A Bayesian Approach to Estimate Parameters for Children from a Pharmacokinetic Model for Methylmercury Exposure to Adults
Rationale (100 words):
Many studies have investigated the association between children’s blood or hair methylmercury concentration and neurological deficits. Other studies have investigated the effect of prenatal exposure based on maternal methylmercury blood or hair concentration. To compare these studies, a model is necessary to equate a methylmercury exposure with a certain blood or hair concentration. While Bayesian methods have been applied for pharmacokinetic models involving women of reproductive age and pregnant women, no such approach has been applied for children. This work describes the development of a model for children, based on application of Bayesian methods and prior information of adult pharmacokinetics.
Approach (100 words):
Our starting point for this work was a previously published 1-compartment pharmacokinetic model (Stern 2005). This model equates a continuous methylmercury exposure to a steady-state blood concentration based on the apparent clearance rate, which is related to the bioavailability, elimination rate, and proportion of methylmercury in blood. To calibrate model parameters for children, estimated methylmercury exposure distributions and biomarker measurements for total mercury in hair, and total mercury and methylmercury in blood were used. The original parameter values, based on data for pregnant women, were used as prior distributions to identify a joint posterior distribution for parameters appropriate for children.
Results & Discussion (100 words):
Model predictions using prior distributions showed substantial overpredictions of blood methylmercury levels. After model calibration the apparent clearance rate (mean ± standard deviation) estimated for children was 122.5 ± 12.0 ml/kg-d, five-fold higher than the estimated value for pregnant women (23.6 ± 10.1). This reflects the calibration correcting the overprediction, and could be related to differences in bioavailability, distribution, or excretion between children and pregnant women. This suggests that a similar biomarker concentration will be associated with a greater exposure in children compared to pregnant women. This model will be applied for quantitative comparison of different neurodevelopmental points of departure.
Why does it matter? (100 words):
Methylmercury is a neurotoxicant commonly associated with the consumption of seafood. It is crucial to understand the risk posed by methylmercury consumption to different populations to reduce the chance of neurodevelopmental toxicity. A better understanding of risk to children could lead to improved guidance on the ideal amount and frequency of seafood consumption for different ages. Application of this model in risk assessment will allow for comparison of methylmercury sensitivity between children and pregnant women and for populations with varying levels of seafood consumption and exposure.
Do you feel this is newsworthy and if so, why? (100 words):
Yes. Methylmercury is a common hazard, and many parents worry about their children’s exposure to methylmercury through frequent consumption of seafood. This work shows that a given biomarker level in a child (e.g., in a hair sample) may imply that the child has experienced a much greater overall exposure than a pregnant woman with the same biomarker level. More work is ongoing to apply this model in risk assessment to compare extrapolated exposure values for many published points of departure and eventually derive reference values.