Childhood cancer in Texas counties with moderate to intense agricultural activity.

With few established risk factors, cancer remains the second leading cause of death for children in the U.S. Agricultural pesticide use is one of many suspected factors that may contribute to the etiology of childhood cancer. This study tests the hypothesis that birth in Texas counties with moderate to intense agricultural activity increases childhood cancer risk. This case-control study analyzed 6974 cases and controls ages 0 to 14, identified through the Texas Cancer Registry and Texas birth records, respectively. Exposure data were obtained from the Census of Agriculture. Percent cropland in the county of birth and total county-specific pesticide exposure incorporating the EPA's carcinogenicity classification served as surrogates for pesticide exposure. Cancer sites examined include: all cancers, leukemia, lymphoma, CNS tumors, and several specific subsites. Elevated, although not statistically significant, ORs for the association between birth in counties with > or =50% cropland were produced for all CNS tumors (OR = 1.3, 95% CI = 0.9-1.8), astrocytoma (OR = 1.4, 95% CI = 0.8-2.2), and PNET (OR = 1.3, 95% CI = 0.7-2.5). A similar pattern was not observed using the index of total county-specific pesticide exposure. Although imprecise, these exposure assessment methods represent novel applications of agricultural census data. Although a pattern of increased risk was observed between percent cropland and CNS tumors, this study's results do not support an association between birth in Texas counties with moderate to intense agricultural activity and childhood cancer. Due to study limitations, such an association should not be ruled out. Future research should incorporate individual-level data from various sources to increase precision and decrease misclassification in the exposure assessment.