Nicotine pharmacokinetic (PK) data are central to the assessment of oral nicotine products, including evaluation of abuse liability, product switching potential, and bridging across related formulations. However, clinical PK studies are costly, time consuming, and generally performed late in development, after key formulation and design decisions have already been made. We developed a product-attribute-based modeling framework to estimate Controlled Use nicotine PK outcomes for nicotine pouch products using measurable inputs, including nicotine content, pH, estimated unprotonated nicotine fraction, pouch matrix composition, in vitro nicotine release kinetics, pouch mass, and other physical design attributes. Model development used a structured workflow that included aggregation of Controlled Use PK data from publicly available product studies, harmonization of PK endpoints, model feature engineering based on mechanistic determinants of oral nicotine absorption, and training and internal validation of machine learning models to predict Cmax, Tmax, and AUC. The resulting model reliably differentiated products with distinct nicotine content, pH, nicotine release, and matrix characteristics, and approximates PK values within typical regulatory confidence bounds (<10%). The framework is intended to support formulation screening, selection of products for clinical abuse liability or bioequivalence studies, and generation of quantitative hypotheses for bridging across related nicotine pouch products. It is not intended to replace clinical PK studies or to predict ad libitum use, where nicotine exposure may be influenced substantially by behavioral compensation and user-controlled titration. This approach provides a scalable, transparent tool for integrating product chemistry, dissolution, and design attributes into nicotine pouch development and regulatory decision-making.

TSRC 2026