Leukaemia in young children living in the vicinity of nuclear power plants

For some years there has been concern that environmental contamination from nuclear power plants might result in an increase in cancer risks among the population living in the vicinity of the plants. The concern was that radioisotopes emanating from air emissions from the plants or pollution of water supplies, might be absorbed by susceptible individuals and result, perhaps particularly, in leukemia. A number of studies in the UK in the 1980s appeared to increase this concern.1-7 However, studies based on data pooled from areas around several nuclear plants, compared with pooled data from control areas, which are less likely to be affected by small area differences in the prevalence of exposure to other risk factors than those of studies based on single sites, did not in general show an increased risk for leukaemia in children living near nuclear installations either in the UK8-12 or in other jurisdictions.13-18 Moreover, the theoretical risk with the extent of measured contamination seemed vanishingly small. In addition, a possible explanation of an increased risk from a postulated infection or infections spread as a result of large-scale mixing of rural and urban populations, particularly in areas where such mixing tends to involve substantial population influx into a sparsely populated area19, 20 seemed to offer a possible alternative explanation. However, in this issue of the IJC, a report by Kaatsch et al. will be bound to reraise the issue.21 In this carefully conducted population-based case-control study, which includes published data for the period 1980–95 as well as previously unpublished data for the period 1996–2003, an increased risk of childhood leukemia was found among children under 5 years living within 5 km of a nuclear power plant at the time of diagnosis, and a lower, but still statistically significant risk, among those living 5–10 km from a plant. The data for the most recent 8-year period are suggestive of a trend although the association was not as strong as the earlier period. Only residence at the time of diagnosis (or corresponding reference date for controls) was considered. As noted by the authors, misclassification resulting from the lack of residential history would be likely to have biased the association towards the null. The authors also point out that excesses of this type were not to be expected under current radiobiological theory. There are potentially 3 explanations for the finding of Kaatsch et al. First, and most likely, this is simply a chance observation, which has persisted (but is possibly diminishing) in these areas of Germany for unknown reasons. In the largest dataset on childhood cancer examined so far (from Great Britain), for example, the spatial and space-time distributions of leukaemia and several other types of childhood cancer were observed to be nonrandom.22 Second, the exposure of some individuals living in these areas was much higher than could be inferred from the available measures, and that they developed leukemia because of this exposure. Third, the Kinlen hypothesis is correct and there is an infectious cause of some cases of leukemia,19, 20 or an alternative but unknown causal factor exists, and one of these was expressed in study areas. How can this be resolved? There appear to have been some difficulties in obtaining direct contact with the subjects in several of these areas, perhaps because of concern surrounding earlier publicity from research findings. In addition, this is now a survivor population, and even if contact could be made (e.g., to request investigation of other risk factors, including possible genetic susceptibility) the results on a selected population would be uninterpretable. We feel, however, that the finding cannot be dismissed. Other issues surrounding low level environmental exposures, for example to nonionizing radiation from cellular phones and transmission towers (“Wi-Fi”), engender concern in several countries, including Canada and Germany. Our ability to investigate the effect of such exposures has been limited because we have not been able to make advances in our ability to interpret associations at the ecological level, even using more sophisticated statistical analyses. We note, however, that there has also been little attempt to use multilevel methods of analysis, in which existing sources of research information available at both the individual and aggregate level could be used, to strengthen the investigation of such exposures. For example, we are currently involved in discussions on establishing a new large cohort study in Canada, and we consider that this will be an opportunity to collect from our participants more detailed data on residential and occupational histories (including for the latter geographic coordinates) that might enable more sophisticated analyses of the type conducted by Kaatsch et al. around potential sources of environmental contamination to be made. We urge other research groups to do likewise. We encourage further discussions on these issues and anticipate that the International Journal of Cancer could be a forum for such discussions.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.