Adaptation of a population pharmacokinetic model to inform tacrolimus therapy in heart transplant recipients.

AIM: Existing tacrolimus population pharmacokinetic models are unsuitable for guiding tacrolimus dosing in heart transplant recipients. This study aimed to develop and evaluate a population pharmacokinetic model for tacrolimus in heart transplant recipients that considers the tacrolimus-azole antifungal interaction. METHODS: Data from heart transplant recipients (n = 87) administered the oral immediate-release formulation of tacrolimus (Prograf®) were collected. Routine drug monitoring data, principally trough concentrations, were used for model building (n = 1099). A published tacrolimus model was used to inform the estimation of Ka , V2 /F, Q/F and V3 /F. The effect of concomitant azole antifungal use on tacrolimus CL/F was quantified. Fat-free mass was implemented as a covariate on CL/F, V2 /F, V3 /F and Q/F on an allometry scale. Subsequently, stepwise covariate modelling was performed. Significant covariates influencing tacrolimus CL/F were included in the final model. Robustness of the final model was confirmed using prediction-corrected visual predictive check (pcVPC). The final model was externally evaluated for prediction of tacrolimus concentrations of the fourth dosing occasion (n = 87) from one to three prior dosing occasions. RESULTS: Concomitant azole antifungal therapy reduced tacrolimus CL/F by 80%. Haematocrit (∆OFV = -44, P < .001) was included in the final model. The pcVPC of the final model displayed good model adequacy. One recent drug concentration is sufficient for the model to guide tacrolimus dosing. CONCLUSION: A population pharmacokinetic model that adequately describes tacrolimus pharmacokinetics in heart transplant recipients, considering the tacrolimus-azole antifungal interaction was developed. Prospective evaluation is required to assess its clinical utility to improve patient outcomes.

A Machine Learning Approach to Predict Interdose Vancomycin Exposure.

INTRODUCTION: Estimation of vancomycin area under the curve (AUC) is challenging in the case of discontinuous administration. Machine learning approaches are increasingly used and can be an alternative to population pharmacokinetic (POPPK) approaches for AUC estimation. The objectives were to train XGBoost algorithms based on simulations performed in a previous POPPK study to predict vancomycin AUC from early concentrations and a few features (i.e. patient information) and to evaluate them in a real-life external dataset in comparison to POPPK. PATIENTS AND METHODS: Six thousand simulations performed from 6 different POPPK models were split into training and test sets. XGBoost algorithms were trained to predict trapezoidal rule AUC a priori or based on 2, 4 or 6 samples and were evaluated by resampling in the training set and validated in the test set. Finally, the 2-sample algorithm was externally evaluated on 28 real patients and compared to a state-of-the-art POPPK model-based averaging approach. RESULTS: The trained algorithms showed excellent performances in the test set with relative mean prediction error (MPE)/ imprecision (RMSE) of the reference AUC = 3.3/18.9, 2.8/17.4, 1.3/13.7% for the 2, 4 and 6 samples algorithms respectively. Validation in real patient showed flexibility in sampling time post-treatment initiation and excellent performances MPE/RMSE<1.5/12% for the 2 samples algorithm in comparison to different POPPK approaches. CONCLUSIONS: The Xgboost algorithm trained from simulation and evaluated in real patients allow accurate and precise prediction of vancomycin AUC. It can be used in combination with POPPK models to increase the confidence in AUC estimation.

Dosing for Personalized Prophylaxis in Hemophilia A Highly Varies on the Underlying Population Pharmacokinetic Models.

BACKGROUND: Model-informed personalized prophylaxis with factor VIII (FVIII) replacement therapy aimed at higher trough levels is becoming indispensable for patients with severe hemophilia A. This study aimed to identify the most suitable population pharmacokinetic (PK) models for personalized prophylaxis using various FVIII products and 2 clinical assays and to implement the most suitable one in open-access software. METHODS: Twelve published population PK models were systematically compared to predict the time above target (TaT) for a reference dosing occasion. External validation was performed using a 5-point PK data from 39 adult patients with hemophilia A with FVIII measured by chromogenic substrate (CSA) and 1-stage assays (OSAs) using NONMEM under 3 different conditions: a priori (with all FVIII samples blinded), a posteriori (with 1 trough sample), and general model fit (with all FVIII samples including the reference dosing occasion provided). RESULTS: On average, the baseline covariate models overpredicted TaT (a priori; bias -3.8 hours to 49.6 hours). When additionally including 1 previous trough FVIII sample before the reference dosing occasion (a posteriori), only 50% of the models improved in bias (-1.0 hours to 36.5 hours) and imprecision (22.4 hours and 60.7 hours). Using all the time points (general model fit), the models accurately predicted (individual TaT less than ±12 hours compared with the reference) 62%-90% and 33%-74% of the patients using CSA and OSA data, respectively. Across all scenarios, predictions using CSA data were more accurate than those using the OSA data. CONCLUSIONS: One model performed best across the population (bias: -3.8 hours a priori, -1.0 hours a posteriori , and 0.6 hours general model fit ) and acceptably predicted 44% (a priori) to 90% ( general model fit ) of the patients. To allow the community-based evaluation of patient-individual FVIII dosing, this model was implemented in the open-access model-informed precision dosing software "TDMx."

Insights into Early Interactions on Innovative Developments with European Regulators.

INTRODUCTION: The European Medicines Agency Innovation Task Force (ITF) acts as early point of contact for medicine and technology developers to enable innovation during early drug development stages through ITF briefing meetings. AIM: To reflect on the current pace of innovation and to assess the potential of ITF stakeholder interactions, a comprehensive analysis of the ITF briefing meetings held between 2021 and 2022 was conducted with a focus on individual questions raised by the developers and the related feedback provided by the European regulators. METHODS: Questions raised during ITF briefing meetings were extracted and categorised into main and sub-categories, revealing different themes across the whole medicine development process such as manufacturing technologies, pre-clinical developments, and clinically relevant questions. RESULTS: There was positive feedback from regulators who gave initial guidance in 85% of the answers, provided concrete examples in 20% of the answers and recommended to continue discussions through additional regulatory procedures in 22% of the answers. CONCLUSION: This analysis frames the content and the type of topics discussed during ITF briefing meetings. Moreover, it describes the type of regulatory feedback provided to medicine developers and identified potential for improvement of these early interactions. Therefore, this analysis emphasises the role of ITF briefing meetings in fostering innovation in medicine.

Optimized sampling to estimate vancomycin drug exposure: Comparison of pharmacometric and equation-based approaches in a simulation-estimation study.

Vancomycin dosing should be accompanied by area under the concentration-time curve (AUC)-guided dosing using model-informed precision dosing software according to the latest guidelines. Although a peak plus a trough sample is considered the gold standard to determine the AUC, single-sample strategies might be more economic. Yet, optimal sampling times for AUC determination of vancomycin have not been systematically evaluated. In the present study, automated one- or two-sample strategies were systematically explored to estimate the AUC with a model averaging and a model selection algorithm. Both were compared with a conventional equation-based approach in a simulation-estimation study mimicking a heterogenous patient population (n = 6000). The optimal single-sample timepoints were identified between 2-6.5 h post dose, with varying bias values between -2.9% and 1.0% and an imprecision of 23.3%-24.0% across the population pharmacokinetic approaches. Adding a second sample between 4.5-6.0 h improved the predictive performance (-1.7% to 0.0% bias, 17.6%-18.6% imprecision), although the difference in the two-sampling strategies were minor. The equation-based approach was always positively biased and hence inferior to the population pharmacokinetic approaches. In conclusion, the approaches always preferred samples to be drawn early in the profile (<6.5 h), whereas sampling of trough concentrations resulted in a higher imprecision. Furthermore, optimal sampling during the early treatment phase could already give sufficient time to individualize the second dose, which is likely unfeasible using trough sampling.

Model-informed precision dosing of vancomycin via continuous infusion: a clinical fit-for-purpose evaluation of published PK models.

BACKGROUND: Model-informed precision dosing is an innovative approach used to guide bedside vancomycin dosing. The use of Bayesian software requires suitable and externally validated population pharmacokinetic (popPK) models. OBJECTIVES: This study aimed to identify suitable popPK models for a priori prediction and a posteriori forecasting of vancomycin in continuous infusion. Additionally, model averaging (MAA) and model selection approach (MSA) were compared with the identified popPK models. METHODS: Clinical pharmacokinetic data were retrospectively collected from patients receiving continuous vancomycin therapy and admitted to a general ward of three large Belgian hospitals. The predictive performance of the popPK models, identified in a systematic literature search, as well as the MAA/MSA were evaluated for the a priori and a posteriori scenarios using bias, root mean square errors, normalised prediction distribution errors and visual predictive checks. RESULTS: The predictive performance of 23 popPK models was evaluated based on clinical data from 169 patients and 923 therapeutic drug monitoring samples. Overall, the best predictive performance was found using the Okada et al. model (bias < -0.1 mg/L) followed by the Colin et al. MODEL: The MAA/MSA predicted with a constantly high precision and low inaccuracy and were clinically acceptable in the Bayesian forecasting. CONCLUSION: This study identified the two-compartmental models of Okada et al. and Colin et al. as most suitable for non-ICU patients to forecast individual exposure profiles after continuous vancomycin infusion. The MAA/MSA performed equally as well as the individual popPK models; therefore, both approaches could be used in clinical practice to guide dosing decisions.

A Model Averaging/Selection Approach Improves the Predictive Performance of Model-Informed Precision Dosing: Vancomycin as a Case Study.

Many important drugs exhibit substantial variability in pharmacokinetics and pharmacodynamics leading to a loss of the desired clinical outcomes or significant adverse effects. Forecasting drug exposures using pharmacometric models can improve individual target attainment when compared with conventional therapeutic drug monitoring (TDM). However, selecting the "correct" model for this model-informed precision dosing (MIPD) is challenging. We derived and evaluated a model selection algorithm (MSA) and a model averaging algorithm (MAA), which automates model selection and finds the best model or combination of models for each patient using vancomycin as a case study, and implemented both algorithms in the MIPD software "TDMx." The predictive performance (based on accuracy and precision) of the two algorithms was assessed in (i) a simulation study of six distinct populations and (ii) a clinical dataset of 180 patients undergoing TDM during vancomycin treatment and compared with the performance obtained using a single model. Throughout the six virtual populations the MSA and MAA (imprecision: 9.9-24.2%, inaccuracy: less than ± 8.2%) displayed more accurate predictions than the single models (imprecision: 8.9-51.1%; inaccuracy: up to 28.9%). In the clinical dataset, the predictive performance of the single models applying at least one plasma concentration varied substantially (imprecision: 28-62%, inaccuracy: -16 to 25%), whereas the MSA or MAA utilizing these models simultaneously resulted in unbiased and precise predictions (imprecision: 29% and 30%, inaccuracy: -5% and 0%, respectively). MSA and MAA approaches implemented in TDMx might thereby lower the burden of fit-for-purpose validation of individual models and streamline MIPD.