Comment on 'Updated investigations of cancer excesses in individuals born or resident in the vicinity of Sellafield and Dounreay': premature all-clear for nuclear power.

Sir,

Bunch report no increased leukaemia/cancer rates in children and young adults below age 25 in Seascale ward near the Sellafield nuclear site, during the period 1991–2006. But the case numbers are very small – 1 observed leukaemia case (O=1) vs 0.26 expected cases (E=0.26), which means a standardised incidence ratio (SIR) of 3.9 with a wide 90% confidence interval (90% CI: 0.28–33.8). Near the Dounreay nuclear installation, in the wards of Thurso and Reay, not a single leukaemia case was registered in 1991–2006, and also not in the respective control region, the rest of Caithness county. Thus, for 1991–2006, no meaningful conclusions about leukaemia risk near Sellafield and Dounreay can be drawn from these numbers.

Bunch et al fail to discuss the leukaemia increases over the full period, 1963–2006. A highly significant increase is found in Seascale ward (O=6, E=0.91, SIR=6.67, 90% CI: 2.9, 13.0). The ratio of the SIR in Seascale (SIR=6.67) to the SIR in Copeland and Allerdale County excluding Seascale (SIR=0.90) yields a relative risk (RR) of RR=6.67/0.90=7.4 (P=0.0002). Near Dounreay the increase in leukaemia risk is not significant (RR=1.64, P=0.227). For all malignancies and over the whole study period 1963–2006 a significantly increased risk is found near Sellafield (RR=3.3, P=0.0004), but the increase is not statistically significant near Dounreay (RR=1.22, P=0.274).

Pooled data from Sellafield, Dounreay, and La Hague: Guizard reported on leukaemia around the La Hague reprocessing plant in France between 1978 and 1998. Leukaemia rates in the 10-km zone were compared with rates in the 10–35-km zone. During the 21-year study period, four leukaemia cases were found among children in the 10-km zone: all were between 1 year and 6 years old. A relative risk of RR=3.4 was found, which was not statistically significant using a two-sided test. However, with the one-sided test, which we use here as we test for an increase, the increase of childhood leukaemia near La Hague is statistically significant (P=0.042).

We have pooled the data for leukaemia in children and young adults aged under 24 in the vicinities of Sellafeld, Dounreay, and La Hague (study areas: the wards Seascale, Thurso/Reay, and Beaumont-Hague). We compared the leukaemia rates in the combined study areas with the rates in the combined respective control areas and found a significantly increased relative risk of 2.40 (P=0.0010). The results are shown in Table 1.

Table 1

Leukaemias in children and young adults aged under 24 in the vicinities of Sellafeld, Dounreay, and La Hague

	Study region			Control region
Site	O	E	SIR	O	E	SIR	RR	P-value
Sellafield	6	0.90	6.67	68	75.93	0.90	7.44	0.0002
Dounreay	8	5.46	1.47	8	8.88	0.90	1.63	0.2317
La Hague	5	2.30	2.17	33	34.63	0.95	2.28	0.0852
Pooled data	19	8.66	2.19	109	119.4	0.91	2.40	0.0010

Abbreviations: E=expected; O=observed; RR=relative risk; SIR=standardised incidence ratio.

In 2012, we analysed the pooled data of leukaemia cases near nuclear power stations in Germany, Great Britain, Switzerland, and France, and found a 37% increased risk in young children living near them (Koerblein and Fairlie, 2012). Since radiation exposures near nuclear reprocessing plants are likely to be greater than those near nuclear power stations, higher leukaemia risks would also be expected. The result in Table 1 is in line with this expectation.