Modeling the exposure history of human individuals to persistent organic pollutants

Combined environmental fate and food chain bioaccumulation models provide the capability to simulate, through a fully mechanistic description of the relevant thermodynamic and kinetic processes, a persistent organic pollutant (POP)’s journey from release into the environment to its accumulation in humans. Non-steady-state models allowing time-variant simulations covering periods of multiple decades have been used to estimate the average exposure history of different population cohorts to POPs. Such assessments may then serve as a complement to empirical
biomonitoring studies evaluating a population\'s exposure to POPs. It is of substantial interest to identify sources of variability in the measured human concentrations. Biomonitoring studies often obtain additional information on the sampled individuals through questionnaires.
Collected data generally include age at sampling, date of sampling, gender, body mass index, number of children and age at birth, breastfeeding habits, and detailed dietary composition. While such data can be used to find statistical associations between measured contaminant exposures and potential influential factors, Statistical analyses cannot identify causal relationships and provide ultimately no
mechanistic understanding of the processes underlying human POP exposure. However, the same questionnaire responses can also be used to individualize the parameterization of the human in mechanistic exposure models. In other words, instead of only calculating average population exposure, the models can be used to simulate the distinct exposure histories of all sampled human participants individually. Here we present the results of three studies simulating the exposure of individuals to polychlorinated biphenyls (PCBs) using the environmental fate and bioaccumulation model CoZMoMan or similar model combinations. The models generally succeed in predicting population average exposure. Also the range of predicted PCB exposures generally agrees with the range of measured exposures.
Trends with key variables, such as BMI and age, were mostly similar for the measured and simulated concentration data. However, in two studies the models attributed more of the variability in PCB exposure to interindividual differences in dietary composition than a statistical analysis of the measured data would support. The measured data indicated that
there were other sources of variability not currently accounted for in the models.