Who's afraid of radiation?
The panic over low-level radiation is hampering medical diagnosis and
treatment, reports Dr David Hall
The threat posed by radiation is rarely out of the news. The government's
decision to transport five kilograms of uranium from a reactor in Georgia
to Scotland led to protests by boiler-suited Greenpeace supporters in Downing
Street and mass media coverage. Articles about leukaemia clusters in areas
near to nuclear power stations pepper both the scientific and popular press.
And, 12 years on, the after-effects of Chernobyl are still being debated.
Two themes stand out. The first is that, testing of atomic bombs by India
and Pakistan aside, the main concern is over the long-term, even intergenerational
effects of prolonged exposure to relatively low doses of radiation rather
than the immediate destructive effects of nuclear conflict. The second is
the tendency to fear the worst. For example, the available scientific evidence
suggests that radiation exposure is not responsible for leukaemia clusters
(GJ Draper et al, British Medical Journal, 8 November 1997). However, as
a recent MORI poll found, 83 per cent of people trust scientists from green
groups, whereas only 47 per cent trust government scientists; so radiation
risks tend to be played up (New Scientist, 18 April 1998).
One consequence of this excessive fear of radiation is that an enormous
amount of money is spent on radiation protection. Preventing one death
at Sellafield by radiation-protection measures during normal use has been
calculated to cost £30 million ('Economic valuation of statistical life
from safety assessment', D Fernandes-Russell et al, University of East
Anglia, 1988). But there are other consequences, and at this point I should
declare a personal interest.
I am a scientist working in the NHS, using radioactive materials. Radiation
is widely used in hospitals today for both diagnosis and treatment. All
of us have x-rays at one time or another, and we are familiar with the
detail that can routinely be seen within our bodies. In addition nearly
one in 10 of us will have radiotherapy for cancer at some time in our lives.
Radioactive materials are also used in a medical context; for example,
as tracers to diagnose the spread of cancer, or in larger doses to treat
overactive thyroids.
In fact modern medicine could not do without radiation, in one form
or another. So while I would certainly question whether £30 million to
avoid the possibility of one death is money well spent, I am more concerned
about how doom-laden proclamations concerning radiation are influencing
both public attitudes and medical use. In rushing to protect patients from
extremely unlikely effects we might compromise everyday diagnosis and treatment.
What are the risks? The radiation doses received in medical tests are
fairly low and very unlikely to be dangerous. However, since many people
receive these low doses it is important that their effects are well understood.
The starting point for an examination of the issue is the fact that low-dose
effects, of any kind, are hard to study. If only a small number of people
suffer from a disease any variation, whether chance or otherwise, can seem
large; so an increase in local leukaemia cases from just 10 to 15 in a
year can be headlined as 'Child cancer soars by 50 per cent'. There are
variations from area to area in the incidence of all sorts of rare diseases,
and connecting these to any cause is fraught with difficulty. However, if
there is any hint of exposure to radiation, it tends to be assumed that
this is the cause of the increase, unless it can be proved otherwise. If
the actual levels of radiation detected are then shown to be too small
to cause an effect, it is easy to conclude that the risk of any particular
dose has been underestimated.
So where do the official risk estimates come from? They are largely derived
from studies of the survivors of the Hiroshima and Nagasaki atomic bombs.
But even here it is hard to come to clear-cut conclusions, since of the
100 000 atomic bomb survivors, '20 000 are going to die of cancer anyway.
We are looking at the difference between 20 000 and 20 400' (Eric J Hall
from the Centre for Radiological Research at Columbia University, quoted
in the Washington Post, 14 April 1990).
People who have received large radiation doses have measurable increases
in rates of cancer, but smaller dose effects are much harder to measure,
since the rates of cancer are close to normal. The simplest way then to
estimate the risk of small doses is to assume that the risk is proportional
to the dose, all the way down to zero. This is called the Linear No Threshold
model (LNT) and is quite influential.
Given that we are all exposed to radiation from natural sources, small
additional amounts from man-made sources should not be much of a cause
for concern. Many environmentalists question this, and argue that the insidious
effects of long-term exposure to artificial radiation have been underestimated.
But we should also consider the other possibility: that the LNT model overestimates
the problem.
The Hiroshima and Nagasaki data refers to individuals who received a
radiation dose at a relatively high rate, which might be expected to overwhelm
the body's defences. Perhaps a lower dose, experienced over a longer time
span, might not overwhelm the defences and so would have little or no effect-even
if the cumulative dose received were the same. Consider this analogy. Taking
one aspirin a day for a month to protect from heart disease has no ill
effects; taking 30 at one go is a different matter.
The effects of low dose and low dose-rate radiation have been considered
in several large-scale studies, some of which have even suggested a health
benefit from low-dose radiation. For example, a Canadian study of breast
cancer rates in 32 000 women who received chest x-ray fluoroscopy for tuberculosis
found that small doses actually decreased the rate of breast cancer compared
with no dose, followed by an increase at higher doses. Another study, comparing
lung cancer rates with average radon concentrations across 1729 US counties,
covering 90 per cent of the population, found a reduction in the rate of
lung cancer proportional to rising dose (New Scientist, 14 March 1998).
The full effects of radiation are still imperfectly understood, but
should become clearer with a better understanding of the effects of radiation
on DNA. The idea of genomic instability, in which the effects of radiation
are propagated to descendants, and the idea that DNA repair mechanisms
are stimulated by low-level radiation (which is used to explain the counterintuitive
results given above), are but two ideas that remain to be properly explored.
Where does this leave us in relation to the medical uses of radiation?
Generally we should be sceptical of those who emphasise risks: if the results
we have in hand are even partly right it seems unlikely that the risks
from low-level radiation have been underestimated. From the point of view
of the individual patient it is worth pointing out that receiving radiation
for medical diagnosis or therapy is a good risk, one in which you risk
'little for the chance of much', as Samuel Johnson put it in the eighteenth
century. I am worried that some people are in danger of forgetting this
today, and of separating risk from benefit.
It is often assumed by clinicians that it is preferable to use a test
that does not involve ionising radiation, such as ultrasound or magnetic
resonance imaging, and that if radiation is used the smallest possible
dose should be employed. However, evidence suggests that in some cases
we should increase rather than reduce the amounts of radiation used.
An editorial in the European Journal of Nuclear Medicine in February
1997 estimated that, on risk-benefit grounds, the amount of radioactivity
used in many common tests should be increased by 10 to 20 times in order
to improve diagnosis. If somebody already has a severe condition the benefit
of learning more now must outweigh any possible small risk some years down
the line. The rules under which radiation is given do not take this difference
into account-they effectively treat everybody as if they were perfectly
healthy and unlikely to gain much benefit from the test.
In healthcare, as in other areas of life, it is often necessary to take
risks-good risks-in order to gain any benefit. With regu- lators adopting
a precautionary approach, and patients showing a worrying trend towards
litigation over the side-effects of treatments that have saved their lives,
this is a message that needs to be put across if both patients and doctors
are to avoid becoming victims of the fear of radiation.
Reproduced from LM issue 114, October 1998