Integrated Model-based Medication Dosing Assist App(Lication) in Klinisch Werk Station (KWS). Proof-of-Concept Validation Study: Tacrolimus in Kidney Transplantation.

Tacrolimus (Prograft™, Advagraf™, Astellas), a calcineurin-inhibitor, is globally used as primary immunosuppressive drug in all types of solid organ transplantation (kidney, liver, lung, heart, pancreas, bowel,....). Tacrolimus is characterized by a strongly variable oral absorption (oral bioavailability) that necessitates continuous therapeutic drug monitoring (TDM) of tacrolimus dosing during clinical follow-up. Tacrolimus has a critical (narrow) therapeutic index whereby clinicians use pre-dose blood tacrolimus levels (C0) to guide dose adaptations aimed at predefined target concentration ranges that vary with time after grafting and according to clinical conditions. Attaining target concentration ranges is important to avoid rejection and toxicity. Clinicians currently execute tacrolimus dose adaptations based on (previous) experience, varying (pharmacological) knowledge about variables that affect tacrolimus disposition and clinical circumstances.

Based on the exploratory analysis of a test set of 315 kidney transplant patients in the University Hospitals Leuven with extensive tacrolimus PK data available, the accuracy and precision of this clinical approach was shown to be suboptimal. At day 2 post-transplantation clinicians underestimate tacrolimus exposure (i.e. overdose) in 63% of patients, overestimate exposure (i.e. underdose) in 23% of patients and correctly predict the dose in 14% of patients. Over time these numbers considerably improve: at day 4, in 30% of patients tacrolimus exposure is correctly predicted/dosed and at day 10, in 24% of patients correctly predicted/dosed, respectively. These numbers appear low, but the clinicians performance is hampered by variable individual experience with the drug, the rapidly changing post-transplantation circumstances (gastrointestinal motility, steroid tapering, peri- and post-transplantation hemodynamic changes, co-medication, comorbidity, ...) and lack of pharmacometric (PM) training, which leads to the use of steady-state approaches to execute dose adaptations (previous tacrolimus concentration and corresponding dose) in patients who are effectively not in steady-state.

These problems are not unique for tacrolimus. In pharmacometrics, algorithms for data-driven identification of pharmacokinetic/pharmacodynamic (PK/PD) parameters in a population, a posteriori identification of individual PK/PD parameters and control of concentration and/or effect targets are available. They can be implemented in clinical patient care via software linked to a graphical user interface operable by clinicians lacking formal pharmacometric training. We hypothesize that such a software will help clinicians, in a safe environment, to significantly increase precision and decrease bias while aiming for the optimal individualized dosing of tacrolimus (personalized or precision dosing in solid organ recipients.

The use of organ-specific predictive pharmacometric PK models (PM) for tacrolimus that take into account genetic, demographic (age, body weight), biological (hematocrit) and other parameters (time after transplantation, calculated tacrolimus oral clearance from previous dosing) that influence ADME (Absorption, Distribution, Metabolism, Elimination) of tacrolimus, can support clinicians in determining the optimal dose adaptations in routine clinical practice. With model-based dosing support or dosing assistance, quality of clinical care can be potentially improved:

1. Faster achievement of target therapeutic concentrations ranges
2. Attenuation of "off-target concentrations" (sub- and supra-therapeutic)
3. Reduction of intra-patient variability of C0 (IPV)
4. Less tacrolimus pre-dose trough concentration measurements (except in initial phase)
5. More accurate dose adjustments in the presence of known (and unknown) inhibitors / inducers of tacrolimus ADME (drug-drug interactions; DDI)
6. Potential pharmaco-economic advantage
7. Improved drug tolerance / quality of life patients (less adverse effects)
8. Optimization of clinical workflow (automation clinical decision tree in KWS)
9. Improved patient safety (traceable validated decision process)
10. Support, education and (learning) feedback for medical trainees