Bayesian toxicokinetic modeling of cadmium exposure in Chinese population

Abstract Cadmium (Cd) is a toxic heavy metal widely present in the environment. Estimating its internal levels for a given external exposure using toxicokinetic (TK) models is key to the human health risk assessment of Cd. In this study, existing Cd TK models were adapted to develop a one-compartment TK model and a multi-compartment physiologically based toxicokinetic (PBTK) model by estimating the characteristics of Cd kinetics based on Cd exposure data from 814 Chinese residents. Both models not only considered the effect of gender difference on Cd kinetics, but also described the model parameters in terms of distributions to reflect individual variability. For both models, the posterior distributions of sensitive parameters were estimated using the Markov chain-Monte Carlo method (MCMC) and the approximate Bayesian computation-MCMC algorithm (ABC-MCMC). Validation with the test dataset showed 1.4–22.5% improvement in the root mean square error (RMSE) over the original models. After a systematic literature search, the optimized models showed acceptable prediction on other Chinese datasets. The study provides a method for parameter optimization of TK models under different exposure environment, and the validated models can serve as new quantitative assessment tools for the risk assessment of Cd in the Chinese population.