COMARE Comment at DECC Meeting Lacks Integrity

As many of us who were at the DECC Discussion on Saturday were not given the time to discuss anything in depth I enclose a letter from Dr Ian Fairlie talking about COMARE

Letter to JRP -July 1, 2009


COMARE Response on the KiKK study (Kaatsch et al, 2008)

Comment by Dr Ian Fairlie



Explanation: COMARE’s undated response (one page: - reproduced below for ease of reference) does not appear on its website as of July 1, 2009. It was sent to me on June 17 by Professor John Haywood, chairman of the Environmental Health Subcommittee of West Cumbria Sites Stakeholder Group. Professor Haywood received his copy from the COMARE chairman, Professor Alex Elliot. It is understood that other SSGs have received copies of the COMARE paper.



COMARE’s response (2009) generally seeks to diminish the significance of the KIKK study, although it acknowledges that “this is an interesting finding from a respected group”. COMARE states that its 10th report (COMARE, 2005) examined leukaemias (and non-Hodgkins lymphomas) near 28 GB nuclear facilities and found no evidence of excess cancers. However its methodology and chosen parameters have been superseded by the KiKK study which revealed the cancer increases were restricted to within 5 km of NPPs and to under 5 year olds. Therefore it is perhaps unsurprising that COMARE’s 10th report, which used 25 km radii and under 15 year olds, did not find an excess.



More important, COMARE’s 10th report was of ecological design, examining only incidence data. Such studies are considered less reliable because of their susceptibility to confounders and their failure to examine individual cancer cases. Generally their findings are treated as preliminary until stronger evidence (such as that provided by KiKK) emerges. The KiKK study was a four year long case-control study in which individual cancer cases were examined and matched with carefully-selected controls. It is invidious to suggest, as the COMARE page seems to do, that the results of an ecological study invalidate those of a case-control study.

Unfortunately the COMARE page omits two relevant studies whose results appear to be consistent with those of KiKK. The first was by the UK Childhood Cancer Research Group (funded mainly by the Department of Health) of cancer increases in under 5 year olds within 5 km of most UK reactors (Bithell et al, 2008). This found a 23% [O=18, E = 14.58 (95% CI= 0.73:1.95)] increase in childhood leukemias, although the numbers were too small to permit statistical significance. Of course, lack of statistical significance does not necessarily mean there is no effect, merely that the data numbers were not large enough to pass the authors’ statistical test. A combined analysis (Körblein, 2009) using the Bithell and KiKK data does provide a statistically significant increased risk within 5 km circles of German and GB NPPs, thus lending support to both studies. For this reason, the statement by Bithell et al that their data do not support the German findings is incorrect.

The second recent study (Laurier et al, 2008) also found a small cancer increase within 10 km zones near French nuclear power stations, although again the numbers were too small to be statistically significant at p=0.05 level. The point is that the findings of these UK and French studies are consistent with those of the KiKK study.

Second, COMARE states that its 11th report (COMARE, 2006) found childhood cancers occurred non-randomly within GB and linked this effect to socio-economic factors. The COMARE page said their 11th report allowed the cancer increases found around UK nuclear installations “to be put into context against the UK as a whole, an analysis … not addressed by the German study”.

This conflation of naturally-occurring cancers arising nation-wide with specific cancers arising near NPPs muddies the issue. Although it is well known that childhood leukaemias are non-randomly distributed, to suggest that this is a cause of the cancers near NPPs is far-fetched in the light of the KiKK study and the over 50 other studies world-wide (Laurier and Bard, 1999; Laurier D, Jacob S, et al, 2008) showing a fairly consistent pattern of increased leukaemias near NPPs.

In addition, socio-economic factors affect most epidemiology studies to some extent or other. Therefore to suggest this confounder is linked to a possible explanation for cancers near NPPs is unconvincing. It is also inconsistent with our knowledge that childhood leukaemia in Western countries occur more often among the offspring of wealthy not poorer families.

Third, COMARE states Kaatsch et al (2008) acknowledged their study could not account for the confounder of socio-economic status, which could influence their results. But Kaatsch et al were less emphatic: they stated their results "may possibly be influenced by confounders (like social class, pesticides, factors influencing immunological factors, exposure to other ionizing radiation)." In other words, social class was merely one of four possible factors. More important, the companion study by the same team (Spix et al, 2008) stated that, as regards uncontrolled confounding "no risk factors of the necessary strength for this [KIKK] effect are known for childhood cancer and specifically childhood leukemia."

The KiKK team actually tried to control for these confounders in a separate analysis but there was some self-selection among the controls interviewed, meaning they might not have been representative of the study population. For this reason, the results of the separate confounder analysis were not presented in their published reports. However the team revealed that "none of them [ie the confounders] changed the distance parameter by more than one standard deviation". In other words, the confounders studied by the KiKK team, including socio-economic status, may have had little effect on the KiKK findings.

In sum, COMARE’s response is rather unsatisfactory. It is untransparent in its omission of relevant studies; unscientific in suggesting that its ecological studies have more ‘weight’ than case-control ones; and its interpretation of KiKK’s treatment of confounders is possibly misleading. This is all very unfortunate and it is not surprising that COMARE have refrained from publishing these (perhaps preliminary) views on its website to date.



Ian Fairlie

115 Riversdale Road

London N5 2SU

United Kingdom

REFERENCES



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



COMARE 2009. COMARE comment on the publication 'Leukaemia in young children living in the vicinity of German nuclear power plants' - Kaatsch et al, 2008.



COMARE. 2006. Committee on the Medical Aspects of Radiation in the Environment. The distribution of childhood leukemias and other childhood cancer in Great Britain 1969–1993. Eleventh report. London: Health Protection Agency.



COMARE. 2005. Committee on the Medical Aspects of Radiation in the Environment. The incidence of childhood cancer around nuclear installations in Great Britain. Tenth report. London: Health Protection Agency.



Kaatsch P, Spix C, Schulze-Rath R, Schmiedel S, and Blettner M. 2008. leukemias in young children living in the vicinity of German nuclear power plants. Int J Cancer 122:721–726.



Körblein A. 2009. Neue Ökologische Studien zu Leukämien bei Kleinkindern um Kernkraftwerke. [In German] Strahlentelex 528-529:1-2. (In German)

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



Laurier D, Hémon D, Clavel J. 2008. Childhood leukemia incidence below the age of 5 years near French nuclear power plants. J. Radiol. Prot. 28: 401–403.



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



Spix C, Schmiedel S, Kaatsch P, Schulze-Rath R, Blettner M. 2008. Case–control study on childhood cancer in the vicinity of nuclear power plants in Germany 1980–2003. Eur J Cancer. 44:275–284.



COMARE comment on the publication 'Leukaemia in young children living in the vicinity of German nuclear power plants' - Kaatsch et al, 2008.

The recent publication by Kaatsch et al in the International Journal of Cancer reports the results of a case-control study of leukaemia in young children under five years of age living in the vicinity of nuclear power stations in western Germany. The study was based upon almost 600 cases diagnosed throughout western Germany during 1980-2003 and just over 1750 control children unaffected by leukaemia, and finds an increased risk of leukaemia among young children living within 5 km of a nuclear power plant. The same research group has also published (in the European Journal of Cancer) the findings of a similar case-control study of all types of cancer in young children and finds that the elevated risk in the vicinity of the nuclear power stations was largely restricted to leukaemia. COMARE acknowledges that this is an interesting finding from a respected group.

COMARE have previously investigated the relationship between childhood cancers and nuclear power installations in the UK and found no evidence of excess numbers of cases of leukaemia or other cancers in any local 25 km area around any of the nuclear power stations, publishing the results in its tenth report (2005). This work utilised a unique data set of over 32,000 cases of childhood cancer covering the time period 1969-1993 for children under the age of 15 years. The UK database is a verified national database and is believed to be the largest national database on childhood cancer in the world. A variety of specific statistical techniques (designed to maximise the chance of a positive finding) were used in the analyses, providing more robust results than previous studies, due to the large size of the database. A subsequent study of childhood leukaemia in areas around French nuclear power plants reported by Evrard et al in the British Journal of Cancer came to the same conclusion.

Importantly, in addition to the analyses around nuclear power stations in the tenth report, the eleventh COMARE report (2006) used the same database to provide an overview of the geographical pattern of the variation of childhood cancers on a nation-wide basis. This allowed patterns found around nuclear installations to be put into context against the UK as a whole, an analysis which was not addressed by the German study. It was found that childhood leukaemias (and some other childhood cancers) occur in a non-random fashion within the population of Great Britain, an effect related to socio-economic factors; there is increasing evidence that socio-economic factors, linked to patterns of infection experienced by children, are important in determining the risk of childhood leukaemia. Kaatsch et al acknowledge that their study could not account for certain factors, such as socio-economic status, which could influence their results; this may be one reason for their findings. A companion paper by the same group comments on a general and significant rise in German childhood leukaemia rates, most notably in the former East Germany, and hypothesises a link to changes in lifestyle factors, which have been especially marked in the former East Germany.

COMARE believes that the German data should be investigated further, but does not believe that they invalidate the conclusions reached in the tenth report concerning British nuclear power stations. Again, the size of the data set used by COMARE gives considerable confidence in the important results for the non-uniform pattern of childhood leukaemia throughout Great Britain published in the eleventh report.