64a,b). Ml sanples of vertebrae are assured to have been obtained at routine autopsies,
and therefore are vertebral body only. The ash fraction - 0.108 g asVg vertebral bod/ was used for the Cblorado-ftermsyl vania set (the oanpoei-
tion of general hospital autopsy populations, obtained frcm the data tables of McI79 is as follows: 20 to 40 years 10.5%, 40 to 60 years 35%, more
than 60 years 54.5%). The ash fraction 31 0.14 g asVg vertebral body was used for the New York City sanples, which in the McI79 series, ^pear to
be a coroner's population with a composition as follows: 20 to 40 years 50%, 40 to 60 years 45%, more than 60 years 5%. The age-related ash frac-
tions of wrtebral body were calculated from data given by Arnold and Wfei (Ar72). The calculated value of 0.14 for the ash fraction of vertebral
wedges of young adilts is supported by the measured ash fraction (0.143 g ash/g wet vertebral wedge) for 4 large oonposites of vertebral wedge
obtained after acute deaths of persons 20 to 25 years old in 4 separate geographic regions of the U.S. (Br80).
kflhe Reference Man skeleton contains 2800 g ash.
cbne set of very hi
-------
Table B-2: U in soft tissues of persons with no known occupational exposure
Tissue
Reported U Concentrations (ng U/g wet tissue)
Best value3
Liver
0.25
(Ha72), 0.13 (We76), 0.30 (Sin83)b
0.23
Lung
1.0
(We67), 0.53 (We76), 1.7 (Sin83)b
1.08
Kidney
0.24
(We76), 0.63 (Sin83)b
0.44
Whole blood
0.72
(Ham70b)c, 0.57 (We76), 0.11 (Lu70)
0.46
Red cells
0.15
(Lu70)
Plasma
<0.08 (Lu70)
Muscle
0.19
(Ha72)
0.19
Heart
0.16
(Ha72)
0.16
Spleen
0.42
(Sin83)b
0.42
Gonad
0.60
(Sin83)b
0.60
TB Lymph Nodes
24.8
(Sln83)b
24.8
Fat
0.60
(Ha72)
0.60
a Best value in the arithmetic mean of the mean values reported by the different
analysts) the value for blood includes separated red cells and plasma (Lu70).
l 238
The values reported as dpm U/kg wet tissue (Sin83) have been recalculated
as in the footnotes to Table B-1.
c Only the data for U.K. samples.
-------
Table B-3: Estimation of U content of soft tissues of Reference Man.
U Concentration Reference Man U Content
Tissue (ng/g wet weight) Tissue Weight (g) (yg)
Liver
0.23
1,800
0.41
Lung
1.03a
1,000
XI
~*•»»
CO
o
Kidr.ey
0.44
310
0.14
Whole blood
0.46
2,600°
1.2
Muscle
0.19
28,000
5.3
5c art
0.16
330
0.053
Spleen
0.42
180
0.08
Gonad
0.60
35*
0.021
T3LN
24. 8e
15
(0.37)
Pat
0. 6f
(13,500)
(8.1)
Res. Soft Tissue^
32,000
5.6
Total (adjusted)13
13
a About 85% of lung 0 is considered the residue of inhalation,
k Adjusted total soft tissue does not include values enclosed in parentheses.
c Assuming only one-half of blood volume is drained from tissues at autopsy.
^ Two testes.
e U in T3LN is considered to have been inhaled.
^ See text.
-------
Table C-1. Parameters
of twc-compar
tment models
of 0 met*
abolism in
hjrcan kidney.
f1
f21
T1 (d)
f 22
T2(d)
v»
ICRP(1979)
0.05
0.12
6
0.00052
1500
3. 5x10"*
ICR?(1979)
modified
0.014
0.12
6
0.0C052
1500
9.3x10
Sp73
0.01
0.11
15
0
0
7.7x10"5
Sp73 (mod-
ified f.)
0.1
0.11
15
0
0
7.7x10
1
0.05
0.11
15
0
0
3. 8x10~4
0.03
0.11
15
0
0
2. 3x10~4
Committee
Consensus
0.014
0.11
15
0
0
1.1x10"
-------
Table C-2. Parameters of two-compartment nodels of U metabolism in human
skeleton.
f, f21 T1 (d) f22 T2(d) A»/I(d)
(days)
ICRP (1979)
0.05
0.2
20
0.023
5000
8.3
modified
0.05
0.2
300
0.023
5000
12.5
mod ifled , ^21' ^22
0.05
0.1
300
0.1
5000
26.5
Cothern et al. '>CoS3b)
0.2
0.2
20
0.023
5000
33
modified f^
0.01
0.2
300
0.023
5000
1.7
Committee Consensus*
0.014
0.1
300
0.1
5000
10.7
* This model predicts 17 times greater amounts of 0 in cortical (compart-
ment 2) vs. trabecular (coapartaent 1) bone, which seems at variance with
reported data (Ea72, We67) on relative concentration of 0 in cortical and
trabecular bone. No single ®odel or combination of parameters chosen is
completely consistent with the experimental data.
-------
F-' 8
Caption
Figure A-1. Gastrointestinal absorption of U (%) plotted as a function of U
intake (^g U/kg/day). Numbers within the symbols are the same as the numerical
order of appearance of the experimental results in Table A-2. The circled
alphabetic notations are gastrointestinal absorption from environmental data of
We67 and S08O. Results from experiments in rats and other experiments in which 0
ingested was more than 10^ ug 0/kg/day have been omitted from the fits to curves
1 and 2 as unlikely to be encountered in environmental exposures.
-------
I02
2 I01
~El
<
cc
z>
Ll.
o
a.
cr
o
CO
CD
<
3
~ man and animals other
than rots
A rats
Points labeled E ore from Tobie A-5
(man),other points from Toble A-2
All data for man and animois other thon rots A(#/0)» 18% I"012
10
-I
| OE3
1.4{%) l"60#(excludes point El)
^,0.11% I-0 06 , rats only
Ab
AC
~AV
Ac Ac
Se - - .
Ae
Ad
Ae
!0*'
lO'^IO"1 10° 10' I02 !03 I04
URANIUM INTAKE (y.g/U/kg/day)
I05 I06
-------
COMMITTEE ON THE HEALTH EFFECTS AND
RISK DUE TO THE INTAKE OF NATURALLY
OCCURING ALPHA PARTICLE EMITTERS
Chairman: Charles W. Mays
Recorder: Frederick Hodge
Committee Members: Robert E. Rowland
Andrew F. stehney
-------
CANCER RISK FROM THE LIFETIME INTAKE OF RADIUM AND URANIUM ISOTOPES"
Charles W. Mays
Radiobiology Division, Bldg. 351, U. of Utah
Salt Lake City, Utah 84112
Robert E. Rowland
Past Associate Director, Argonne National Lab
Present address; Box 281B, RR #1, Princeton, KY 42445
Andrew F. Stehney
Center for Human Radiobiology, Argonne National Lab
Argonne, Illinois 60439
ABSTRACT: From extensive human data on the induction of skeletal cancers
(bone sarcomas and carcinomas of the head sinuses) by 226Ra, 228Ra, and
221+Ra, the cumulative lifetime risk to 1 ,000,000 people, each ingesting 5
pCi of a radium isotope per day, was calculated to be 9 bone sarcomas plus
12 head carcinomas for 226Ra, 22 bone sarcomas for 228Ra, and 1.6 bone sar-
comas for 22l>Ra. Assuming that the risk per rad of average skeletal dose is
equal for 226Ra and the uranium isotopes with half-lives exceeding 1000
years, and that the equilibrium skeletal content is 25 times the daily in-
gestion of 22SRa, but 11 times the daily ingestion of long-lived uranium,
the cumulative lifespan risk to 1,000,000 persons, each ingesting 5 pCi per
day of 233U, 231*U, 235U, 236U or 233U, is estimated to be about 1.5 bone sar-
comas. The uranium risk is not well established and additional research is
needed on the metabolism of uranium in humans and its carcinogenicity in lab-
oratory animals. These estimates assume linear dose-responses. However, if
incidence varies with the square of dose, virtually no induced cancers would
be expected from these levels of radioactivity.
Prepared for an EPA-sponsored workshop on Radioactivity in Drinking Water,
24-26 May 1983- The research on the effects of 22l+Ra in man is supported by
DOE Contract DE-AC02-76EV-00119 and EURAT0M Contract D1 -D-461 —0 (B). The re-
search on the effects of 225Ra and 228Ra in man is supported by DOE Contract
W-3i-i09-Eng-38.
-------
- 2 -
INTRODUCTION
We made this analysis at the request of C. Richard Cothern and William
L. Lappenbusch of the Office of Drinking Water, U.S. Environmental Protec-
tion Agency, to provide the best presently available scientific information
on the cancer risk from the daily ingestion of radium and uranium. The
predicted cancer risk can then be compared to the financial costs and pos-
sible adverse health effects of procedures to decrease the radium or uranium
concentrations in a local water supply.
In this paper we will estimate the lifetime risk of skeletal cancers in
humans from the daily intake of 22I+Ra, 226Ra, and 228Ra. Using the risk co-
efficient for 226Ra-induced bone sarcomas, the lifetime risk will then be
calculated for the daily intake of 233U, 23t+U, 235U, 236U, and 238U. The
chemical toxicity of uranium is presented in a companion paper (Wr 84).
TYPES OF CANCER INDUCED BY RADIUM ISOTOPES IN MAN
At low and medium doses of internally-deposited radium, the most severe
biological damage is cancer arising from skeletal tissue. For 226Ra, the
following two types of malignancy are induced: (1) Bone sarcomas (mostly
osteosarcomas, fibrosarcomas, and chondrosarcomas), and (2) head carcinomas
(carcinomas of the paranasal sinuses and the mastoid air cells).
Among some 3700 located U.S. persons" who were exposed to 226Ra and
228Ra intake by dial painting, medical administration, and other means, a
*Different subgroupings of this population have been analyzed by the follow-
ing quoted authors: Spiers, et at. (83), Rowland, et at. (83), and
Stebbings, 3t at. (84). We give the composition of each subgroup as it is
discussed in the text.
-------
- 3 -
total of 85 cases of bone sarcoma and 36 cases of head carcinoma had been ob-
served as of 31 December 1982 (Gu 83). However, n£ head carcinomas have yet
been observed in the follow-up of 2324 German patients injected with 22ltRa,
although 55 have developed bone sarcomas (May 84, Wi 83). Similarly, no
head carcinomas have been observed in persons internally contaminated with
228Ra, unless the dose from 226Ra was also high (Ev 66, Ro 78). This sug-
gests that when 226Ra decays to 222Rn within the body, the accumulation of
222Rn gas in the head cavities is the major inducer of these carcinomas
(Ev 66, Ro 78, Ma 84). For pure 22I+Ra and 228Ra, which do not produce 222Rn
gas, the risk from head carcinomas is regarded as trivial compared to the
risk from bone sarcomas.
For a-emitters deposited in mineral bone, the risk from radiation-
induced leukemia in humans has been insignificant relative to that from bone
sarcoma. In the subgroup of 2940 located males and females who were radium
dial workers before 1970, there were 63 cases of bone sarcoma observed com-
pared to 1 case expected naturally, in contrast to 10 cases of leukemia ob-
served compared to 9 cases expected naturally (Spier 83, Ro 83). Further-
more, 4 of the 10 observed leukemias were chronic lymphocytic leukemia, a
type of leukemia not found to be increased in any study of irradiated people
(NAS 80).
The lifespan risk coefficient for leukemia-induct ion by a-particles in
humans is very much lower than obtained by multiplying the ICRP risk coeffi-
cient for sparsely-ionizing radiation by a quality factor of 20 (ICRP 77)-
ICRP risk coefficient _ <¦ 20 leukemias i r20 rertN _ 400 leukemias /j\
for a-particles ~ 106 person*rem rad 106 person«rad
-------
- k -
Spiers, et al. (Spier 83), have calculated the a-particle dose to red marrow
in a special subgroup of 693 located women who were dial workers before 1930,
measured while living, surviving without bone sarcoma for 2 years after
their first measurement, and alive in 1957- Their a-particle dose to red
marrow averaged about 25 rads. Assuming a similar average dose to the total
measured and unmeasured subgroup of 1285 located women who were dial workers
prior to 1930, the ICRP risk coefficient would predict about 13 cases of
radiation-induced leukemia among these 1285 women.
[1285 persons] [25 rad] = '3 1 eukemias (2)
10° person«rad
The total number of predicted leukemias would then have been about 18 cases
(13 radiation-induced plus 5 naturally-expected)(Spier 83). However, the pre-
diction of 18 leukemias is strongly rejected by the observation of only h leu-
kemias among these 1285 women. But if the risk coefficient were 10-fold
lower, the predicted total of about 6 leukemias (1 radiation-induced plus 5
naturally-expected) would be in reasonable agreement with the observed k leu-
kemias.
Independent evidence that the ICRP risk coefficient for a-particle-
induced leukemia is about 10 times too high comes from Robin Hole's analysis
of leukemias in the European Thorotrast patients (Mo 78). There were kk leu-
kemias observed among 4000 Thorotrast patients, injected with an average of
25 ml of Thorotrast, for which the dose rate to red marrow was 9 rads per
year, giving a dose of 270 rads in 30 years of follow-up.
Leukemia risk coefficient _ M leukemias b0 leukemias
for Thorotrast a-particles (4000 persons)(270 rad) 106 person-rad
-------
- 5 -
Why is the observed risk coefficient for leukemia induction by a-particles
in humans so much lower than the ICRP risk coefficient given in Equation 1?
Possible explanations might include: (1) Nonuniform irradiation which allows
nonirradiated (undamaged) stem cells to replace cells killed by irradiation,
(2) low frequency of potentially leukemogenic cells in the regions most
heavily irradiated by a-particles such as regions near the bone surfaces or
near the phagocytic cells that take up colloidal Thorotrast, (3) a leukemia
RBE less than 20 for uniform a-particle vs. sparsely ionizing radiation at
low dose rate, (4) a 1 ifespan risk coefficient less than 20 leukemias/106
person-rem for sparsely ionizing radiation at low dose rates, or (5) other
factors. Regardless of which theoretical possibility is correct, it is now
firmly established by direct observation of humans exposed to graded levels
of a-radiation emitted from mineral bone that the risk from induced leukemia
is very small compared to the risk from induced bone sarcomas (Spier 83, May
83). We have given considerable attention to the risk from leukemia, pri-
marily because a recent calculation suggested that the lifetime ingestion of
225Ra might induce more leukemias than bone cancers (Co 83). However, the
opposite is indicated by actual observation on the effects of radium iso-
topes in humans (Spier 83, Ro 83, May 83).
Multiple myeloma has occurred in 6 of the 1285 located women who were
dial workers prior to 1930, compared to 2.1 cases naturally-expected (Steb
84). While the 3-fold increase is statistically significant (P = 0.042),
the increase is not necessarily due to internally-deposited 226Ra and 228Ra.
The externa] y-ray exposure from the radium paint pots has been estimated at
8-20 rad/year (Steb 84). Of the 6 myeloma cases, 4 worked as dial painters
at least 50 weeks, one worked only 2 weeks, and the duration of employment
-------
- 6 -
for the remaining one is unknown (Steb 84). Even if the 4 excess cases of
myeloma were caused by skeletally-deposited radium, induced myeloma would be
only 6% as frequent as the 63 cases of bone sarcoma observed among these 1285
women. Among the 2324 German 221+Ra patients, the 1 case of multiple myeloma
(plasmacytoma) was only 2% as frequent as the 55 cases of bone sarcoma ob-
served among these patients (Ma 84, W? 83). Multiple myeloma appears to be
only slightly inducible by radium in human bone.
For the 1285 located women who were dial workers before 1930, Stebbings,
et at. (Steb 84), have compared the numbers of observed vs. expected cancers
for the 10 following types of highest naturally-expected frequency: breast,
colon, lung, stomach, pancreas, cervix uteri, corpus uteri, rectum, leukemia,
and liver. For none of these cancers was the observed incidence significantly
different (P<0.05) from the natural expectation based on rates in the rele-
vant counties. Summing the 10 sites, there were 101 observed cancer cases vs.
94 expected, based on country rates. The excess of only 7 cases indicates
that the internally-deposited radium and the external y-ray exposure received
by these women were relatively ineffective in inducing nonskeletal cancers.
It is clear that the observed 63 bone sarcomas and 23 head carcinomas
represent the vast majority of radium-induced cancers among the 1285 located
women who were radium dial workers prior to 1930, although a small number of
additional cancers might possibly have been induced in other tissues (Fin 69,
Steb 84).
Similarly, among 612 adult German patients of known dose, injection
span, and health status (as of June 1974), 38 soft-tissue neoplasma were ob-
served compared to about 33-40 naturally-expected cases during the follow-up
averaging 17 years for the adults (Spies 78). Among a similar 204
-------
- 7 -
patients Injected as juveniles and followed an average of 20 years, 5 soft-
tissue neoplasms were observed, similar to natural expectation for their
ages (Spies 78). If an excess of soft-tissue neoplasms had occurred, it was
small relative to the b~J of these patients of known dose that developed bone
sarcoma. (Among the 899 patients of known and unknown dose in the Spiess
series, 53 have developed bone sarcomas.)
DOSE-RESPONSE RELATIONSHIPS
1. Head carcinomas
The incidence rate I for head carcinomas, due to 226Ra intake to blood,
was fitted to a variety of dose-response relationships of the general form
I = (C + a D + B D2)e for the measured female dial workers employed be-
fore 1930 (Ro 78). A minimal latent period of 10 years was assumed. Based
on Chi-squared statistics, the best fit (probability P = 0.86) was to the fol-
lowing linear relationship (Ro 78):
Yearly incidence , , ,n-5 „ /,¦>
/ . ' , \ = 1.6 x 10 D (4)
(carcinomas/person*yr)
where the radium intake to blood is D_ yCi 226Ra. The intake of 226Ra to
blood was back-calculated from the measured body content of 226Ra at "t"
days after intake using the Norris retention function (No 55), retention R =
— 0 52
0.5^ t ' . The intake of 228Ra was disregarded, since head carcinomas have
not been associated with 228Ra.
2. Bone sarcomas
Two groupings were made for analysis of bone sarcoma incidence in the
measured female dial workers employed before 1950 (Ro 83).
-------
- 8 -
Group A consisted of all measured female dial workers. Their years at
risk were assumed to begin 5 years after the start of employment (allowing
for a 5-year minimal latent period for bone sarcoma) and to continue until
death, the diagnosis of a bone sarcoma, or the end of follow-up. Group A
contained kZ bone sarcomas among 1468 persons averaging kQ.O years at risk.
It may underrepresent persons who died early and were not measured for ra-
dium, but overrepresent those who ultimately suffered from the effects of
radium and were identified and measured as a consequence.
Subgroup B, a part of Group A, was selected to eliminate persons possi-
bly measured for radium as a consequence of their cancers. It included only
the female dial workers who were measured for radium while living and ex-
cluded any person who died or was diagnosed with a bone sarcoma within 2
years after the first measurement, since the symptoms of bone sarcoma usu-
ally first appear within 2 years before diagnosis or death. Thus, the years
of monitored risk for Subgroup B started 2 years after the first measurement
and continued until death, the diagnosis of a bone sarcoma, or the end of
follow-up. Subgroup B contained 13 bone sarcomas among 1257 persons averag-
ing S.k years at monitored risk.
For Group A, the incidence rate for bone sarcomas was best fit by Chi-
squared analysis (P = 0.73) by the following dose-squared - exponential equa-
tion (Ro 83):
Yearly incidence 7 0 x 10"8 D2 e"0.0011 0 (5)
(sarcomas/person*yr)
Where for bone sarcomas, D_ is the intake to blood in uCi 22®Ra plus 2.5
times the intake in yCi 228Ra, since 1 yCi 228Ra with high retention of
-------
- 9 -
daughters was estimated to be equivalent for bone sarcoma-induct ion to about
2.5 yCi 226Ra with low retention of daughters (Ro 78).
No bone sarcomas have yet been identified among any of the exposed
measured persons with radium intakes to blood below 50 yCi (Ro 83). Among
the measured 1321 female dial workers below 50 yCi intake, the preceding
equation p1 us the natural incidence predict 0.67 total expected bone sarcomas
(0.29 induced plus 0.38 natural). For an expectation of 0.67 sarcomas, zero
sarcomas could occur with a 51% chance (P = 0.51).
At low values of intake, the exponential in Equation 5 approaches unity,
and the calculated yearly incidence from 22SRa varies as the square of the
intake. For an intake of 1 yCi 226Ra, the calculated yearly incidence of
-8 -5
7 x 10 is less than 1% of the natural yearly incidence of about 1 x 10
At lower doses the calculated yearly incidence from 225Ra rapidly approaches
zero. We will use zero risk as a lower limit for radium-induced bone sar-
comas at low intakes.
For the epidemiologically more suitable Subgroup B, acceptable fits
were observed for the following equations (Ro 83):
Linear Model (P = 0.26)
Yearly incidence x -5 0 (6)
(sarcomas/person«yr)
Dose-Squared - Exponential Model (P = 0.27)
Yearly incidence l0 .--8 _2 -0.0015 D /-,*
, 7 , \ = 18 x 10 D e (7)
(sarcomas/person«yr)
Where for bone sarcomas, D_ is the intake to blood in yCi 226Ra plus 2.5
times the intake in yCi 228Ra.
-------
- 10 -
The linear equation (Eq. 6) from Subgroup B predicts a total of 1.25
bone sarcomas (1.06 induced plus 0.19 natural) in the Subgroup B female dial
workers with intakes to blood below 50 yCi. However, applying this equation
to the 1321 female dial workers below 50 pCi in Group A, with their much
longer assigned years at risk, predicts a total of 4.45 bone sarcomas (4.07
induced plus O.38 natural). For an expectation of 4-^5 sarcomas there is
only a 1.2? chance that zero sarcomas would occur (P = 0.012). Hence, we re-
gard linear equation (Eq. 6) 2 x 10 5 D as a reasonable upper 1imit to the
risk from radiation-induced bone sarcomas at low intakes.
We regard the actual risk from bone sarcoma at low intakes of radium to
be somewhere between our upper limit (linear Eq. 6) and the lower limit of
zero. Tentatively, we arbitrarily assume it to be midway between our upper
and lower limits.
Provisional best estimate of bone sarcoma risk from 226Ra
Yearly incidence _ j ^-5 ^ /gx
(bone sarcomas/person*yr)
The above provisional best estimate predicts 2.41 total expected bone sar-
comas (2.03 induced + 0.38 natural) among the 1321 female dial painters of
Group A with intakes below 50 pCi. For an expectation of 2.41 sarcomas
there is a 3% chance that zero sarcomas would occur. Hence, Eq. 8 cannot be
statistically rejected (P > 0.05).
It is possible that 1-or-more new bone sarcomas may occur in the low-
intake persons under present surveillance. In fact, 1-or-more sarcomas at
low-intake may already exist among the 24 known cases of bone sarcoma who
were buried without an evaluation of radium content (Gu 83; see p. 162).
-------
- 11 -
Therefore, It is important to exhume and measure as many as possible of the
Ik buried but unmeasured bone sarcoma patients, and to continue to follow the
living exposed persons. If no additional sarcomas are found, it would sug-
gest that our provisional best estimate of bone sarcoma risk may be too high,
whereas finding more sarcomas may suggest that our tentative best estimate
is about right or perhaps too low. The present study is unique in that it
offers the only known opportunity to resolve this question of scientific and
humane importance.
CUMULATIVE RISK FROM LIFELONG INTAKE OF RADIUM
The following cumulative risk estimates for long-lived 226Ra and 228Ra
correspond to a constant concentration of radium in bone throughout life —
at the equilibrium concentration resulting from the adult intake to blood of
1 pCi of a radium isotope per day. The daily intake required to maintain a
constant concentration of radium in bone may differ for infants, children,
adolescents, and adults. The general finding that the 22SRa/Ca ratio in
bone remains roughly constant with age (Fis 79, Ra 64) probably reflects a
more-or-less constant ratio of 226Ra/Ca in the diet at environmental levels
during much of the lifespan (Wr 77, Wr 84). Even the recent results of Muth
and Globel (Mu 83), which suggest possible peaks in 226Ra concentration at
about 1 and 10 years of age, correspond to an average lifespan concentration
only about 20% higher than the average of their adult values. For simplic-
ity, we assume that the daily intake of dietary 226Ra and 228Ra at environ-
mental levels result in constant Ra/Ca ratios in bone. For long-lived 225Ra
(Tj = 1600 yr) and 228Ra (Ti = 5.77 yr), one might expect the high rate of
Ra accretion in the growing bones of a child to be opposed by the enhanced
rate of Ra removal from resorbing bone.
-------
- 12 -
However, short-lived 221+Ra (T£ = 3-62 days) should decay before appreci-
able resorption occurs. Therefore, we assume that skeletal concentrations
of 22tfRa in children aged 0-15 and in adolescents, aged 16-20, respectively,
would be 3 and 2 times the adult concentrations, based on the fractions of
injected 221fRa estimated to decay in bone (Spies 70).
Now the risks from the lifelong intakes to blood of 1 pCi Ra/day will
be estimated for 225Ra (head carcinomas and bone sarcomas), 228Ra (bone sar-
comas only), and 221*Ra (bone sarcomas only).
226Ra: Cumulative risk from HEAD CARCINOMA:
The average life expectancy in the U.S. has increased from 50 years in
1910, to 60 years in 1930, to 70 years in I960, and was Jk years in 1980
(Ub 84). We assume it to be about 75 years in the mid-1980's. For a life-
time of 75 years, only the 226Ra intake during the first 65 years is consid-
ered effective for the induction of observable head carcinomas because of
the assumed minimal latent period of 10 years for these carcinomas (Ro 78).
cumulative . (I0'6 uCt)(365 dars)(65 yr) . 0.0237 uCi (9)
intake to blood day yr
The risk rate for observable head carcinomas builds up linearly from
zero at age 10 years (the minimal latent period) to a maximum 65 years later
at the end of life. From equations k and 3:
-------
- 13 -
Risk rate at _ ,1.6 x 10~5W„ 0.380 x lO-6 /l<%x
end of life = ( nCi-yr )(°-Q237 uC.) (10)
• 6 • 6
Av. risk rate during = 0 + 0.368 x 10 /yr _ 0.190 x 10
expression period 2 yr
,-6
(11)
Cumulative risk ,0.190 x 10 v ,£c >> ..-6 ,,-x
("6Ra head carc.) ¦ ( -r )<«S yr) = 12x10 (12)
In words, if 1 million persons each had the equivalent intake throughout life
of 1 pCi 226Ra per day to blood, about 12 cases of induced head carcinoma
would be predicted. For a natural incidence rate averaging 0.5 x 10 5 per
year, about 375 naturally-occurring head carcinomas should occur among
1,000,000 persons during their lifespan, averaging about 75 years.
226Ra: Cumulative risk from BONE SARCOMA;
For a lifetime of 75 years, only the 226Ra intake during the first 70
years is considered effective for the induction of observable bone sarcomas
because of the assumed minimal latent period of 5 years for bone sarcomas
(Ro 78).
<» V'J - »¦"* UCI (.3)
The risk rate from induced bone sarcomas builds up linearly from zero at
age 5 years (the minimal latent period) to a maximum 70 years later at the
end of life. From equations 8 and 13:
-------
- 14 -
Risk rate at . (' * '°~5)(0.02S6 „CT) - °'256 x l0"6 (H)
end of life • yCi»yr yr
Av. risk rate during = 0 + 0.256 x lO'Vyr _ 0.128 x 10~6 /._x
expression period = 2 ~ yr
Cumulative risk _ ,0.128 x 10 ^ x „ ,„-6
(225Ra bone sar.) = ( 7? } (7° yr) = 9 X 10 °6)
In words, among 1 million persons so exposed, about 9 cases of induced bone
sarcoma would be predicted. For a natural incidence rate, averaging 1 x 10 5
per year, about 750 naturally-occurring bone sarcomas should occur among
1,000,000 persons during their lifespans, averaging 75 years.
A total of 21 226Ra-induced cancers (12 carcinomas + 9 sarcomas) are pre-
dicted for this exposed population.
228Ra: Cumulative risk from BONE SARCOMA:
Following the above procedures, but assuming that each uCi of 22®Ra is
equivalent to 2.5 yCi 226Ra for bone sarcoma-induct ion (Ro 78), about 2.5 x
9 = 22 bone sarcomas should accumulate among the lifespan of 1,000,000 per-
sons, each with an intake to blood of 1 pCi 228Ra per day.
221*Ra: Cumulative risk from BONE SARCOMA:
The intravenous injection of 1 yCi 22lfRa gives a cumulative a-particle
dose of about 0.2 rad averaged over the 7-kg marrow-free skeleton of adult
reference man (Spies 70)- Thus, the adult intake to blood of 1 pCi 221+Ra/day
— S -s
or 365 x 10 uCi/year, gives an average skeletal dose rate of 73 x 10 rad/
year to adult man. We assume, however, that the skeletal concentrations of
short-lived 221tRa are 3 times higher in children under age 16, and 2 times
higher in adolescents 16 thru 20 years than in adults, based on the fractions
-------
- 15 -
of intravenous 1y-injected 22ltRa estimated to decay in juvenile, adolescent,
and adult bone (Spies 70). The dose within each age interval is:
Years in Dose rate Rads in
Age interval (rads/yr) interval
under 16 16 3 x 73 x l(f® 0.00350
16 thru 20 5 2 x 73 x 10"® 0.00073
21 to 70 h3 1 x 73 x 10 0.00358
Cumulative 70 0.00781
Using the cumulative risk coefficient of 200 bone sarcomas/106 person«rad of
average skeletal dose from 22**Ra, which life-table analysis has shown to be
fairly constant for injection ages ranging from 1-70 years (May 83):
Cumulative risk _ ,200 bone sar. > fn ,\ , , ,„-6
(224Ra bone sar.) = <10* person-rad'(°'00781 rad) - '-6 x ,0
-------
- 16 -
results in the intake to blood of 1 pCi Ra. If the concentration of radium
in a local water supply is 5 pCI Ra per liter, and 1 liter per day of
local water is consumed, the intake to blood would be 1 pCi Ra per day from
this source.
DAILY FLUID INTAKE FROM THE LOCAL WATER SYSTEM
We assume an average adult consumption of 1 liter per day from the local
water system (tap water), based on the average reported consumption of 32.7
ounces (0.97 liters) per day from a survey of some 200 residents of New
Mexico, published by Johnathan M. Mann, State Epidemiologist for the State
of New Mexico (Man 83). Consumption of local water includes the direct drink-
ing of tap water, the tap water in home-prepared drinks such as coffee and tea,
and the tap water in locally-prepared foods such as soups and cooked vegeta-
bles. It excludes water from sources outside the local water supply such as
milk, bottled drinks, canned food, and the water formed within the body by
the metabolic oxidation of food.
The daily drinking of tap water alone by reference persons is given in
ICRP Publication 23 as 150 ml for adult man, 100 ml for adult woman, and 200
ml for a 10-year-old child (ICRP 75). However, the total fluid intake from
the local water supply exceeds the direct drinking of tap water.
The Environmental Protection Agency commonly assumes a daily consump-
tion of 2 liters of water per day (Co 83). This was based on an upward
rounding from the average per capita water (liquid) consumption of 1.63
liters per day, as calculated from a survey of 9 different literature
sources (NAS 77). Examination of these sources revealed that the 1.63
liters per day included not only local tap water but nonlocal sources such
-------
- 17 -
as bottled drinks. Thus, from the local water supply alone, a lower average
fluid intake of about 1 liter per day seems reasonable, although improved
evaluations of the true average are needed.
Assuming a linear dose-response relationship for cancer induction at
environmental levels, the population risk is determined by the average intake,
not by the maximum intake to the highest-exposed individual. For example,
for a local water supply containing 100 pCi 226Ra per liter, the predicted
number of induced cancers is the same in a population of 1,000,000 persons,
each drinking 1 liter per day from the system, as in 1,000,000 similar per-
sons of whom 500,000 drink 0.5 liters per day and the other 500,000 drink
1.5 1iters per day.
In contrast, to prevent a presumed threshold type of chemical toxicity,
it is appropriate to consider a higher-than-average daily intake such as the
1.7 liters per day used by Wrenn, et at., to evaluate the threshold for chem-
ical damage to the kidney from uranium (Wr 84).
To limit the risk of radionuclide-induced cancer, some regulators may
prefer inflating the average daily intake from a local water supply to 2
liters per day to provide a "hidden" safety factor. We prefer to estimate
the risks as realistically as possible. We believe that when safety factors
are desired, they should be given openly rather than be disguised.
ESTIMATED RISK FROM LIFELONG INTAKE OF URANIUM
Whereas the risk from the radium isotopes, 226Ra, 228Ra, and 22l*Ra, is
based on the observed toxicity in humans and can be expressed per yCi intake,
no direct information exists on cancer-induction by uranium in people.
Therefore, the risk from uranium will be inferred from the risk for long-
1ived, bone-volume-seeking 226Ra.
-------
- 18 -
For the lifespan dietary intake of uranium at environmental levels,
most of the retained uranium in the skeleton is uniformly distributed through-
out the volume of mineralized bone (Wr 84). Thus, the bone sarcoma risk per
rad of average skeletal dose from bone-volume-seeking uranium isotopes with
half-lives exceeding 1,000 years, is more likely to resemble that from long-
lived, bone-volume-seeking 226Ra (which produces no bone-surface-seeking
daughters) than from either 22l*Ra (much of which quickly decays while still
on bone surfaces) or 228Ra (which decays to surface-seeking 228Th and 22ltRa).
In adult and adolescent males, the skeletal content of 226Ra at equilib-
rium is about 25 times the daily ingestion (Steh 56, Wr 84). Thus, the
adult ingestion of 5 pCi 226Ra per day should produce an equilibrium skeletal
content of about 125 pCi 226Ra. The 226Ra, with 30% retention of its
daughters, 222Rn, 218Po, 21tfPo, and 210Po (tCRP 79), released a summed 12.11
MeV of a-particle energy per 226Ra disintegration.
For the five uranium isotopes with half-lives exceeding 1000 years, the
individual half-lives and ot-energies per decay are as follows (ICRP 83):
Half-life
a-energy
1sotope
(yr)
(MeV)
233J
1.59 x 10s
4.82
23<*u
2.44 x 105
4.76
23 5jj
7.OA x 108
4.40
236u
2.3^ x 107
4.50
238u
4.47 x 109
4.19
The individual a-energies cluster closely around the group average of 4.5
MeV, which we will use.
For consistency with our example for radium, consider the ingestion of
5 pCi of uranium per day. The equilibrium skeletal content of long-lived
-------
- 19 -
uranium lies between 1 and 35 times the daily ingestion, with 11 as a best
present estimate (Wr 8k). Thus, in our example, the best present estimate
of the equilibrium skeletal content would be 11 x 5 = 55 pCi of uranium.
The cumulative lifespan risk from the daily ingestion of 5 pCi uranium
per day, can be computed from the cumulative lifespan risk of 9 bone sar-
comas/106 persons each ingesting 5 pCi 22SRa per day (Eq. 13) and the equi-
librium skeletal burdens and summed a-energies:
Lifetime risk _ r9 bone sar.i > 55 pCi Uran. in skel.w k.5 MeV from Uran.
(Long-lived U) ~ 106 persons 125 pCi 226Ra in ske1.12.11 MeV 226Ra + dau.
_1.5 bone sarcomas /.qx
105 persons
In 1 million persons so exposed, about 1.5 cases of uranium-induced bone sar-
coma are predicted, assuming that the ratio of skeletal content to daily in-
gestion is 11 (best present estimate). Since the true ratio is within the
range of 1-35, the corresponding number of predicted bone sarcomas could
range from 0.1-5. In contrast, about 750 naturally-occurring bone sarcomas
are expected among these 1,000,000 persons during their lifetimes.
232U has been omitted from these calculations because it decays to
surface-seeking 228Th. The toxicity per rad from 232U with its decay pro-
ducts might be similar to that for 228Ra, which also decays to 228Th.
The build-up of 222Rn gas is not a problem with pure uranium isotopes,
and thus the risk from head carcinomas is considered negligible compared to
that from bone sarcomas. Head carcinomas are not a problem except when 226Ra
deposits in bone.
-------
- 20 -
Whereas the risk from induced soft-tissue cancers is known to be small
for internally-deposited 226Ra, 228Ra, and Z2kRa, the possible risk of soft-
tissue cancers induced by uranium isotopes is unknown, but by analogy with
radium is not predicted to be large.
The predicted numbers of induced cancers for radium and uranium iso-
topes, based on a linear dose-response, is shown in Table 1. However, if
the true dose-response for a-particle-induced cancers in humans at low doses
varies with the square of dose, virtually no cancers would be expected from
environmental levels of radium or uranium.
THE COST OF REMEDIAL MEASURES
There are two kinds of costs in remedial measures: (a) the direct fi-
nancial cost, and (b) the biological cost which, in part, becomes a financial
cost. Both are important and should be considered. Remedial action at en-
vironmental levels should not be taken unless the benefit of the remedial ac-
tion exceeds the total cost of this action (financial and biological).
Specifically, an important practical consideration is to assess the
total costs of alternative methods of removing radium from drinking water.
One way is to use ion exchange columns that remove divalent cations such as
|—^ | ^ -j, | ^
Mg , Ca , Sr , Ba , and Ra from water, replacing them with sodium ions,
Na+. But what is the health impact of increasing the sodium concentration in
drinking water, especially on highly susceptible individuals such as persons
with high blood pressure who must minimize their sodium intake? Is it proper
to make a small reduction in an assumed cancer risk from radium at the ex-
pense of killing a much larger number from sodium-augmented cardiovascular
disease? It would be much more difficult to detect a moderate numerical
-------
- 21 -
increase in deaths from cardiovascular disease, presently about 50% of U.S.
deaths (Am 8l) than from bone sarcomas and head carcinomas (only about 0.1%
of U.S. deaths). We urge caution until the total costs are better known.
Recently the Environmental Protection Agency sponsored a symposium on
Drinking Water and Cardiovascular Disease (Ca 80). It is our impression,
that relative to the effects of radium at low doses, the health effects of
sodium addition, while difficult to assess, could be substantially greater.
Often, it may be more feasible to switch to a less-contaminated source
of drinking water than to remove the radioactivity from a local supply.
RECOMMENDATIONS FOR FUTURE RESEARCH
1. Continue to follow the health of the U.S. radium persons internally con-
taminated with 226Ra and 228Ra, and the German patients injected with
22kRa, until virtually all of them have died. Until this is done, uncer-
tainty will remain on the total health effects from radium isotopes in
human beings.
2. Reliably establish the ratio between uranium in the skeleton at equilib-
rium and the daily ingestion of uranium.
3. Using laboratory animals, evaluate the bone sarcoma risk per rad from
high-specific activity uranium such as 233U, and compare this risk with
that from 225Ra. Is the assumption of equal risks per rad really valid?
4. Emphasize basic research on the nature of radiation-induction of cancer.
An important endpoint should be better prediction of the actual effects
at low doses, based on observed effects at higher doses.
5. Explore alternative ways to reduce intake of radioactivity by the popula-
tion if, indeed, such reductions are to be seriously considered. The
-------
- 22 -
biological, as well as engineering costs, of these alternatives should
be established well enough that the decisions should have a net positive
benefit.
ACKNOWLEDGMENT
We thank C. Richard Cothern and William L. Lappenbusch for their dedi-
cated scientific interest in evaluating the risk from environmental levels
of radioactivity, and their encouragement to us in making our analysis. We
also thank Frederich A. Hodge for his valuable input as a member of the Sub-
committee on Risk and Nancy Peixhot for her skillful typing of the manuscript.
-------
- 23 -
REFERENCES
Am 81 American Cancer Society, 1981, Cancer Facts and Figures, 777 Third
Avenue, New York City, p. 7*
Ca 80 Calabrese E.J., Moore G.S., Tuthill R.W., and Siegler T.L., Eds.,
1980, Drinking Water and Cardiovascular Disease, 326 pages,
Pathotox Publishers, Park Forest South, IL 60466, USA.
Co 83 Cothern C.R., Lappenbusch W.L., and Cotruvo J.A., 1983, "Health
effects guidance for uranium in drinking water," Health Phys. 44,
Sup. 1, 377-384 (see Table 3f p. 38l).
Ev 66 Evans R.D., 1966, "The effect of skeletally deposited alpha-ray emit-
ters in man," Brit. J. Radiol. 39, 881-895 (see p. 892).
Fin 69 Finkel A.J., Miller C.E., Hasterlik R.J., 1969, "Radium-induced
malignant tumors in man," in: Delayed Effects of Bone-Seeking Radio-
nuclides (Edited by C.W. Mays, W.S.S. Jee, R.D. Lloyd, B.J. Stover,
J.H. Dougherty, and G.N. Taylor), pp. 195-225 (Salt Lake City: Uni-
versity of Utah Press).
Fis 79 Fisenne I.M. and Keller H.W., 1979, "The world wide distribution of
radium-226 in ashed human bone," in: EML-356, Environmental Measure-
ments Laboratory, 376 Hudson Street, New York City, pp. 1-47 to 1-58.
Gu 83 Gustafson P.F. and Stehney A.F., 1983, "Radium-induced malignancies,"
in: Environmental Research Division Annual Report, ANL-83-IOO,
Part II, Argonne National Laboratory, Argonne, IL, pp. 159-165.
ICRP 75 International Commission on Radiological Protection, 1975, Report of
the Task Group on Reference Man, ICRP Publication 23 (Oxford:
Pergamon Press), p. 360.
-------
ICRP 77
ICRP 79
ICRP 84
Mai 69
Man 83
May 83
- 24 -
International Commission on Radiological Protection, 1977, Recommen-
dations of the ICRP, ICRP Publication 26, Annals of the ICRP, Vol. 1;
No. 3 (Oxford: Pergamon Press), p. 10.
International Commission on Radiological Protection, 1979, Limits
for Intakes of Radionuclides by Workers, ICRP Publication 30, Part 1,
Annals of the ICRP, Vol. 1, No. 3/4 (Oxford: Pergamon Press), pp.
98-99.
International Commission on Radiological Protection, 1983, Radionu-
clide Transformations, Energy and Intensity of Emissions, ICRP Pub-
lication 38, Annals of the ICRP, Vol. 11-13 (Oxford: Pergamon Press).
Maletskos C.J., Keane A.T., Telles N.C., and Evans R.D., 1969,
"Retention and absorption of 22I+Ra and 23itTh and some dosimetric
considerations of 22lfRa in human beings," in: Delayed Effects of
Bone-Seeking Radionuc1ides (Edited by C.W. Mays, W.S.S. Jee, R.D.
Lloyd, B.J. Stover, J.H. Dougherty, and G.N. Taylor), pp. 29-49
(Salt Lake City: University of Utah Press).
Mann J.M., 1983, "Report of the working group on the public health
impact of uranium mill tailings," Office of Epidemiology, Health
and Environmental Department, State of New Mexico, P.O. Box 968,
Sante Fe, NM, p. 2.
Mays C.W. and Spiess H., 1983, "Epidemiological studies of German
patients injected with 22I+Ra," in: Epidemiology Applied to Health
Physics, Proceedings of the 16th midyear topical meeting of the
Health Physics Society, Albuquerque, New Mexico, 9—13 Jan. 1983,
pp. 159-166 (National Technical Information Service, CONF 83OIOI,
Springfield, VA).
-------
- 25 -
May 84 Mays C.W. and Spiess H. , 1984, "Bone sarcomas in 221+Ra patients,"
in: Radiation Carcinogenesis: Epidemiology and Biological Signif-
icance (Edited by Boice J.D. and Fraumeni J.F.), pp. 241-252
(New York: Raven Press).
Mo 78 Mole R.H., 1978, "The radiobiological significance of the studies
with 221+Ra and Thorotrast," Health Phys. 35, 167-174.
Mu 83 Muth H. and Globel B., 1983, "Age dependent concentration of 226Ra
in human bone and some transfer factors from diet to human tissues."
Health Phys. 44, Sup. 1, 113-121.
NAS 77 National Academy of Sciences, 1977, Drinking Water and Health,
(Washington, D.C.: National Academy Press), p. 11.
NAS 80 National Academy of Sciences, 1980, The Effects on Populations of
Exposure to Low Levels of Ionizing Radiation ("BE 1R III") (Wash., D.C.:
National Academy Press), p. 332.
No 55 Norris W.P., Speckman T.W., and Gustafson P.F., 1955, "Studies on
the metabolism of radium in man," Am. J. Roentgenol. Radium Ther.
Nucl. Med. 73, 785-802.
Ra 65 Rajewsky B., Belloch-Zimmermann V., Lohr E., and Stahlhofen W.,
1965, "226Ra in human embryonic tissue, relationship of activity to
the stage of pregnancy, measurement of natural 226Ra occurrence in
the human placenta," Health Phys. 11, 161-169.
Ro 78 Rowland R.E., Stehney A.F., and Lucas H.F., 1978, "Dose-response
relationships for female radium dial workers," Radiat. Res. 76,
368-383.
Ro 83 Rowland R.E., Stehney A.F., and Lucas H.F., 1983, "Dose-response
relationships for radium-induced bone sarcomas," Health Phys. 44,
Sup. 1, 15-31.
-------
Spier 83
Spies 70
Spies 78
Steb 84
Steh 56
Ub 84
Wi 83
- 26 -
Spiers F.W. , Lucas H.F., Rundo J., and Anast G.A., 1983, "Leukemia
incidence in the U.S. Dial Painters," Health Phys. 44, Sup. 1,
65-72 (see pp. 67 and 69).
Spiess H. and Mays C.W. , 1970, "Bone cancers induced by 22ltRa (ThX)
in children and adults," Health Phys. 19, 713-729.
Spiess H., Gerspach A., and Mays C.W., 1978, "Soft-tissue effects
following 22lfRa injections into humans," Health Phys. 35, 61-81
(1978).
Stebbings J.H., Lucas H.F., and Stehney A.F., 1984, "Mortality from
cancers of major sites in female radium dial workers," Am. J.
Industrial Med, (in press).
Stehney A.F. and Lucas H.F., 1956, "Studies on the radium content
of humans arising from the natural radium of their environment,"
Proc. First Conf. on Peaceful Uses of Atomic Energy, Vol. II,
pp. 49-54 (Geneva: United Nations).
Ubell E., 15 January 1984, "How to live longer at any age," Parade
Magazine, 750 Third Avenue, New York, pp. 15-18 (Quoting estimates
from the National Center for Health Statistics, U.S. Dept. of Health
and Human Services).
Wick R.R. and Gossner W., 1983, "Incidence of tumors of the skeleton
in 22^Ra
-treated ankylosing spondylitis patients," in: Biological
Effects of Low-Level Radiation, International Atomic Energy Agency,
Vienna, pp. 281-288.
-------
- 27 -
Wr 77 Wrenn M.E., 1977, "Internal dose estimates," in: International
Symposium on Areas of High Natural Radioactivity (Rio de Janeiro:
Brazilian Academy of Sciences), pp. 131-157-
Wr 8A Wrenn M.E., Lipzstein, Durbin P.W., Still E., Willis D.L., Howard B.,
and Rundo J., 1984, "Uranium and radium metabolism," Health Phys.
(this issue).
-------
- 28 -
TABLE 1. Best present estimates of induced cancers during
the lifetimes of 1,000,000 persons, each with a
constant skeletal concentration equal to the
adult equilibrium level resulting from the daily
ingestion of 5 pCi of a radium or uranium isotope
Radionuclide
rBone sarcomas-.
106 persons
rHead carcinomas-,
10b persons
226Ra
9
12
228Ra
22
*
1.6
0
22H a
0
233,j X
23"U j
235u L
1.5
0
236u V
238u J
*The skeletal concentration of short-lived 22t*Ra was assumed 3
times higher in children under 16 years of age, and 2 times
higher in adolescents 16-20 years of age, than in adults.
**For these uranium isotopes, the estimated number of induced bone
sarcomas/106 persons ranges from 0.1-5-0 with a best estimate of
1.5.
-------
COMMITTEE ON COMPLIANCE AND POLICY
Chairman: John E. Regnier
Recorder: Robert Sullivan
Committee Members: Richard Blanchard
Melvin U. Carter
Edward Cowan
David Duncan
James Martin
-------
COMPLIANCE AND POLICY ISSUES AND RECOMMENDATIONS
RELATED TO REVISION OP THE NATIONAL INTERIM PRIMARY
DRINKING WATER REGULATIONS FOR RADIONUCLIDES
John E. Regnier,
Alabama Rural Water Association
Montgomery, AL 36105,
Richard Blanchard
U.S. Environmental Protection
Agency
P.O. Box 3009
Montgomery, AL 36193,
Melvin W. Carter
Georgia Institute of Technology
Atlanta, GA 30332,
Edward Cowan
U.S. Environmental Protection
Agency
Seattle, WA 98101,
James Martin
2604 Bedford Road
Ann Arbor, MI 48104,
and
Robert Sullivan
U.S. Environmental Protection
Agency
Office of Radiation Programs
Washington, D.C. 20460
ABSTRACT
This paper summarizes the deliberations and conclusions
of the Compliance and Policy Committee of the National Workshop on
Radiation in Drinking Water held in Easton, Maryland, May 24, 25 &
26, 1983. Prior to and during the workshop the Committee considered
a total of 32 possible compliance and policy issues and determined
that 22 were valid. The Committee developed positions on seven of
these and these positions are presented herein. The remaining 25
issues which were considered are also listed with the Committee's
evaluation of each indicated.
-------
The purpose of this paper is to present the results of
the deliberations of The Compliance and Policy Committee prior to
and during the workshop. Prior to the workshop, this committee
exchanged written and verbal communications on possible policy
issues. From these exchanges and from input from other commit-
tees and EPA a composite list of tentative issues was drawn up and
circulated for comment. During the workshop, this list was analyzed
as follows:
1) Was a true issue involved rather than a
matter requiring a technical assessment?
2) If a true issue was involved, was it of
a policy matter or a technical matter and
therefore, better addressed in a different
committee?
3) Were there additional issues that needed
consideration?
Subsequent to this analysis, the committee addressed those policy
issues it felt were of highest priority and developed a policy
statement and/or a recommendation on each.
In the remainder of this discussion, all issues considered
are presented, an indication is given of whether the committee felt
each was a valid policy issue, and policy statements are provided on
those issues judged to be most pressing.
POSSIBLE ISSUES
Prior to the workshop a tentative list of 29 possible
issues had been identified. During the workshop, as a result of
-2-
-------
input from EPA, other committees, and observers, three additional
issues were added. These issues, posed as questions, are tabulated
in the Appendix.
As previously mentioned, the committee deliberated the
merit of each question to decide if it required more than a simple
technical assessment, and if so, were policy considerations involved.
The results of these deliberations are briefly stated in the remarks
column of the tabulation.
Although it is not summarized in the tabulation, the out-
come of the foregoing procedure was that 22 of the questions were
deemed valid policy issues, two were judged to be issues, but more
in the purview of other committees, and several were identified as
technical assessment problems or invalid issues.
POLICY STATEMENTS
Of the 22 valid policy questions, the committee had time
to develop a position on seven. In the following pages, each of
these seven and its accompanying policy statement are briefly
stated.
ISSUE: Is there a compliance problem?
REPORT: Monitoring data available from the
first nationwide monitoring of public water supplies shows no
compliance problems with man-made radionuclides. Substantially
all cities impacted by nuclear facilities are in compliance with
-3-
-------
the Maximum Contaminant Level (MCL) for man-made radioactivity.
Current monitoring data for Ra-226 indicate that about
150 supplies exceed the MCL of 5 pCi/1. However, it has also been
projected that about 500 of the 59,812 total water supplies may
exceed this MCL which constitutes a compliance problem.
For Ra-228 a possible compliance probelm exists because
of undetected violations. Ra-228 concentrations are not determined
unless Ra-226 concentrations exceed 3pCi/l. A limited study indi-
cated that 40-50% of Ra-228 violations could go undetected because
of this coupling of Ra-228 with Ra-226.
ISSUE: Should EPA adopt a total dose equiva-
lent for the purpose of regulating radionuclides in the drinking
water?
REPORT: The Committee regards this as an
attractive alternative to setting of MCLs for individual radionu-
clides and recommends its adoption by the Agency. The total dose
equivalent selected should be based on the risk to health but
should take into consideration costs and technical feasibility.
Secondary limits in the form of radionuclide concentra-
tions yielding this dose equivalent should be provided by EPA.
ISSUE: Should ICRP-30 calculations and assump-
tions be used as a basis for revised regulations?
REPORT: The Office of Drinking Water should
-4-
-------
avail itself of the pertinent evaluations and recommendations of
recognized radiation protection authorities, such as the National
Academy of Science, National Council on Radiation Protection and
Measurements, and International Commission on Radiological Protec-
tion, for appropriate guidance. An example is ICRP-30 which
describes state-of-the-art metabolic and dosimetric models for
application to certain radiation protection problems.
ISSUE: Should drinking water standards for
radioactivity consider the broad issue of performance of waste
disposal sites?
REPORT: The Committee recognized that
drinking water standards become a reference point for waste disposal
regulations, but its considered view is that for policy reasons
laid out in the statute, it is inappropriate to derive drinking
water standards to control waste disposal. Equally significant to
the Committee is the corollary to this: that EPA in its standards
for RCA sites and radioactive waste sites make it clear that the
protection needs and authorities for those problems should derive
from those problem areas. EPA should not inappropriately apply
the drinking water standards to solve waste disposal problems
rather than develop requirements appropriate to those problems.
ISSUE: Should gross alpha and beta be used
as an MCL or a screening method?
-5-
-------
REPORT: The Committee believes that the
use of gross measurements as legally enforceable MCLs has inherent
problems because those measurements do not reflect appropriate
accuracy. On the other hand, they are useful as a screening tool.
EPA should provide a framework of MCLs for major specific radionu-
clides and discontinue the use of gross MCLs. The gross MCLs
should be replaced with nuclide-specific MCLs for which monitor-
ing and regulatory actions may be taken on the basis of gross
measurements used as a screening mechanism.
ISSUE: Can a method other than monitoring
radionuclide concentrations be used to obtain compliance with MCLs
such as inspection and certification of components, maintenance
records, or demonstration that the source of the drinking water is
of high quality?
REPORT: The Committee discussed this issue at
some length, expressing that whereas it may be desirable to allow
flexibility in achieving compliance, certain practical problems
such as system fluctuation and the certainty and public relations
worth of measured values may not make it reasonable. Considerable
concern was expressed about the technical validity of relying on
surrogates such as softness, pH turbidity, etc., as indicators of
systems performance and also whether one might not perform more
measurements to demonstrate system efficiency than would be required
by water-tap-measurements. Having no specific technical factors
to allay these concerns, the Committee reached no consensus as to
-6-
-------
whether this procedure may be achievable despite its seeming desir-
ability. This issue was left unresolved; EPA should examine it
further with regard to compliance programs, especially if the
frequency of compliance measurements and the number of nuclide-
specific MCLs increases.
ISSUE: Two waste-handling aspects exist regarding
removal of radioactivity in treatment of drinking water: (1) Does
radiation exposure of operators from treatment systems argue against
removal of such material and (2) disposal of waste products?
REPORT: For operators, OSHA regulates workplaces,
and have to assure that proper procedures are followed, including
keeping exposure records. Fundamentally, this is the same issue
faced by all standards related to processes that may require filtration
or on-line treatment. Control levels are struck on the basis that
public health protection is essential even though occupational
protection procedures may need to be augmented.
Similarly waste disposal is an accepted necessity
associated with all standards requiring on-stream treatment.
The Committee believes management of a concentrated waste better
than dispersion of material into the environment. For naturally
occurring radioactive waste, EPA's RCRA program would need to
assure that such collected materials are disposed of properly.
-7-
-------
SUMMARY
The Compliance and Policy Committee feels that the seven
issues it addressed are of significant importance in assuring that
the health of the public is adequately and equitably protected from
insult by radionuclides in drinking water. Further, the brief
statements presented herein do not reflect the extent of discussion
that led to these statements nor do they reflect the many ramifica-
tions which would be involved in implementation of these policies.
The Committee also wishes to emphasize that the fact
that policy statements were not developed for the other 15 policy
issues considered does not imply that they are necessarily less
important. This omission resulted because of limited time avail-
able and we recommend that EPA address these issues by subsequent
workshops or alternate mechanisms.
-8-
-------
APPENDIX
TENTATIVE ISSUES CONSIDERED
ISSUE
1. Is there a compliance problem with existing
regulations?
2. If there is a compliance problem, is it
measurable in terms of risk, extent, dis-
tribution, and costs to correct?
3. Are treatment methods available for non-
compliance instances?
4. Does EPA have flexibility within the law
to regulate or not regulate?
5. Does EPA have a responsibility regarding
public education?
6. Do small systems require special treatment?
7. Do private single family well supplies
require special treatment?
8. What criteria should be used in making a
determination to regulate?
9. If EPA has the flexibility to do both,
when should health advisories versus MCLs
be used?
10. Should EPA use a linear or non-linear dose
response curve in assessing risk for
purposes of establishing standards?
11. Is there a threshold for the dose response
curve?
12. Should 2 liters/day or some other value be
used for water consumption?
13. Should there be protective action guides
for water exclusive of other media? If so,
how should they be set?
14. Are current monitoring frequencies accep-
table?
REMARKS
Valid Policy Issue
Technical Assessment
Technical Assessment
Valid Policy Issue
Valid Policy Issue
Valid Policy Issue
Valid Policy Issue
Valid Policy Issue
Valid Policy Issue
Valid issue but more
in purview of other
committees.
Technical Assessment
Technical Assessment
Valid Policy Issue
Valid Issue but more
in purview of other
committees.
-------
15. Should gross alpha be a screening method
or an MCL?
16. Can a method other than monitoring radio-
nuclide concentrations be used to ascer-
tain compliance with MCLs? For example,
is there a surrogate such as maintenance
of treatment plant efficiency?
17. Should ICRP-30 calculations and assump-
tions be used as a basis for revised regu-
lations?
18. Should recommended limits be protected of
groundwaters at waste disposal sites?
19. What is the proper framework for deciding
a justifiable level of risk for a given
radionuclide in drinking water?
20. Should chemical toxicity or radio toxicity
be the determining criterion for assessing
uranium exposure risk?
21. Should regulations be cast in the framework
of a total dose equivalent for members of
the public rather than incremental exposures
(MCLs) for individual radionuclides?
22. Is there a reasonable rationale for treating
man-made and naturally-occurring nuclides
differently?
23. Should EPA use a conservative, middle-of the
-road, or liberal philosophy in establishing
risk criteria on which to base regulations?
24. Is individual, point-of-use, treatment a
viable alternative to system treatment for
situations of non-compliance?
25. Should population risk or individual risk be
the determining criterion in standard
setting?
26. Should internal accumulation models use a
time dependent or equilibrium dose rate?
27. Is there a "propagation of conservatism" in
the rationale behind current standards?
28. Should only human subject validated models
be used for intake and accumulation?
Valid Policy Issue
Valid Policy Issue
Valid Policy Issue
Valid Policy Issue
Valid Policy Issue
Technical Assessment
Valid Policy Issue
Valid Policy Issue
Valid Policy Issue
Valid Policy Issue
Valid Policy Issue
Technical Assessment
Technical Assessment
Not a valid issue
-2-
-------
29. Should standards for different nuclides Valid Policy Issue
be based on equivalent risks or on vary-
ing criteria depending on economics and
other factors?
30. Should non-community supplies be regu-
lated for radionuclides
Valid Policy Issue
31. How can predictive modelling which would
estimate that radionuclides would not be
expected in geographic portions of the
U.S. be used in implementation and design
of regulations?
32. Does the concept of de minimus risk have
application in design of drinking water
regulations as developed under the Safe
Drinking Water Act?
Valid Policy Issue
Valid Policy Issue
-3-
-------
COMMITTEE ON SAMPLING AND ANALYTICAL METHODS
Chairman: Bernd Kahn
Recorder: Seymour Gold
Committee Members: Rolf M.A. Hahne
David McCurdy
William S. Moore
Jacob Sedlet
Stanley Ualigora,
Earl Vittaker
-------
Radiological Sampling and Analytical Methods for
National Primary Drinking Water Regulations
Richard L. Blanchard
Eastern Environmental Radiation Facility
U.S. Environmental Protection Agency
Montgomery, Alabamaa 36193
Rolf M.A. Hahne*
University Hygienic Laboratory
University of Iowa
Iowa City, Iowa 52242
Bernd Kahn
School of Nuclear Engineering and Health Physics
Georgia Institute of Technology
Atlanta, Georgia 30332
David McCurdy and Russell A. Mel lor
Yankee Atomic Environmental Laboratory
Yankee Atomic Electric Company
Framingham, Massachusetts 01701
Willard S. Moore
Department of Geology
University of South Carolina
Columbia, South Carolina 29208
Jacob Sedlet
Occupational Health and Safety
Argonne National Laboratory
Argonne, Illinois 60439
Earl Whittaker
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Las Vegas, Nevada 89114
* Currently with Dow Chemical Co., Midland, Michigan
-------
Abstract
Radiological sampling and analysis performed under the National
Interim Primary Drinking Water Regulations were evaluated for the EPA
Office of Drinking Water to consider whether any changes should be recom-
mended. The authors reviewed the analytical screening scheme; sample
collection, storage, and analysis procedures; selection of analytical
methods; reliability of results; and possible future needs.
The main problem in the program has been dependence on a screening
scheme of gross alpha particle activity measurement and 226r3 analysis
for predicting elevated 228r3 levels to determine compliance with the
maximum contaminant level (MCL) for radium. In some aquifers, 228r3
levels have been found to be unrelated to 226r3 levels. Several alterna-
tives are discussed to eliminate this problem. A secondary problem is
that the measurement for assuring compliance with the MCL for gross
alpha particle activity minus radium, radon, and uranium utilizes
chemical uranium analysis and assumes equilibrium of 238j ancj 234j#
Because some ground waters are known to be at disequilibrium,
radiometric uranium analysis is needed for those gross alpha particle
activities and chemical uranium values that could result in an erroneous
conclusion relative to the MCL. In addition, studies were recommended
for determining analytical uncertainties and assuring reliable sampling
and sample maintenance; improvements in the system for accepting methods
were suggested; and methods were identified for several radionuclides
not currently in the analytical program that may be needed to assure
absence of elevated radiation doses and could be useful for identifying
trace contaminants.
-------
Introduction
The National Interim Primary Drinking Water Regulations (NIPDWR,
40 CFR Part 141), which require determination of specified radionuclides
in public water supplies, have been in effect since 1977. Analytical
results for tens of thousands of supplies are now available. In 1982,
the Office of Drinking Water, U.S. EPA, established committees to advise
it concerning the effectiveness of the current Regulations and any
changes that might be needed. Reviewed here is the sampling and
analysis program.
The regulations can be met by collecting water samples at specified
intervals and analyzing either individual samples or composites from
each supply according to two analytical schemes. One of these, applied
to all samples, determines gross alpha particle activity, 226Raj an(j
228r3 in that order if the preceding result exceeds the specified screen-
ing level. Elevated gross alpha particle activity levels not attribut-
able to 226r3 may a-|so be analyzed for uranium. The other scheme is for
relatively few surface water supplies in large cities and communities
downstream from nuclear facilities to determine gross beta particle
activity as well as and 90$rj then, if screening levels are exceeded,
certain other man-made radionuclides must be analyzed. The schemes are
described in detail by another paper in this series (La84).
The results reported so far indicate the extent of the analytical
effort required for the continuing program. In 170 of the approximately
47,000 public water supplies for which results have been reported, 226r3
concentrations exceeded the Maximum Contaminant Level (MCL) of 5 pCi/1;
about 350 additional elevated values are expected, although only 13,000
supplies remained unreported (Co83a). Very few supplies exceeded the 15
3
-------
pCi/1 MCL for alpha particle activity less radon and uranium.
Occasionally, 232jh, 210po or 224r3 contributed to the alpha particle
activity. Data in a few reports suggest that approximately 250 water
supplies may contain 228r3 in excess of 2 pCi/1 (Co83a, Gi81, Ki82,
Kr82, Mi80). Sufficient results have been accumulated to indicate that
concentrations of 228r3 are not related to 226r3 in some aquifers (Gi81,
Ki82, Kr82, Mi80). Several investigators have reported that radium
concentrations in specific wells remain fairly constant (Gi81, Ho81,
Ki82, Kr82, Mi80), with variations only as high as 20%.
In 60,000 drinking water supplies, uranium is projected to occur at
the following concentrations (Co83): 25 to 650 exceed 20 pCi/1, 100 to
2,000 exceed 10 pCi/1 and 2,500 to 5,000 exceed 5 pCi/1. Most water
supplies in which uranium concentrations exceed 5 pCi/1 serve small
communities.
No man-made radionuclide has so far exceeded its MCL (Co83b).
Only a single water supply with a man-made radionuclide at detectable
levels — tritium at a concentration of 3,000 pCi/1 (CI83) -- has been
reported.
Elevated concentrations of 222rp have been found in a number of
underground water supplies (Ki82, He84) in a study unrelated to these
Regulations. Some concentrations greatly exceeded levels supported by
dissolved 226r3> /\ few measured values of 210pb and 210po were general-
ly below 1 pCi/1 (Ho83). 224r3 has been looked for and found unsuppor-
ted by 228yh (Mo83). Elevated thorium levels have not been found to any
extent despite the abundance of thorium in the earth's crust.
The results to date indicate that 226r3j 228Ra> 224r3> 222r0j an(j
uranium are the only natural radionuclides likely to be in public water
4
-------
supplies at concentrations that may be of health concern. Elevated con-
centrations of natural radionuclides occur mostly in underground water
supplies associated with certain geological formations.
The many results of these analyses (Co83a) reported by individual
states, together with published descriptions of and comments on these
activities, (Kr82, Mi80, C183) permit evaluation of the applicability of
these schemes. Results and observations were also reported at the
National Workshop on Radioactivity in Drinking Water sponsored by EPA on
May 24 - 26, 1983, in Easton, Maryland. Based on this information,
consideration was given to the reliability of the currently prescribed
sampling and analysis program, its effectiveness in using the best
techniques available for the purpose, and the extent to which it
provides complete radiological information. Where problems are believed
to exist, recommendations are given for changes, additions, or
developmental studies.
Discussion
Screening Procedures
The currently prescribed regulations for radionuclides in drinking
water are found in 40CFR Parts 141.26(a)(l)(i) and (ii) and
141.26(b)(l)(i) and (4)(i) (La84). These separately address naturally-
occurring radioactivity and man-made radionuclides. Several problems
concerning the screening procedure for the former are discussed here.
In the general absence of detectable levels of man-made radionuclides,
no recommendations are presente'd concerning this screening scheme.
The problems concerning the gross alpha particle activity screening
scheme are, in order of importance; (1) the scheme does not necessarily
indicate whether 228r3 is at concentrations that contribute to exceeding
5
-------
the MCL for radium; (2) the gross alpha particle activity and 226Ra con-
centrations may differ by more than the random error due to counting
statistics when only 226r3 is present; and (3) the gross alpha particle
activity and uranium concentrations may differ by more than the random
error due to counting statistics when only uranium is present. Several
approaches to replacing the screening scheme for 228r3 are discussed
briefly in this section; some procedures that may be sufficiently cost-
effective for widespread use as replacements are described in the Analy-
tical Procedures section. Possible improvements for gross alpha parti-
cle activity measurement are also discussed, although it should be recog-
nized that this measurement is inherently less reliable than indicated
by the counting error. For comparing uranium concentrations with gross
alpha particle activity, possible nonequi 1 ibrium between 238j antj 234y
is an additional problem.
Radium-228. In the screening procedure, a value of the gross alpha
particle activity not in excess of 5 pCi/1 is taken to imply a radium
concentration not in excess of 5 pCi/1, and a 226r3 concentration not in
excess of 3 pCi/1 implies a 228r3 concentration not in excess of 2 pCi/1.
In the absence of other information, the 226Ra:228Ra ratio was expected
to be 3:2 or greater, so that water systems with 226r3 not exceeding 3
pCi/1 would not exceed the 5 pCi/1 MCL for these two radium isotopes of
interest. This has proven not to be the case.
Ratios of 226Ra;228Ra ranging from over 14 to less than 0.07 have
been reported (Mi80, Kr82, Ki820 and values below 3:2 are relatively
common. Thus, the screening scheme can not be depended on to assure the
absence of total radium concentrations above 5 pCi/1 unless the 226r3:
228r3 ratio can be predicted for a particular aquifer.
6
-------
The screening scheme was developed to avoid requiring 228r3 analy-
ses for numerous samples. The 228r3 method (Kr76) specified in Part
141.25 is relatively long and expensive. Neither 228r3 nor its short-
lived daughter, 228/\c^ emit alpha particles for detection by the gross
alpha particle activity measurement. The next radionuclide in the decay
chain, 228-n^ emits alpha particles but accumulates slowly because of
its 1.9-year half life. It may also be deposited on container walls
because of the insolubility of thorium. Its daughter, 224r3> emits
alpha particles but is generally associated with 228jh because of its
relatively short half life (3.64 d).
In the absence of any nationwide 226Ra;228Ra ratio, one of the fol-
lowing three approaches should be considered for replacing the 3:2 ratio
assumed in the screening scheme:
(1) A national study of aquifers that contain 228Rn could permit
implementation of the requirement that 228r3 be measured where
its presence is expected. Even a few regional studies might
be sufficient to demonstrate patterns of elevated 228r3 levels
for predicting aquifers with such elevated levels.
(2) Review of available 228r3 or total radium procedures (see
Analytical Procedures) may permit selection of a method
that has acceptable levels of effort and cost for process-
ing every sample.
(3) One or more screening procedures for 228r3 or total radium may
be found for replacing*the current screening scheme. For
example, gross beta particle activity measurement or total
radium measurement after selected intervals for ingrowth of
7
-------
226Ra and 228Ra progeny could be applied if their detection
limits were sufficiently low.
Radium-226. The results of the gross alpha particle activity meas-
urements may differ from 226r3 results in the absence of other radio-
nuclides by more than the specified 95% confidence level value based on
random counting error because several sources of error are currently not
considered. These errors can occur because (1) the detector is calibra-
ted with a different radionuclide, (2) the correction for self-absorp-
tion is only approximate, or (3) short-lived progeny of 226r3 accumulate
to various degrees in the sample for gross alpha particle activity.
Hence, the requirement in Section 141.26(1)Cii) that 226r3 be measured
if the gross alpha particle activity exceeds 5 pCi/1 may result in over-
looking some water supplies in which the 226r3 concentration exceeds
the MCL by several pCi/1. Suggestions for improved measurements when
the total solid content of a sample is high and for using specific stand-
ard radionuclides are given below. Beyond this, the gross alpha parti-
cle activity must be recognized as an estimate.
Uranium. For 238u pius 234j radioactivity determined by chemical
uranium analysis and calculated by multiplying the analytical result in
micrograms by the equilibrium factor of 0.67 pCi/pg, the result will be
in error if the two isotopes are not in equilibrium. The actual factor
can range from 0.33 pCi/yg (no 234u) to at least 7 pCi/^g (234y;238y =
20). Hence, the calculated gross alpha particle activity, excluding
226Raj 222r0 an(j uranium, could" be in error if based on chemical determi-
nation of uranium and assumption of equal activity of 238(j an(j 234^ At
234u:238u ratios below 1.0, the uranium activity is overestimated and
the resulting "gross alpha particle activity minus uranium" is falsely
8
-------
low. The above comments concerning the reliability of the gross alpha
particle activity also apply here.
Analytical Procedures
Currently approved analytical methods for drinking water compliance
monitoring are referenced in the NIPDWR and are published in EPA's
Interim Radiochemical Methods Manual (Kr76), Standard Methods (APHA71,
ASTM75), or the HASL-300 Manual (Ha75). The following improvements are
suggested for consideration.
Gross alpha and gross beta particle activity. The national program
has relied heavily on gross alpha and gross beta particle activity
screening of water samples. The current practice of distributing two
unspecified radionuclides combined for gross alpha and beta particle
activity in the EPA cross-check program for radionuclides in water, how-
ever, can result in consistent calibration error. Performance evalua-
tion studies (De83) show that numerous laboratories have had problems
with gross beta activity measurements of cross-check samples that con-
tain the usual 241/\m ancj 137qs# Reliability of results might be improv-
ed if the radionuclides in these cross-check samples were selected for
calibrating detectors, and were identified so that the user could deter-
mine their suitability for a specific isotope calibration.
A pure alpha-particle emitter such as 230jh is recommended as a
standard and cross-check solution for gross alpha particle activity.
The 241/\m currently used in these cross-check samples emits numerous
conversion electrons and soft X rays that are detected as beta particles
and can be attributed to the beta particle emitter. The 137qs usecj in
the same solution for gross beta particle activity is acceptable if the
9
-------
radionuclide is identified so that the presence of an additional 9.5
percent conversion electrons from 137mga can be considered in calibrat-
ing for other isotopes.
Some public drinking water supplies have relatively high dissolved
solids levels, which cause self-absorption of alpha particles and great-
ly reduce the sensitivity of the gross alpha particle activity measure-
ment. For a gross alpha activity screening test with a 20-cm2 planchet,
a dissolved solids concentration limit of 500 mg/1 (25 mg/cm^) is recom-
mended. Barium sulfate/iron hydroxide coprecipitation (Li82) may be
considered for gross activity measurement if samples with higher dissolv-
ed solids are involved. This method is being evaluated by EMSL-Las
Vegas. In the range 0 -.25 mg/cm^, EPA could assist in improving data
reliability either by determining self-absorption factors for the more
common alpha-particle activity detectors or by distributing calibrated
solutions for performing this determination.
Radium-228. At present, two approved test procedures can be used
nationwide for measuring 228r3 concentrations in drinking water samples.
Both methods are technique-dependent, time-intensive and costly. Many
laboratories that participate in the EPA EMSL-Las Vegas intercomparison
and performance studies for 228r3 have been using the method referenced
in the NIPDWR (Kr76). Tables 1 and 2 show the high fraction of unaccept-
able 228r3 results during four-year periods of intercomparison and
performance tests.
Another problem with the 228r3 method is possible contamination by
the 90y daughter of 90sr. Tables 1 and 2 show the interference when the
samples are analyzed by the approved method. Performance samples in
Table 1 contain a complex mixture of radionuclides, including 90Sr-90y.
10
-------
Intercomparison samples in Table 2 contain only 226r3 ancj 228r3 p-|us
their progeny, except when 25 pCi of 90$r per liter were added in the
9/79 study to examine the effects of 90y daughter interference.
Other methods that are becoming available appear to be improvements
over the approved methods (No81) and should be evaluated. Many of them
simultaneously determine 226r3. one such method uses beta-gamma coinci-
dence counting after radium coprecipitation with barium sulfate (Mc81a).
The method requires a 1-liter sample and minimal chemistry in order to
meet the required sensitivity of measurement for 228r3# jhe instrument
consists of a well-type Nal(Tl) detector for gamma rays, a thin plastic
scintillator for detecting beta particles, and a coincidence pulse
analyzer.
Another method also coprecipitates radium with barium sulfate, then
dissolves the precipitate in alkaline EDTA, mixes the solution with li-
quid scintillator solution, and counts the low-energy beta particles of
228r3 (As81). The author claims adequate sensitivity with one liter
samples. A modification of this method has been tested by EPA EMSL-CIN
(Ve83).
A third method uses a 3-liter sample and the same radium coprecipi-
tation. It collects the precipitate on a filter, allows two days for
228Ac
ingrowth, and then analyzes for 228r3 a gamma-ray spectro-
meter. The 986-keV gamma-ray emitted by the ingrown 228/\c is measured.
The State of Arkansas is seeking approval for use of this method (He82).
In a fourth method, radium, is collected from a multiliter sample on
a column of fibers coated with manganese dioxide (Mi81). Radium is then
dissolved from the column, coprecipitated with barium sulfate, collected
on a filter paper, and analyzed for 228r3 by counting gamma rays emitted
11
-------
by the ingrown 228ac> as in the above-cited method. After a period of
weeks to months, the ingrown 228jh plus 224r3 can also be measured with
an alpha-particle detector. As an alternative, the fibers are compress-
ed and 228/\c -jS counted directly with a gamma-ray spectrometer.
Uranium. According to Section 141.15 part (b) of the NIPDR, if the
gross alpha particle activity exceeds the 226r3 concentration by 15
pCi/1, then a uranium determination is necessary. The uranium
measurement referenced in the NIPDWR is a fluorometric method which
measures uranium mass (ASTM75).
Because 234u:238j isotopic ratios have been found to vary in ground-
water systems from less than 1 to 20 (0s82), the factor of 0.67 pCi/yg
for a 1:1 ratio may not apply. Therefore, the fluorometric method can
only be reiied upon to give the concentration of uranium in yg/L. If
the radioactivity level is needed, the revised regulations should refer
to a radiochemical method that measures the total uranium alpha particle
activity of the sample. Such a method has been published (Kr80). The
method has been multilaboratory tested in a collaborative study and
precision and accuracy have been estimated.
If the gross alpha particle activity in excess of the measured
226r3 concentration is consistent with the uranium fluorometric (mass)
measurement at a 0.67 pCi/yg conversion factor, then results of the fluo-
rometric method are acceptable. Moreover, some uranium measurements in
ground water suggest that ratios greater than 5:1 usually occur at low
concentrations (1 pCi/1 and les^) (0s82). Under these circumstances,
the ASTM method or several alternatives are usable, although any such
alternatives would need official method approval. The Prescribed Proce-
dures (Kr80) have a carbonate/fluoride fusion fluorometric method for
12
-------
uranium with a detection limit of 0.1 pg/1. With laser excitation, the
detection limit without purification or concentration is 0.05 yg/1
(Pe82).
Iodine-131. The MCL for 131j in the NIPDWR is 3 pCi/1 and the re-
quired sensitivity of measurement is 1 pCi/1. Gamma spectrometry analy-
sis of drinking water samples usually is not sufficiently sensitive with-
out concentrating the in the sample. A radiochemical concentration
method for 131j that has sufficient sensitivity was included in EPA's
Interim Radiochemical Methodology (Kr76) and repeated in Prescribed
Procedures (Kr80). When that method was multilaboratory tested in a
collaborative study, however, low yields were identified as a problem.
StrOntium-90. The 90$r method (Kr76) still appears to be difficult
for many laboratories (EPA81) despite many years of use. Removal of
calcium interference with concentrated nitric acid may be the major prob-
lem because the acid concentration is crucial in separating strontium
from calcium by solubility difference. Other published methods, for
example, purification of strontium by ion exchange (Po67), could be eval-
uated as alternatives.
Sensitivity and Uncertainty
The NIPDWR (EPA76) addresses these topics only briefly, indicating
in Section 141.25 (c) that the listed detection limit is the concentra-
tion which can be counted with a precision of ± 100% at the 95% confi-
dence limit (1.96 times the standard deviation of the net counting rate
of the sample.) Section 141.25 (c) also specifies that the detection
limit for the combined 226r3 ancj 228r9 concentration (MCL is 5 pCi/1)
13
-------
shall not exceed 1 pCi/1 and for gross alpha activity (MCL is 15 pCi/1)
the detection limit shall not exceed 3 pCi/1.
Additional information on sensitivity, accuracy, and uncertainty
is given in the EPA's Prescribed Procedures (Kr80). For each procedure
(except actinides), the calculated detection limit and the precision and
bias determined in a collaborative or a single-laboratory replicate test
study are given. The equation for calculating the counting error and
three methods for determining the counting time required to obtain the
required sensitivity are given in appendices to that report.
Under the current drinking water regulation, uncertainty is not a
factor in determining compliance, except in one instance. According to
Section 141.26(a)(l)(i), gross alpha particle activity may be substitu-
ted for radium analysis if the result does not exceed 5 pCi/1 at a confi-
dence level of 95 percent. Otherwise, Section 141.25 (d) states that
"to judge compliance with the maximum contaminant levels ... averages of
data shall be used and shall be rounded to the same number of signifi-
cant figures as the maximum contaminant levels."
To characterize the reliability of results, detection limits and
uncertainty of methods applied under normal conditions should be deter-
mined and reported by each participating laboratory. The methods of
calculating uncertainty (Be69, Cu78, Ma78) and the detection limit need
to be described by the EPA in greater detail than given in Section
141.25(c) to assure consistent application by individual laboratories.
It is important to include not only the random counting error as is now
done, but also random and systematic errors in sample collection and
storage, chemical procedures, and detector calibration and use.
Analytical uncertainty can be determined from the laboratory cross-
14
-------
check program, method validation studies, and tests with known sources,
blanks, and replicate samples within a laboratory. When sufficient
empirical data have been obtained, the EPA should replace its currently
specified detection limits based on counting precision with limits based
on these measurements. It should also use this information to determine
whether analytical laboratories report results with acceptable analyti-
cal reliability, and whether the results for cross-check samples in the
quality assurance program are good indicators of reliability.
Acceptance Criteria for Alternative Methods
Within Section 141.25, referenced analytical methods are stipulated
which must be employed in order to comply with regulations. Section
141.27 states that an alternative method is acceptable "if it is substan-
tially equivalent to the prescribed test in both precision and accuracy
as it relates to the determination of compliance with any maximum conta-
minant level." Other than the exceptions noted in the regulation, no
other radiochemical procedure is considered acceptable without obtaining
the written permission of the Administrator of the U.S. EPA and the
appropriate agency of the State where the laboratory is located.
Initially, the EPA only considered approving alternative techniques
according to a very strict protocol, i.e., NIPDWR nationwide approval so
that all laboratories throughout the country could benefit and utilize
the technique.
Within the past few years, granting alternative method approval has
m
been modified to embrace "limited" approvals. For testing a "limited
alternative technique", the total number of analyses is 120 (60 analyses
by the alternative technique and 60 analyses by the approved methodolo-
15
-------
gy) as compared to 240 analyses on a national basis. This "limited"
approval permits that laboratory to utilize the alternate methodology
for water samples from the same watersheds as the test water. In both
types of approvals, the data are evaluated for equivalency according to
standardized statistical methods (EPA77). This, however, evaluates the
accuracy of the alternative technique only if the originally approved
technique is adequate.
Only one radiochemical method has been given an NIPDWR nationwide
alternative method approval status by the EPA since the promulgation of
the regulations. Approval for this method, to analyze 228r3> was based
on its technical equivalency to the approved technique and not on the
comparability testing criteria.
Statements concerning measured precision and accuracy parameters
are presented for most procedures in the Prescribed Procedures manual
(Kr80) intended to update the earlier EPA manual (Kr76), although the
former has not been officially accepted as a replacement. These state-
ments indicate that several of the procedures have only been evaluated
for accuracy and precision by a single laboratory test involving, in
general, nine analyses of a single solution containing a known amount of
the radioisotope of interest in a matrix of unspecified chemical composi-
tion. Thus, in the case of Methods 900.1, 901.0, and 907.0, the accura-
cy and precision capabilities of the methods may not be well known or
acceptable.
No definite precision and accuracy criteria seem to have been appli-
ed to the originally approved methods. Moreover, no relationship ap-
pears to exist between precision and accuracy criteria for the EPA Inter-
comparison Studies Program (EPA81) currently used as a compliance
16
-------
measure, and the measured parameters for the procedures given in the
Prescribed Procedures manual.
The EPA has also utilized the protocol for interim certification
(EPA78) to test the capability of laboratories and ensure data quality.
This program consists of inspecting the radiochemical laboratory every
three years and requiring participation in the EPA interlaboratory cross
check program. The results of this program, however, have not been used
for laboratory certification or decertification.
In brief, the existing system makes it extremely difficult to intro-
duce new methods (EPA77a), although currently approved methods are not
as well defined for accuracy, precision, and sensitivity as would be
required for a new method. The inspection and cross-check program iden-
tifies unsatisfactory conditions but does not necessarily eliminate them.
A much more desirable framework for this extensive national program
would encourage development and application of improved methods while
assuring continued reliability and uniformity of data produced in numer-
ous laboratories. Two alternatives for the current system of methods
approval are suggested here, one that substitutes laboratory accredita-
tion and the other that provides a definite test of reliability.
The preferred option is a change to laboratory accreditation. A
currently active Federal activity that appears applicable is the Nation-
al Voluntary Laboratory Accreditation Program (NVLAP)(DC76, DC79) admin-
istered by the National Bureau of Standards. This program is being
applied in the radiometric measurements field to personnel radiation
dosimetry processing and is being evaluated for accrediting in-vivo and
in-vitro bioassay processors whose clients are regulated by the Nuclear
Regulatory Commission. The concept of the NVLAP for radiation dosimetry
17
-------
processing is particularly pertinent since the ability to process person-
nel radiation dosimeters routinely is being accredited, regardless of
the techniques used to measure the radiation dose.
The general NVLAP program encompasses all efforts necessary to en-
sure that routinely produced data reflect the capability of the labora-
tory which was demonstrated under performance testing. These efforts
include administering the performance testing program according to an
established standard or methodology, on-site peer evaluations of labora-
tories to assess their capabilities, and publishing a list of accredited
facilities. Procedures exist for submitting recommendations by a peer
review committee, removing accreditation for inadequate performance,
performing additional onsite evaluations, and reaccrediting facilities.
As a secondary option, the EPA can modify the existing protocols
for alternate method approval to expedite the process for both single-
laboratory and nationwide-use approvals. In this approach, each proce-
dure is submitted by the applicant in a detailed format for review.
For single laboratory approval, a set of representative water sam-
ples containing known quantities of the radionuclide to be measured are
supplied by the EPA Quality Assurance group to the applicant laboratory
for replicate analysis. These samples contain the radionuclide (1) near
the detection limit, (2) near the MCL, and (3) at a higher concentration
to determine accuracy and precision reliably. Also included in this set
are (4) a blank sample and (5) one or more samples that contain the
radionuclide plus potentially interfering stable and radioactive
materials. The result of these analyses must satisfy published criteria
of acceptance for accuracy, precision, and sensitivity.
18
-------
For approval nationwide under this option, a second laboratory must
qualify for approval in the use of the same method. Subsequently, a
multi-laboratory collaborative study must be conducted, with the results
again meeting predetermined levels of acceptability.
Sampling Frequency
The monitoring frequency for naturally-occurring radionuclides is
specified in 40CFR Part 141.26(a). Although there is evidence that
radium in some ground-water sources remains at relatively constant
levels (H08I, Ki82, Kr82, Mo83), significant changes in the
radiochemical composition of the water could occur for a given well due
to changes in pumping rates or ground-water levels. Examination of data
from repeated samplings is recommended to determine whether the
currently specified 4-year cycle of reanalysis is appropriate.
Sample Stability
The NIPDWR allow quarterly sampling and analysis, averaging for
the annual concentration, or compositing quarterly samples and an analy-
sis of the composite for the annual average concentration. The integri-
ty (stability) of such samples stored before analysis needs to be evalu-
ated for the radionuclides specified in the Regulations. Physical chan-
ges such as precipitation have been observed by EPA (EMSL-Las Vegas)
personnel during on-site laboratory evaluations for certification in
some stored drinking water samples, even when samples were acidified to
pH 2 or lower.
Sample preservation procedures have been recommended in methods
manuals that are referenced in the NIPDWR and the EPA certification
manual (EPA78). The preface to the EPA methods manual (Kr76) suggests
19
-------
that "when quarterly composites are set aside for future analyses, the
samples should be acidified with 1 ml 16 HNO3 per liter to minimize
losses caused by adsorption on container walls." Chapter 6 of the EPA
certification manual recommends that drinking water samples be acidified
to pH 2 at the time of collection. If this cannot be done in the field
because samples are to be separated into soluble and insoluble
fractions, then samples must be acidified upon arrival at the laboratory
followed by a holding period of at least 16 hours (overnight) before the
samples are analyzed.
Section 105 of Standard Methods (APHA80) and several EPA publica-
tions (EPA79, Be82) recommend preservation procedures while indicating
that complete sample stability by preservation treatment is practically
impossible. Some preservation treatments are effective for specific
substances, but in many cases they only retard chemical and biological
processes that continue after the sample is collected.
Some information on the effects of long-term sample storage is
available. A survey of the literature on long-term storage of
environmental samples to be analyzed for trace elements (Ma76) suggests
that information on sampling and storing trace elements in environmental
samples should also apply to radionuclides. One study (Ei65) used
dilute aqueous solutions of radionuclides to determine the relative
adsorption of radionuclides on glass and plastic surfaces. A study of
reactor water and waste solutions collected for analyses demonstrated
the fractionation of radionuclide ions among the soluble and insoluble
portions and the walls of the container (Be80).
Although the studies contain some conflicting details, they show
that loss of trace elements, including radionuclides, to container walls
20
-------
can be a serious problem for water samples. Losses have been reduced
with reagents that tend to maintain a substance in its soluble form,
such as strong acids or complexing agents; with isotopic carriers to
prevent "radioedloidal" behavior; and with substances that tend to satu-
rate ("pacify") adsorption surfaces. Containers with a less adsorptive
surface or surface coatings that reduce adsorption have also been found.
Such remedial measures, however, usually apply to a specific radionuc-
lide rather than the entire group of radionuclides to be analyzed.
Worse, a good preservation technique for one radionuclide may cause
losses for another.
Sampling Reliability
In view of the time and cost of sample analyses, commensurate
efforts should be applied to assuring that the sample adequately repre-
sents the water supply. The water must be collected at uncontaminated
taps in uncontaminated containers, from a freely flowing supply. The
sample must represent the water used throughout the distribution system,
from the usual sources. Conditions of reproducibility and absence of
contamination can be encouraged by preparing collection protocols.
Sampling each source of water in a multi-source system, although
an additional effort, would provide definition of all possible combina-
tions of sources. At present, Section 141.26(a)(3)(iii) requires analy-
sis of all sources only if ordered by the State. Quantification of the
radionuclide content in the various sources would permit calculation of
concentrations after the application of measures such as mixing, special
treatment, or termination of some sources to obtain acceptable levels.
21
-------
Possible effects of the distribution system on radionuclide levels
at the tap should be evaluated for systems in which radionuclide levels
are elevated at the source. Differences in radionuclide levels through-
out a system can be caused by connections to sources at various points;
removal of radionuclides by deposition, volatilization and radioactive
decay; or remobilization of radionuclides accumulated at some points in
the system.
Expanded Analytical Program
If the EPA considers potential radiation doses from elevated levels
of uranium and radon, it may specify MCL values for these radionuclides
and require analyses with prescribed methods. Other reports in this
series discuss the radiation doses associated with uranium and 222^n in
water. The dose from 234y p]us 238|j appears to be of the same order of
magnitude as from 226r3 at the same concentration. Although the dose
per pCi of 222^n -js much lower, the concentration of 222rp -jn some
ground waters is much higher than 226Ra. Persons may be exposed to radi-
ation from 222Rn -jn the water supply by drinking the water or inhaling
222r0 that has escaped from water into room air. In either case, the
important source of radiation exposure is alpha particles emitted by
short-lived 218p0 an(j 214p0 pr0geny of 222Rn>
The volatility and short half life of 222r0 differentiate this radi-
onuclide from the natural radionuclides now measured under the Regula-
tions. These characteristics v^ill result in lower 222r0 concentrations
at the tap than measured at the source; permit application of relatively
simple removal techniques; and introduce sources of error in analysis.
22
-------
Methods for analyzing the naturally occurring radionuclides of
thorium and polonium may be needed on a few occasions to examine water
with an elevated gross alpha particle activity not attributable to radi-
um, uranium, or radon. Because of their relative insolubility in water
under most environmental conditions, these radionuclides (234jh5 2-32j|1>
231m, 230-rh, 228-rn, 227yh, and 210po) would normally not be encountered
at concentrations that exceed the alpha particle screening levels. In
only a few instances among the tens of thousands of samples analyzed so
far were 232jh and 210po detected at levels exceeding the screening
value for the gross alpha particle activity.
The following methods can be considered for these analyses:
Uranium. Fluorometric methods for determining uranium by mass and
radiometric methods have been discussed in a preceding section. Al-
though no MCL for uranium is currently given, these methods appear to
have the necessary sensitivity and accuracy if the MCL were 10 pCi/1 or
higher.
Radon. Methods for the determination of 222pn in water include
de-emanation into a scintillation flask or Lucas cell (Lu57), gamma spec-
trometry (Lu64), high volume extraction followed by liquid scintillation
counting (No64, Ho77), and direct low-volume liquid scintillation count-
ing (Pr77, Ho83). The latter method is probably the most rapid and
simple, while other methods may exhibit higher sensitivity. Concentra-
tions of 222Rn at several hundred pCi/1 have been measured precisely and
accurately with the direct, low-volume, liquid scintillation counting
procedure (Ho83). It would be especially suited for large numbers of
samples if the MCL were in the hundreds or thousands of pCi/1. This
method has been used extensively by EPA (EPA78a).
23
-------
One possible defect in the present method being used by the EPA
is the collection of a 10 ml aliquot by drawing (negative pressure) the
aliquot into a syringe and then transfering the aliquot to a liquid scin-
tillation vial which contains the mineral-oil-based scintillator solu-
tion. This may draw radon out of solution and cause some loss during
transfer to the scintillation vial. Filling the vial to a 10 ml pre-
marked level under positive pressure by transfering the aliquot from a
free flowing tap with a needle attachment should be considered.
A standard 222Rn generating source is being prepared for EPA
(EMSL-LV) by the NBS. The above-cited method (EPA78a) will be multilab-
oratory tested when the standard becomes available.
Thorium. A method for alpha-particle-emitting actinides with indi-
vidual elements separated sequentially is in the EPA Prescribed Methods
manual (Kr80). Separation of thorium by coprecipitation and solvent
extraction is part of this procedure and can be used separately or to
determine several actinides simultaneously. Testing this procedure to
determine reliability and precision is planned by EPA.
Polonium. Most 210po procedures, with minor variations, recover
polonium from solution by electrochemical displacement onto a metal plan-
chet that is more electropositive than polonium. Silver or nickel are
the metals generally used (Ba66, Ba75, B166, Ho66, Hu70, Li66). Separa-
tion of polonium by other techniques, such as precipitation (Ru66) or
ion exchange (Da57), are rarely used. Spontaneous deposition of poloni-
um onto nickel or silver is a reliable technique with recoveries general-
ly exceeding 90 percent. This procedure is also rapid and simple.
In the recommended procedure (B166), a 1,000-ml sample of acidified
drinking water is evaporated to 100 ml. A nickel planchet is placed in
24
-------
the sample. The polonium deposits on it while the solution is being
stirred at 85° C for 3 hours. The alpha particle count rate is measured
on the rinsed planchet, and the disintegration rate is computed for an
estimated yield.
Summary and Conclusions
The radionuclide analyses of public water supplies performed under
the NIPDWR and parallel state regulations have yielded extensive nation-
wide information concerning the absence of elevated levels in most sys-
tems and the occurrence of elevated radium and uranium levels in a few
areas. These efforts and some ancillary studies also provide informa-
tion that may be applied to evaluating the reliability of the data,
suggesting improvements in procedures, planning studies to resolve uncer-
tainties, and possibly expanding the program for better defining the
radiation dose.
The gross alpha particle activity screening scheme for naturally
occurring radionuclides in water has been very useful. It may miss
elevated 228r3 concentrations, however, lead to an erroneous value of
"gross alpha particle activity excluding uranium and 222Rn»s ancj give a
different gross alpha particle activity than 226r3 concentration even in
the absence of any other radionuclides. The first problem is the most
significant one associated with NIPDWR results. A development effort is
recommended to replace the screening scheme with a procedure that will
assure determination of 228Ra> such procedures are available, but
factors of time, cost, and applicability by numerous laboratories must
be considered in selecting an improved approach. The second problem is
due to disequilibrium between ?38u an(j 234y -jn some aqUifers, whereas
the recommended chemical analysis of uranium determines only 238y an(j
25
-------
assumes equal activity of 234(j# use of any one of several available
radiometric methods for the few samples in which this source of error
could lead to a value exceeding the MCL would eliminate the problem.
The third problem is due to the lesser reliability of the gross alpha
particle activity measurement compared to 226r3 analysis. Its
reliability can be improved by use of defined alpha particle activity
standards and assistance in determining self-absorption corrections.
Little can be said concerning gross beta activity screening for
man-made radionuclides. No MCL values were exceeded — in fact, posi-
tive results were rare.
A number of approved analytical methods can bear improvement, espe-
cially the method for 228r3< jhe existing system for approving new meth-
ods is so cumbersome, however, that it discourages improvement efforts.
Two alternatives are offered to replace the existing approval system.
The preferred replacement is laboratory accreditation, which could be
combined with the current EPA laboratory evaluation program through
inspection and distribution of cross-check samples. The other alterna-
tive requires that laboratories obtain acceptable analytical results for
sets of approval samples distributed by the EPA quality assurance pro-
gram. Either approval program needs to maintain a careful balance be-
tween assuring that analytical results from all participating laborato-
ries are reliable and encouraging improvements in methods.
Several aspects of the program were identified for further atten-
tion in the future. These include preparing a sample collection proto-
col to assure reliable sampling; evaluating the impact of source and
distribution system variables on radionuclide levels at the tap; examin-
ing the stability of the radionuclides of interest in stored water
26
-------
samples to assure minimal losses; arid better defining the collected
information in terms of standard deviations and detection limits applic-
able to the reporting laboratory.
Analytical procedures were recommended for 222Rn> radiometric urani-
um, thorium, and polonium. The first two radionuclides would require
routine analysis if EPA recommends an MCL for each. The latter two have
been found only rarely in water supplies, but analyses would be useful
if elevated gross activity levels can not be attributed to other natural-
ly occurring radionuclides.
Acknowledgements. This work was supported in part by the Office of
Drinking Water, USEPA, through Dynamac International, Inc. We thank
Stanley Waligora, Jr., Eberline Corp., C. Richard Cothern, ODW, USEPA,
and Seymour Gold, EMSL-Cincinnati, USEPA, for their participation in
discussions.
References
APHA71 American Public Health Association, 1971, Standard Methods for
the Examination of Water and Wastewater, 13th Edition, (New York:
APHA).
APHA80 American Public Health Association, 1980, Standard Methods for
the Examination of Water and Wastewater, 15th Edition, (New York:
APHA).
As81 Asikainen, M., 1981, Radium Content and the Radium-226/Radium-228
Activity Ratio in Groundwater from Bedrock, Geochim. Cosmoshim.
Acta 45, 1375.
ASTM75 American Society for Testing and Materials, 1975, Annual Book of
Standards, Part 31, Water, D-2907-75, "Microquantities of Uranium
in Water by Fluorometry", (Philadelphia: ASTM).
Ba66 Bagnall, K.W., 1966, "Tha Chemistry of Selenium, Tellurium, and
Polonium" (New York: Elsevier Publishing Company).
Ba75 Bale, W.F., Helmkamp, R.W., Hrynyszyn, V. and Contreras, M.A.,
1975, "The Determination of 210po in urine", Health Physics 29,
663.
27
-------
Be69 Bevington, P.R., 1969, Data Reduction and Error Analysis for the
Physical Sciences, (New York: McGraw Hill).
Be80 Bernabee, R.P., Percival, D.R. and Martin, D.B., 1980,
"Fractionization of Radionuclides in Liquid Samples from Nuclear
Power Facilities", Health Phys. 39, 57.
Be82 Berg, E.L., 1982, "Handbook for Sampling and Sample Preservation
of Water and Wastewater", U.S. Environmental Protection Agency
Report EPA-600/482-029, September 1982.
B166 Blanchard, R.L., 1966, "Rapid Determination of Lead-210 and
Polonium-210 in Environmental Samples by Deposition on Nickel",
Anal. Chem. 38, 189.
C183 Cline, W., Adamovitz, S., Blackman, C., and Kahn, B., 1983,
"Radium and Uranium Concentrations in Georgia Community Water
Systems", Health Phys. 44, 1.
Co83 Cothern, C.R. and Lappenbusch, W.L., 1983, "Occurrence of Uranium
in Drinking Water in the U.S.," Health Phys. 45, 89.
Co83a Cothern, C.R. and Lappenbusch, W.L., 1983, "Compliance Data for
the Occurrence of Radium and Gross Alpha Particle Activity in the
United States," submitted to Health Phys.
Co83b Cothern, C.R., 1983, Private Communication, U.S. Environmental
Protection Agency, Office of Drinking Water, Washington, D.C.
20460.
Cu78 Currie, L.A., 1978, "Sources of Error and the Approach to
Accuracy in Analytical Chemistry," in: Treatise on Analytical
Chemistry (Edited by I.M. Kolthoff and P.J. Elving), Part I, Vol.
1, Chapt. 5 (New York: John Wiley and Sons).
Da57 Danon, J. and Zamith, A.A.L., 1957, "Ion-Exchange and Solvent
Extraction Studies with Polonium", J. Phys. Chem. 61, 431.
DC76 Department of Commerce, 1976, "Proceedures for a National Volun-
tary Laboratory Accreditation Program", Federal Register 41,
8163 - 8168.
DC79 Department of Commerce, 1979, "National Voluntary Laboratory
Accreditation Program: Federal Government", Federal Register 44,
12982-12990.
De83 DeLullo, G., Stevenson, 0. and Jarvis, A., 1983, "Intercomparison
Program for Radiation Quality Assurance: Annual Report" U.S.
Environmental Protection Agency Report EPA-600/X-83-010.
Ei65 Eichholz, G.G. Nagel, A.E., and Hughes, R.B., 1965, "Absorption
of Ions in Dilute Aqueous Solutions on Glass and Plastic
Surfaces", Anal. Chem. 37, 863.
28
-------
EPA76 U.S. Environmental Protection Agency, 1976, "National Interim
Primary Drinking Water Regulations," U.S. Environmental Protec-
tion Agency Report EPA-570/9-76-003.
EPA77 Environmental Protection Agency, 1977, "Alternative Test
Procedure Equivalency Testing System (Interim)", Internal
document, Environmental Monitoring and Support Laboratory,
Cincinnati, Ohio.
EPA77a Environmental Protection Agency, 1977, "Environmental Protection
Agency's Methods Equivalency Program for Water Quality and Water
Supply", Internal document, Environmental Monitoring and Support
Laboratory, Cincinnati, Ohio.
EPA78 Environmental Protection Agency, 1978, "Manual for the Interim
Certification of Laboratories Involved in Analyzing Public
Drinking Water Supplies", U.S. Environmental Protection Agency
Report EPA-600/8-78-008, May 1978.
EPA78a Environmental Protection Agency, 1978, "Radon in Water Sampling
Program", U.S. Environmental Protection Agency, Montgomery, AL,
Eastern Environmental Radiation Facility, Manual 78-1.
EPA79 Environmental Protection Agency, 1979, "Methods for Chemical
Analysis of Water and Wastes", U.S. Environmental Protection
Agency Report EPA-600/4-79-020, p. xv.
EPA81 Environmental Protection Agency, 1981, "Environmental
Radioactivity Laboratory Intercomparison Studies Program, Fiscal
Year 1980-1981", U.S. Environmental Protection Agency Report
EPA-600/4-81-004.
Gi81 Gilkeson, R.H. and Holtzman, R.B., 1981, "226Ra an(j 228r3 -jn
Ground Water of the Cambrian-Ordovician Aquifer System in North-
ern Illinois," Radiological and Environmental Research Division
Annual Report 1980, Argonne National Laboratory, Arqonne, IL,
ANL-81-85, p. 157.
Ha75 Harley, J., 1975, "Procedure Manual", Health and Safety
Laboratory, U.S. Energy Research and Development Administration
Report HASL-300.
He82 Henry, J., 1982, Private Communication, Arkansas State Department
of Health, Little Rock, Arkansas.
He84 Hess, T., 1984, [This series].
Ho66 Holtzman, R.B. and Ilcewicz, F.H., 1966, "Lead-210 and Polonium-
210 in Tissues of Cigarette Smokers", Science 153, 1259.
Ho77 Homma, Y. and Murakami, Y., 1977, "Study on the Applicability of
the Integral Counting Method for the Determination of 226r3 in
Various Sample Forms Using a Liquid Scintillation Counter", J.
Radioanalyt. Chem. 36, 173.
29
-------
H08I Holtzman, R.B. and Gilkeson, R.H., 1981, "226Ra Concentrations in
Some Illinois Well Waters," Radiological and Environmental
Research Division Annual Report 1980, Argonne National Labora-
tory, 9700 S. Cass Ave., Argonne, IL, ANL-81-15, p. 148.
Ho83 Horton, T.R., 1983, "Methods and Results of EPA's Study of Radon
in Drinking Water," Proceedings of 15th Annual National
Conference on Radiation Control, 65 Fountain Place, Frankfort,
KY, May 16-19, 1983, to be published.
Hu70 Hunt, V.R., Radford, E.P., and Segal 1, A., 1970, "Naturally
Occurring Concentrations of Alpha-Emitting Isotopes in a New
England Population," Health Phys. 19, 235.
Ki82 King, P.T., Michel. J., and Moore. W.S., 1982, "Ground Water
Geochemistry of 228Ra> 226r3 an(j 222Rn>« Qeochim. Cosmochim. Acta
46, 1173.
Kr76 Krieger, H.L., 1976, "Interim Radiochemical Methodology for
Drinking Water", U.S. Environmental Protection Agency Report
EPA-600/4-75-008.
Kr80 Krieger, H.L. and Whittaker, E.L., 1980, "Prescribed Procedures
for Measurement of Radioactivity in Drinking Water," U.S. Envi-
ronmental Protection Agency Report EPA-600/4-80-032.
Kr82 Kriege, L.B. and Hahne, R.M.A., 1982, "226Ra an£j 228r3 in jowa
Drinking Water,"Health Phys. 43, 553.
La84 Lappenbusch, W.L. and Cothern, C.R., 1984 [This series].
Li66 Little, J.B., and McGandy, R.B., 1966, "Measurement of Polonium-
210 in Human Blood," Nature 211, 842.
Li82 Lieberman, R. 1982. "Radiochemical Determination of Gross Alpha
Activity in Drinking Water by Coprecipitation," presented at the
27th Annual Meeting, Health Physics Society, June 27 - July 1.
Eastern Environmental Radiation Facility U.S. E.P.A., P.O. Box
3009 Montgomery, AL 36193
Li82a Lieberman, R., 1982, Private Communication, U.S, Environmental
Protection Agency, Eastern Environmental Facility, P.O. Box 3009,
Montgomery, Alabama 36193.
Lu57 Lucas, H.F., 1957, "Improved Low-Level Alpha-Scintillation
Counter for Radon," Rev. Sci. Inst. 28, 680.
Lu64 Lucas, H.F., 1964, "A Fast and Accurate Survey Technique for Both
Radon-222 and Radium-226," in The Natural Radiation Environment,
U. of Chicago Press, p. 315.
Ma76 Maienthal, E.J. and Becker, D.A., 1976, "A Survey of Current
Literature on Sampling, Sample Handling, and Long-Term Storage
30
-------
for Environmental Materials", National Bureau of Standards, NBS
Technical Note 929, October 1976.
Ma78 Mandel, J., 1978, "Accuracy and Precision: Evaluation and
Interpretation of Analytical Results," in: Treatise on Analyti-
cal Chemistry (Edited by I.M. Kolthoff and P.J. Elving), Part I,
Vol. I, Chapt. 5 (New York: John Wiley and Sons).
Mc81 McCurdy, D.E. and Mellor, R.A., 1981, "Determination of Radium-
224, -226, and -228 by Coincidence Spectrometry," Anal. Chem.
53, 2212.
Mc81a McCurdy, D.E. and Mellor, R.A., 1981, "The Concentration of
226r3 ancj 228r3 -jn Domestic and Imported Bottled Waters," Health
Phys. 40, 250.
Mi80 Michel, J. and Moore, W.S., 1980, "228Ra and 226r3 Content of
Groundwater Fall Line Aquifers," Health Phys. 38, 663.
Mi81 Michel, J., Moore, W.S. and King, P.T., 1981, "Gamma-ray
Spectrometry for Determination of Radium-228 and Radium-226 in
Natural Waters," Anal. Chem. 53, 1885.
Mo83 Moore, W.R., 1983, Personal Communication, U. of South Carolina,
Columbia, SC.
No81 Noyce, J.R., 1981, "Evaluation of Methods for the Assay of
Radium-228 in Water", National Bureau of Standards, Technical
Note 1137.
No64 Noguchi, M., 1964, Radioisotopes 13, 362 (in Japanese).
0s82 Osmond, J.K. and Cowart, J.B., 1982, "Ground Water" in: Uranium
Series Disequilibrium: Applicat ions to Environmental Problems
(Edited by M. Ivanovich and R.S. Harmon), Chapter 9 (Oxford:
Clarendon Press), 202 - 245.****Cothern's copy not have these pp.
Pe82 Pembroke Laboratory, 1982, Private Communication, Pembroke,
Florida.
Po67 Porter, C.R., Kahn, B., Carter, M.W., Rehnberg, G.L., and Pepper,
E.W., 1967, "Determination of Radiostrontium in Food and Other
Environmental Samples", Env. Science Tech. 1, 745.
31
-------
Pr77 Prichard, H.M. and Gesell, T.F., 1977, "Rapid Measurements of
Rn-222 Concentrations in Water with a Commercial Liquid
Scintillation Counter," Health Phys. 33, 577.
Ru66 Rushing, D.E., 1966, "Detrmination of Dissolved Polonium-210 in
Water by Coprecipitation with Tellurium by Stannous Chloride,"
Anal. Chem. 38, 900.
Ve83 Velten, R., 1983, Private Communication, U.S. Environmental
Protection Agency, EMSL-CINCINNATI, St. Clair Street, Cincinnati,
OH 45268.
32
------- |