Predictive performances of different therapeutic drug monitoring approaches to assess Tacrolimus and mycophenolic acid exposure

Objective: Mycophenolic acid (MPA) and Tacrolimus (TAC) are immunosuppressive drugs routinely used after solid organ transplantation. MPA and TAC have a rather narrow therapeutic window and their pharmacokinetics show considerable variability. In most centers, TDM is frequently performed based only on trough concentrations (C₀). The main objective of the present study was to compare performances of three different approaches of C ₀ based TDM of these two drugs, in their ability to predict MPA and TAC area under the curve (AUC). Methods: Linear regression models and Bayesian estimators were developed based on solely on C₀ levels. Kappa and McNemar concordance and divergence tests were used to see the ability of different C₀ TDM approaches to predict the same dosage recommendation as AUC TDM. Results: 52% and 57% of the observed AUC values from patients receiving a fixed dose were outside the recommended therapeutic range for MPA and TAC, respectively. For both TAC and MPA, the Bayesian estimators performed much better than the two other approaches, followed by linear regression models.