Population pharmacokinetic models for unbound mycophenolic acid

Mycophenolic acid (MPA) is a first-line immunosuppressant for organ transplant recipients. MPA therapeutic drug monitoring relies on total drug exposure which poorly reflects the biologically-active unbound MPA in clinical conditions with altered renal function and protein-binding. Population pharmacokinetic-Bayesian forecasting models (i.e. model-informed precision dosing [MIPD]) incorporating unbound MPA are needed to advance personalized MPA dosing.

PubMed and Embase were searched up to 22 July 2026 for population pharmacokinetic models of unbound MPA. Eight models were in adult kidney transplantation and three in hematopoietic stem cell transplantation. We evaluated their designs, sampling schedules, structural models, protein-binding, handling of the glucuronide metabolite (MPAG), incorporation of entero-hepatic recirculation (EHC), covariates, pharmacodynamics, and evaluation/validation methods.

The studies exhibit heterogenous sampling schedules, analytical methods, structural models, and covariates. The handling of MPAG and EHC modeling is inconsistent, and protein-binding, a crucial component defining the relationship between total and unbound MPA, is not well characterized. No MIPD strategy is yet available. Over the next five to ten years, unbound MPA therapeutic and toxicity exposure thresholds should be established, which, in conjunction with validated population pharmacokinetic models that capture the complex pharmacokinetics-dynamics of MPA can form a more effective, personalized dosing strategy.