The International Agency for Research on Cancer, IARC, claims to have done a six-year evaluation of combined cancer incidence data on over 95,000 workers in the nuclear industry and to have verified that IARC risk estimates are correct. See
Several studies have found that nuclear workers have a lower incidence of cancer than the average elsewhere. This is cited as an example of hormesis (health benefit from low-level radiation) or as a "healthy worker effect,'' depending upon the writer's bias regarding the no-threshold linear hypothesis. This hypothesis has been badly damaged by Cohen's finding that radon radiation exposure reduces lung cancer incidence (see
Access to Energy 21-4, December 1993, p 4), but the IARC study is potentially valuable, especially for external radiation.By using only nuclear workers, the IARC risks a bias against hor-mesis. Their study, however, includes 10,000 nuclear workers classified as having received zero dose, so we were eager to read their paper.
Alas, the
Lancet publication contains no figures or data whatever from which the reader can evaluate this work. In a five-page scientific publication with three tables and one figure, this is unacceptable. They promise to publish a better account elsewhere.We are left to "believe'' their final result, which does little to support their conclusion that current warnings about the dangers of low level radiation are acceptable. For leukemia, they report an excess relative risk of cancer of 2.2 per Sv (the International Commission on Radiological Protection proposes a dose limit of 0.001 Sv per year for the public) and, for all other cancers, they report an excess relative risk of cancer of - 0.07 per Sv (less cancer with higher radiation exposure).
IARC reports a 90% confidence range (the range within which there is a 90% chance that the actual values fall) of their results for leukemia of 0.1 to 5.7 excess relative risk per Sv and for all other cancer of - 0.39 to 0.30 excess relative risk per Sv.
In the absence of data, the best we can do is to consider their statistical result. With 119 leukemia deaths and 3830 deaths from other cancers and averages of 2.2 and - 0.07 excess relative risk respectively, we calculate (119)(2.2)+(3830)(- 0.07) = - 6.3. Therefore, the IARC study actually reports a slight decrease in total cancer deaths with increased low-level radiation exposure (with a confidence range that makes this result statistically insignificant).< /FONT >
We wonder how the IARC study group and writing committee of 18 coauthors managed to rationalize the publication of a six-year study of 96,000 people purporting to measure a correlation between low-level radiation and cancer deaths without even a simple graph of radiation exposure versus deaths from cancer.
|
|
Vol. 22, No. 5
Newsletter: Access to Energy Newsletter Archive Volume: Issues Issue/No.: Vol. 22, No. 5 Date: January 01, 1995 03:10 PM Title: Information
Copyright © 2004 - Access to Energy Newsletter Archive
|