Childhood Leukaemia Near Nuclear Installations

A recent meta-analysis by Baker and Hoel (2007) documented consistently elevated leukaemia incidence and mortality in children, especially those under age 10 years, near nuclear installations. Although a consistent dose– response association was not found, results suggest more detailed investigation is in order. The report extends an investigation of low-dose radiation exposure and childhood leukaemia risk that began in the late 1950s, when a near-doubling of leukaemia mortality by age 10 years from in utero pelvic X-rays was documented (Stewart et al. 1958). The studies cited by the authors indicate that more current data may be needed. Of the 17 studies in the meta-analysis, 12 were published before 1994, raising the question of whether the findings accurately represent present patterns of childhood leukaemia. Only one study examined US nuclear plants, even though the USA is home to nearly one-fourth of all nuclear power reactors worldwide. This report examined cancer mortality rates near US plants that began operating before 1982, before and after startup, but ended with 1984 data (Jablon et al. 1991). The availability of historical mortality data on the US Centers for Disease Control and Prevention web site makes an update of this study feasible. The prior study, conducted by the US National Cancer Institute, presented mortality data for childhood leukaemia (age 0–9 and 10–19 years) near 51 US nuclear power plants. It used a Standard Mortality Ratio (SMR), defined as the proportion of the local to national death rate, to analyse temporal changes near nuclear plants after startup. (The one or two closest counties to each plant were selected as the local area.) It is now possible to observe any changes in SMR for childhood leukaemia as nuclear plants age. Table 1 compares the ratio for the year after startup through 1984 to the period 1985–2004. The 51 plants are also divided into three categories: older plants (startup from 1957 to 1970 and still operating), newer plants (startup from 1971 to 1981 and still operating) and plants permanently closed. The local areas constitute a total of 67 counties, with a current population of about 25 million, or 8% of the US total. We observe a uniform pattern of increase in childhood leukaemia SMR from the earlier period to the most recent 20 years for the plants that remain in operation. The greatest changes occurred in the older plants; the leukaemia SMR for children aged 0–19 years rose 13.9%, from 0.986 to 1.123 (P < 0.02). Areas closest to the newer plants had a smaller increase of 9.4% (SMR from 0.897 to 0.981, not significant). For both groups of plants, the SMR rose more rapidly for the 10–19 age group compared with the 0–9 group, a pattern that is inconsistent with the Baker and Hoel findings. The areas near the closed plants experienced an insignificant 5.5% decrease in SMR, from 1.028 to 0.971. In the most recent two decades, a total of 1037 childhood leukaemia deaths occurred near the plants still operating, while 255 occurred near the closed plants. Current (1985–2004) local childhood leukaemia mortality near older US plants still operating is above the US rate (SMR > 1.00), while mortality near newer plants is below the US (SMR < 1.00). While it is feasible that higher emissions of radioisotopes into the environment from older plants may account for the observed trends, caution should be used when interpreting the data. There may be demographic differences between the two groups that can include factors affecting mortality risk such as poverty, proximity to medical facilities and presence of other environmental pollutants. Prudence should also be used when reviewing results for the areas near closed reactors. It is possible that reduced emissions after closing are associated with reduced childhood leukaemia mortality, but other possible confounding factors should be considered. The analysis is also affected by the time frames used in the early years after nuclear plant startup. Anywhere from 3 to 27 years after startup was used by the US National Cancer Institute in the earlier period, according European Journal of Cancer Care, 2008, 17, 416–418 Letter to the editor