PWI-003 - ADVANCEMENTS IN PEDIATRIC PHYSIOLOGICALLY-BASED PHARMACOKINETIC (PBPK) MODELING OF DRUGS IN LACTATION FOR GUIDING NEONATAL EXPOSURE RISK ASSESSMENT

S. Wang¹, S. Suryavanshi², D. Hajducek¹, A. Hamadeh¹, C. Yeung³, P. Maglalang⁴, S. Ito³, J. Autmizguine⁵, D. Gonzalez⁶, A. Edginton⁷; ¹University of Waterloo, Kitchener, Ontario, Canada, ²Altasciences, Lethbridge, AB, Canada, ³Division of Clinical Pharmacology and Toxicology, The Hospital for Sick Children, Toronto, Ontario, Canada, ⁴University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, ⁵Department of Pediatrics & Department of Pharmacology and Physiology, Université de Montréal, Montréal, Québec, Canada, ⁶Duke University, Durham, USA, ⁷University of Waterloo, Kitchener, ON, Canada.

Background: Current methods in predicting infant risk to maternal drug exposure through breastfeeding use data on breastmilk drug concentrations for modeling infant drug pharmacokinetics (PK). These models do not account for prenatal drug exposure in the early neonatal period and tend to underpredict exposure in this vulnerable phase of life [1]. We apply umbilical cord (UC) drug concentrations in model predictions to holistically represent all sources of drug exposure to the breastfed infant. Levetiracetam and sodium benzoate are used as case examples.

Methods: A pediatric PBPK model was built for the study drugs to predict infant drug exposure from breastfeeding. Infant exposures were accounted for using literature-based UC drug levels, model-based milk-intake volumes [1,2], and breastmilk drug levels from literature or clinical collaborators. Model parameterization was informed by published in vitro assays and clinical PK studies. Independent PK studies were used for evaluation. The plasma area under the curve (AUC) for three age groups was simulated and the upper AUC ratio (UAR) between the 95th percentile AUC, and the median maternal AUC was calculated to assess the relative risk [1]. Model performance was compared before and after the addition of prenatal drug levels through boxplot analyses.

Results: The PBPK models accurately captured the observed AUC0-24h in pediatric populations, providing confidence in the adult to pediatric extrapolation. Neglecting UC data resulted in underpredictions in drug exposure of breastfed neonates for levetiracetam where observed data were available. Accounting for this improved predictions for the earliest age group (Figure 1) and UARs approximated low-end therapeutic exposures in clinical studies. Due to prenatal exposure, infant drug levels were highest at birth despite continued exposure through breastfeeding. The influence of prenatal exposure on neonatal drug levels was more prevalent with increased half-life and depends on neonatal clearance.

Conclusion: Prenatal drug exposure largely contributes to neonatal plasma levels and including them improves PBPK model-based predictions in the early neonatal period. These results suggest that most infants should not experience adverse outcomes from breastfeeding exposure to the study drugs.