Collaboratrices et collaborateurs

Introduction: Perfluoroalkyl acids (PFAAs), a subgroup of per- and polyfluoroalkyl substances (PFAS), are highly persistent, often bioaccumulative and toxic. This has led to the establishment of maximum levels for 4 PFAAs in animal food in the EU. Understanding transfer kinetics using modelling is critical for human health and safe livestock production systems. This study develops a physiologically based toxicokinetic (PBTK) model to simulate the absorption, distribution and excretion of PFAAs in dairy goats and their products. Recent models address pharmaceutical kinetics in goats (Ai et al., 2024) or PFAS kinetics in cows (Mikkonen et al., 2023), but not isomer-specific PFAS kinetics in goats. We examine here how chain length and functionality of 30 branched (br-) and linear (n-) PFAAs affect kinetics in 12 tissues and excreta (e.g., milk). Chain branching and functionality are expected to influence kinetics by altering serum binding and thus renal and milk excretion. The model integrates the experimental data to predict PFAA levels in goats across growth and lactation stages. Material and methods: A dynamic, compartmental PBTK model was developed based on experimental data from 8 late stage lactating dairy goats (White German Noble Goat) housed at the BfR experimental farm. The goats were divided into a control and an exposure group (4 animals + 1 reserve per group). The exposure group received PFAS-contaminated hay from a contamination incident in Brilon/Scharfenberg for 8 weeks, followed by a 12-week depuration phase with PFAS-free hay. The contaminated hay had a total quantified PFAS content of 497 µg/kg (88 % dry matter), including perfluorobutanoic acid (PFBA, 167 µg/kg), perfluorooctanesulfonic acid (PFOS, 80 µg/kg) and perfluorooctanoic acid (PFOA, 60 µg/kg). The milk yield was 0.2-1.6 L/day. Milk was collected 1–3 times per week individually plus a weekly pooled sample. Individual hay intake was recorded daily. Hay, feces, urine, milk, and serum were sampled periodically. On weeks 1, 5, and 9, urine and feces were collected individually. At the end of the study, goats were slaughtered and muscle, liver, kidney, lung, heart, brain and spleen samples collected for analysis. The proposed PBTK model consists of 12 interconnected compartments representing organs, including blood, liver, kidney, spleen, mammary gland, gastrointestinal tract (stomach and intestine), lungs, brain, muscle, heart and a rest compartment plus 3 excretion pools (urine, feces, and milk). The model was implemented in Python 3.14 and uses an analytical solution of the differential equations for computational efficiency. Model parameterization is currently based on average physiological and intake data from exposed animals. Although the model does not yet incorporate interindividual variability or dynamic physiology (e.g., changes in milk yield), the framework is designed to support such extensions. Results and discussion: Preliminary in vivo feed-to-milk transfer rates (TRs) ranged from less than 1 % (e.g., n- and br-PFOA) to up to 15 % for n-PFOS. Depuration half-lives ranged from <1 day (n-perfluoropentanesulfonic acid, PFPeS) to 61 days (br-perfluoroundecanoic acid, PFUnDA). The br-isomers consistently showed lower TRs than their n-counterparts, but this trend did not extend to half-lives. On average, milk TR for br-isomers were 63 % lower than those for n-isomers, with reductions ranging from ∼18 % for n-PFOS (12.7 %) vs. br-PFOS (10.4 %) to over 80 % for n-PFOA (5.4 %) vs. br-PFOA (1.2 %). For most compounds, the exposure period (8 weeks) was sufficient to approach steady-state TRs, so that the estimated values are reliable for modelling. Further lab results on serum, urine, feces and tissues are expected and will be incorporated into the model. Conclusion and implications: The proposed PBTK model represents an initial framework for simulating the kinetics of linear and branched PFAAs in dairy goats. While still under development, the model is designed to support quantitative assessment of accumulation and elimination processes, including excretion via milk. Its modular structure allows future adaptation to different goat breeds, milk yield, and exposure scenarios.