Are human exposure assessment the same for non-persistent organic chemicals? -from the lens of urinary variability and predictability

Assessment of human exposure to mixtures of non-persistent chemicals from food matrices and consumer products requires accurate characterization and estimation of their preceding exposure levels, and assessment sampling approaches for these varying chemicals remain disputable. Here, we used high-throughput targeted method to quantify urinary concentrations of 59 most common non-persistent chemicals (6 parabens, 14 bisphenols, 1 triclosan, 7 benzophenones, 2 dichlorophenols, 13 phthalate metabolites and 16 antioxidants) in 158 consecutive spot samples from 11 participants over three consecutive days, 33 samples of which were first morning voids (FMVs). We found 49 chemicals with detection frequencies over 70 % in all urine samples. Principal component analyses showed greater inter-person variations than each person's inter-day variations. Intraclass correlation coefficient (ICC) to assess the reproducibility of targeted chemicals demonstrated that regardless of sampling approaches, dichlorophenols, most parabens, benzophenones and triclosan showed moderate to high reproducibility (0.445 < ICC < 0.969), with relatively high predictive power of FMVs for 24-h collections. Notably, most phthalates, bisphenols and antioxidants showed low ICC values. Together, our work demonstrates that FMV samples may be adequate for assessing human exposure to parabens, benzophenones, triclosan and dichlorophenols, whereas multiple consecutive urine collections may be advantageous for evaluating exposure to most phthalates, bisphenols and antioxidants.

Automated summary

Assessment of human exposure to mixtures of non-persistent chemicals requires accurate characterization and estimation of preceding exposure levels, and assessment sampling approaches remain disputable. This study used a targeted method to quantify urinary concentrations of 59 common non-persistent chemicals in consecutive spot samples from 11 participants. The results showed that certain chemicals had high reproducibility, while others had low reproducibility, suggesting that different sampling approaches may be needed for different chemicals.