Actualités du Registre Suisse du Cancer de l’Enfant

Spatial clustering of childhood leukaemia in Switzerland: A nationwide study

Summary of results

Our research question

We investigated whether children who develop leukaemia tend to live closer to each other than expected by chance (spatial clustering) and whether there are any areas in Switzerland with a particularly high risk (spatial cluster). We investigated both the place of residence at birth and at diagnosis.

Why is this important?

The causes of leukaemia in children are largely unknown. However, studies indicate that environmental factors such as traffic related air pollution or agricultural pesticides might increase the risk. Because exposure to such factors depends on where children live, leukaemia risks might vary from place to place. The presence of spatial clustering or clusters could thus provide new clues about possible environmental risk factors.

What did we do?

We included children aged less than 16 years who developed leukaemia in Switzerland during 1985-2015. For almost all children we had geocoded place of birth and diagnosis. To mimic the scenario in which the risk of leukaemia did not depend on where children lived, we obtained places of residence of children who had the same age and sex as the leukaemia cases, but were randomly selected from the national censuses. We performed several statistical tests for spatial clustering and spatial clusters. We also separately examined different age groups (0-4, 5-15) and different leukaemia subgroups (acute lymphoid leukaemia, acute myeloid leukaemia).

What did we find?

We included 1871 children who developed leukaemia. After accounting for the fact that we had performed several tests, we found no statistical evidence of spatial clustering. However the results of individual tests did indicate spatial clustering when considering distances smaller than 100 m. The test for spatial clusters suggested a cluster consisting of 5 children with acute lymphoid leukaemia in a small rural area of Switzerland.

What does this mean?

Overall, this study provides little support for spatial clustering or spatial clusters of childhood leukaemia. Although individual tests did indicate spatial clustering and identified a cluster, these results must be seen in the context of the multiple tests performed. Performing multiple tests increases the likelihood that any single test will show positive results by chance alone. If indeed risks are higher in certain locations due local pollution sources, then our study suggest that these pollution sources effect only small areas extending at most few hundred metres from the source. 

More information:

Garyfallos Konstantinoudis, Email :

Konstantinoudis et al., International Journal of Cancer 2017

© ISPM - Université de Berne 2019