Monitoring immunosuppression after liver transplantation : development of individualized Bayesian limited sampling monitoring

Abstract

After using C0-monitoring as the tool for therapeutic drug monitoring of cyclosporine for many years, studies suggested that C2-monitoring might be better. After switching 31 liver transplant patients using cyclosporine from C0 via C2 to flexible limited sampling models (LSM), combinations of blood sampling time points 0+2h (r_=0.94); 0+1+2h (r_=0.94); 0+1+3h (r_=0.92); 0+2+3h (r_=0.92) and 0+1+2+3h (r_=0.96) showed excellent correlation with AUC0-12h with acceptable precision and bias. When evaluating the LSM0+1+2+3h model in the 18 months after introduction there was no significant change in average cyclosporine dose and creatinine clearance, compared to previous C2-monitoring. Especially LSM0+2h was optimal in terms of accuracy, ease-of-use and intrapatient variability. When optimizing tacrolimus monitoring after calculating limited sampling formulas (LSF) and LSM single and multiple-point combinations showed good correlations with AUC0-12h. The best single point calculation in terms of estimating systemic tacrolimus exposure using LSM were LSM 4h (r_=0.97) and LSM 6h (r_=0.97). During the study of the pharmacokinetic behaviour of MMF we found a wide range in MPA clearance in the population (8.08__57.47 L/h). Looking at possible sources of this variability in MPA clearance we divided our group, based on clinical selection, into two groups (with and without calcineurin inhibitors). These groups were used for further development of LSM for monitoring MPA. The combination 0-_-1-2h showed very good correlations with trapezoidal AUC0-12h for both models, with acceptable bias and precision.