A machine learning approach to predict daptomycin exposure from two concentrations based on Monte Carlo simulations.

Daptomycin is a concentration-dependent lipopeptide antibiotic for which exposure/effect relationships have been shown. Machine learning (ML) algorithms, developed to predict the individual exposure to drugs, have shown very good performances in comparison to maximum a posteriori Bayesian estimation (MAP-BE). The aim of this work was to predict the area under the blood concentration curve (AUC) of daptomycin from two samples and a few covariates using XGBoost ML algorithm trained on Monte Carlo simulations. Five thousand one hundred fifty patients were simulated from two literature population pharmacokinetics models. Data from the first model were split into a training set (75%) and a testing set (25%). Four ML algorithms were built to learn AUC based on daptomycin blood concentration samples at pre-dose and 1 h post-dose. The XGBoost model (best ML algorithm) with the lowest root mean square error (RMSE) in a 10-fold cross-validation experiment was evaluated in both the test set and the simulations from the second population pharmacokinetic model (validation). The ML model based on the two concentrations, the differences between these concentrations, and five other covariates (sex, weight, daptomycin dose, creatinine clearance, and body temperature) yielded very good AUC estimation in the test (relative bias/RMSE = 0.43/7.69%) and validation sets (relative bias/RMSE = 4.61/6.63%). The XGBoost ML model developed allowed accurate estimation of daptomycin AUC using C0, C1h, and a few covariates and could be used for exposure estimation and dose adjustment. This ML approach can facilitate the conduct of future therapeutic drug monitoring (TDM) studies.

In Silico Pharmacokinetics Evaluation of Forgiveness for Doravirine and Rilpivirine.

BACKGROUND: This study aimed to evaluate the concentrations of rilpivirine (RLP) and doravirine (DOR) after 3 days-off using simulations from population pharmacokinetics models. METHODS: The authors conducted a series of 500 sets of 10,000 Monte Carlo simulations to examine the steady-state conditions for 2 common dosage levels: 25 mg/d for RLP and 100 mg/d for DOR. These simulations were conducted under 2 scenarios: 1 without drug cessation and another after a 3-day break. The validity of the implementation was established through a comparison of median trough concentrations (C24h) with previously reported data. Subsequently, the proportion of simulated patients with C24h and C72h after 3 days-off (C72h/3do) that exceeded the inhibitory concentration 50 (IC50), 5.2 mcg/L for DOR and 20.5 mcg/L for RLP respectively, was calculated. The inhibitory quotient (IQ) was also computed, which was 6 times IC50 for DOR and 4.5 times IC50 for RLP. Finally, nomograms were constructed to estimate the probability of having C72h/3do > IC50 or > IQ for different ranges of C24h. RESULTS: Simulated C24h median ± SD for RLP were 61.8 ± 0.4 mcg/L and for DOR 397 ± 0 mcg/L. For RLP, 99.3 ± 0.1% exceeded IC50 at C24h, 16.4 ± 0.4% at C72h/3do, and none surpassed the IQ threshold. In contrast, DOR had 100% ± 0% above IC50 at C24h, 93.6 ± 0.2% at C72h/3do, and 58.6 ± 0.5% exceeded the IQ. CONCLUSIONS: These findings suggest that treatment with DOR may offer a more forgiving therapeutic profile than RLP, given the larger proportion of patients achieving effective drug exposure with DOR. However, it is important to acknowledge a significant limitation of this study, namely, the assumption that drug concentration is a perfect surrogate for drug effectiveness.

Population Pharmacokinetics and Limited Sampling Strategy of Mycophenolate Mofetil for Indian Patients with Lupus Nephritis.

BACKGROUND: Mycophenolic acid is widely used to treat lupus nephritis (LN). However, it exhibits complex pharmacokinetics with large interindividual variability. This study aimed to develop a population pharmacokinetic (popPK) model and a 3-sample limited sampling strategy (LSS) to optimize therapeutic drug monitoring in Indian patients with LN. METHODS: Five blood samples from each LN patient treated with mycophenolic acid were collected at steady-state predose and 1, 2, 4, and 6 hours postdose. Demographic parameters were tested as covariates to explain interindividual variability. PopPK analysis was performed using Monolix and the stochastic approximation expectation-maximization algorithm. An LSS was derived from 500 simulated pharmacokinetic (PK) profiles using maximum a posteriori Bayesian estimation to estimate individual PK parameters and area under the curve (AUC). The LSS-calculated AUC was compared with the AUC calculated using the trapezoidal rule and all the simulated samples. RESULTS: A total of 51 patients were included in this study. Based on the 245 mycophenolic acid concentrations, a 1-compartmental model with double absorption using gamma distributions best fitted the data. None of the covariates improved the model significantly. The model was internally validated using diagnostic plots, prediction-corrected visual predictive checks, and bootstrapping. The best LSS included samples at 1, 2, and 4 hours postdose and exhibited good performances in an external dataset (root mean squared error, 21.9%; mean bias, -4.20%). CONCLUSIONS: The popPK model developed in this study adequately estimated the PK of mycophenolic acid in adult Indian patients with LN. This simple LSS can optimize TDM based on the AUC in routine practice.

Limited sampling strategy and population pharmacokinetic model of mycophenolic acid in pediatric patients with systemic lupus erythematosus: application of a double gamma absorption model with SAEM algorithm.

INTRODUCTION: Mycophenolic acid (MPA), the active metabolite of mycophenolate mofetil (MMF), is widely used in the treatment of systemic lupus erythematosus (SLE). It has been shown that its therapeutic drug monitoring based on the area under the curve (AUC) improves treatment efficacy. MPA exhibits a complex bimodal absorption, and a double gamma distribution model has been already proposed in the past to accurately describe this phenomenon. These previous population pharmacokinetics models (POPPK) have been developed using iterative two stage Bayesian (IT2B) or non-parametric adaptive grid (NPAG) methods. However, non-linear mixed effect (NLME) approaches based on stochastic approximation expectation-maximization (SAEM) algorithms have never been published so far for this particular model. The objectives of this study were (i) to implement the double absorption gamma model in Monolix, (ii) to compare different absorption models to describe the pharmacokinetics of MMF, and (iii) to develop a limited sampling strategy (LSS) to estimate AUC in pediatric SLE patients. MATERIAL AND METHODS: A data splitting of full pharmacokinetic profiles sampled in 67 children extracted either from the expert system ISBA (n = 34) or the hospital Saint Louis (n = 33) was performed into train (75%) and test (25%) sets. A POPPK was developed for MPA in the train set using a NLME and the SAEM algorithm and different absorption models were implemented and compared (first order, transit, or simple and double gamma). The best limited sampling strategy was then determined in the test set using a maximum-a-posteriori Bayesian method to estimate individual PK parameters and AUC based on three blood samples compared to the reference AUC calculated using the trapezoidal rule applied on all samples and performances were assessed in the test set. RESULTS: Mean patient age and dose was 13 years old (5-18) and 18.1 mg/kg (7.9-47.6), respectively. MPA concentrations (764) from 107 occasions were included in the analysis. A double gamma absorption with a first-order elimination from the central compartment best fitted the data. The optimal LSS with samples at 30 min, 2 h, and 3 h post-dose exhibited good performances in the test set (mean bias - 0.32% and RMSE 21.0%). CONCLUSION: The POPPK developed in this study adequately estimated the MPA AUC in pediatric patients with SLE based on three samples. The double absorption gamma model developed with the SAEM algorithm showed very accurate fit and reduced computation time.

Synergizing pharmacometrics and pharmacovigilance for medication error management: the case of secukinumab.

PURPOSE: To demonstrate the effective integration of pharmacometrics and pharmacovigilance in managing medication errors, highlighted by a case involving secukinumab in a patient with hidradenitis suppurativa. METHODS: We present the case of a 41-year-old male with progressive hidradenitis suppurativa, unresponsive to multiple antibiotic regimens and infliximab treatment. Due to a medication error, the patient received 300 mg of secukinumab daily for 4 days instead of weekly, totaling 1200 mg. The regional pharmacovigilance center assessed potential toxicity, and a pharmacometric analysis using a population pharmacokinetic model was performed to inform dosing adjustments. RESULTS: Clinical data indicated that the received doses were within a non-toxic range. No adverse effects were observed. Pharmacometric simulations revealed a risk of underexposure due to the dosing error. Based on these simulations, it was recommended to restart monthly secukinumab injections on day 35 after the initial dose. Measured plasma concentrations before re-administration confirmed the model's accuracy. CONCLUSION: This case highlights the crucial collaboration between clinical services, pharmacovigilance, and pharmacometrics in managing medication errors. Such interdisciplinary efforts ensure therapeutic efficacy and patient safety by maintaining appropriate drug exposure levels.

Tacrolimus population pharmacokinetics in adult heart transplant patients.

INTRODUCTION: Tacrolimus is an immunosuppressant largely used in heart transplantation. However, the calculation of its exposure based on the area under the curve (AUC) requires the use of a population pharmacokinetic (PK) model. The aims of this work were (i) to develop a population PK model for tacrolimus in heart transplant patients, (ii) to derive a maximum a posteriori Bayesian estimator (MAP-BE) based on a limited sampling strategy (LSS) and (iii) to estimate probabilities of target attainment (PTAs) for AUC and trough concentration (C0). MATERIAL AND METHODS: Forty-seven PK profiles (546 concentrations) of 18 heart transplant patients of the Pharmacocinétique des Immunosuppresseurs chez les patients GREffés Cardiaques study receiving tacrolimus (Prograf®) were included. The database was split into a development (80%) and a validation (20%) set. PK parameters were estimated in MONOLIX® and based on this model a Bayesian estimator using an LSS was built. Simulations were performed to calculate the PTA for AUC and C0. RESULTS: The best model to describe the tacrolimus PK was a two-compartment model with a transit absorption and a linear elimination. Only the CYP3A5 covariate was kept in the final model. The derived MAP-BE based on the LSS (0-1-2 h postdose) yielded an AUC bias ± SD = 2.7 ± 10.2% and an imprecision of 9.9% in comparison to the reference AUC calculated using the trapezoidal rule. PTAs based on AUC or C0 allowed new recommendations to be proposed for starting doses (0.11 mg·kg-1 ·12 h-1 for the CYP3A5 nonexpressor and 0.22 mg·kg1 ·12 h-1 for the CYP3A5 expressor). CONCLUSION: The MAP-BE developed should facilitate estimation of tacrolimus AUC in heart transplant patients.

A Hybrid Algorithm Combining Population Pharmacokinetic and Machine Learning for Isavuconazole Exposure Prediction.

OBJECTIVES: Maximum a posteriori Bayesian estimation (MAP-BE) based on a limited sampling strategy and a population pharmacokinetic (POPPK) model is used to estimate individual pharmacokinetic parameters. Recently, we proposed a methodology that combined population pharmacokinetic and machine learning (ML) to decrease the bias and imprecision in individual iohexol clearance prediction. The aim of this study was to confirm the previous results by developing a hybrid algorithm combining POPPK, MAP-BE and ML that accurately predicts isavuconazole clearance. METHODS: A total of 1727 isavuconazole rich PK profiles were simulated using a POPPK model from the literature, and MAP-BE was used to estimate the clearance based on: (i) the full PK profiles (refCL); and (ii) C24h only (C24h-CL). Xgboost was trained to correct the error between refCL and C24h-CL in the training dataset (75%). C24h-CL as well as ML-corrected C24h-CL were evaluated in a testing dataset (25%) and then in a set of PK profiles simulated using another published POPPK model. RESULTS: A strong decrease in mean predictive error (MPE%), imprecision (RMSE%) and the number of profiles outside ± 20% MPE% (n-out20%) was observed with the hybrid algorithm (decreased in MPE% by 95.8% and 85.6%; RMSE% by 69.5% and 69.0%; n-out20% by 97.4% and 100% in the training and testing sets, respectively. In the external validation set, the hybrid algorithm decreased MPE% by 96%, RMSE% by 68% and n-out20% by 100%. CONCLUSION: The hybrid model proposed significantly improved isavuconazole AUC estimation over MAP-BE based on the sole C24h and may improve dose adjustment.

Longitudinal Exposure to Tacrolimus and New-Onset Diabetes Mellitus in Renal Transplant Patients.

PURPOSE: Tacrolimus is an immunosuppressant widely used in transplantations requiring mandatory concentration-controlled dosing to prevent acute rejection or adverse effects, including new-onset diabetes mellitus (NODM). However, no relationship between NODM and tacrolimus exposure has been established. This study aimed to evaluate the relationship between cumulative tacrolimus exposure and NODM occurrence. METHODS: A total of 452 kidney transplant patients were included in this study. Sixteen patients developed NODM during the first 3 months after transplant. We considered all tacrolimus concentration (C0) values collected until the diagnosis of NODM in these patients and until 3 months after transplant in the others. New tacrolimus cumulative exposure metrics were derived from the time profile of the tacrolimus morning predose concentration, C0: the percentage of C0 values > cutoff, the average of C0 values above the cutoff, and the percentage of the area under C0 versus time curve, AUCC0, above the cutoff. The cutoff chosen was 15 ng/mL, corresponding to the higher end of the therapeutic range for the early post-transplant period. The influence of these metrics on NODM and other clinical and biological characteristics was investigated using the Cox models. RESULTS: The percentage of C0 > 15 mcg/L was statistically different between patients with and without NODM (P = 0.01). Only these tacrolimus C0-derived metrics were significantly associated with an increased risk of NODM [HR: 1.73 (1.43-2.10, P < 0.001)]. CONCLUSION: This study shows that tacrolimus concentrations >15 mcg/L affect the incidence of NODM.

Mycophenolate Mofetil Dose Adjustment in Pediatric Kidney Transplant Recipients.

BACKGROUND: The Immunosuppressant Bayesian Dose Adjustment web site aids clinicians and pharmacologists involved in the care of transplant recipients; it proposes dose adjustments based on the estimated area under the concentration-time curve (AUCs). Three concentrations (T 20 min , T 1 h , and T 3 h ) are sufficient to estimate mycophenolic acid (MPA) AUC 0-12 h in pediatric kidney transplant recipients. This study investigates mycophenolate mofetil (MMF) doses and MPA AUC values in pediatric kidney transplant recipients, and target exposure attainment when the proposed doses were followed, through a large-scale analysis of the data set collated since the inception of the Immunosuppressant Bayesian Dose Adjustment web site. METHODS: In this study, 4051 MMF dose adjustment requests, corresponding to 1051 patients aged 0-18 years, were retrospectively analyzed. AUC calculations were performed in the back office of the Immunosuppressant Bayesian Dose Adjustment using published Bayesian and population pharmacokinetic models. RESULTS: The first AUC request was posted >12 months posttransplantation for 41% of patients. Overall, only 50% had the first MPA AUC 0-12 h within the recommended 30-60 mg.h/L range. When the proposed dose was not followed, the proportion of patients with an AUC in the therapeutic range for MMF with cyclosporine or tacrolimus at the subsequent request was lower (40% and 45%, respectively) than when it was followed (58% and 60%, respectively): P = 0.08 and 0.006, respectively. Furthermore, 3 months posttransplantation, the dispersion of AUC values was often lower at the second visit when the proposed doses were followed, namely, P = 0.03, 0.003, and 0.07 in the 4 months-1 year, and beyond 1 year with <6-month or >6-month periods between both visits, respectively. CONCLUSIONS: Owing to extreme interindividual variability in MPA exposure, MMF dose adjustment is necessary; it is efficient at reducing such variability when based on MPA AUC.

Doravirine Exposure Decreased by Dialysis in a HIV Patient: A Grand Round.

BACKGROUND: The authors report the case of a 66-year-old male patient who was hemodialyzed 3 times per week for chronic renal failure and treated with 100 mg of doravirine once daily in combination with dolutegravir for HIV-1. No dose adjustment is required for doravirine in cases of severe renal injury, but the effect of dialysis on its exposure is poorly understood. METHODS RESULTS: Two series of 2 samples were drawn before and after 4-hour hemodialysis and showed an average doravirine concentration decrease of 48.1 ± 6.7%. The effects of hemodialysis were important, contrary to what was expected and has been previously reported. In addition, intraindividual variability was low. Nevertheless, because the concentrations reported were largely above the inhibitory concentration 50 (IC 50 ), no dose adjustment was required. CONCLUSIONS: The decrease in doravirine concentration due to hemodialysis observed in this case report was quite significant. Therefore, therapeutic drug monitoring might be recommended in certain patients undergoing doravirine treatment also on hemodialysis.

Tacrolimus Exposure Before and After a Switch From Twice-Daily Immediate-Release to Once-Daily Prolonged Release Tacrolimus: The ENVARSWITCH Study.

LCP-tacrolimus displays enhanced oral bioavailability compared to immediate-release (IR-) tacrolimus. The ENVARSWITCH study aimed to compare tacrolimus AUC0-24 h in stable kidney (KTR) and liver transplant recipients (LTR) on IR-tacrolimus converted to LCP-tacrolimus, in order to re-evaluate the 1:0.7 dose ratio recommended in the context of a switch and the efficiency of the subsequent dose adjustment. Tacrolimus AUC0-24 h was obtained by Bayesian estimation based on three concentrations measured in dried blood spots before (V2), after the switch (V3), and after LCP-tacrolimus dose adjustment intended to reach the pre-switch AUC0-24 h (V4). AUC0-24 h estimates and distributions were compared using the bioequivalence rule for narrow therapeutic range drugs (Westlake 90% CI within 0.90-1.11). Fifty-three KTR and 48 LTR completed the study with no major deviation. AUC0-24 h bioequivalence was met in the entire population and in KTR between V2 and V4 and between V2 and V3. In LTR, the Westlake 90% CI was close to the acceptance limits between V2 and V4 (90% CI = [0.96-1.14]) and between V2 and V3 (90% CI = [0.96-1.15]). The 1:0.7 dose ratio is convenient for KTR but may be adjusted individually for LTR. The combination of DBS and Bayesian estimation for tacrolimus dose adjustment may help with reaching appropriate exposure to tacrolimus rapidly after a switch.

Diagnostic, prognostic and clinical value of left ventricular radial strain to identify paradoxical septal motion in ventilated patients with the acute respiratory distress syndrome: an observational prospective multicenter study.

BACKGROUND: Acute cor pulmonale (ACP) is prognostic in patients with acute respiratory distress syndrome (ARDS). Identification of paradoxical septal motion (PSM) using two-dimensional echocardiography is highly subjective. We sought to describe feature-engineered metrics derived from LV radial strain changes related to PSM in ARDS patients with ACP of various severity and to illustrate potential diagnostic and prognostic yield. METHODS: This prospective bicentric study included patients under protective ventilation for ARDS related to COVID-19 who were assessed using transesophageal echocardiography (TEE). Transgastric short-axis view at mid-papillary level was used to visually grade septal motion, using two-dimensional imaging, solely and combined with LV radial strain: normal (grade 0), transient end-systolic septal flattening (grade 1), prolonged end-systolic septal flattening or reversed septal curvature (grade 2). Inter-observer variability was calculated. Feature engineering was performed to calculate the time-to-peak and area under the strain curve in 6 LV segments. In the subset of patients with serial TEE examinations, a multivariate Cox model analysis accounting for new-onset of PSM as a time-dependent variable was used to identify parameters associated with ICU mortality. RESULTS: Overall, 310 TEE examinations performed in 182 patients were analyzed (age: 67 [60-72] years; men: 66%; SAPSII: 35 [29-40]). Two-dimensional assessment identified a grade 1 and grade 2 PSM in 100 (32%) and 48 (15%) examinations, respectively. Inter-rater reliability was weak using two-dimensional imaging alone (kappa = 0.49; 95% CI 0.40-0.58; p < 0.001) and increased with associated LV radial strain (kappa = 0.84, 95% CI 0.79-0.90, p < 0.001). The time-to-peak of mid-septal and mid-lateral segments occurred significantly later in systole and increased with the grade of PSM. Similarly, the area under the strain curve of these segments increased significantly with the grade of PSM, compared with mid-anterior or mid-inferior segments. Severe acute cor pulmonale with a grade 2 PSM was significantly associated with mortality. Requalification in an upper PSM grade using LV radial strain allowed to better identify patients at risk of death (HR: 6.27 [95% CI 2.28-17.2] vs. 2.80 [95% CI 1.11-7.09]). CONCLUSIONS: In objectively depicting PSM and quantitatively assessing its severity, TEE LV radial strain appears as a valuable adjunct to conventional two-dimensional imaging.

Within-subject pharmacokinetic variability has a strong influence on individual exposure ratios in bioequivalence studies, hence on drug formulation interchangeability.

INTRODUCTION: Bioequivalence between a reference and a generic drug is based on the hypothesis that a ± 20% change in blood exposure (or ± 10% for drugs with narrow therapeutic index, NTI) following the generic/reference switch will not have any therapeutic consequences. However, the individual exposure ratio between generic and reference can be higher than 1.20 (or 1.10). This study aims to analyse the different parameters influencing the individual exposure ratio, hence the conditions for reference/generic interchangeability. METHODS: Bioequivalence studies with a double cross-over design for a virtual drug were simulated using 100 random sets of 12, 24, 48 or 100 pairs of areas under the curve (AUC), varying the generic/reference AUC geometric mean ratios between 0.80 and 1.25 and the within-subject exposure variance of the reference and the generic formulations. RESULTS: The proportion of subjects with an exposure generic/reference ratio outside the ± 10% or ± 20% acceptance intervals increases when (1) the reference within-subject variance increases; (2) the ratio of the generic within-subject variance on the reference within-subject variance increases; and (3) the generic/reference mean AUC ratio diverges from 1.0. When only considering replicated administrations of the reference, the individual exposure ratio increases with the within-subject variance, yielding values outside the usually accepted individual exposure ratio range of 0.5 to 2 for drugs with narrow therapeutic index as soon as the within-subject variance standard deviation is ≥ 0.25 (equivalent to within-patient CV% > 25%). CONCLUSIONS: Interchangeability between reference and generic formulations, especially for drugs with narrow therapeutic index can only be assumed if, the within-subject variance of generic is less or equal to the within-subject variance of reference or, if this is not the case, if the distribution of the generic/generic individual exposure ratios is included within the therapeutic margins of the reference drug.

Application of machine learning to predict tacrolimus exposure in liver and kidney transplant patients given the MeltDose formulation.

PURPOSE: Machine Learning (ML) algorithms represent an interesting alternative to maximum a posteriori Bayesian estimators (MAP-BE) for tacrolimus AUC estimation, but it is not known if training an ML model using a lower number of full pharmacokinetic (PK) profiles (= "true" reference AUC) provides better performances than using a larger dataset of less accurate AUC estimates. The objectives of this study were: to develop and benchmark ML algorithms trained using full PK profiles to estimate MeltDose®-tacrolimus individual AUCs using 2 or 3 blood concentrations; and to compare their performance to MAP-BE. METHODS: Data from liver (n = 113) and kidney (n = 97) transplant recipients involved in MeltDose-tacrolimus PK studies were used for the training and evaluation of ML algorithms. "True" AUC0-24 h was calculated for each patient using the trapezoidal rule on the full PK profile. ML algorithms were trained to estimate tacrolimus true AUC using 2 or 3 blood concentrations. Performances were evaluated in 2 external sets of 16 (renal) and 48 (liver) transplant patients. RESULTS: Best estimation performances were obtained with the MARS algorithm and the following limited sampling strategies (LSS): predose (0), 8, and 12 h post-dose (rMPE = - 1.28%, rRMSE = 7.57%), or 0 and 12 h (rMPE = - 1.9%, rRMSE = 10.06%). In the external dataset, the performances of the final ML algorithms based on two samples in kidney (rMPE = - 3.1%, rRMSE = 11.1%) or liver transplant recipients (rMPE = - 3.4%, rRMSE = 9.86%) were as good as or better than those of MAP-BEs based on three time points. CONCLUSION: The MARS ML models developed using "true" MeltDose®-tacrolimus AUCs yielded accurate individual estimations using only two blood concentrations.

Machine learning-supported interpretation of kidney graft elementary lesions in combination with clinical data.

Interpretation of kidney graft biopsies using the Banff classification is still heterogeneous. In this study, extreme gradient boosting classifiers learned from two large training datasets (n = 631 and 304 cases) where the "reference diagnoses" were not strictly defined following the Banff rules but from central reading by expert pathologists and further interpreted consensually by experienced transplant nephrologists, in light of the clinical context. In three external validation datasets (n = 3744, 589, and 360), the classifiers yielded a mean ROC curve AUC (95%CI) of: 0.97 (0.92-1.00), 0.97 (0.96-0.97), and 0.95 (0.93-0.97) for antibody-mediated rejection (ABMR); 0.94 (0.91-0.96), 0.94 (0.92-0.95), and 0.91 (0.88-0.95) for T cell-mediated rejection; >0.96 (0.90-1.00) with all three for interstitial fibrosis-tubular atrophy. We also developed a classifier to discriminate active and chronic active ABMR with 95% accuracy. In conclusion, we built highly sensitive and specific artificial intelligence classifiers able to interpret kidney graft scoring together with a few clinical data and automatically diagnose rejection, with excellent concordance with the Banff rules and reference diagnoses made by a group of experts. Some discrepancies may point toward possible improvements that could be made to the Banff classification.

Population pharmacokinetics and Bayesian estimation of mycophenolate mofetil in patients with autoimmune hepatitis.

AIMS: Mycophenolate mofetil (MMF) is the most widely used second-line agent in autoimmune hepatitis (AIH). Individual dose adjustment of MMF may avoid adverse outcomes while maximizing efficacy. The aim of the present study was to develop population pharmacokinetic (popPK) models and maximum a posteriori Bayesian estimators (MAP-BEs) to estimate mycophenolic acid interdose area under the curve in AIH patients administered MMF using nonlinear mixed effect modelling. METHODS: We analysed 50 mycophenolic acid PK profiles from 34 different patients, together with some demographic, clinical, and laboratory test data. The median number of plasma samples per profile, immediately preceding and following the morning MMF dose, was 7. PopPK modelling was performed using parametric, top-down, nonlinear mixed effect modelling with NONMEM 7.3. MAP-BEs were developed based on the best popPK model and the best limited sampling strategy selected among several. RESULTS: The pharmacokinetic data were best described by a 2-compartment model, Erlang distribution to describe the absorption phase, and a proportional error. The mean (relative standard error) of popPK parameter estimates of clearance, intercompartmental clearance, central volume and absorption rate with the final model were: 21.6 L h-1 (11%), 22.7 L h-1 (19%), 35.9 L (21%) and 8.7 h-1 (9%), respectively. The peripheral volume was fixed to 300 L. The best MAP-BE relied on the limited sampling strategy at 0.33, 1 and 3 hours after MMF dose administration and was very accurate (bias = 5.6%) and precise (root mean squared prediction error <20%). CONCLUSION: The precise and accurate Bayesian estimator developed in this study for AIH patients on MMF can be used to improve the therapeutic management of these patients.

A single Bayesian estimator for iohexol clearance estimation in ICU, liver failure and renal transplant patients.

AIMS: Iohexol clearance has been proposed to estimate the glomerular filtration rate (GFR). A population pharmacokinetics (popPK) model was developed from heterogeneous patients. A Bayesian estimator (MAP-BE) based on a limited sampling strategy (LSS) was derived and evaluated in external patients. METHODS: Full pharmacokinetic data (7-12 samples) from 172 patients receiving iohexol for measurement of their GFR (unstable and stable ICU patients, liver failure patients and kidney transplant patients) were split into development (n = 136) and validation (n = 36) datasets. A PopPK model was developed in Monolix and was used to develop MAP-BE based on LSS. Its performance for GFR estimation was evaluated in the validation set. RESULTS: A two-compartment model with first-order elimination best described the data. The final model included the type of patients on volume of distribution (Vd), clearance and intercompartmental constants, serum creatinine on clearance and body weight on Vd. The best LSS included samples at 0.1-1-9 h exhibiting a relative mean prediction error (MPE) (RMSE) = -3.7% (14.3%) and better performance than the Bröchner-Mortensen formula (-3.0%/17%). Split by type of patients, the highest interindividual variability and imprecision was observed in unstable ICU patients (MPE (RMSE) = 3.7% (18.8%)) while the best performances were obtained for renal transplant patients (MPE (RMSE) = 1.0% (5.8%)). All LSS that included samples before 9 hours for the third sample were associated with an increased imprecision. CONCLUSION: A single MAP-BE of iohexol based on a three-sample LSS for four heterogeneous populations was developed and allowed accurate estimation of GFR in kidney transplant patients, slightly biased in stable ICU patients and slightly imprecise in unstable ICU patients.

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.

Optimization of Vancomycin Initial Dose in Term and Preterm Neonates by Machine Learning.

INTRODUCTION: Vancomycin is one of the antibiotics most used in neonates. Continuous infusion has many advantages over intermittent infusions, but no consensus has been achieved regarding the optimal initial dose. The objectives of this study were: to develop a Machine learning (ML) algorithm based on pharmacokinetic profiles obtained by Monte Carlo simulations using a population pharmacokinetic model (POPPK) from the literature, in order to derive the best vancomycin initial dose in preterm and term neonates, and to compare ML performances with those of an literature equation (LE) derived from a POPPK previously published. MATERIALS AND METHODS: The parameters of a previously published POPPK model of vancomycin in children and neonates were used in the mrgsolve R package to simulate 1900 PK profiles. ML algorithms were developed from these simulations using Xgboost, GLMNET and MARS in parallel, benchmarked and used to calculate the ML first dose. Performances were evaluated in a second simulation set and in an external set of 82 real patients and compared to those of a LE. RESULTS: The Xgboost algorithm yielded numerically best performances and target attainment rates: 46.9% in the second simulation set of 400-600 AUC/MIC ratio vs. 41.4% for the LE model (p = 0.0018); and 35.3% vs. 28% in real patients (p = 0.401), respectively). The Xgboost model resulted in less AUC/MIC > 600, thus decreasing the risk of nephrotoxicity. CONCLUSION: The Xgboost algorithm developed to estimate the initial dose of vancomycin in term or preterm infants has better performances than a previous validated LE and should be evaluated prospectively.

Pharmacokinetics of mycophenolic acid and external evaluation of two limited sampling strategies of drug exposure in patients with juvenile systematic lupus erythematosus.

INTRODUCTION: Mycophenolate mofetil (MMF), a pro-drug of mycophenolic acid (MPA), has become a major therapeutic option in juvenile systemic lupus erythematosus (jSLE). Monitoring MPA exposure using area under curve (AUC) has proved its value to increase efficacy and safety in solid organ transplantation both in children and adults, but additional data are required in patients with autoimmune diseases. In order to facilitate MMF therapeutic drug monitoring (TDM) in children, Bayesian estimators (BE) of MPA AUC0-12 h using limited sampling strategies (LSS) have been developed. Our aim was to conduct an external validation of these LSS using rich pharmacokinetics and compare their predictive performance. METHODS: Pharmacokinetic blood samples were collected from jSLE treated by MMF and MPA plasma concentrations were determined using high-performance liquid chromatography system with ultraviolet detection (HPLC-UV). Individual AUC0-12 h at steady state was calculated using the trapezoid rule and compared with two LSS: (1) ISBA, a two-stage Bayesian approach developed for jSLE and (2) ADAPT, a non-linear mixed effects model with a parametric maximum likelihood approach developed with data from renal transplanted adults. RESULTS: We received 41 rich pediatric PK at steady state from jSLE and calculated individual AUC0-12 h. The external validation MPA AUC0-12 h was conducted by selecting the concentration-time points adapted to ISBA and ADAPT: (1) ISBA showed good accuracy (bias: - 0.8 mg h/L), (2) ADAPT resulted in a bias of 6.7 mg L/h. The corresponding relative root mean square prediction error (RSME) was 23% and 43% respectively. CONCLUSION: According to our external validation of two LSS of drug exposure, the ISBA model is recommended for Bayesian estimation of MPA AUC0-12 h in jSLE. In the literature focusing on MMF TDM, an efficacy cut-off for MPA AUC0-12 h between 30 and 45 mg h/L is proposed in jSLE but this requires additional validation.