Answer to Question #6416 Submitted to "Ask the Experts"

Q

I'm a little confused when reading various radiation studies such as BEIR VII. It seems like every time results at certain dose levels show no statistical increases in adverse health risks such as cancer and genetic effects on children, they almost seem to dismiss these findings and instead use risk estimates derived from animal studies, etc. For example, the studies show no genetic problems to offspring from A-bomb survivors for which they reference but then go on to provide several pages of various, complicated formulas based on assumptions, etc., to generate risk estimates which actually are not showing up in the real studies, those of the actual people. The same goes for risk estimates on the possible development of cancer in those exposed to the bomb. It seems like the real numbers of cancer incidence are actually lower than the risk estimates they derive from various formulas. Why don't they use actual percentages based on actual numbers observed? Lastly, they state that 42 out of 100 people will develop cancer sometime during their lifetimes and then state that every 100 mSv will generate one extra cancer per 100 exposed. This seems to contradict many of the answers seen here where many responses state that under 100 mSv there is no statistical risk observed, and a couple of your ATE answers state "no observed risks were easily seen under 200 mSv in atomic bomb survivors."

A

You have zeroed in on one of the most controversial topics being debated in the radiation sciences—the effects, if any, of exposure to low doses of ionizing radiation. The Health Physics Society's position paper on this issue, "Radiation Risk in Perspective,"  explains the problem with using highly uncertain risk models to predict risk at low doses.

There are several hypotheses about what happens at low doses. These range from no effect, which should be the default assumption or null hypothesis in well-designed scientific studies, to increases in cancer risk in proportion to the dose (linear no-threshold hypothesis, or LNT), to a decrease in the incidence of cancer (hormesis).

The problem is that epidemiological studies that look at populations of people exposed to radiation (for example, the Japanese atomic bomb survivors) don't have the statistical power to definitively support one model and definitively exclude competing models. This has led to a vigorous debate in the scientific community, with some expert bodies (such as BEIR VII) adopting LNT as a conservative model for radiation protection and others (such as the French National Academy of Sciences and Medicine) rejecting LNT at low doses. My advice is to get information from a variety of sources (including the Health Physics Society!) and stay tuned—relevant new research is coming out very rapidly.

I hope this helps.

Brant Ulsh, PhD, CHP