[Electromagnetic residential fields and childhood cancers: state of epidemiological research]

Thirteen epidemiological studies of the relations between electromagnetic fields and childhood cancers have been published. Eleven have shown some associations between the presumed intensity of exposure to residential magnetic fields and the childhood cancers incidence, either for all cancers or for the 3 most frequent types (leukaemias, brain tumours and lymphomas). These associations are not often significant because of the weak statistical power of these studies in relation with both the low incidence of cancers in childhood, and specially of each particular type, and the little number of subjects considered to be exposed at a high level exposure for residential magnetic fields. All these studies fall in with the same difficulties, particularly for identification and assessment of the exposure to magnetic fields and for the manner to take into account all the potential confounders. So even significant associations do not imply their causality, all the more as the carcinogenicity of extremely low frequency electromagnetic fields, although biologically conceivable, has never been experimentally proven. Today it is impossible to claim or invalidate the influence of residential magnetic fields in the genesis of childhood cancer. Setting up of new epidemiological studies based on large number of cases issued from population based registries and conducted with a best defined methodology seems to be highly desirable.