A case-control pilot study of traffic exposures and early childhood leukemia using a geographic information system

The scientific debate on risk relationships between proximity to electric a nd magnetic fields and the development of childhood leukemia has recently f ocused on the role of other factors that may be strongly correlated with po wer lines. Proximity to high traffic density, as defined by major roadways or automobile counts, and associated socioeconomic neighborhood characteris tics have been suggested as potentially important confounders. For traffic or socioeconomic status (SES) to confound any EMF effect these factors woul d need to have their own independent impact on leukemia risk. This study wa s designed to use geographic information system (GIS) technology to empiric ally examine the relationship between traffic density and socioeconomic ind icators to early childhood leukemia in an urban area of California. Ninety cases of childhood leukemia diagnosed under the age of five between 1988 an d 1994 among children born in San Diego County were matched by gender and b irth date to a total of 349 children also born in the county and not known to have developed any cancer. Case-control differences were assessed via co nditional logistic regression. No significant differences were observed for the neighborhood median family income of the birth residences. When compar ing neighborhoods with median annual income greater than or equal to $56,00 0 to those with incomes <$18,000 the odds ratio was 0.86 (95% confidence in terval 0.31, 2.38). Traffic density was measured using a variety of methods , including information on average daily traffic counts and road characteri stics. None of the measures of traffic were associated with case status. Ne ither SES or traffic density near the birth address as assessed with GIS me thods are strong enough risk factors for leukemia to be confounders which c ould totally explain the effect of another variable (such as wire code). As sociations with the diagnosis address or with more direct exposure measures may differ from those reported here. Bioelectromagnetics Supplement 5:S58- S68, 2001. (C) 2001 Wiley-Liss. Inc.