IRSST - Institut de recherche Robert-Sauvé en santé et en sécurité du travail

Assessing the impact of coexposure on biomarkers of exposure to pyrethroid pesticides in farm workers


There is little published data on the impact of combined exposure to multiple pesticides (coexposure) on pesticide toxicokinetics in farm workers, making adequate interpretation of biomonitoring exposure data difficult. This research looks at the impact of pesticide coexposure on levels of biomarkers of exposure to pyrethroid pesticides in farm workers. The pyrethroid lambda-cyhalothrin (LCT) and the fungicide captan were used as sentinel pesticides as they are widely used in agriculture.  
In part 1, individual time profiles of biomarkers of exposure to LCT in strawberry-field workers were compared following a spray event using LCT alone or in combination with captan. Study participants provided all urine voided over a three-day period after application of a pesticide formulation containing LCT alone or mixed with a captan and, in some cases, after reentering the treated field. LCT metabolites were measured in all urine samples, in particular 3-(2-chloro-3,3,3-trifluoroprop-1-en-1-yl)-2,2-dimethyl-cyclopropanecarboxylic acid (CFMP), 3-phenoxybenzoic acid (3-PBA) and 4-hydroxy-3-phenoxybenzoic acid (4-OH3PBA). Urinary excretion profiles were obtained for 25 spray events for 14 recruited workers. Comparative analysis showed no obvious differences in time profiles or cumulative excretion of individual metabolites (CFMP, 3-PBA and 4-OH3BPA) after exposure to LCT alone or in combination with captan. For most workers and exposure scenarios, CFMP was the main metabolite excreted, but the time course of CFMP in the urine was not always the same as that of 3-PBA and 4-OH3BPA. As the latter two metabolites are common to other pyrethoids, this suggests some workers were coexposed to pyrethroids other than LCT. For several workers and exposure scenarios, CFMP values increased in the hours following spraying. For many pesticide operators, however, other CFMP peaks were observed at later times, indicating that tasks other than spraying of LCT-containing formulations contributed to the increased exposure. According to questionnaire responses, these tasks were mainly work/inspection in treated fields or handling/cleaning of equipment used for spraying.  
In part 2, a cross-sectional study was conducted of the impact on biomarker levels measured in workers of coexposure compared to the impact of other factors. A larger number of workers were included in the sample in part 2 than in part 1, but fewer biological measurements of each worker were taken. Multivariate analyses were used to assess the contribution of coexposure to the variability in levels of biomarkers of exposure compared to that of other factors. Unlike in part 1, which mainly targeted applicators, 87 workers assigned to different tasks (application, weeding and picking) were recruited for part 2, the goal being homogeneity of the workers exposed to LCT alone and those exposed to LCT in combination with captan. The recruited workers provided two-consecutive 24-hour urine collections following an application event involving LCT alone or in combination with captan or following tasks in the treated fields. These were provided in addition to a control collection. As in part 1, concentrations of biomarkers of exposure to LCT were measured in the samples.  
Considered as well were potential determinants of exposure established in a previous study, including task performed and personal factors documented by questionnaire. Multivariate analyses showed that coexposure did not have a statistically significant effect on observed urinary levels of 3-PBA and CFMP. The “time” variable (representing repeated biological measurements and defined as a within-subject variable) was, however, a significant predictor of observed biological levels of 3-PBA and CFMP. Only main occupational task was significantly associated with urinary levels of 3-PBA and CFMP. Compared to weeding and picking, pesticide application was associated with higher urinary 3-PBA and CFMP concentrations.  
In part 3, a toxicokinetic model specific to LCT was used to simulate kinetics data for biomarkers in workers exposed to a single pesticide or a combination of pesticides. Model parameters were determined using a new combined method for numerical solution of differential equations and model parameter search. Model parameters were optimized using published data from volunteers exposed to LCT alone under controlled conditions. A sensitivity analysis was used to determine variations in values of key model parameters necessary to obtain simulations that matched observed data in coexposed workers. The model was able to adequately simulate time profiles of urinary metabolites in participants after coexposure without having to modify the model parameters for exposure to LCT alone, demonstrating a lack of any significant impact of coexposure on biological variability. Monte Carlo simulations were also performed with the model to reconstruct possible absorbed doses for each worker based on quantities of CFMP metabolite measured in the worker’s urine over the entire biomonitoring period. Reconstructed daily doses in pesticide applicators and farm workers assigned to weeding and picking were compared to the Acceptable Operator Exposure Level (AOEL) reference value established by the European Food Safety Authority (EFSA). The percentage of values exceeding the AOEL was calculated for each worker based on the reconstructed daily doses for that worker. The simulations indicate that the AOEL was probably exceeded at some points during the biomonitoring period by the applicators with the highest urinary concentrations. On the other hand, the number (percentage) of workers assigned to weeding and picking showing a probability of exceedance of the AOEL in days following tasks in a treated field was very low. The modelling also made it possible to establish a reference dose of 116 ng CFMP/kg bw/d or 7.5 µg CFMP/L urine corresponding to the AOEL that could serve as an action threshold.  
In sum, all three parts of the project demonstrated that coexposure to agricultural pesticides in strawberry fields had no impact on concentrations of biomarkers of exposure at the exposure levels observed in the workers studied. The study also confirmed earlier data suggesting applicators were more exposed than workers assigned to field tasks such as weeding and picking. Some applicators were also more exposed than the general population, with concentration levels that probably exceeded the acceptable operator exposure limit (AOEL) specified by the European Food Safety Association (EFSA). In workers assigned to weeding and picking, however, there seemed little probability of this limit being exceeded.