LNT and whose junk science?

[Darryl Kaurin <kaurin@mail.sep.bnl.gov>]

Dear James Muckerheide,

I appreciate your comments with regard to Luckey's, Cohen's, and other work
with regard to no observable adverse health effects (possibly even hormetic
effects) and problems with the linear-no-threshold hypothesis.  I personally
think it is irresponsible for the U.S. EPA to not take the work of Cohen
seriously and fund radon research to determine if the current
recommendations are not well founded.

Nevertheless, I think we should point out the short comings of the work, and
refrain from calling "junk science" only those results we disagree with.  As
an example, in the reference: Luckey, T.D.  "The Evidence for Radiation
Hormesis".  Nuclear Report, 21st Century, p.12-20 Fall 1996.  

1)  In Table 3 and Figures 5 and 6, Luckey uses a study of British Nuclear
Workers to show that those receiving the highest doses had the lowest
leukemia mortality.  Since the British results were not corrected for age,
the greatest confounder in epidemiology research, he used United States
leukemia death rates to correct the British results.  By not using British
death rates, he introduces a potential error in differences between US and
UK leukemia death rates.  Also, he used the death rates for all US citizens,
which are not as healthy as workers, introducing the problem of the healthy
worker effect.  This is extremely sloppy science (junk science?) and in my
opinion, this example should be thrown out.  For a person that believes all
radiation is harmful, this appears to be manipulating the data to get the
result you want.

2) In a study of U.S. Nuclear Workers (p.16-17 of the reference I am
referring to), Luckey compares the SMR (standardized mortality ratio) for
workers with doses of greater 50 mSv with the SMR for unexposed workers.
Anyone who has stayed awake during a beginning epidemiology class should
know that you can't compare SMRs - unlike normal math rules, a SMRs do not
have the property of transitivity (ie: if a is greater than b, and b is
greater than c, then a is greater than c).  This is extremely sloppy - and
indefensible (junk science?).  What should have been done was to compare the
death rates between the workers with the exposure and the workers with out
the exposure.   The result is most likely the same, and if my memory serves
me correctly, was done in the UNSCEAR 1994 report.

As for the work of Cohen, I was extremely impressed with his analysis.
Since the results are extremely surprising, one must question if an
ecological study is appropriate, since the radon doses assigned are average
values for the entire county.  Cohen's results are not unique, the UNSCEAR
1994 report lists numerous other radon ecological studies - some show a
negative lung cancer rate with dose, some non-significant, and some a
significant positive lung cancer rate with dose.  I have not read each of
the other ecological studies, but one must question the fundamentals of
ecological studies for radon, since the results are so diverse.

As for the mudslinging at BEIR 5 for not recognizing the lack of harm at low
doses, the studies of areas with high background radiation are listed in the
book, with the results.  In my opinion, this is not ignoring the data.  The
UNSCEAR 1994 report also has an entire section devoted to adaptive response,
with a reference to the work of Luckey in the text.  The  activation of the
immune system by radiation is suggested for some of the results.  Thus, I
think some of your accusations are not entirely true.

Perhaps the dose limits are too conservative.  Perhaps there is a hormetic
effect.  But we still must analyze the data correctly, refute sloppy or junk
science, keep our minds open, and the dialog continuing.

Standard disclaimers.

Sincerely,

Darryl Kaurin, Ph.D.
Brookhaven National Laboratory
Safety and Environmental Protection Division
Building 535A
Upton, NY 11973-5000
(516)344-3166
kaurin@mail.sep.bnl.gov