Updating our quantification of non-occupational pesticide exposure in agricultural settings: A revised algorithm for the agricultural health study

Understanding associations between pesticide exposure and chronic disease requires long-term, chemical-specific estimates. To characterize non-occupational pesticide exposure in the Agricultural Health Study (AHS), we previously developed a quantitative algorithm that estimated active ingredient-specific exposure to estimate exposure for the take-home, agricultural drift, and residential use pathways, with weights derived from indoor pesticide dust concentrations in agricultural homes from ten studies published through 2015. Here, we describe an update. First, we expanded our meta-analysis to include data from nine additional studies published through 2024. We then calculated relative differences (RD) in pesticide concentrations, with sensitivity analyses for AHS data, for known use of the specific pesticide, and for herbicides vs. insecticides. For take-home, the RD was 2.6 (95% confidence interval (CI): 1.9-3.4) between homes with a resident likely vs. not likely to have occupational exposure to pesticides, increasing to 3.7 (95%CI: 1.9-7.4) with known use of the pesticide. For agricultural drift, we observed non-linear trends with distance from treated fields; the RD was 1.9 (95%CI: 1.2-2.8) for homes located 440 yards from field, increasing to 4.2 (95%CI: 1.7-10) with known use of the pesticide. For residential use, we observed trends with increasing probability the pesticide was used, with a RD of 1.6 (95%CI: 1.3-2.0) with ≥20% probability. Our updated findings were used to re-conceptualize the algorithm; we illustrate its application for chlorpyrifos and atrazine. These refinements provide increased confidence in our exposure characterization and will help advance our knowledge of potential health risks related to non-occupational pesticide exposures.