Space–time clustering of childhood cancers: a systematic review and pooled analysis

Abstract

The aetiology of childhood cancers remains largely unknown. Space–time clustering of cases might imply an aetiological role of infections. We aimed to review the evidence of space–time clustering of specific childhood cancers. We searched Medline and Embase for population-based studies that covered a pre-defined study area, included cases under 20 years of age and were published before July 2016. We extracted all space–time clustering tests and calculated the proportion of positive tests per diagnostic group. In a pooled analysis, we performed a Knox test of the number of pairs of cases close to each other in time and space pooled across studies. 70 studies met our eligibility criteria, 32 of which reported Knox tests. For leukaemia, the proportion of positive tests was higher than expected by chance at both time of diagnosis (26%) and birth (11%). The pooled analysis showed strong evidence of clustering at diagnosis for children aged 0–5 years for a spatial and temporal lag of 5 km and 6 months, respectively (p < 0.001). The evidence was mixed for lymphoma and tumours of the central nervous system. The current study suggests that leukaemia cases cluster in space–time due to an aetiological factor affecting children under 5 years of age. The observed pattern of clustering of young children close to time of diagnosis is compatible with Greaves’ delayed-infections-hypothesis.Electronic supplementary materialThe online version of this article (10.1007/s10654-018-0456-y) contains supplementary material, which is available to authorized users.

Risk Factors for Childhood Leukemia: Radiation and Beyond

“…The first systematic review and pooled analysis of such space-time clustering studies was presented and discussed at the workshop. This study showed strong evidence of clustering of CL at time of diagnosis for children aged 0-5 years, an age range including the peak incidence for leukemia at 2-4 years (68). Results were similar for ALL.…”

“…Such clustering in space and time could be explained by "mini-epidemics" of a single infection leading to local clusters of leukemia cases (68), which are observed from time to time (69)(70)(71)(72). For children aged 5 to 15 years no clustering at both birth and diagnosis was observed, and results for lymphoma and CNS tumors provided only weak evidence for space-time clustering (68). Nevertheless, the systematic review was restricted to studies using a certain methodology only and there have been other approaches showing no evidence for clustering, e.g., in Germany (73,74).…”

Clustering of non-leukemia childhood cancer in Colombia: a nationwide study

“…Spatial and spatio-temporal analysis have been previously used in CC, mainly in the study of the geographic pattern of leukemias 8 , 22 . A recent systematic review of space-time analysis identified 70 studies published up to 2016 of which 47 reported results for leukemias, 26 for lymphomas, 13 for CNS tumors and 12 for other types of tumors 23 . All 32 analyses used for the meta-analysis were from Europe and United States; and the analysis showed evidence of leukemia clustering in children between 0 and 5 years of age.…”

Childhood Leukemia in Small Geographical Areas and Proximity to Industrial Sources of Air Pollutants in Three Colombian Cities

“…In Latin America, there is also some evidence of clustering for childhood leukemia: in Argentina, significant clusters of childhood leukemia were identified in the province of Cordoba using the residential address at the time of diagnosis between 2004 and 2013 [37]; in Mexico, clustering of ACL at the time of diagnosis between 2010 and 2014 was identified in the city of Guadalajara [38]; in Colombia a nationwide study at municipality level and time of diagnosis identified ACL clusters in five regions [17]. A recent systematic review and pooled analysis of space-time clustering studies of childhood cancer included 47 studies of childhood leukemia published before July 2016 and concluded that significant clusters are present at both time of diagnosis and birth; the clusters were identified especially for children aged 0-5 years for a spatial lag of 5 km and temporal lag of 6 months, suggesting that the pattern of clustering close to the time of diagnosis might be compatible with an infection cause to be identified [39].…”