Cancer Risks at Low Doses of Ionizing Radiation
or Chemicals.
F. Granath,
Dept. of Cancer Epidemiology, Radiumhemmet, Karolinska Hospital, 171 76
Stockholm, fax +46 8 32 61 13; and L. Ehrenberg, Dept. of Radiobiology,
Stockholm University, 106 91 Stockholm, fax +46 8 16 64 88
The
true shape of dose-response (dose-risk) relationships for cancer caused by low
doses of radiations or chemicals are still debated.
This
is due to the impossibility, for statistical reasons, to obtain informative
observations of the incremental incidence below some dose. Below this detection
limit, which will always exist although it can be decreased by increasing the
size of the studied populations, the dose-risk curve may have any shape. The
estimation of cancer risks in this "dark" region by linear
extrapolation from higher doses, where informative observations can be made, may
then be seen as an administratively practical measure rather than a
scientifically based effort to describe "true" relationship. The
uncertainty about the true dose-risk relationship at low doses recently led the
Health Physics Society to recommend that risks should not be estimated at doses
below 50 mSv (5 rem) in a year or 10 rem in a lifetime.
Recent
cell-genetic studies have given proof of earlier assumptions that mutation is
the key event in the carcinogenic action of ionizing radiation and so-called
genotoxic chemicals. The mutations, e.g. in suppressor genes, caused by an
exposure then interact with other inherited or acquired factors with influences
on the probability of tumour development.
The
key role of mutation in tumorigenesis renders it possible to elucidate the dose
response relationships at the high resolving power of mutational systems. Such
studies indicate, throughout, that the true dose-response relationships for
induced mutation are linear (or somewhat, and perhaps negligeably, superlinear).
It belongs to the rules of statistics that the absence of a no-effect threshold
cannot be proved. However, the demonstration that the upper confidence limit for
a possible no-effect threshold is below the lowest dose a cell can receive, viz.
one hit, may serve as evidence of linearity.
According
to the present state of the art, average attributable risk increments are best
estimated by a linear multiplicative model. This is already applied in
radiological protection, and should be used also for genotoxic chemicals.
Go to:
Table of Contents of the 1997
SRA-Europe Conference
Program of the 1997 SRA-Europe Conference
RiskWorld Welcome Page
Tec-Com, Inc.