Linear, No-Threshold Model: A reply to Mark Winslow, EPA

[dj_strom@ccmail.pnl.gov]

There are many members of the Health Physics Society who don't believe that the 
linear, no-threshold model is incorrect *for radiation protection purposes*.  I 
count myself among them.

As a professional holding the public's trust for environmental protection, you 
should look at *ALL* the evidence regarding the validity of the linear, 
no-threshold dose response model as a basis for radiation protection.  Leave it 
to the tobacco companies and the opponents of food irradiation to look at *only 
one side* of an issue, to examine only the evidence that supports their favored 
conclusion.

For balance, consider all the evidence that supports the linear, no-threshold 
model, as well as the evidence that is equivocal and the evidence that would 
seem to contradict it.  After you discover that there are no data in the realm 
of 10 microsieverts (1 millirem) per year, and that there never will be, then 
you are left with a science policy choice, not a science choice.  How do you 
extrapolate?

Radiation does not act alone, in a vacuum.  Given that we live in a sea of 
initiators, promoters, and tumor progression agents, and given that the 
incidence of cancer in Western populations is roughly 1/3 and the mortality is 
roughly 1/6, why would a small increment of any insult *not* be a linear 
perturbation (a la perturbation theory in physics) on this enormous background 
signal?  In this view, additional amounts of any agent that affects DNA should 
produce an increment of cancer risk.  Bear in mind that such increments are 
above what we normally get in our diets, air, and water.  Such incremental 
agents include radiation, chemicals, irritants that cause mitogenesis [Ames & 
Gold's "mitogenesis causes mutagenesis" observation, a clear example of a 
threshold effect], viruses, or other microbes.

Having read the report referenced below, I agree with this 1994 conclusion:

"It is concluded, therefore, that data relating to the role of gene mutations in
tumorigenesis, the monoclonal origin of tumors, and the relationship between DNA
damage repair, gene/chromosomal mutation and neoplasia are well established and 
broadly consistent with the thesis that, at low doses and low dose rates, the 
risk of induced neoplasia risses as a simple function of dose and does not have 
a DNA damage or DNA repair related threshold-like component.  Whilst adaptive 
responses or other protective mechanisms may influence the risk of tumor 
development, they do not provide a sound basis for judgement that tumorigenic 
response at low doses and low dose rates of radiation is likely to have a 
non-linear component which might result in a dose threshold below which the risk
may approach zero.  These michanisitic studies, in addition to the 
epidemiological information, indicate that for radiation protection purposes 
there is little basis for arguing that low radiation doses (about 10 mGy [1 
rem]) would have no associated cancer risk and that, in the present state of 
knowledge, it is appropriate to assume an increasing risk with increasing dose."

R. Cox, C.R. Muirhead, J.W. Stather, A.A. Edwards, M.P. Little.  Risks of 
Radiation-Induced Cancer at Low Doses and Low Dose Rates for Radiation 
Protection Purposes.  Documents of the NRPB 6(1); National Radiological 
Protection Board (NRPB); 1995; p. 75 (Conclusions).

It is estimated that in normal humans, each genome receives 10,000 hits per day 
to DNA that must be repaired (Susan S. Wallace, Hanford Life Sciences Symposium,
10-24-95).  Furthermore, it has been estimated by others that all but 1 in 
1,000,000 are correctly repaired.  At this rate, one cell out of 100 accumulates
one mis-repaired or unrepaired damage site per day.  Should it be surprising 
that any increment in the hit rate may cause an increment in cancer risk?

See also Strom, D. J.  1995.  "Support for the Linear, No-Threshold Model."  The
Health Physics Society's Newsletter 24(10):6-8; and Charlie Meinhold's comment 
following this.

I write as a private individual; this information has not been cleared or 
approved by my employer.

- Dan Strom <dj_strom@pnl.gov
Pacific Northwest National Laboratory