Shooting the Miners' Canary
The utter inadequacy of the CERRIE Majority Report's treatment of
the increase in infant leukaemia after Chernobyl
EXTERNAL PARADIGM PROBLEMSThe question CERRIE existed to answer was Does ICRP’s methodology accurately predict the health consequences of radioactivity inside the human body? The basic message of the Minority Report is firstly that it does not, and secondly that although the Committee’s discussions covered an acceptably wide range of topics (with the notable exception of Depleted Uranium) its reporting process was biased and ultimately failed to meet the requirements that the final report should be agreed by all Committee members and should explain the reasons for any disagreements remaining at the end of the Committee’s life. The COMARE 9th Report, which advises the Departments on CERRIE’s findings, confirms that the Majority Report does not identify the reasons for the lack of consensus.
Here we explain the inadequacy of CERRIE’s treatment of an unequivocal body of evidence which indicates a massive error in the estimation of risk. While the Majority Report and the media spin that accompanied both its publication and the earlier leak of its contents (in July 2004) emphasise a range of uncertainty, there is credible epidemiological evidence that certain types of exposure uniquely associated with man’s activities consistently incur risks greater than those predicted by conventional models. The implied error is at least 100 times and maybe up to 1000 - well within the uncertainties identified by the Majority Report (see www.cerrie.org).
The conventional view of radiation risk is that, irrespective of whether radiation originates inside or outside the body, cancer risk is in strict proportion to the average energy it delivers into body tissue. For external radiation it is valid to assume an average, because body tissue is uniformly irradiated – for example by cosmic rays, X-rays, or an exploding A-bomb. Averaging is also valid for some radioactive substances even when they are inside the body because they are evenly distributed in tissue. Some are natural (e.g. Potassium 40) and some are man-made (e.g. Caesium 137) However, some other internal emitters, Plutonium and Uranium for example, create a very different type of exposure, since they exist in the form of microscopic hot particles. These do not exist naturally but they are now common in the environment as a result of routine discharges from power stations, reactor accidents like Chernobyl, nuclear weapons, and the military use of Depleted Uranium ammunition. With diameters of less than a few millionths of a metre they are highly mobile and can be inhaled or swallowed. Once embedded in body tissue they deliver their energy so locally that the few cells immediately next to them are chronically irradiated at very high energies while the rest of the body gets no dose at all. In such circumstances the concept of an “average dose” is nonsensical.DOSE IS “MEANINGLESS”CERRIE’s Majority Report has conceded that dose averaging is a false assumption. It states that for internal radiation sources the concept of dose is sometimes meaningless. This is massively significant because official agencies universally quantify risk in terms of dose, which allows them to reassure the public by comparing nuclear industry discharges with the 2 millisieverts we get every year from natural radiation. According to their logic, annual exposures from the industry are no worse than the cosmic rays you’d receive while flying to and from Tenerife for a holiday. But if the discharges involve types of radiation for which “dose” means nothing, the agencies know nothing and can give no valid advice or reassurance. It’s as if they were equating the risks of shopping in sniper’s alley with the health effects of average levels of lead pollution just because the sniper’s bullets are also made of lead.SECOND EVENT THEORY STILL AFLOATA more complex consideration is that, unlike the natural radionuclides, some man-made substances are sequential emitters each atom of which can decay more than once in a relatively short period. Since many of them also mimic elements vital to biological functions this means that repeated decays are taking place in fixed locations. Strontium-90 is an outstanding example – being a calcium mimic it will tend to bind to DNA; its daughter isotope Yttrium-90 is also unstable with a half life of a few hours, so that once the Strontium atom has decayed, possibly damaging the DNA and initiating DNA repair, there is a relatively high probability that the Yttrium will emit a second radiation track before the repair is complete. It is known that the repair process contains a time period during which radiation is highly likely to induce a fixed mutation. This is Chris Busby’s “Second Event” theory (or “SET”) which concerns the probability of two radiation events from internal sources relative to two events from external sources in the same time frame and affecting the same tissue of interest, which in some cases will be as small as a DNA molecule. Since each year on average natural background confers on each cell in the human body one radiation track (which may not hit the DNA) the hazard enhancements for sequential emitters bound to DNA can be very large; the implications for incorporated hot or warm particles are obvious. This was the subject of lengthy debate in CERRIE which while it did lead to refinements of the theory did not destroy it. We will update the SET section of this site when we can. In the meantime readers could compare the CERRIE Majority Report’s account of the theory with the Minority Report. One salient point is that the theory’s critics seem (both from the Majority Report and from their scoffing in the final meeting, June 24 2004) to have failed to understand what Busby had said about the relevance of experiments showing enhanced cell sensitivity to radiation (see Minority Report pp 65-6). To their mind the experiments were at too high a dose to be relevant to the very small doses with which SET is concerned, but they ignored the fact that the experiments were using external radiation so doses had to be high to produce multiple hits; the microscopic tissue volumes receiving the radiation from hot particles or from single organically bound atoms of sequential emitters experience energy deposition densities equivalent to the average external doses in the experiments. (This is one reason why “dose“ is meaningless.)INFANT LEUKAEMIA: AN ACID TESTDuring the summer of 2004 Michael Meacher repeatedly drew attention to a sharp increase in infant leukaemia observed in several countries after the Chernobyl accident in 1986. This was one of the main reasons why he established CERRIE and was a key piece of evidence laid before the Committee. The Majority Report’s treatment of it is a scandal.Leukaemia is recognised as an early indicator of radiation damage; more specifically, infant leukaemia (i.e. diagnosed before a baby’s first birthday) signals damage acquired in the womb. Scientific journals have published papers by different research teams showing post-Chernobyl increases of between 20% and 330% in various countries as far apart as Belarus and the USA (Busby 2000; Gibson 1988; Ivanov 1998; Mangano 1997; Michaelis 1997; Petridou 1996).
This is a crucial challenge to conventional radiation risk estimates because only the Chernobyl fallout can have caused the disease in this very precisely defined subset of the population. The number of sick babies was small but they are in effect miners’ canaries, suggesting that radioactive discharges are contributing to the global epidemic of cancer. Their tragedies will become a far greater scandal if the world’s governments and their advisers fail to learn the lesson.
 |
The various studies
The vertical axis of this graph shows percentage increase in incidence of leukaemia diagnosed between birth and the 1st birthday for children born during the 20 months between the accident (26 April 1986) and January 1988, compared with the period before it and from 1988 onwards.
The horizontal axis shows doses in millisieverts.
|
ReferencesAt between 0.02 and 0.2 millisieverts (mSv) the doses shown on the graph are small fractions of average annual natural background radiation, but the big caveat here is that dose may mean nothing since it is calculated from Caesium fallout. Caesium emits highly penetrating gamma rays, so it is easy to monitor. For this reason it was extensively mapped after the accident, giving researchers a way of estimating doses. But the very fact that it is so penetrating means that its energy deposition (in the form of ionisations) is spatially well distributed in tissue, so its health effects are likely to conform with the external irradiation models. It is, moreover, soluble and does not form particles. The Chernobyl reactor fire produced other isotopes (including Sr-90) as well as microscopic Uranium fuel particles which travelled across Europe and beyond, exposing everyone in the path of the cloud to inhalation and ingestion. There is no reason why the health effects should conform with expectations based on Caesium deposition.FOUR-WAY SPLIT NOT EXPLAINEDThe dotted line shows the percentage increase in leukaemia CERRIE’s majority would expect for any given dose. It slopes up towards a point representing a 40% increase at a dose of 10 millisieverts - five times natural background (not show here; the page would need to be 2 metres wide on this scale). The origin of this yardstick is a survey of cancer in children (Stewart 1975). The dose is what they received when their mothers were X-rayed during pregnancy. X-rays are external exposures, so the yardstick itself is part of the model CERRIE existed to investigate.
All the post-Chernobyl studies show between 150 and 800 times more leukaemia than expected. We argued that this was prima facie evidence against the external risk model. Our opponents on CERRIE set out to show it could be ignored.
They argued that the statistical power of the individual studies was so low that no reliance could be placed on the overall observation. One strand of this argument depends on ignoring studies of Scotland and Wales. They can be seen in the graph; excess risks were high and statistically significant.
The second strand began with using wrong data and ended in nonsense. The first draft of the CERRIE Majority Report said radiation doses in Germany were the same as in Greece. This had the effect of reducing the apparent significance of the German study and CERRIE concluded that
… the only study to show a large discrepancy with the predictions of external radiation risk estimates is the Greek … study.However, we knew from UN monitoring that fallout in Greece was roughly four times higher than in Germany (Savchenko 1995). We told CERRIE’s Chairman that the stated doses were obviously wrong and the Majority Report was changed. It now contains the correct doses but, untenably, still concludes that only the Greek study is out of line with expectation.The third strand of the CERRIE case is an unsubstantiated slur on the quality of data collection in Greece. This was never discussed in Committee, but the Report’s implication is that the Greek study can therefore be ignored.
The fourth concerns the study from Belarus (not shown on the graph), where fallout levels from Chernobyl were more than seven times higher than in Greece ‘though the increase in infant leukaemia was smaller than anywhere else in mainland Europe. The Majority Report says the Greek study (the only one supposed to be an anomaly) is
... statistically inconsistent with … the study in Belarus where the highest doses from Chernobyl contamination were received.Statistically inconsistent with … is code, meaning that the observations challenge the dogma that dose and effect are always strictly proportional or “linear”, or in other words: “twice as much dose means twice as much cancer”. We had consistently argued in CERRIE that there are good reasons why disease may not always show linear relationships with dose. Infant leukaemia is just one example. It starts in the womb, so babies carried by pregnant women in high fallout areas will suffer more damage than in low fallout areas. As a result more babies will be miscarried or stillborn or will die before leukaemia is diagnosed. A high fallout area will therefore inevitably have a lower incidence of leukaemia per unit dose than a low dose area and possibly an absolutely lower incidence as well, as in the case of Belarus.Thus, by a combination of wrong and selective data, innuendo and dependence on assumptions which CERRIE had been set up to test, the Majority Report finds that the increase in each country could have happened by chance, so the overall increase could have happened by chance. This ignores the classic scientific dictum of “instance confirmation”; that is, studies which consistently show a trend increase our confidence that the trend is real. The Chernobyl infants’ studies satisfy Professor Sir Austin Bradford Hill’s famous features of reliable epidemiological studies (Bradford Hill 1965):-
“Biological gradient” is not satisfied, but Bradford Hill envisages circumstances in which a linear dose response would not be seen, and we have already given good reason why infant leukaemia would not display one. As the CERRIE Minority Report’s appendix of studies from the Chernobyl affected territories of Russia, Belarus and Ukraine shows, many disease phenomena show non-linear relationships with dose (see Minority Report Appendix 3).
- “strength” (Is the observed increase in risk large enough, relative to unexposed people, to draw a firm inference about causation? On this Bradford Hill cautions We must not be too ready to dismiss a cause-and-effect hypothesis merely on the grounds that the observed association appears to be slight.)
- “consistency” (Has it been repeatedly observed by different persons, in different places, circumstances and times?)
- “specificity” (Is there a specific association between the disease and the type of exposure?),
- “temporality” (Does the disease follow the exposure?)
- “plausibility” (Is the causation we suspect biologically plausible, bearing in mind that the association we observe may be one new to science or medicine and we must not dismiss it too light-heartedly as just too odd.)
- and “coherence” (Does the cause-and-effect interpretation of our data …. seriously conflict with the generally known facts of the natural history and biology of the disease?)
“Statistical significance”, so crucial to CERRIE’s dismissal of the infant leukaemia, is of minor importance according to Bradford Hill. Nonetheless, we can amalgamate the statistical tests contained in the various studies. The Scottish, Greek and German studies combined, for example, have a p value of 0.00065, meaning that an event on this scale occupying a two year period would not happen by chance in more than two thousand lifetimes. Common sense says that if the events were truly random, at least some of the data points would have been below the dotted line.
On this key issue the Majority Report shows a bizarre four-way split. The reasoning behind some of the views remains a mystery even to CERRIE members, since the Committee never went through an open process of identifying members’ opinions. The weird account in CERRIE's Majority Report is the italicised text on the left. On the right we translate it.
In the judgement of a large majority of Committee members, it is likely that radioactive fallout from the Chernobyl accident resulted in an increased risk of infant leukaemia in the exposed populations. A substantial fraction of members thinks that this increase is at the level anticipated from current risk models. However, another substantial fraction feels that these models may have underestimated the level of this increased risk. Of this latter group, two members further believe that the evidence for infant leukaemia suggests that the current risk estimates are appreciably in error. The remainder of the Committee believes that there exists relatively little evidence that lends support to this view. There is a consensus within the Committee that leukaemia incidence in infants post-Chernobyl merits further study. [CERRIE Majority Report Chapter 4 para. 26] One faction - its size is not stated - seems to think Chernobyl had no effect on how many babies got leukaemia. Most members thought it did have an effect, but were split three ways on how big it was; some thought all the data points could be interpreted as lying on the dotted line in the graph, others thought the risks might have been higher, but the report doesn’t say how much higher, nor who thought so, nor why. Two members (LLRC's representatives – at least we know who they were) insisted that the various scientific papers unequivocally show radiation is at least 100 times and maybe up to 1000 times more dangerous than conventional estimates. The majority felt there was relatively little evidence of this. The Committee’s remit required differences of opinion to be explained, but the Majority Report leaves the reader only to guess whether there were any scientific grounds for this wide divergence (as opposed to the obvious political motivations). The implications are huge, for if the risk factors are so grossly in error we have an explanation not only for the Seascale leukaemia cluster but for the global epidemic of cancer which started when the nuclear industry began to spread radioactive pollution around the planet. On this key issue, as on many others, the Majority Report completely fails to produce any reliable advice for policy makers.The Committee on Medical Aspects of Radiation in the Environment has colluded with the cover-up. Professor Bryn Bridges, Chairman of COMARE, attended every meeting of CERRIE as an observer. Professor Eric Wright was a member of both CERRIE and COMARE and also sat on the COMARE sub-Committee shadowing CERRIE, so despite the opacity of CERRIE’s treatment of the infant leukaemia, COMARE was fully aware of its importance. However, its 9th Report, which advises ministers on CERRIE’s findings, contains not a word about it.
If you know any members of COMARE or of CERRIE we urge you to ask them where they stand on the post-Chernobyl infants' leukaemia,
especially if they're Greens who start bleating Oh dear. There's a lot of uncertainty - we'd better invoke the Precautionary Principle! You could say If the increase in infant leukaemia was a matter of the Uncertainties, surely you'd expect at least some of the studies would show smaller risks than expected. Why are they all greater? Why do you use a weak argument when there's a much stronger one? Whose side are you on?
BRADFORD HILL (1965) “The Environment and Disease: Association or Causation?” Proceedings of the Royal Society of Medicine, 58, 295–300.BUSBY C and Scott Cato SM(2000) “Increases in leukaemia in infants in Wales and Scotland following Chernobyl” Energy and Environment 11 127-139
GIBSON B E S, Eden O B, Barrett A, et al.., (1988) “Leukaemia in young children in Scotland”, The Lancet, 630.
Ivanov E, Tolochko GV, Shuvaeva LP, et al (1998). “Infant leukemia in Belarus after the Chernobyl accident.” Radiat Environ Biophys, 37, 53–5.
MANGANO J, (1997) “Childhood leukaemia in the US may have risen due to fallout from Chernobyl”, British Medical Journal, 314: 1200
MICHAELIS J, Kaletsch U, Burkart W and Grosche B, (1997) “Infant leukaemia after the Chernobyl Accident” Nature 387, 246.
PETRIDOU E, Trichopoulos D, Dessypris N, Flytzani V, Haidas S, Kalmanti M, Koliouskas D, Kosmidis H, Piperolou F, Tzortzatou F, (1996) “Infant Leukaemia after in utero exposure to radiation from Chernobyl”, Nature, 382:25, 352.
SAVCHENKO VK (1995) “The Ecology of the Chernobyl Catastrophe”. Paris UNESCO
If you are seeing this page full screen (i.e. without a navigation bar on the left) you can't see how the rest of the site is organised.
This Home page link takes you to the index page, which has links to all the topics we discuss on the site [only use it if this page is full screen]