Risk assessment of pesticide use in agriculture is easier by integration of traditional monitoring approaches and new computational tools

Introduction: Variable climatic and working conditions in open field, use of complex mixtures of agrochemicals and high cost make traditional approaches of occupational hygiene hardly feasible in small and medium agricultural enterprises. This calls for alternative strategies allowing risk assessment without/with minimal field measurement. This can be achieved through the use of predictive models in combination with environmental and biological monitoring measurements to build physically based ‘exposure and risk profiles’ to which specific, experimentally determined values of relevant parameters need to be assigned to finally derive specific exposure scenaria for the individual crops.

Methods: To test feasibility of this approach, we pooled environmental and biological monitoring data collected in a pilot study of rice and maize herbicide applicators in Lombardy region, Italy. Exposure calculation based on monitoring data showed most workers are below the Acceptable Operator Exposure Level defined in the authorization process. Moreover, we set up a prototype method to calculate a provisional biological exposure limit to perform individual risk assessment based on biological monitoring in 24-hr urine.

Results: The study estimated propanil median levels of workers’ whole-body over-clothing exposure as 0.032 (0.01-1.24) parts-per-million.Farmers who used an open tractor did not show a significantly higher exposure than those who used closed or air-conditioned tractor cockpit. Measurements above and below clothing show ˜60% barrier efficiency of working clothes. Twenty-four-hour post-application, urine excretion of 3,4-dichloroaniline was ˜0.3 parts-per-million of applied propanil which is toxicologically irrelevant based on AOEL of 20 microg/kgbw/day.

Discussion: We identified the main variables affecting exposure intensity in each of the main work phases of pesticide application. To each variable score were assigned ranking from excellent protection from excess exposure to risky condition leading to unacceptable exposure. Presumptive exposure levels and the corresponding protection factors were calculated and were in agreement with those calculated from real-life exposure measurements.