Quantification of biological variability.

Pharmacokinetic models are usually developed to describe the kinetic behaviour of the chemical compounds in an average human body. There are however many situations were the effects of variations in a given pharmacokinetic parameter would be of interest. For example, in biological monitoring, the variability of biological data is a critical parameter when relating air and biological measures of an individual exposure. This paper reviews two possible approaches for the simulation of such variability in occupational pharmacokinetics. In the first, a simple one-compartment model is used together with statistical distributions for the intake and elimination of the compounds involved. Based on realistic distributions, this simple model is applied to the comparison of biological and air monitoring to estimate workers' exposure. The second involves a seven-compartment physiologically based pharmacokinetic model, which includes provision of the means to input satistical distributions for some of the parameters: exposures, physical workload, body height, body weight, liver function and renal function. It makes it possible to predict the pharmacokinetic response of groups of workers, who differ in their exposures and in their physiological parameters. Realistic statistical distributions are then used to describe biological monitoring variability. The advantages and disadvantages of both the simple pharmacokinetic model and the physiological model are discussed in the context of predicting and understanding variability in occupational situations. Other potential developments are also considered.