Model-Informed Precision Dosing of Everolimus: External Validation in Adult Renal Transplant Recipients

Abstract

Background and Objective The immunosuppressant everolimus is increasingly applied in renal transplantation. Its extensive pharmacokinetic variability necessitates therapeutic drug monitoring, typically based on whole-blood trough concentrations (C 0). Unfortunately, therapeutic drug monitoring target attainment rates are often unsatisfactory and patients with on-target exposure may still develop organ rejection. As everolimus displays erythrocyte partitioning, haematocrit-normalised wholeblood exposure has been suggested as a more informative therapeutic drug monitoring marker. Furthermore, model-informed precision dosing has introduced options for more sophisticated dose adaptation. We have previously developed a mechanistic population pharmacokinetic model, which described everolimus plasma pharmacokinetics and enabled estimation of haematocrit-normalised whole-blood exposure. Here, we externally evaluated this model for its utility for model-informed precision dosing. Methods The retrospective dataset included 4123 pharmacokinetic observations from routine clinical therapeutic drug monitoring in 173 renal transplant recipients. Model appropriateness was confirmed with a visual predictive check. A fit-forpurpose analysis was conducted to evaluate whether the model accurately and precisely predicted a future C 0 or area under the concentration-time curve (AUC) from prior pharmacokinetic observations. Bias and imprecision were expressed as the mean percentage prediction error (MPPE) and mean absolute percentage prediction error (MAPE), stratified on 6 months post-transplant. Additionally, we compared dose adaptation recommendations of conventional C 0-based therapeutic drug monitoring and C 0-or AUC-based model-informed precision dosing, and assessed the percentage of differences between observed and haematocrit-normalised C 0 (∆C 0) and AUC (∆AUC) exceeding ± 20%. Results The model showed adequate accuracy and precision for C 0 and AUC prediction at ≤ 6 months (

Volumetric microsampling for simultaneous remote immunosuppressant and kidney function monitoring in outpatient kidney transplant recipients

“…Similarly, others showed that tacrolimus, MPA, everolimus, creatinine and iohexol can be quantified reliably in VAMS samples with haematocrit values between approximately 0.20 and 0.60 43,48,58 . As haematocrit values outside these ranges are seldomly encountered in adult kidney transplant recipients, 93 also in the clinically unstable phase, this probably allows for DBS‐ and VAMS‐based monitoring without pronounced haematocrit interference for most patients. Nevertheless, DBS‐ and VAMS‐based monitoring should still be considered carefully in the first weeks after transplantation and in otherwise clinically unstable patients, as these studies set their maximum tolerable haematocrit effect thresholds at ±15%, which still allows for substantial haematocrit‐dependent assay variability.…”

“…Similarly, others showed that tacrolimus, MPA, everolimus, creatinine and iohexol can be quantified reliably in VAMS samples with haematocrit values between approximately 0.20 and 0.60. 43,48,58 As haematocrit values outside these ranges are seldomly encountered in adult kidney transplant recipients, 93…”

Tacrolimus Exposure Prediction Using Machine Learning

“…Following the publication of a recent PopPK model for everolimus AUC estimation, integrating previous exposure, 23 we checked whether taking into account anteriorities could improve AUC estimation. For this, we selected the patients with at least two requests (data not shown).…”

Development and evaluation of uncertainty quantifying machine learning models to predict piperacillin plasma concentrations in critically ill patients

“…Bayesian prior) and compared to the new ML models. For a posteriori evaluation, individual PK parameter estimates, as opposed to population PK parameter estimates, can be used to generate a posteriori predictions (the Bayesian posterior), and compared to the prev ML models [44][45][46]. All used ML models provide deterministic predictions.…”