Childhood leukemia in metropolitan regions in the United States: a possible relation to population density?

Following recent research in Great Britain, the geographic incidence of leukemia and non-Hodgkin's lymphoma among White children in three metropolitan regions of the United States (San Francisco-Oakland, CA; Detroit, MI; and Atlanta, GA) during 1978-82 has been analyzed using census tract-specific data. There was no evidence of a general tendency for cases to cluster geographically, in contrast to results from Britain. Further, rates did not vary with median income or education levels for census tracts. However, there was a statistically significant increasing trend in incidence rates with increasing population density: relative risk for highest relative to lowest category = 1.4 (95% percent confidence interval [CI] = 1.1-2.0) for White population density, and 1.4 (CI = 1.0-2.0) for total population density. The interpretation of these findings is unclear and further investigation is required. It is possible that population density is acting as a surrogate for some virus-related factor.