Melissa J. Perry, Anne Marbella, Peter M. Layde, “Nonpersistent Pesticide Exposure Self-report versus Biomonitoring in Farm Pesticide Applicators,” Annals of Epidemiology, 2006, 16:9, DOI: 10.1016/J.ANNEPIDEM.2005.12.004.
PURPOSE: Few studies using biologic markers to examine nonpersistent pesticide exposure among pesticide applicators were conducted in field settings. This study compares self-reported dermal, inhalation, and ingestion exposures with urinalysis results after one-time application of the commonly used herbicide atrazine to field crops. It was hypothesized that: i) applicator reports of exposure would be associated positively with detection of urinary atrazine, and ii) applicator reports of personal-protective-equipment (PPE) use would be associated negatively with detection of urinary atrazine metabolites.
METHODS: Wisconsin dairy farmers were randomly selected to participate in 1997 to 1998 and were instructed to collect a urine sample 8 hours after the first pesticide application of the season. Farmers then were interviewed within 1 week of their first application to report on application practices. Eighty-six urine samples were analyzed for deethylatrazine, a major atrazine.
RESULTS: Comparing urinalysis results with self-reported dermal, inhalation, and ingestion exposure showed poor agreement between self-reported exposure and urinary deethylatrazine detections (all κ < 0.40). Multivariate linear regression modeling with deethylatrazine level as the outcome showed that self-reported practices did not significantly predict atrazine metabolite levels.
CONCLUSIONS: Possible explanations for the discrepancies between urinalysis results and self-reported data include: i) inaccuracies in self-reported data and ii) substantial interpersonal variation in atrazine, resulting in major differences in body burden for similar exposures. Either explanation poses challenges for epidemiologic studies of the health effects of pesticides, which rely solely on self-reported measures of exposure. Additional evaluation of determinants of accuracy in self-assessed occupational and environmental exposures is needed.