As many agencies and individuals have requested my comments, this is an initial response to the new Bithell et al study on childhood leukemias near nuclear power stations. Its methodology raises several questions to which answers are not yet available. This reply should therefore be viewed as interim.
Contrary to many other studies, the new report fails to find an increase in leukemias and non-Hodgkin lymphomas near nuclear power stations. However the authors admit that their negative findings are not statistically significant and that their study is statistically underpowered compared with many other studies, especially the powerful KiKK study, which do find increases. The normal rule is that weak epidemiological studies which are not strong enough to pick up effects should be careful about making negative conclusions, as this one does unfortunately.
Bithell et al (2013) (see references below) have published a new study on the numbers of cases of childhood leukaemia + non-Hodgkin lymphoma (LNHL) in Great Britain between 1962 and 2007. These were compared with matched controls and analysed by regression to estimate the risk of residential proximity to UK nuclear power plants (NPPs). The comparisons were made in two ways – by residence at the time of birth and by residence at the time of diagnosis.
The authors stated their results showed little evidence of an increase in risk of LNHL to children from living in the vicinity of 13 UK NPPs. In essence, the authors found 13 cases of LNHL among under fives living within five km of the UK NPPs. These 13 cases were fewer than the 14 matched controls near NPPs, so the authors concluded there was “little evidence of an increased risk”.
However there are several problems apparent with the new study.
First is the discrepancy between their observed 13 LNHL cases (residence at time of diagnosis) and the 20 LNHL cases found in the 14th COMARE report (COMARE, 2011). This is so even though their new study period (1962 to 2007) is 10 years longer than the 14th COMARE report (1969 to 2004). Other things being equal, the new study should have found more, not fewer, cancers.
I have contacted Dr Bithell about this and will amend this interim post in the light of his reply.
Second, the results of the new study are inconsistent with the very large number (over 60) of epidemiology studies of cancers near NPPs around the world, the large majority of which have found increased risks. These studies have been discussed by me previously (Fairlie, 2010).
In particular, the new study’s results conflict with several recent powerful studies including the German KiKK study (Kaatsch et al, 2008) which found increased risks. The authors themselves state that their derived risk estimates are incompatible with the KiKK risk estimates.
In addition, their results do not agree with a recent meta-analysis of four European studies (Körblein and Fairlie, 2012) nor with the COMARE 14 report itself which found a 22% increase in LNHL. Some discussion of possible reasons for this inconsistency with previous studies is necessary.
Third, there are several problematic technical points such as the authors’ use of “other cancers” instead of no cancers for the controls in their diagnosis analysis. This depends on their hypothesis (that there is no association between increased risks and proximity to NPPs) being true when it might not be true.
Fourth, the authors themselves admit their study is statistically underpowered, certainly when compared with the more powerful KiKK study. They also admit that their findings are not statistically significant. The usual rule is that weak epidemiological studies which are not strong enough to pick up effects should be careful about making negative conclusions. As Altman and Bland (1995) have stated, absence of evidence – as found in this study – should not be construed to mean evidence of absence.
Altman DG, Bland J. Absence of evidence is not evidence of absence. BMJ. 1995;311:485.
Bithell JF, M F G Murphy, C A Stiller, E Toumpakari, T Vincent and R Wakeford. (2013) Leukaemia in young children in the vicinity of British nuclear power plants: a case–control study. Br J Cancer. advance online publication, September 12, 2013; doi:10.1038/bjc.2013.560.
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.
COMARE (2011) Committee on Medical Aspects of Radiation in the Environment Fourteenth Report. Further Consideration of the Incidence of Childhood Leukaemia Around Nuclear Power Plants in Great Britain. HMSO: London. (http://www.comare.org.uk/).
Fairlie I (2010) Hypothesis to explain childhood cancer near nuclear powerplants. Int J Occup Environ Health 16: 341–350.
Körblein A and Fairlie I (2012) French Geocap study confirms increased leukemiarisks in young children near nuclear power plants. Int J Cancer 131: 2970–2971.