Device-Drug Interaction Mechanisms

Abstract

Device-drug interactions (DDIs) -- clinically significant modifications of drug pharmacokinetics, pharmacodynamics, or device performance arising from the co-presence of an implanted or wearable device and a systemically or locally administered drug -- represent an under-recognised safety domain that falls between the traditional silos of pharmacovigilance and medical device surveillance. From MRI-induced heating of metallic implants altering local tissue drug distribution to drug-eluting stent coatings interacting with concurrent antiplatelet pharmacology, DDIs create adverse effects that neither the device nor the drug would produce independently. This study presents the Device-Drug Interaction Assessment Framework (DDIAF), evaluating five interaction investigation methodologies -- pharmacokinetic alteration studies, device material-drug compatibility testing, computational interaction modelling, clinical pharmacovigilance data mining, and AI-integrated multi-database signal detection -- across four interaction categories: absorption modification by implanted devices, distribution alteration by device-induced tissue changes, device material-drug chemical interactions, and device function alteration by concurrent medications. Our DDI Detection Performance Score (DDIPS) integrates interaction detection sensitivity, mechanistic characterisation depth, clinical relevance assessment, detection timeliness, and cross-database coverage. AI-integrated multi-database signal detection achieved the highest DDIPS (0.924) through NLP and knowledge graph models that linked adverse events across FAERS, MAUDE, and EHR databases, identifying 38% more DDI signals than single-database surveillance, while computational interaction modelling achieved the highest mechanistic depth (0.955) through molecular dynamics simulation of drug-material interfaces that predicted adsorption-mediated DDIs with 91% accuracy.