A translational physiologically based pharmacokinetic approach to predict amoxicillin exposure in the cerebrospinal fluid of pigs with meningitis
PBPK modelling; translational reverse model; Streptococcus suis; cerebrospinal fluid; swine meningitis.
Amoxicillin is commonly used to treat Streptococcus suis meningitis in pigs, but dosing regimens rely solely on plasma exposure data. To the authors' knowledge, no study has quantified amoxicillin concentrations in the cerebrospinal fluid (CSF) of this species, leaving current dosing recommendations for meningitis unsupported by target-site data. The aim of this study was (1) to develop and validate a whole-body physiologically-based pharmacokinetic (PBPK) model for amoxicillin plasma profiles in pigs following intravenous, intramuscular, and oral administration; (2) to build a compartmental model, using paired human plasma-CSF data from three independent meningitis studies, to estimate the permeability-surface area product (PSc) reflecting altered blood-CSF barrier passage in meningitis; and (3) to integrate the human-derived PSc into the pig PBPK model, proposing a translational, physiologically-based framework to predict amoxicillin CSF exposure in pigs with meningitis. The whole-body PBPK model showed adequate predictive performance (global mean and absolute fold error within 0.8–1.25). Nonlinear mixed-effects modeling of the human data estimated PSc at 14.95 mL/h (relative standard error, 31%). Coupling this parameter to the pig PBPK model, via a new CSF compartment linked to brain and peripheral venous plasma, predicted a later, lower CSF concentration peak and slower decline than the systemic plasma profile, consistent with the expected kinetics of a hydrophilic beta-lactam under a compromised blood-CSF barrier. These findings provide a first mechanistic estimate of amoxicillin central nervous system exposure in pigs and allow proposing protocols with potential therapeutic efficacy against meningitis, to be tested in future in vivo studies.