Pharmacokinetic-Pharmacodynamic Integration in Devices

Abstract

Drug delivery devices are not passive carriers -- they are active participants in pharmacokinetics and pharmacodynamics, determining where drug goes, how fast it gets there, and what concentration profile the target tissue experiences. The integration of pharmacokinetic and pharmacodynamic modelling into device design and control has moved from an academic niche to a clinical and regulatory priority, driven by the recognition that optimised device-drug integration can substantially improve therapeutic outcomes beyond what either the drug formulation or the device mechanics can achieve independently. This study presents the PK-PD Device Integration Assessment Framework (PKPDIF), evaluating five integration approaches -- model-informed precision dosing with therapeutic drug monitoring (MIPD-TDM), closed-loop PK feedback control (CL-PK), population PK-PD modelling for device parameter optimisation (PopPK-PD), physiologically-based PK modelling for device design (PBPK-DD), and AI-driven real-time PK-PD adaptation (AI-PKPD) -- across four drug-device system contexts: insulin pump and closed-loop delivery systems, implantable drug infusion pumps, drug-eluting stents and vascular implants, and transdermal and microneedle delivery systems. Performance was scored using the Integration Effectiveness Score (IES), a weighted composite of dosing accuracy (0.30), patient outcome improvement (0.20), regulatory model acceptance (0.20), implementation complexity-inverse (0.15), and safety profile (0.15). MIPD-TDM achieved the highest IES (0.894), combining the best regulatory acceptance (0.920) with strong dosing accuracy (0.900) and safety (0.900). Closed-loop PK feedback control ranked second (0.883) with the best dosing accuracy (0.940) and patient outcomes (0.920), driven by the artificial pancreas literature. PBPK-based device design (0.876) and AI-driven adaptation (0.872) both showed strong technical performance but face regulatory acceptance gaps that limit current deployment.