Pierce et al. 1996 article: Atomic Bomb Survivor Cancer Mortality Data Update

[Ken Mossman <icklm@asuvm.inre.asu.edu>]

Radsafe Colleagues:

I do not know whether the paper by Pierce et al. "Studies of the Mortality of Atomic Bomb Survivors. Report 12, Part I. Cancer: 1950-1990" (Radiation Research 146: 1-27; 1996) has been previously discussed on Radsafe but I would be interested in any comments from anyone who is familiar with the paper. Pierce's paper is an imprortant contribution to the Japanese cancer mortality database. The paper reports an extended  follow-up of 5 years (1986-1990) and an analysis of an additional 10,500 survivors (mostly individuals who were children in 1945). The extended observational period adds about 550,000 person-years of follow-up.

I call for discussion of this paper because certain members of the NCRP and ICRP, and others,  have used it as further support for the linear no-threshold theory claiming that the additional mortality data provides evidence for a significant excess risk at 50 mSv (5 rem). If true, this would be an extraordinary finding since all previous epidemiological studies of adult populations show statistically significant risks at 200 mSv (20 rem) or more.
  
In my view, the paper makes no such claim. The reference to the 50 mSv limit refers to the minimum dose for which there is a statistically significant dose-response, not statistically significant excess relative risk. Clearly, a statistically significant dose-response can be shown using data points which do not reflect statistically significant excess risk.

I compared the cancer mortality data in the Pierce paper with data from the previously published Shimizu paper (Radiation Research 121: 120-141; 1990) which analyzed data to 1985. The addiitonal years of follow-up reported by Pierce confirms the general conclusions in the Shimizu paper that excess relative risk does not become statistically significant until 200 mSv or more. In fact, Pierce's own Figure 5 supports this conclusion.

The significant findings in the Pierce  paper relate to the strong age and gender dependency on excess relative risk. 

Any comments?

Thank you.

Ken Mossman