Safety assessment of drugs in pregnancy: An update of pharmacological models.

Abstract

Pregnant women are largely excluded from clinical drug trials due to ethical concerns, leading to a critical knowledge gap in medication safety during pregnancy. Indeed, over 90 % of approved drugs lack pharmacokinetic data in pregnant patients, and only two new therapies have been specifically developed for use in pregnancy in the past three decades. Traditional pharmacological models (animal studies and static 2D cell cultures) often fail to predict maternal-fetal drug transfer and toxicity due to interspecies differences and an inability to mimic the dynamic physiology of pregnancy-particularly the pivotal role of the placenta. Consequently, clinicians must frequently weigh maternal treatment needs against uncertain fetal risks. Recent technological advances have begun to bridge these gaps: placental organoids and microfluidic placental-on-chip now serve as placental pharmacology platforms, and AI-based predictive models can integrate complex datasets to forecast drug disposition in pregnancy. This mini-review provides an updated overview of these emerging approaches for pregnancy drug safety assessment. It highlights how these innovative models recapitulate key aspects of placental structure and function, enabling more accurate evaluation of drug pharmacokinetics and toxicity at the maternal-fetal interface. Integrating advanced placental models with computational tools offers transformative potential for pregnancy pharmacotherapy. Future efforts should focus on combining experimental models with computational approaches to improve translational relevance. Standardized protocols, clinical biomarker validation, and ethical governance are essential to advance these technologies from experimental platforms to regulatory and clinical decision-making tools.