Predicting Cumulative Neonatal Sodium Benzoate Exposures During Breastfeeding with Physiologically Based...

Sodium benzoate (SB) is used as a second-line therapy to treat rare urea cycle disorders (UCDs) in paediatric and adult patients. However, lactation data for SB do not exist, leading to uncertainties regarding the extent of infant drug exposure due to breastfeeding while taking this medication. These uncertainties may lead to cessation of breastfeeding, unnecessarily depriving both mother and infant of its numerous benefits. Thus, this study aims to develop and apply a paediatric physiologically based pharmacokinetic (PBPK) model to predict the compounded neonatal SB exposure from both in utero and from breastfeeding.

An adult whole-body PBPK model for SB was first developed and validated with literature-based plasma concentrations after oral or intravenous doses. To account for age-related changes in SB pharmacokinetics, the adult model was scaled to paediatric populations using age-dependent algorithms to capture physiological and anatomical changes, while ontogeny functions were developed to capture maturation of enzymes and transporters where possible. A published cord-coupled PBPK modelling workflow was followed to simulate the cumulative pre- and post-natal drug exposure levels in breastfed infants, accounting for variability in milk ingestion volumes and assuming a conservative milk-to-plasma ratio of 1:1. The upper area under the curve (AUC) ratio (UAR) was used for risk assessment, taking the ratio between the AUC for the highest risk (95^th percentile simulated AUC) infants and the median maternal AUC following a therapeutic SB dose and multiplying by 100% to report as a percentage.

The paediatric PBPK model captured observed SB plasma concentrations in both neonates and children with acceptable bias and precision, supported by the average-fold error (AFE: 1.17) and the absolute AFE (AAFE: 1.24) evaluation metrics, respectively. Neonatal SB exposure levels were the highest on the first day of life (UAR: 14.2%) due to the presence of residual prenatal drug levels. Subsequent UAR values were low (1.5-3.3%) on postnatal days 1-31, suggesting rapid neonatal clearance and negligible contribution of prenatal exposure relative to breastfeeding exposure for the infant's total drug burden.

This study is the first to report a paediatric PBPK model of SB pharmaceutical use, incorporating prenatal and lactational exposure to assess the cumulative risk of exposure for the breastfed infant. Model predictions suggest that SB exposure through breastfeeding is minimal and unlikely to cause adverse outcomes when compared with clinical studies. These findings may support clinical decision making in the absence of clinical lactation data; however, empirical studies are needed to validate the predictions given the limited information on milk transfer, enzyme ontogeny, and neonatal concentrations.