Evaluation of the predictive performance of physiologically based pharmacokinetic models for intramuscular injections of therapeutic proteins.

Several physiologically-based pharmacokinetic (PBPK) models have been reported for intravenous (IV) and subcutaneous (SC) injections, but there has been a paucity of work for intramuscular (IM) injections. The primary objective of this work was a wide-scale evaluation of the predictive performance of IM PBPK models of therapeutic proteins. PBPK models for all administration routes available in the literature have regarded muscle as the total muscle (TM) in the body; however, anatomically, the body is composed of discrete muscle groups. Clinically, IM is administered to a specific muscle (SM). We explored the predictive performance of IM PBPK models with an SM or TM dosing site. The plasma concentration-time profiles of seven therapeutic proteins after an IM dose in humans served as the clinically observed data for model evaluation - this was a diverse group ranging from 30 to 149 kDa from six protein classes. Pharmacokinetic parameters Cmax, tmax, AUC0-∞, and ka were estimated. SM and TM IM PBPK approaches were compared using Average Fold Error (AFE) and Pearson Chi-Square LineShape analyses. This work represents the first wide-scale validation of IM PBPK models and suggests that these models predict IM PBPK reasonably well. The SM and TM approach provided comparable performance.