gms | German Medical Science

Objective: In Germany, articles on the association between childhood leukaemia incidence and residential distance to nuclear power plants (NPPs) have been published since the 1990s [1]. Increased incidence rates were reported in the late 1990s among children residing in several municipalities within ten kilometres from the NPP-Krümmel, located in the urban district of Geesthacht. Further articles confirmed this observation in the following years [2], [3]. Even most recently, the increase compared with the German population was confirmed with data up to 2019 [4], [5]. Most of these publications are based on ecological data, without information on individual radiation exposure or other potential risk factors.

Methodological challenges arise when analysing the association of radiation exposure and leukaemia incidence at an ecological level. They include the use of a suitable reference population, unstable estimates of standardised incidence ratios (SIRs) and the potential role of prevailing winds. The aim of our study is to disentangle these three major challenges.

Material and methods: The German Childhood Cancer Registry provided data on leukaemia in children under 15 years of age for the years 2004–2019 by residence at the time of diagnosis. The German Federal Statistical Office provided the corresponding population data. The study region included all municipalities with ≥75% surface area within 50 kilometres from the Krümmel NPP.

We calculated SIRs and used two reference populations: the complete German population and the population of federal States with at least one municipality within 50 km from the NPP-Krümmel: Schleswig-Holstein, Lower Saxony, Hamburg, and Mecklenburg-Western Pomerania

To overcome the issue of unstable estimates, smoothed relative rates (RRs) were calculated using the hierarchical Bayesian Poisson-gamma method. We used Tango’s index to detect spatial clustering.

Finally, we mapped the RRs calculated for each community to check for potential patterns associated with the wind direction prevailing in the region under study, i. e. from west to east.

Results: Overall 356 cases of childhood leukaemia were observed in 321 municipalities within a radius of 50 kilometres from the NPP-Krümmel from 2004 to 2019 (person-years contributed up to age 15=6,349,938). SIRs calculated assuming Germany as reference population show nearly no difference to those calculated assuming the four federal States as reference. In the study area we observed a SIR=1.0 (95% confidence interval: 0.9-1.1). In Geesthacht a SIR=2.3 was observed (observed-cases=eight; person-years=65,394; 95% CI: 1.0-4.5). The smoothed RR calculated for Geesthacht was still higher than the reference (RR=1.8; 95% credibility interval: 0.9-3.0). Tango’s index did not show any evidence of spatial correlation. The map reporting RRs does not show patterns associated with prevailing wind directions.

Conclusion: Using a regional population as reference in place of the national one, we found in 2004-2019 an increased SIR for childhood leukaemia in Geesthacht. RRs do not show any pattern associated with prevailing wind direction. The increased risk observed in Geesthacht even after smoothing the estimates deserves epidemiological surveillance.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.