[ RadSafe ] BBC Reports "Nuclear workers' cancer risk rise

[Dukelow, James S Jr]

Re the 15 nation pooling of epi studies of workers in nuclear facilities
of various sorts, available online from www.bmj.org

Fred Dawson wrote:
-----Original Message-----
From: radsafe-bounces at radlab.nl [mailto:radsafe-bounces at radlab.nl] On
Behalf Of Fred Dawson
Sent: Wednesday, June 29, 2005 9:12 AM
To: radsafe at radlab.nl
Subject: [ RadSafe ] BBC Reports "Nuclear workers' cancer risk rise

BBC Reports "Nuclear workers' cancer risk rise 

http://news.bbc.co.uk/1/hi/health/4629461.stm

also see

http://www.guardian.co.uk/nuclear/article/0,2763,1516929,00.html


Exposure to a low level of radiation is linked to a small increase in a
person's cancer risk, a study of nuclear power station workers found. An
international team studied over 407,000 workers in 15 countries, who
were followed up for around 13 years. 

The British Medical Journal study estimates up to 2% of cancer deaths
were due to radiation exposure.  

But they said the increased risk did not apply to people living near to
power stations. 

Ionising radiation is a well known cancer-causing agent. 

Current radiation protection recommendations are to limit occupational
doses to 100 millisieverts (mSv) over five years and doses to the public
to 1 mSv per year. 

These guidelines were based mainly on data from survivors of the atomic
bomb in Japan and the extrapolation of risks to the general population
and radiation workers is controversial.  

Researchers studied the thousands of nuclear industry workers in order
to get a better idea of their risk. 

Hiroshima 
Most were men and had been employed for at least one year in nuclear
power production facilities, or in related activities such as research,
waste management or fuel and weapons production.  

Ninety per cent of workers were exposed to a cumulative dose of under
50mSv, and less than 1% over 500mSv.  

Factors such as age, duration of employment, and socioeconomic status
were taken into account when the researchers looked at the workers.  

Just under 200 died from leukaemias, and 6,519 from other cancers.  

The researchers say, that from their evidence, 1 to 2% of deaths from
cancer among workers in this study may be attributable to radiation. 

The risk estimates from the study are consistent with those used for
current radiation protection standards, they say.  

And they add that many of the workers in this study worked in the early
years of the industry when doses tended to be higher than they are
today. 

'Vigilence' 
Dr Colin Muirhead, of the Radiological Protection Division of the Health
Protection Agency, who worked on the study, told the BBC News website:
"This is what we expected to see, because even with a low does of
radiation, there would be a cancer risk.  

He added: "The levels of exposure we saw in this study are much lower
than were seen in Hiroshima and Nagasaki.  

"In absolute terms, it is a fairly small increase in risk."  

He said workers in the industry should be reassured by the study's
findings. 

"At an individual level it will make a very small difference."  

And he said the results also fitted in with studies which had found no
link between cancer risk and living near a nuclear power station.  

"It's certain that for the population, exposure would be much lower than
what we're talking about here.  "There is no inconsistency."  

Professor John Toy, Medical Director at Cancer Research UK, said:
"Radiation is a very well known carcinogen.  

"This extremely large study shows an increased risk, albeit small, of
cancer and most types of leukaemia associated with occupational low-dose
radiation exposure."  

He added: "The radiation risk estimates are statistically comparable
with those used for current radiation protection standards.  

"The nuclear industry must remain ever vigilant to ensure these
standards are not breached and constantly endeavour to reduce the
exposure of its workers to radiation." " 

Fred Dawson

====================

The report itself, available online as noted above, has some interesting
features.

It is entirely a report of dose-response coefficients using pooled epi
studies for various groups of facilities, comprising 154 facilities in
15 countries, involving 407391 workers, 5192710 person-years, 24158
all-cause deaths, 6519 cancer (excluding leukemia) deaths, 196 leukemia
(excluding chronic lymphocytic leukemia) deaths, and an average
individual cumulative dose of 19.4 mSv.

I am guessing that most of these studies and pooled results would show
the workers exhibiting a healthy-worker effect, with lower cancer and
all-cause death rates than the appropriate general population control
population.  The authors do not give us any information about incidence
and mortality rates and do not provide a URL for that information.

They report that confounding due to smoking may be partly, but not
entirely, responsible for the positive dose-response coefficients they
report.

They do an interesting jackknife analysis, excluding one country at a
time from the analysis, and reporting excess relative risk per Sv
ranging from 0.58 (excluding Canada) to 1.25 (excluding the UK).  When
Canada is excluded the ERR/Sv = 0.58 with 95% CI (-0.22 to 1.55), that
is, the weakly positive ERR/Sv is no longer statistically significantly
different from zero slope.

Figure 2 tells a similar story.  It gives error bars showing ERR/Sv and
the 95% CIs for Canada, Sweden, UK, US Hanford, US nuclear power plants,
US ORNL, and finally all combined.  Canada has ERR/Sv approx. equal to
6.8 with 95% CI approx. equal to (2.8 to more than 10).  All combined is
approx. (1.0; (0.2,1.9)).  All of the others are not statistically
significantly different from zero slope dose-response and the slopes for
Sweden and US nuclear power plants are actually negative.  US ORNL has
ERR/Sv approx. = 4.4 with 95% CI approx. (-0.3, more than 10).  The ORNL
results may reflect epi studies produced by the Steve Wing et al. epi
studies factory, although none of the relevant reports is cited in the
BMJ report.

Since the authors/compilers of these data chop it up into many pieces
and perform many calculations of confidence intervals, we would expect a
number of spurious rejections of the null hypothesis (that is,
spuriously positive ERR/Sv values) if it were actually true.  There is
no indications that the authors/compilers used any of the standard
approaches to minimize this source of error.

Finally, what do the calculated ERR/Sv values mean in the real world?
For a worker with a lifetime dose of 100 mSv (according the report's
Fig. 1 approximately 20000 of the 407000 workers), the ERR would be
approximately 0.1.  For workers receiving the average 19.4 mSv, the ERR
would be approximately 0.02.

Best regards.

Jim Dukelow
Pacific Northwest National Laboratory
Richland, WA
jim.dukelow at pnl.gov

These comments are mine and have not been reviewed and/or approved by my
management or by the U.S. Department of Energy.