Correlation Analysis of Pesticide Use Data and Cancer Incidence Rates in California Counties

Abstract

California, the leading agricultural state in the United States, has maintained a population-based cancer registry since 1988, and it also maintains a comprehensive, statewide pesticide reporting system. Data on cancer incidence and pesticide use reporting are available, by county, for all 58 counties in California. Average annual age-adjusted cancer incidence rates (1988–1992), on a county-, sex-, and race/ethnicity-specific basis, were obtained from the California Cancer Registry (CCR), which maintains the population-based cancer registry throughout California. Pesticide use data (i.e., pounds of active ingredient applied annually in each county) were obtained from the California Department of Pesticide Regulation for 1993. Investigators used Pearson product-moment correlation coefficients (r) to correlate age-adjusted incidence rates for selected cancers with the use data for selected pesticides. For most sex- and race/ethnicity-specific groups, the correlation coefficients were very close to zero or negative in sign, indicating no correlation between pesticide use and cancer incidence. There were, however, several exceptions, particularly in Hispanic males for whom the following correlations were observed: leukemia and atrazine (r = .40), leukemia and 2,4-dichlorophenoxyacetic acid (r = .41), leukemia and captan (r = .46), atrazine and brain cancer (r = .54), and atrazine and testicular cancer (r = .41). For black males, we observed the following: atrazine and prostate cancer (r = .67) and Captan and prostate cancer (r = .49). In females, only a few of the correlations were elevated. Although most of the correlations examined in this analysis were not elevated, several of those in the Hispanic and black male populations were. These segments of the population have traditionally been employed as farm workers in California and have had the greatest potential for exposure to pesticides. This was an ecological study for which no data about exposure to pesticides at the individual level were available for analysis. In addition, no latency period was allowed between potential exposure and diagnosis with cancer. However, the results obtained in two minority groups who represented the majority of farm workers in the fields suggested that additional research studies, in which more rigorous study designs are used, should be conducted in those groups.