Miao-Chan Huang(1), Julia Macente(1), Bart Lammens(2), Domenico Ventrella(3), Pieter Annaert(1,2)
(1) KU Leuven, Leuven, Belgium; (2) BioNotus, Niel, Belgium; (3) University of Bologna, Bologna, Italy.
Objectives: Antibiotics are frequently prescribed to postpartum women and nursing mothers for prophylaxis and treatment of infections. In the latter population, the infant exposure to the maternal antibiotics is a safety issue commonly addressed. To minimize the knowledge gap on the safety of medicines used in lactating women and their breastfed infant(s), work package 3 in the Innovative Medicines Initiative (IMI) project ConcePTION has the goal to establish a non-clinical platform for the safety assessment of medicines by predicting exposure levels of medicines. The Göttingen minipig can serve as a non-clinical species for in vivo lactation studies with good bio-relevance to humans1. As the first step to investigate the transfer of amoxicillin (AMX) from maternal plasma to milk, the present study aimed to develop a population pharmacokinetic (Pop PK) model to capture the PK characteristics of intramuscular (IM) AMX in postpartum minipig plasma.
Methods: Data on plasma concentration of AMX in postpartum Göttingen minipigs were provided by the research team at University of Bologna under the IMI ConcePTION project. A total of 125 plasma concentration data were available from the 3 minipigs receiving once daily IM administration of AMX (Clamoxy® RTU; 7 mg/kg). The demographics of the minipigs are summarized in Table 1. Seven plasma concentrations collected on Day 1 from one of the minipigs were excluded because the animal had developed fever 4 hours post-dose. This episode resolved on the same day. A total of 118 plasma concentrations were included in the analysis. Pop PK analysis was performed using nonlinear mixed-effects modeling approach in Monolix® version 2020R1. The initial structural model was optimized to the plasma concentrations within 24 hours, as Day 1 had richest data to describe the absorption and distribution phase. This model was then modified to fit the multiple-dose data. The covariate analysis was performed with the following selection criteria: a reduction in objective function value (OFV) of >3.84 (p <0.05) in the forward selection and an increase in OFV of >6.63 (p <0.01) in the backward elimination. The predictive performance of the final model was evaluated by internal validation based on the goodness-of-fit (GOF) plots and visual predictive checks (VPCs). The population and individual parameters derived from the final model were used to simulate the plasma concentrations of AMX in a population of 1000 individuals. The simulation result was visualized against the observations using R version 4.0.3 and RStudio version 1.3.1093.
Results: The basic structural model comprised a two-compartment model with zero-order absorption and linear elimination. Given the trough concentrations increased over time until the end of the study, the clearance (CL) and the central volume of distribution (V1) were implemented to be time-varying (TV) parameters using days postpartum (DPP) as the regressor (equation 1 and 2). β1 and β2 were the coefficients for the effect of the time after delivery. As the samples were collected before and at 2-6 hours post-dose from Day 2 onwards, there were insufficient data to estimate the duration of the zero-order absorption process (Tk0), the intercompartmental clearance (Q), and the peripheral volume of distribution (V2). These parameters were therefore fixed to stabilize the model.
CLTV = CL*(1/DPP)β1 (1)
V1TV = V1*log(DPP)β2 (2)
No significant covariate effect was identified in the covariate analysis. Overall, the relative standard error (RSE) values for the Pop PK parameters were within 30%, except for V1 (42.7%). Most of the observations fell within the 90% prediction intervals in VPCs and the simulation range.
Conclusions: The developed model demonstrated the ability to describe the PK profiles of AMX in postpartum minipig plasma with the interindividual-variability captured. The next step will be connecting this model to the milk data, and the simulation results will be compared with the published human model to obtain information for the inter-species extrapolation.
Acknowledgements: This work is supported by the EU/EFPIA Innovative Medicines Initiative Joint Undertaking ConcePTION grant No. 821520 and Taiwan Scholarship Programme. The research leading to these results was conducted as part of the ConcePTION consortium. This abstract only reflects the personal views of the stated authors.
References:
[1] Ventrella D, Ashkenazi N, Elmi A, et al. Animal Models for In Vivo Lactation Studies: Anatomy, Physiology and Milk Compositions in the Most Used Non-Clinical Species: A Contribution from the ConcePTION Project. Animals. 2021;11(3):714. doi:10.3390/ani11030714
Reference: PAGE 30 (2022) Abstr 10123 [www.page-meeting.org/?abstract=10123]
Poster: Drug/Disease Modelling - Infection