\
15

/

UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, 0,C, 20460

September 5, 1986

SaB-RAe~B6-028

Hon. Lee M. Thomas	OF(S)<:et

Administrator	?«e admix.si

U.S. Environmental Protection.

Agency
401 M Street, SW
Washington, D.C. 20460

Dear Mr. Thomas:

The Science Advisory Board's Radiation Advisory Committee was
requested by the Office of Research and Development to review the
scientific merit of a proposal to conduct an epidemiological study of
radon in indoor air. The Board accepted this request and formed a
Radioepidemiology Subcommittee to carry out the review. The Sub-
committee has submitted its report in which it responded to two
overriding scientific issues:

o Can further epidemiological study contribute to an understanding
of the risks oŁ lung cancer associated with household exposures?
The Subcommittee concludes that scientific uncertainties in
current epidemiological studies (chiefly studies of uranium
miners) could be further reduced through direct investigations
of the domestic population.

o Is the proposed study under review by the Office of Research
and Development, entitled Health Effects of Waterborne Radon,
appropriately designed to address this risk? For reasons
cited in the attached report, the Subcommittee concludes that
it is not appropriately designed-

The Subcommittee's report has been approved by both the Radiation
Advisory Committee and the Executive Committee of the Science Advisory
Board. We hope that our conclusions will assist the Agency in determin-
ing the merit of the,proposed study and are prepared to provide any
additional assistance that is requested. We request that the Agency
formally respond to the report and indicate which recommendations the
Office of Research and Development plans to accept or reject, providing
the reasons in cases where the Subcommittee's advice is not accepted.

William J, Schuiy

Chair, Radiation Advisory Committee
Science Advisory Board

Norton «eison

Chair, Executive Committee
Science Advisory Board


-------
Review of a Proposed Study—Health Effects of tfaterborne Radon
Radioepidemiology Subcommittee
Radiation Advisory Committee

On February 18, 1986 Dr. Hugh McSirwoa, the Acting Director of the Office
of Health Research in the Office of Research and Development (ORD)» requested the
Radiation Advisory Committee to review a proposed epidemiological study of indoor
radon. The EPA's Office of Drinking Water and the Office of Radiation Programs
also requested the review. The questions the Agency wished the Committee to
address included:

1.	Can further epidemiological study contribute to an understanding of the
risks of lung cancer associated with household radon exposures?

2.	If so, is the proposed study under consideration by the Office'of Health
Research, Health Effects of Waterborne Radon, appropriately designed to
determine this risk?

At its January 21-22, 1986 meeting, Mr. Gunther Craun of GRD's Health Effects
Research Laboratory briefed the Committee on the background and content of the
proposed study. At this meeting, the Committee formed a Radioepidemiology Sub-
committee, chaired by Mr. Seymour Jablon of the National Research Council, to
carry out the review. The Subcommittee met on April 18, 1986 in Denver for
briefings by the investigators proposing the study. The Subcommittee provided
written comments and short verbal summary of its activities at the June 20, 1986
Radiation Advisory Committee meeting. The contents of this letter report were
subsequently developed, circulated and unanimously approved by mail.

The Subcommittee concludes that further epidemiological studies could clar-
ify our understanding of the risks of lung cancer associated with radon in indoor
air. Appropriately designed and well conducted epidemiological studies of those
segments of the general population exposed to indoor radon are urgently needed.


-------
2

There are at least two important questions that epidemiological studies
can address: do radon exposures in indoor air pose a significant health risk
and, if so, what is the magnitude of that risk? While the Subcommittee
believes that exposure to radon in indoor air may be the most significant
radiation exposure that affects the health of the general population, this
conclusion is based On the extrapolation to the general population of lung
cancer risks derived from studies of uranium miners. This extrapolation
implies that a substantial fraction of all lung cancers in non-smokers may
be attributable to radon exposures from indoor air. Because the actual
radon exposures experienced by some miners and some families exposed to
radon in the home may be quite similar, this is not primarily an issue of
extrapolation from high radon doses to low, but of extrapolation between
two human populations, in different environmental settings, that differ in
ways that are known to affect the likelihood of certain diseases occurring*

While the Subcommittee is confident that these miner studies support the
conclusion that radon causes lung cancer in humans, studies addressing the
relationship between lung cancer and radon in domestic populations are only
now being proposed and/or conducted. The differences between occupationally
exposed miners and people exposed to radon at home may be sufficient to
yield different risks of lung cancer for the domestic population. Mines
typically contain other pollutants in addition to radon. The same may be
said for homes, but the mixtures of other pollutants and levels of radon
exposure in the two environments may not be the same.., Miners and the
population exposed to radon at home also differ in such factors as age, sex,
ethnicity, race, smoking habits, nutritional practices, activity levels and
general health- These differences in target populations and exposure patterns


-------
make a study directed specifically at residential exposures very worthwhile*
Such a study can reduce many of the uncertainties inherent in the present
extrapolation from uranium miners to the general public.

Such a study must, of course, be appropriately designed and conducted
to successfully measure these differences in risk between the mining and
home environments. It is important to note that improperly designed
studies of factors that may adversely affect public health pose a real
danger because, by failing to identify risks that actually are present,
such studies provide false reassurance. Such studies, especially those
with small populations and uncertain dosimetry, may obscure effects which
are quite real. While the scientific community is familiar with the fact
that a negative study in a small population does not necessarily mean
there is no problem, it is wasteful to conduct a study with little chance of
answering the questions asked.

The Subcommittee's critical comments on the proposed study of the Health
Effects of Waterborne Radon should not be construed by the reader to mean
that epidemiological studies of radon in the home are not needed, but as
an indication that those studies which ate conducted should be designed
to have adequate power to answer the important scientific questions.

The Subcommittee members confined their review to a single proposal, but
they are individually familiar with other proposed and ongoing radon
studies. For example, members are aware that the National Cancer Institute
and the New Jersey Department of Health are undertaking a study that includes
a larger population with lung cancer. Completed 30-page questionnaires are.
already available on each person which address diet, occupation, and smoking
(both active and passive)- Each individual was a long-term resident of a


-------
single household and each of these homes will be measured for radon.

Other studies are planned or underway in Pennsylvania, Sweden, Canada
and China.

The fact that other studies of Indoor radon are being conducted does
not imply that a study sponsored by EPA would be unproductive. In a
new research area, multiple studies with different populations, exposure
levels, and investigators are especially useful. The results, if
complementary, greatly strengthen the conclusions.

The study the Subcommittee was asked to review, Health Effects of
Waterborne Radon, was proposed by investigators at the Maine Medical
Center. What potential does the Maine study have for making a unique
contribution to the current understanding of radon in homes? Two
possibilities have arisen. One""'is that, by assessing a middle range of
radon exposures, a successful study in Maine, together with studies in
Pennsylvania (where the levels of radon in homes are thought to be higher)
and in New Jersey (where the levels of radon are thought to be lower) , could
provide complementary results that would clarify the present understanding
of risks associated with various levels of radon exposure.

•' The other possibility is that the study will add to what is known about
the risks posed by radon in water. As is made clear by the title of the
proposal, Health Effects of Waterborne Radon, the study was originally
spurred by the observation that many wells serving as jsources of household
water in Maine are characterised by high levels of waterborne radon.

However, the preliminary work by the investigators has shown that, even
in Maine, radon in water is usually a minor contributor to radon in
indoor air. While the title and the water measurements indicate the
motivation for the study, it must now be evaluated exclusively in teres of


-------
its utility as an investigation of lung cancer in women in relation
to indoor air concentrations of radon.

Given the importance of evaluating the lung cancer risks that result
from radon in household air, the key question is whether the proposed study-
in Maine gives reasonable assurance of providing useful information on this
subject. The Subcommittee's conclusion is that it will not.

Although the proposal is for a case-control study, which is the design
of choice for addressing this issue, there are certain overriding problems
with the study that seriously compromise its potential utility. The Sub-
committee highlights the following difficulties:

1.	There are two problems with the proposed location: 1} the number of
cases available is barely adequate and will not allow for dropping
a possibly substantial number for whom dosimetry is inadequate;
and 2) there is the related difficulty of measuring exposure for
that portion of the study population that has lived in multiple
dwellings. Consequently, estimates of levels of individual radon
exposures will be especially uncertain.

2.	The study is designed to answer the question, "Is there an increased
risk of lung cancer in white women who are exposed to radon in the
home?" It is not primarily aimed at the question of how these
risks differ from those encountered in studies of miners. The
Subcommittee recognizes that there are some divergent views in the
scientific community on whether radon exposures in indoor air do,

in fact, pose a significant health risk—a question which would be
addressed by the proposed study. The Subcommittee believes,


-------
¦however, chat the critical scientific question, concerns the
determination of levels of risk in relation to levels of household
exposures- The proposal does not address this issue and is not of
sufficient quality to have a high liklihood of testing the null
hypothesis.

The assumptions and sample size calculations underlying the selection
of 600 cases and 800 controls in the proposal were not appropriate or
adequate to address the question of levels of risk posed above.

Whether an absolute or relative risk model is used, whether continuous
or dichotomous exposures are employed, and which question the study
addresses all afEect the number of cases needed in the study and the
desirable distribution between cases and controls. The overall number
of people in the study may actually be adequate to address levels of
risk associated with radon in indoor air, but the proposal is not as
clear on this issue as a study of this importance warrants, A study
of radon in indoor air is important, as is the need to ensure that there
is a reasonable probability that the study can answer the questions
posed. The Subcommittee is not convinced that this study, as currently
designed, can meet this need.

Although the investigating team appears to have most of the necessary
combination of scientific skills, the in-house statistical expertise is
weak.

The difficulties and uncertainties in estimating radon exposures over
thirty years are large. The proposal gave no estimate of tfhe accuracy


-------
7

or precision needed for estimating even total annual exposure based
on short-term integrated sampling. The individual estimates of
exposure to radon over thirty years are to be made from recalled
activity patterns and current measurements of radon levels in
homes which may have "been altered over that time. The uncertainties
in the exposure estimates may be so large, and the expected increase
in the number of cancers in the study population so small, as to
obscure the effects of radon on lung cancer rates. Errors in estimates
of individual exposures will bias downward the risk estimates calculated
and, if sufficiently large, may obscure them completely. Because the
difficulties posed by historical exposure assessment are large for
radon studies generally, spending the tin® now to conduct methodological
studies for assessing radon exposure would provide a sound basis for
the conduct of better epidemiological studies on this subject at a
later date.

6,	The effort to characterize radon in water does not seem useful since,
according to the proposal itself, radon in water is not a significant
source of radon in houses even when levels of radon in water are
elevated.

7.	The appropriate scientific method of dealing with deceased individuals
is a matter of gome controversy among epidemiologists. Simply excluding
them from the study risks introducing selection bias if deceased
persons systematically differ from living cases in terms of exposure

or confounders. On the other hand, the use of proxies for controls
(as proposed) risks introducing differential misclassification bias.

This loss of validity is of particular concern for cigarette smokers


-------
since smoking is a major cause of lung cancer, and comparable
information on this point is especially important. A third
alternative, using deceased controls for proxy interviews risks
introducing other forms of selection Mas and adding further
aisclassification, while not guaranteeing that the misclassification
will now he noti-differential. There are no universal guidelines.
The alternatives need to be considered very carefully.
8, The proposed dietary questionnaire appears to be too brief to
obtain satisfactory data concerning retinal and beta-carotene.

This approach is unlikely to be informative and would appear
not to be worth the effort.

Individual members of the Subcommittee also made a number of detailed
recommendations that have been made available to the Office of Research
and Development. While supporting the need for epidemiological studies
on radon in indoor air, the Subcommittee recommends that the Agency not
undertake the study reviewed in this report as it is presently planned.


-------
U.S. ENVIBONMEMAL PBDTE3CTICN AGENCY
SCIENCE ADVISORY BOMD
RADIATION ADVISORY COMMITTEE
RADICEPIDEMIOLOGY SUBCOMMITTEE

ROSTER

Chairman

Mr# Sepiour Jabloti

Director	.

Medical Follow-up Agency-
National Research Council
2101 Constitution Avenue, N.W.

Washington, D* C. 20418

Members/Consul tants

Dr. Jonathan M. Samet
New Mexico Tumor Registry
¦ 900 Camino Ete Salud, N.E.
Albuquerque, New Mexico 87131

Dr. Richard SextrQ
Building Ventilation and

Indoor Air Quality Program
Lawrence Berkeley Laboratory
Building 90, Room 3058
Berkeley, CA 94"720

Dr. Duncan C. Thomas

University of Southern California

School of Medicine

Department of Preventive Medicine

2025 Zonal Avenue

Los. Angeles, California 90033

Executive Secretary

Ms. Kathleen W. Conway
Deputy Director
Science Advisory Board
Radiation Advisory Committee
401 m Street, S.w. (A-1Q1F)

Washington, D. C. 20460


-------