You are here

Leukemia increases near nuclear power stations

Nuclear Monitor Issue: 
Ian Fairlie

The main aim of the 14th COMARE Report was to undertake a further review of the incidence of childhood leukemia near most UK nuclear power plants (NPPs). In 2008, a previous study(i) commissioned by the Department of Health had found a 36% increase in acute childhood leukemia's between 1969 and 2004 within 5 km of 13 of the 14 UK nuclear power stations. The observed increase was considered not to be statistically significant as there was a >5% probability that it could have arisen by chance.

In November 2009, the Department of Health requested COMARE (Medical Aspects of Radiation in the Environment) to extend the 2008 study to include more recent data in order to increase the statistical strength of its findings. This is reflected in COMARE’s Press Release which states that the new study examines data “…for the extended period…”

However the COMARE Report actually does not do this: it uses the same 1969 to 2004 time period as the 2008 study. The Report states (para 6.6) “….any significant amount of later information would have entailed a delay in carrying out the analysis. Later years are also becoming increasingly difficult to analyze satisfactorily because of the ways in which recent census data are made available.”

This excuse is not transparent and leaves several questions unanswered. One question is - what is therefore the difference between the 2008 study and the present one? The apparent difference is that the COMARE study now includes non-Hodgkins lymphomas (NHL), chronic myeloprolific diseases, and unspecified leukemia's - as well as the acute leukemia's examined in 2008. These are strange inclusions as there are no actual cases of these extra diseases in the 5 km circles near British NPPs in the study period, and these disease categories were used in neither the 2008 Bithell nor the KiKK study which were supposed to be replicated.

The Report finds a 22% increase in childhood acute leukaemia + non-Hodgkins lymphoma (NHL) + chronic myeloprolific disease + unspecified leukaemia. Therefore the net result of adding the new disease categories is to reduce the apparent increase in leukemia's/lymphomas near NPPs from 36% in the 2008 study to 22%. The Report states (para 6.40) that its study had a “negative finding”. But in statistics, it is incorrect to make negative conclusions merely because a study lacks statistical significance.

Evidence on leukemia's near NPPs
The COMARE Report refutes the clear pattern of epidemiological evidence across the world indicating increased leukemia risks near NPPs. In a study(ii) not cited by COMARE, Laurier and Bard examined the literature on childhood leukemia's near NPPs world-wide. They listed a surprising total of 50 studies (29 ecological, 7 case-control and 14 multi-site studies) the majority of which revealed small increases in childhood leukemia near NPPs although most were not statistically significant. In a later study, Laurier et al(iii) reviewed epidemiological studies on childhood leukemia at 198 nuclear sites in 10 countries, including 25 major multi-site studies. They found that increased risks of childhood leukemia near nuclear installations were a recurrent issue. The authors, employees of the French Government’s Institut de Radioprotection et Sûreté Nucléaire (IRSN), confirmed that clusters of childhood leukemia cases existed locally near NPPs but they declined to generalize their findings.

In fact, the 2008 Laurier et al study, taken together with Laurier and Bard’s 1999 study, indicate over 60 studies world-wide on increased childhood cancers near nuclear facilities, most of them finding cancer increases. It is hard to think of any other toxicity studies, eg with chemicals or biological agents, which remotely approach this number.

The findings of all these studies have been discussed by Fairlie and Körblein(iv) who concluded that “the copious evidence indicating increased leukemia rates near nuclear facilities, specifically in young children, is quite convincing, at least to independent observers.” The COMARE Report comes to the opposite conclusion and fails to discuss the preponderance of the evidence of the above-stated studies, ie the finding of increased leukemia's near NPPs.

Most important, is the German KiKK study(v, vi) (Kinderkrebs in der Umgebung von KernKraftwerken = Childhood Cancer in the Vicinity of Nuclear Power Plants) which found a 120% increase in leukemia's and a 60% increase in solid cancers among children under 5 years old living within 5 km of all German nuclear power plants. The KiKK report is significant because it is a large well-conducted study; because it is scientifically rigorous; because its evidence is particularly strong; and because the German Government, which commissioned the study, has confirmed its findings. The COMARE Report gives a number of reasons for refusing to acknowledge the KiKK study. These reasons are disingenuous and unconvincing.

The COMARE Report chooses to downplay the KiKK study, but it simply cannot invalidate the more sophisticated and rigorous KiKK study, as it attempts to do. First, the KiKK study found statistically significant cancer increases. Second, the KIKK study determined precise distances between the homes of cancer cases and NPPs to within 25 meters. In contrast, the COMARE study measured the distances between NPPs and the population centroids of irregularly-shaped electoral wards.

Finally KiKK is a case-control study, that is, it examined 593 leukemic children together with 1,766 controls. On the other hand, the COMARE study used geographical averages rather than parameters characterizing individual cases and controls. Such studies are termed ‘geographical’ or ‘ecological’ and they are much less reliable than case-control studies. Policy makers who should be guided by the best available scientific evidence should rely on the better KiKK study rather than the COMARE study.

Exclusion of Calder Hall reactors
Only 13 of the 14 UK nuclear power plants were used in COMARE’s leukemia study. The Report states (para 6.12) that the former Calder Hall nuclear power station at Sellafield was excluded from its study. This raises the question as to why. This is an important matter because in the 1980s and 1990s several epidemiology studies revealed relatively large numbers of excess leukemia's (> 7) at Seascale a small village less than 5 km from Sellafield. If these had been included, the Report acknowledges (para 6.13) “…the result would have yielded a higher estimate of risk…”.

The COMARE Report (para 6.12) gives the following reasons for the exclusion

(a) “The observation of an excess of childhood leukaemia near Sellafield was the ‘hypothesis-generating’ observation and good scientific practice proceeds by attempting to test hypotheses on independent sets of data.

(b) Power generation has always been an incidental part of the activities on the Sellafield site, which have included nuclear operations (eg reprocessing) that release considerably more radioactivity into the environment than Calder Hall.

(c) The well-known excess of childhood leukaemia cases in the village of Seascale adjacent to the Sellafield site would have an undue influence on the overall results, and distort the findings for the group of NPPs.”

These reasons do not stand scrutiny. As regards (a), the purpose of the COMARE study was to ascertain the number of increased leukemia's near all UK nuclear facilities, not to test a hypothesis. The phenomenon of increased leukemia's near NPPs had already been convincingly shown by KiKK and many other studies: scientifically speaking, there was little reason to have to test any such “hypothesis” again.

Reasons (b) and (c) are largely the same. Reason (b) contains an interesting admission that the release of radioactivity into the environment may be a causative factor for the increased leukemia's. However its attempt to divorce reprocessing from nuclear power is disingenuous: most UK nuclear power generation would be impossible without a means for dealing with spent nuclear fuel - the large majority of which is still reprocessed. Reprocessing is therefore an integral part of nuclear power in Britain and its radioactive discharges should logically be included in any reckoning of its health effects. From the point of view of the health of nearby citizens, it does not matter whether the radiation emanates from a reprocessing plant or from nuclear reactors.

This problem could have been addressed by presenting the data with and without Calder Hall: in other words, by widening the study to include all nuclear installations not just NPPs. Indeed this was indicated by the title of the Bithell et al (2008) study “Childhood Leukemia near British Nuclear Installations”. Reason (b) states that reprocessing releases considerably more radioactivity than NPPs: this is true for sea discharges but not necessarily for air emissions which are responsible for the majority of the collective dose to local people. Annual air emissions for some nuclides (especially C-14) from the four Calder Hall reactors could be of similar magnitudes to those from reprocessing. For example, using data available to the author, C-14 releases from Calder Hall in 1995 were 1.4 TBq(vii) compared with 2.62 TBq(viii) for all Sellafield facilities (including Calder Hall) in 1998.

20. Para 6.13 states “.. had the data from this site been included – the results would certainly have yielded a higher estimate of risk, but it would have been entirely unclear what implications this had for purpose-built power-generating plants.” But surely COMARE was established to concern itself primarily with the health of people living near NPPs rather than the need to construct purpose-built power-generating plants?

Resurrection of discredited reason for leukaemia increases
The new Report states (para 1.3) "There is growing epidemiological evidence that childhood leukemia is linked to infections...either a rare response to a common infection...or a rare response to general exposure to infectious agents...however the biological mechanism underlying these hypotheses remain the subject of considerable scientific debate."

No such agent has been remotely identified, and the source, pathway and receptor for any such infectious agent are unknown. This myth is periodically recycled but it has been comprehensively criticized(ix, x) in the past. The resurrection of the evidence-free notion of an infectious agent being responsible for the increased leukemia's is an embarrassment, and will act to discredit UK science in other countries.

The data in the COMARE Report indicate a 22% increase in various types of leukemia's and non-Hodgkins lymphoma. However it concluded “that the latest British data has (sic) revealed no significant evidence of an association between risk of childhood leukemia … and living in proximity to an NPP”.

This statement pivots on the equivocal meaning of the word “significant”. COMARE rejects the 22% increase by incorrectly implying that, as its findings did not meet a significance test, the findings were negative - a type II error in statistics. COMARE’s Report is regrettable as it may mislead members of the public into thinking there are no increases in leukemia's near UK nuclear power stations when in fact this may not be the case. The Report should have said that it found increases ranging between 22% and 47%; that these increases did not meet the statistical test used by COMARE; but that this could be due simply to the low numbers in the study and not to lack of effect.

In three areas, the COMARE Report’s handling of epidemiological data is not transparent

* it excludes recent data on child leukemia's near NPPs after 2004, despite being established to do precisely that

* it includes new categories of lymphomas and leukemia's although none were actually observed and although neither KiKK nor the 2008 Bithell study examined these types

* it excludes data from the Calder Hall nuclear power station although they state "...their inclusion would certainly have yielded a higher estimate of risk.”

This irregular handling of data unfortunately lays the COMARE Report open to accusations of selecting or ‘cherry picking’ their data. In order to dispel any doubts in this area and increase transparency, it is recommended COMARE should release its data 

* on the observed numbers of childhood acute lymphoblastic leukemia's within 5 km of NPPs between 2004 and 2010, and

* on the observed numbers of childhood acute lymphoblastic leukemia's within 5 km of Sellafield between 1969 and 2010.


(i) Bithell JT, Keegan TJ, Kroll ME, Murphy MFG and Vincent TJ. 2008. Childhood leukaemia near British Nuclear Installations: Methodological Issues and Recent Results. Radiation Protection Dosimetry. vol 45:1–7.

(ii) Laurier D, Bard D (1999) Epidemiologic studies of leukaemia among persons under 25 years of age living near nuclear sites. Epidemiol Rev.1999;21(2):188-206.

(iii) Laurier D, Jacob S, Bernier MO, Leuraud K, Metz C, Samson E, Laloi P. Epidemiological studies of leukaemia in children and young adults around nuclear facilities: a critical review. Radiat Prot Dosimetry. (2008) 132(2):182-90.

(iv) Fairlie I and Körblein A. A Review of epidemiology studies of childhood leukaemia near nuclear facilities: commentary on Laurier et al. Radiat Prot Dosimetry. 2010 Feb;138(2):194-5; author reply 195-7.

(v) Kaatsch P, Spix C, Schulze-Rath R, Schmiedel S, Blettner M. leukaemias in young children living in the vicinity of German NPPs. Int J Cancer. 2008;122:721–726.

(vi) Spix C, Schmiedel S, Kaatsch P, Schulze-Rath R, Blettner M. Case–control

study on childhood cancer in the vicinity of nuclear power plants in Germany

1980–2003. Eur J Cancer. 2008;44:275–284.

(vii) BNFL Annual Report on Radioactive Discharges and Monitoring of the Environment 1995.

(viii) Radioactivity in Food and the Environment (RIFE 4) MAFF 1999.

(ix) Russell Jones R (1993) Infective cause of childhood leukaemia. Chapter in: Childhood cancer and nuclear installations. edited by Valerie Beral, Eve Roman and Martin Bobrow. BMJ Publication Group: 1993. London.

(x) Hewitt H. 1994.The Gardner hypothesis: old infective theory discredited. BMJ Jan 1;308 (6920):60.

Dr Fairlie is an independent consultant on radiation in the environment. He was a member of CERRIE - the independent government ‘Committee Examining Radiation Risks of Internal Emitters’.He expresses his thanks to Dr Alfred Körblein for his help, particularly on statistical tests.

(This article has been shortened and edited by WISE Amsterdam. An unedited version is published as Policy Briefing 82 by the Nuclear Free Local Authorities and available at:

Emails reveal why COMARE report was delayed.
The publication of the long-awaited and fiercely-disputed COMARE report on the radiation risks of nuclear power stations was accelerated but then delayed at the insistence of the government’s Department for Energy and Climate Change (DECC), internal emails reveal. The Department of Health (DoH) released 79 pages of emails about the publication of the COMARE report in response to a request under freedom of information legislation.

DECC initially insisted that the DoH publish the report as soon as possible to help combat a court challenge on nuclear power. The government was being sued by an anti-nuclear activist for failing to take account of radiation risks. On 10 March 2011 Peter McDonald, from DECC’s Office for Nuclear Development emailed health officials urging publication. “In the interests of transparency, the best possible thing is that the report is published as soon as possible and that, if anything, greater urgency is needed precisely because of the pending court action.”

But the day after his email, an earthquake and tsunami in Japan knocked out the Fukushima nuclear plants and their back-up safety systems, triggering the world’s worst nuclear accident since Chernobyl 25 years ago. As a result, DECC changed its tune. In an email on 14 March, a DoH official said a submission to ministers to publish the report on 21 March had been withdrawn. “DECC have asked us to delay the publication of the COMARE report for a while given the current nuclear issues in Japan due to the earthquake,” wrote the official, whose name has been blacked out. “We ask that DECC keep us informed as to the situation and when it might be appropriate to publish the report.”

Ten days later, on 24 March, DECC changed its mind again. “DECC have just rang and are now content for us to seek permission to publish," wrote a DoH official. But then the problem became intruding on the election campaign that had just begun in Scotland, offending the civil service tradition of not interfering with elections. In the end, after seeking permission from the Prime Minister, David Cameron, the report was published on 6 May, the day after the Scottish election., 4 July 2011

Source and Contact: Dr Fairlie, 115 Riversdale Road, London N5 2SU, United Kingdom