Abstract

STUDY OBJECTIVE: To assess spatial clustering of childhood leukaemias and lymphomas in New Zealand, using a national dataset from a country with no nuclear installations. DESIGN: New Zealand Map Grid coordinates, derived from the birth addresses of cases and controls were used in clustering analyses that applied Cuzick and Edwards' method. SETTING: The whole of New Zealand. PARTICIPANTS: The cases were ascertained from the New Zealand Cancer Registry. They were diagnosed with leukaemia or lymphoma at ages 0-14 years during the period 1976 to 1987. For Hodgkin's disease, the age range was extended to include those aged from 0-24 years. The cancer registrations were linked with national birth records, to obtain the birth addresses of the cases. The controls were selected at random from birth records, with matching to cases (1:1) on age and sex. The analyses included 600 cases and 600 controls. MAIN RESULTS: There was no statistically significant spatial clustering for any tumour group overall, including acute lymphoblastic leukaemia, acute nonlymphoblastic leukaemia, other leukaemias, non-Hodgkin's lymphomas, Hodgkin's disease, and all these combined. Significant clustering was found in a sub-analysis for one of three age specific subgroups of acute lymphoblastic leukaemia (ages 10-14 years, p = 0.003). CONCLUSION: The subgroup finding may have been real or a chance association, as several comparisons were made. This study found little evidence for spatial clustering of leukaemias or lymphomas in a population with no nuclear installations.