To the Editor: When discussing causes of childhood leukaemia, Ziegler et al stated, “Exposure to electromagnetic fields has been ruled out as playing any significant role”. 1 They cited one large study2 in support of this statement, but overlooked two independent pooled analyses that showed the opposite. Greenland et al analysed 12 studies involving 2656 patients and 7084 controls, 3 and Ahlbom et al analysed nine studies involving 3247 patients and 10 400 controls.4 Each analysis found an association with a doubling of risk of childhood leukaemia at levels of household exposure at and over 0.4 microtesla (4 milligauss). Confounders and sources of bias to explain these findings have been sought without success. Consequently, in 2002, the International Agency for Research on Cancer classified 50 and 60 Hz magnetic fields as a “possible carcinogen” (Group 2b)5 even though the mechanism of an effect is not clear.

The role of magnetic fields in childhood leukaemia cannot be “ruled out”, given the substantial epidemiological evidence, the international classification of magnetic fields as a possible carcinogen, and the subtlety of gene–environment interactions. Moreover, although exposures to magnetic fields are low within most households, there is opportunity to easily prevent or treat the uncommon situations where household exposures exceed 0.4 microtesla by means of electrical engineering, household wiring and town planning.

In reply: Hocking states that electromagnetic fields cannot be ruled out as a cause of childhood leukaemia. However, several large studies have all failed to find any association between childhood exposure to electromagnetic radiation and leukaemia.1-4 The two pooled meta-analyses Hocking refers to both found no increased incidence of leukaemia with exposure to electromagnetic fields of < 0.4 microtesla.5,6 Although there was an increased risk of leukaemia with exposure to ≥ 0.4 microtesla, 99.2% of children with leukaemia had not received such a high level of exposure. 5 In addition, both studies acknowledged the potential for selection bias. As such, for the overwhelming majority of children with leukaemia, exposure to electromagnetic fields does not play any significant causative role. Although we agree its effect cannot be ruled out for the remaining < 1% of patients, it should not be given undue epidemiological weight.