Search bioRxiv⌕ Search

Biology subjects

Krekounian, O.

Publications and source records attributed to Krekounian, O..

2 recordsLinked to original sources
bioRxiv2026

Replacing In Vivo Experiments for PK/PD Target Determination Through In Vitro Time-Kill Experiments and PK/PD Modelling Incorporating Inter-strain Variability: Application to Meropenem Against Pseudomonas aeruginosa

Background. Optimal antibiotic dosing regimens depend on the pharmacokinetic/pharmacodynamic (PK/PD) index that best predicts antibacterial efficacy. PK/PD targets are traditionally determined using murine infection models based on a limited number of bacterial isolates. Objective. This study aimed to investigate whether animal experiments could be replaced by in vitro time-kill experiments performed on a large collection of clinical isolates and analyzed using a modelling approach accounting for inter-strain variability. The proposed framework was evaluated using meropenem against Pseudomonas aeruginosa. Materials and Methods. In vitro time-kill experiments were performed on 66 clinical isolates of P. aeruginosa. A population pharmacodynamic model was developed from experimental data. A murine pharmacokinetic model was reproduced from literature and combined with the pharmacodynamic model to simulate in vivo bacterial burden over time. The relationships between simulated bacterial counts at 24 h and the three main PK/PD indices (fCmax/MIC, fAUC/MIC and %fT>MIC) were characterized using nonlinear mixed-effects Imax models. Results. The PK/PD index showing the strongest correlation with meropenem efficacy at 24 h was %fT>MIC (R2 = 0.989), compared with fAUC/MIC (R2 = 0.373) and fCmax/MIC (R2 = 0.284). These findings are consistent with previous studies using murine thigh infection models. The %fT>MIC target required to achieve a 2-log CFU reduction was estimated at 44%, with substantial inter-strain variability (10th and 90th percentiles: 27% and 71%, respectively). Conclusions. Using meropenem against P. aeruginosa as a proof of concept, we demonstrate that in vitro time-kill experiments combined with pharmacometric modelling can identify the same PK/PD efficacy targets as animal infection models. Moreover, performing experiments on a large panel of clinical isolates enables the quantification of inter-strain variability in PK/PD targets, providing information that may improve their translation to clinical dosing optimization.

pharmacology and toxicology↗
bioRxiv2026

Model Ensembling and Machine Learning Approaches to Predict the First Dose of Amoxicillin in Intensive Care

A priori model informed precision dosing (MIPD) recommends an appropriate first dose based solely on the patients covariates enabling faster target attainment without required concentration measurements. Population pharmacokinetic model ensembling and machine learning (ML) approaches were developed and evaluated to predict a first dose of amoxicillin in intensive care. Following a bibliographic review, a virtual patient population was simulated based on cohorts from four published adult amoxicillin PopPK models. Model-development cohorts were reproduced, and steady-state trough concentrations were simulated using cohort-specific dosing regimens. As reference methods, weighted model ensembling (WME) and classification tree (CT)-informed ensembling were implemented. Two novel ensembling strategies were developed: regression tree (RT)-informed ensembling, using RT to predict the log individual prediction/observation ratio, and factor analysis of mixed data (FAMD), assigning model weights based on patient similarity to original model cohorts. In parallel, four ML algorithms (support vector machine, k-nearest neighbors, random forest, and XGBoost) were trained to predict the dose achieving target concentrations based on covariates and dosing scheme. All approaches were compared with single-model PopPK dosing, standard dosing, and a nomogram, and externally validated using clinical data. Most MIPD methods outperformed standard dosing. On simulated data, ensembling (30-42 % correct predictions) and ML (36-39 %) exceeded single-model approaches (14-32 %). RT-informed and FAMD ensembling improved performance by 6-10 % over uninformed ensembling on clinical data. In clinical patients receiving continuous infusion, ensembling further improved performance, with FAMD achieving 49 % correct predictions. ML-based ensembling eliminates the need for model selection and increase target attainment.

pharmacology and toxicology↗