Keywords: 
• trichloroethylene /
• metabolite /
• PBPK model /
• Monte Carlo simulation /
• cancer risk

Abstract: Trichloroethylene (TCE) has been used as a degreasing and cleaning agent in the hardware, electroplating, and electronic industries. Occupational exposure to TCE produces a range of health risks, including hypersensitivity syndrome and carcinogenicity. In 2012, TCE was classified as a Group 1 carcinogen to humans by the US Environmental Protection Agency (USEPA) and the International Agency for Research on Cancer. TCE air levels from manufacturing process of an enterprise in Dalian were analyzed by adsorption tube sampling and thermal desorption coupled with gas chromatography/mass spectrometry. The dynamic distribution, metabolite production and carcinogenic risk of TCE in tissue levels of occupational workers by inhalation exposure were predicted with physiologically based pharmacokinetic (PBPK) model. The predicted maximal tissue levels of TCE presented an accumulation trend of fat tissue > gut tissue > richly perfused tissue > tracheo-bronchial tissue > poorly perfused tissue > liver tissue > venous blood > arterial blood. The predicted maximal levels of metabolites with carcinogenicity presented an accumulation trend of trichloroacetate > dichloroacetate > chloral > S-(1,2-dichlorovinyl)-L-cysteine. The estimated mean cancer risk of occupational workers was 1.31×10^-5 based on external exposure at the monitored TCE level of (39.2±24.4) μg·m^-3 with an exposure duration of 8 h d^-1 for 20 years. Under the detected exposure level, comparable risks were estimated by the internal tissue levels of TCE based on PBPK model and by the external exposure level of TCE, however, 1.17 to 1.73 times of risk values were estimated based on internal exposure to the main carcinogenic metabolites than that of external exposure to TCE. The higher the exposure level of TCE, the greater the discrepancy between the assessment results of carcinogenic risk from internal exposure and external exposure methods. Sensitivity analysis showed that values of cardiac output and blood flow of richly perfused tissue highly affected the predicted results by PBPK model. Uncertainty analysis presented that the variation of model parameters significantly affected the output of cancer risk assessment. Nonetheless the variation was acceptable. It was indicated that it was necessary to consider in vivo distribution and metabolism of TCE in the assessment of the carcinogenic risk for human being.