OFFICE OF
THE ADMINISTRATOR
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
' WASHINGTON, D,C, 20460
September 9, 1988 SAB-RAC-88-G42
Honorable Lee M. Thomas
Administrator
U. S. Environmental Protection Agency
401 M Street SW
Washington, DC 20460
Dear Mr. Thomas:
The Dose and Risk Subcommittee of the Science Advisory
Board's Radiation Advisory Committee has completed•its review of
the Office of Radiation Programs1 Radon Risk Estimates. This
review was requested on May 25, 1988 by the Director of the
Office of Radiation Programs and was conducted on June 20, 1988,
at an open meeting in Washington, DC, at which Dr. Douglas
Chambers of SENES Consultants and Dr. Leonard Hamilton of the
Broofchaven National Laboratory presented oral and written
comments on behalf of the public. Because the Office of
Radiation Programs is considering using its new estimate as part
of the technical basis for the revised radionuclides NESHAP, the
Science Advisory Board was asked to respond as early as possible.
The Office of Radiation Programs1 proposed risk estimate,
based upon an adjusted BEIR-IV model, consists of a range of 120-
750 fatal lung cancers per million working level months of radon
exposure with a central estimate of 300, The Environmental
Protection Agency's current central value is 460 fatal lung
cancers per million working level months of radon exposure with a
range of 230-920, For radon exposure, risk to the lung is
considered the only relevant cancer risk,
Three important technical consensus reports on radon risks
are of potential use to the Agency; (1) ' the 1987 International
Commission on 'Radiological Protection's Report entitled "Lung
Cancer Risks from Indoor EKposures to Radon Daughters" (ICRP 50};
(2) the 1988 report of the National Academy of Sciences'
Committee on the Biological Effects of Ionizing Radiation Report
entitled "Health Risks oŁ Radon and other Internally Deposited
Alpha-emitters" (BEIR-IV) ; and (3) the 1984 report of the
National Council on Radiation Protection and Measurements1 Report
78 entitled "Evaluation of Occupational and Environmental
Exposures to Radon and Radon Daughters in the United states"
(NCRP 78). None of these reports is clearly pre-eminent; each has
its strengths and limitations. These limitations reflect the
uncertain state of our knowledge in particular areas and the
-------
- 2 -
consequent legitimate basis for differences in the interpretation
of the limited data available. Of the three reports cited,
ICRP 50 and BEIR-IV appear to have the greater utility for the
Agency. Both use the relative risk projection model arid
generally their projections are quite similar. Their models
differ in certain ways, however; for example, BEIR-IV, in its
projections diminishes the excess relative risk with time since
exposure, whereas ICRP 50 uses a constant excess relative risk,
The Dose and Risk Subcommittee does not recommend that the
Agency attempt to develop still another model for radon and radon
daughter risk estimation and projection; there is not time nor
sufficient personnel to do it better than existing technical
consensus groups have done before the radionuclides NESHAP is
scheduled to be published in final form. Moreover, such a model
would not have the weight of authority of the consensus documents
since it could not be scrutinized carefully by the scientific
community. Instead the Agency should use the ICRP 50 and BEIR-IV
reports to determine the range of risk estimates that exists, to
identify the ambiguities in the data on which-the risk estimates
rest, to determine the , consequences of the projecting of these
risks to a lifetime, and to understand the uncertainties of the
models. The ICRP 50 and BEIR-IV projection values could be
averaged to provide a central estimate except where there are
good reasons to prefer one result to the other. One such case
may be risks to children as the BEIR-IV report is primarily
concerned with a working population, whereas the ICRP 50 report
attempted to address risks to the general population whose
susceptibility to cancer may differ by age at exposure. The
Subcommittee suggests that the risk estimate used in the
projection of the lifetime risks with the BEIR-IV model be the
BEIR estimate basedi on the internal rather than the external
comparisons, but strongly urges that the description of the
projections be tempered by a discussion of the limitations of the
epidemiological data.
The National Council on Radiation Protection Report on radon
and radon daughter products differs from the ICRP and BElR-lv in
a number of ways. Specifically these are: (a) NCRP 78 uses ar.
absolute model for risk projection; (b) NCRP 78 projects the sar.e
risks for males and females, for different age at exposure
groups, and for smokers and nonsmokers; (c) NCRP presumes -2
expression of the increased risk until after the age of 40, *rd
(d) NCRP 78 assumes that a loss half-time (20 years) exists fcr
the stem cells in the lung transformed by exposure to alpr.a
irradiation, so that the cancer risk diminishes with time s:~re
exposure. Nevertheless, it should be noted that the NCR? "
model, for non-smokers, leads . to conclusions within the rar.,7^
resulting from the ICRP 50 and BEIR-IV models.
Generally speaking, the ICRP 50 and BEIR-IV models w;;,
lower the level of risk understood to be associated with radcr.
-------
- 3 -
exposure somewhat below the levels estimated by the Agency in
1985. No additional major technical assessments of radon and
radon daughter risks are currently underway that would contribute
to the Agency's understanding of these risks in the next two
years or so, but several studies are in progress or have been
recently published that will provide the basis for -an improved
evaluation later. Such studies include an update on the Ontario
miners, further work with the Czechoslovakian miners,
experimental studies of radon as a cancer promoter, and the
National Cancer Institute sponsored studies of radon exposure and
lung cancer.
Besides recommending that the Agency rely on the ICRP 50 and
BEIR-IV reports, the Subcommittee offers comments on certain
aspects of the modelling of risk.
The Subcommittee supports the relative risk model used by
ICRP 50 and BEIR-IV because the experience in Japan clearly shows
that for survivors aged 20 and older at the time of exposure, the
excess relative risk of cancer changes very, little with time
since exposure. For those under age 20 at the time of exposure,
neither the absolute nor the relative risk model works very well.
The atomic bomb -survivors were exposed primarily to external,
low-LET radiation at a high dose rate, not radon or its
daughters; but because the survivors represent the only well-
studied large general population exposed to radiation, the
experience in Japan for predicting population risks is
compelling.
The Subcommittee endorses the use of a linear model for low-
dose extrapolation of risks from radon exposure' although it is
not possible to determine statistically whether the linear,
linear-quadratic, or quadratic dose-response model best fits the
available human population studies. The Ontario and Colorado
miner data sets have a linear-quadratic quality to them; however,
controlled experiments in animals exposed to high-LET radiation
indicate linearity in the low-dose range, and the Subcommittee
places greater reliance in these latter findings, for the
present, given the limitations in the human studies.
The Subcommittee believes that risks established for mining
populations can be transferred to a general population including
the young, females, and nonsmokers since the levels of exposure
are similar, at least in some instances. Although the dose to
the lung is determined by the unattached fraction of the
daughters of radon, and this fraction appears somewhat higher in
homes than in mines (which implies a slightly higher dose in rad
or gray in .the former than the latter environment), the lower
breathing rate, on average, in the home would offset this effect.
Accordingly, the Subcommittee believes the current state of
knowledge is such that the risks established for miners provide
an acceptable basis to predict risks for the general population.
-------
- 4 -
With regard to age, the Subcommittee recommends that where
there is a specific need to predict risks from radon and radon
daughters in a young population, such as risks of radon in
schools, "the Agency should rely on the ICRP 50 rather than the
BEIR-IV report because children may be more susceptible*
The Subcommittee finds the data on the interrelationship of
smoking and radon exposure to be complex but recommends that
under these circumstances the Agency treat the interrelationship
as multiplicative at the present time. If the effects of smoking
and exposure are additive, this will have the effect of
overestimating the risk, but given the uncertainties in their
relationship, this would seem the conservative course at present.
In summary, the Subcommittee endorses the range 'of risk
estimates proposed by the Office of Radiation Programs. It does
recommend that the Agency not attempt to develop still another
model for radon and radon daughter risk, estimation and
projection, but urges the use of the existing technical consensus
reports described above.
The Subcommittee looks forward to the opportunity to review
the Background Information Document on which the final
regulations, will be based.
We appreciate the opportunity to share our views with you
and look forward to a written response from the Agency.
Sincerely,
Norton
Chairman, Executive Committee
Science Advisory Board
William J, Sc
Chairman, Radiation Advisory Cc T -.--
and Dose and Risk Subcommittee
Attachment; Roster
CCi R. Guimond
J. Cotruvo
------- |