Biomonitoring-Based Risk Assessment of Pyrethroid Exposure in the U.S. Population: Application of High-Throughput and Physiologically Based Kinetic Models

Abstract

Pyrethroid insecticides have been extensively utilized in agriculture and residential areas in the United States. This study evaluated the exposure risk by age using available biomonitoring data. We analyzed pyrethroid metabolite concentrations in urine using the National Health and Nutrition Examination Survey (NHANES) data. Reverse dosimetry was conducted with a high-throughput model and a physiologically based kinetic (PBK) model integrated with a Bayesian inference framework. We further derived Benchmark Dose (BMD) values and systemic points of departure in rats using Bayesian BMD and PBK models. Margins of exposure (MOE) were calculated to assess neurotoxic risk based on estimated daily oral intake and dose metrics in plasma and brain. Results from both models indicated that young children have higher pyrethroid exposure compared to other age groups. All estimated risk values were within acceptable levels of acute neurotoxic effect. Additionally, MOEs calculated from oral doses were lower than those derived from internal doses, highlighting that traditional external exposure assessments tend to overestimate risk compared to advanced internal dose-based techniques. In conclusion, combining high-throughput and PBK approaches enhances the understanding of human health risks associated with pyrethroid exposures, demonstrating their potential for future applications in exposure tracking and health risk assessment.

Keywords: bayesian; exposure risk; pyrethroids; reverse dosimetry; urinary metabolites.

Figure A1

Cumulative distribution of urinary pyrethroid metabolite concentrations across different age groups.

Figure A2

Evaluation of model performance of refined PBK model with human toxicokinetic experiments [37,38,48,49,74].

Figure 1

Schematic diagram of the overall study workflow. The study consists of four key steps: (1) using Bayesian reverse dosimetry to estimate daily intake from NHANES biomonitoring data, (2) applying the Bayesian benchmark dose approach to reanalyze dose-response data from animal neurotoxicity studies, (3) simulating systemic exposure dose and determining the point of departure based on external daily intake and benchmark dose thresholds, and (4) probabilistically estimating the margin of internal exposure by integrating predicted exposure levels with effect thresholds. Note: BMD, Benchmark Dose; BMDL, Benchmark Dose Lower Confidence Limit; MOE, Margin of Exposure; MOIE, Margin of Internal Exposure; NHANES, National Health and Nutrition Examination Survey.

Figure 2

Refined generic PBK model for pyrethroids and their metabolites based on Mallick et al. [42] and Quindroit et al. [34]. The diagram illustrates the PBK model structure, depicting the absorption, distribution, metabolism, and elimination processes of pyrethroids and their metabolites in the human body.

Figure 3

Goodness–of–fit evaluation for the total population across NHANES cycles. These plots compare observed and predicted urinary concentrations of pyrethroid metabolites for the total population across NHANES cycles. Predictions are based on parameters from the final Markov Chain Monte Carlo iterations. The solid line represents a perfect match between observed and predicted values, while dashed lines indicate error margins within a ten-fold difference.

Figure 4

Comparison of exposure predictions from the PBK model and high-throughput approaches, highlighting similarities and discrepancies. The solid line represents a perfect match between the two methods, while dotted and dashed lines indicate errors within three- and ten-fold differences, respectively.

Figure 5

Variability in exposure estimates across age groups. Variability in exposure is shown using P95/P50 ratios for 3PBA and permethrin across age groups. Greater variability is observed in the high differences between permethrin and 3PBA in the top panel, as well as in the larger dots depicted in the lower panel.

Figure 6

Trends in urinary 3PBA (top) and derived permethrin daily intake (bottom) from 1999 to 2016. Predicted intakes are shown separately for PBK and high-throughput models.

Figure 7

Exposure similarity across age groups. Age groups are arranged by class, highlighting similar exposure levels across NHANES cycles.

Figure 8

Distribution of the MOE and MOIE for acute neurotoxic effects calculated from daily oral and corresponding systemic exposure, respectively. Higher values indicate a lower exposure risk.