Space-time interactions in childhood cancers.

STUDY OBJECTIVE The aim of the study was to examine a cohort of cases of childhood cancer occurring in a defined geographical area to try to identify clustering and possible causative factors. DESIGN Data were analysed using the close pair method developed by Knox for signs of clustering in relation to date and place of onset or date and place of birth. SETTING Cases were those occurring in the 8 year period between 1953 and 1960 in four old counties of the Midlands of England (Worcestershire, Warwickshire, Staffordshire, Shropshire). CASES 418 children under 10 years developed cancers during the study period: leukaemia 228, cerebral tumour 99, neuroblastoma 45, nephroblastoma 46. It is thought that data collection was complete. Cases were matched for region, sex and date of birth with live controls. RESULTS There was evidence among some age groups and diagnoses of an unexpectedly high number of close pairs of onsets, and some indication of similar patterns among births of children who later developed cancer. The pre- and postnatal experiences of children involved in close pairs were examined to see whether they differed from those of control children. Measles appeared to occur more often in the 2-3 years before the onset of leukaemia in children who were later involved in close pairs than in their matched controls. CONCLUSIONS Common infectious diseases of childhood may play a minor role in the development of some cancers. Epidemics of these diseases may then be reflected on a greatly reduced scale in the subsequent distribution of cancer cases.