Population Pharmacokinetics Research Articles

Population pharmacokinetics of oxcarbazepine active metabolite in Chinese paediatric patients with epilepsy: Model-based dose optimization.

The pharmacological activity of oxcarbazepine (OXC) is primarily exerted through its active 10-monohydroxy metabolite (MHD). Nonetheless, there is limited pharmacokinetic information available regarding paediatric patients with epilepsy treated with OXC, especially in infants and toddlers. Concurrently, this drug exhibits substantial variability in pharmacokinetics and therapeutic response across different individuals. We aimed to develop a model to quantitatively investigate factors that affect MHD pharmacokinetics to formulate a dosage guideline for OXC in Chinese paediatric patients. A total of 297 MHD trough concentrations were obtained from 287 epileptic children. Six body weight (BW)-based allometric models were used for population pharmacokinetic modelling, while investigating the impact of other covariates on the apparent clearance. The one-compartment model and age cut-off model for the apparent clearance (CL/F) were established to describe the pharmacokinetics of MHD. The probability to obtain target trough concentration ranges (TTCRs) of MHD between 3 and 35 mg/L was determined by Monte Carlo simulations for doses ranging from 8 to 90 mg/kg/day. A new dose optimization strategy combining the dosage guidelines and Bayesian method provides a tailored approach for Chinese paediatric epileptic patients based on their individual BW and desired TTCRs of MHD, and also supports current dose recommendations, with the exception of children weighing ≤5kg.

Towards optimizing cefepime/tazobactam (WCK 4282) exposure to achieve efficacy against piperacillin/tazobactam-resistant ESBL infections: dose recommendations for various renal functions, including intermittent haemodialysis, in healthy individuals.

WCK 4282 is a novel combination of cefepime 2 g and tazobactam 2 g being developed for the treatment of infections caused by piperacillin/tazobactam-resistant ESBL infections. The dosing regimen for cefepime/tazobactam needs to be optimized to generate adequate exposures to treat infections caused by ESBL-producing pathogens resistant to both cefepime and piperacillin/tazobactam. We developed pharmacokinetic population models of cefepime and tazobactam to evaluate the optimal dose adjustments in patients, including those with augmented renal clearance as well as various degrees of renal impairment, and also for those on intermittent haemodialysis. Optimal doses for various degrees of renal function were identified by determining the PTA for a range of MICs. To cover ESBL-producing pathogens with an cefepime/tazobactam MIC of 16 mg/L, a dosing regimen of 2 g q8h infused over 1.5 h resulted in a combined PTA of 99% for the mean murine 1 log10-kill target for the cefepime/tazobactam combination. We found that to adjust for renal function, doses need to be reduced to 1 g q8h, 500 mg q8h and 500 mg q12h for patients with CLCR of 30-59, 15-29 and 8-14 mL/min (as well as patients with intermittent haemodialysis), respectively. In patients with high to augmented CLR (estimated CLCR 120-180 mL/min), a prolonged 4 h infusion of standard dose is required. The suggested dosing regimens will result in exposures of cefepime and tazobactam that would be adequate for infections caused by ESBL-producing pathogens with a cefepime/tazobactam MICs up to 16 mg/L.

Integrative model-based comparison of target site-specific antimicrobial effects: A case study with ceftaroline and lefamulin

ObjectivePredictions of antimicrobial effects typically rely on plasma-based pharmacokinetic-pharmacodynamic (PK-PD) targets, ignoring target-site concentrations and potential differences in tissue penetration between antibiotics. In this study, we applied PK-PD modelling to compare target site-specific effects of antibiotics by integrating clinical microdialysis data, in vitro time-kill curves, and antimicrobial susceptibility distributions. As a case study, we compared the effect of lefamulin and ceftaroline against methicillin-resistant Staphylococcus aureus (MRSA) at soft-tissue concentrations. MethodsA population PK model describing lefamulin concentrations in plasma, subcutaneous adipose and muscle tissue was developed. For ceftaroline, a similar previously reported PK model was adopted. In vitro time-kill experiments were performed with six MRSA isolates and a PD model was developed to describe bacterial growth and antimicrobial effects. The clinical PK and in vitro PD models were linked to compare antimicrobial effects of ceftaroline and lefamulin at the different target sites. ResultsConsidering minimum inhibitory concentration (MIC) distributions and standard dosages, ceftaroline showed superior anti-MRSA effects compared to lefamulin both at plasma and soft-tissue concentrations. Looking at the individual antibiotics, lefamulin effects were highest at soft-tissue concentrations, while ceftaroline effects were highest at plasma concentrations, emphasising the importance of considering target-site PK-PD in antibiotic treatment optimisation. ConclusionGiven standard dosing regimens, ceftaroline appeared more effective than lefamulin against MRSA at soft-tissue concentrations. The PK-PD model-based approach applied in this study could be used to compare or explore the potential of antibiotics for specific indications or in populations with unique target-site PK.

Population pharmacokinetics of a novel PCSK9 antisense oligonucleotide.

The aim of this study was to characterize the population pharmacokinetics of AZD8233, an antisense oligonucleotide (ASO) that targets the PCSK9 transcript to reduce hepatocyte PCSK9 protein production and plasma levels. AZD8233 utilizes generation 2.5S-constrained ethyl motif (cET) chemistry and is conjugated to a triantennary N-acetylgalactosamine (GalNAc3) ligand for targeted hepatocyte uptake. A non-linear mixed-effect modelling approach utilizing NONMEM software was applied to AZD8233 concentration-time data from 3416 samples in 219 participants from four phase 1-2 studies, one in healthy volunteers (NCT03593785) and three in patients with dyslipidaemia (NCT04155645, NCT04641299 and NCT04823611). The final model described the AZD8233 plasma concentration-time profile from four phase 1-2 studies in healthy volunteers or participants with dyslipidaemia, covering a dose range of 4 to 120 mg. The pharmacokinetics of AZD8233 were adequately described by a two-compartment model with first-order absorption. The supra-proportional increase in maximum plasma concentration (Cmax) across the observed dose range was described by non-linear Michaelis-Menten elimination (maximum elimination rate, 9.9mg/h [12% relative standard error]; concentration yielding half-maximal elimination rate, 4.8mg/L [18% relative standard error]). Body weight, sex, estimated glomerular filtration rate and disease status (healthy participant vs. patient with dyslipidaemia) were identified as factors affecting exposure to AZD8233. Covariate analysis showed body weight to be the main factor affecting exposure to AZD8233, which largely explained the higher Cmax observed in the Asian population relative to non-Asians.

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.

Efficacy, Safety, and Population Pharmacokinetics of Eltrombopag in Children with Different Severities of Aplastic Anemia.

Eltrombopag was approved as a first-line treatment for patients older than 2 years old with severe aplastic anemia (SAA). However, data on eltrombopag in children with different types of aplastic anemia (AA), especially non-severe AA (NSAA), are limited. We performed a prospective, single-arm, and observational study to investigate eltrombopag's efficacy, safety, and pharmacokinetics in children with NSAA, SAA, and very severe AA (VSAA). The efficacy and safety were assessed every 3 months. The population pharmacokinetic (PPK) model was used to depict the pharmacokinetic profile of eltrombopag. Twenty-three AA children with an average age of 7.9 (range of 3.0-14.0) years were enrolled. The response (complete and partial response) rate was 12.5%, 50.0%, and 100.0% after 3, 6, and 12 months in patients with NSAA. For patients with SAA and VSAA, these response rates were 46.7%, 61.5%, and 87.5%. Hepatotoxicity occurred in one patient. Fifty-three blood samples were used to build the PPK model. Body weight was the only covariate for apparent clearance (CL/F) and volume of distribution. The allele-T carrier of adenosine triphosphate-binding cassette transporter G2 was found to increase eltrombopag's clearance. However, when normalized by weight, the clearance between the wild-type and variant showed no statistical difference. In patients with response, children with NSAA exhibited lower area under the curve from time zero to infinity, higher CL/F, and higher weight-adjusted CL/F than those with SAA or VSAA. However, the differences were not statistically significant. The results may support further individualized treatment of eltrombopag in children with AA.

Application of machine learning combined with population pharmacokinetics to improve individual prediction of vancomycin clearance in simulated adult patients.

Vancomycin, a glycopeptide antimicrobial drug. PPK has problems such as difficulty in accurately reflecting inter-individual differences, and the PPK model may not be accurate enough to predict individual pharmacokinetic parameters. Therefore, the aim of this study is to investigate whether the application of machine learning combined with the PPK method can improve the prediction of vancomycin CL in adult Chinese patients. In the first step, a vancomycin CL prediction model for Chinese adult patients is given by PPK and Hamilton Monte Carlo sampling is used to obtain the reference CL of 1,000 patients; the second step is to obtain the final prediction model by machine learning using an appropriate model for the predictive factor and the reference CL; and the third step is to randomly select, in the simulated data, a total of 250 patients for prediction effect evaluation. XGBoost model is selected as final machine learning model. More than four-fifths of the subjects' predictive values regarding vancomycin CL are improved by machine learning combined with PPK. Machine learning combined with PPK models is more stable in performance than the PPK method alone for predicting models. The first combination of PPK and machine learning for predictive modeling of vancomycin clearance in adult patients. It provides a reference for clinical pharmacists or clinicians to optimize the initial dosage given to ensure the effectiveness and safety of drug therapy for each patient.

Proof of concept of physiologically based pharmacokinetic modelling in paediatric acute lymphoblastic leukaemia.

Physiologically based pharmacokinetic (PBPK) modelling is an alternative modelling technique that is increasingly used in pharmacokinetics. Due to its nature, it can be complementarily employed to population pharmacokinetics, especially when it comes to small population size. Here, we report the proof of concept of its application to accurately describe the pharmacokinetics of a recombinant L-asparaginase in paediatric patients with acute lymphoblastic leukaemia. Data from two randomized, double-blind, phase II/III clinical studies (MC-ASP.4/ALL; MC-ASP.5/ALL) were included to setup and evaluate the final model, respectively. Final population values for basic pharmacokinetic parameters were calculated (clearance: 0.0569 L/h/19.5 kg, volume of distribution: 1.251 L, half-life: 18.5 h, trough concentration: 140.9 IU/L). Pharmacokinetic parameter prediction as well as predictive performance of the model proofed to be comparable to a separately developed population pharmacokinetic model with 13% deviation in predicted median L-asparaginase trough levels. To the best of our knowledge, this is the first whole-body PBPK model of a non-antibody therapeutic protein.

Optimization of Ganciclovir and Valganciclovir Starting Dose in Children by Machine Learning.

Ganciclovir (GCV) and valganciclovir (VGCV) show large interindividual pharmacokinetic variability, particularly in children.The objectives of this study were (1) to develop machine learning (ML) algorithms trained on simulated pharmacokinetics profiles obtained by Monte Carlo simulations to estimate the best gancicloviror valganciclovirstarting dose in children and (2) to compare its performances on real-world profiles to previously published equation derived from literature population pharmacokinetic (POPPK) models achieving about 20% of profiles within the target. The pharmacokinetic parameters of four literature POPPK models in addition to the World Health Organization (WHO) growth curve for children were used in the mrgsolve R package to simulate 10,800 pharmacokinetic profiles. ML algorithms were developed and benchmarked to predict the probability to reach the steady-state, area-under-the-curve target (AUC0-24within 40-60 mg × h/L) based on demographic characteristics only. The best ML algorithm was then used to calculate the starting dose maximizing the target attainment. Performances were evaluated for ML and literature formula in a test set and in an external set of 32 and 31 actual patients (GCV and VGCV, respectively). A combination of Xgboost, neural network, and random forest algorithms yielded the best performances and highest target attainment in the test set (36.8% for GCV and 35.3% for the VGCV). In actual patients, the best GCV ML starting dose yielded the highest target attainment rate (25.8%) and performed equally for VGCV with the Franck model formula (35.3% for both). The ML algorithms exhibit good performances in comparison with previously validated models and should be evaluated prospectively.

Population pharmacokinetics of prophylactic cefoxitin in elective bariatric surgery patients: a prospective monocentric study

BackgroundThis study describes the population pharmacokinetics of cefoxitin in obese patients undergoing elective bariatric surgery and evaluates different dosing regimens for achievement of pre-defined target exposures. MethodsSerial blood samples were collected during surgery with relevant clinical data. Total serum cefoxitin concentrations were measured by chromatographic assay and analysed using a population PK approach with Pmetrics®. The cefoxitin unbound fraction (fu) was estimated. Dosing simulations were performed to ascertain the probability of target attainment (PTA) to achieve cefoxitin fu above minimum inhibitory concentrations (MIC) from surgical incision to wound closure. Fractional target attainment (FTA) was calculated against MIC distributions of common pathogens. ResultsA total of 123 obese patients (median BMI 44.3 kg/m2) were included with 381 cefoxitin concentration values. Cefoxitin was best described by a one-compartment model, with a mean clearance and volume of distribution of 10.9 ± 6.1 L/h and 23.4 ± 10.5 L, respectively. In surgery <2 h, a 2 and a 4 g doses were sufficient for an MIC up to 4 and 8 mg/L (fu 50%), respectively. In prolonged surgery (2–4 h), only continuous infusion enabled optimal PTA for an MIC up to 16 mg/L. Optimal FTAs were obtained against Staphylococcus aureus and Escherichia Coli only when simulating with 50% cefoxitin protein binding (intermittent regimen) and regardless of the protein binding for the continuous infusion. ConclusionIntermittent dosing regimens resulted in optimal FTAs against susceptible MIC distributions of S. aureus and E. coli when simulating with 50% cefoxitin protein binding. Continuous infusion of cefoxitin may improve FTA regardless of protein binding. Study registrationRegistration on ClinicalTrials.gov, NCT03306290

Model-based exploration of the rationality of off-label use of cetirizine in Chinese pediatric patients: a prospective cohort study.

Aims: Cetirizine is frequently administered at an increased dosage in clinical practice and recommended by several guidelines. Nonetheless, the pharmacokinetic (PK) profile and real-world safety data remain insufficient in the Chinese pediatric population. The objective of the current study is to develop a population pharmacokinetic (PPK) model for cetirizine in Chinese pediatric patients and to investigate the rationale behind its off-label usage. Methods: A prospective cohort study was conducted, enrolling children who had been diagnosed with allergic diseases and prescribed cetirizine. The outcomes were safety and pharmacokinetic (PK) parameters. Cetirizine concentrations were measured using a pre-established analytical method. Subsequently, a PK model was developed, followed by model evaluation and simulation. The developed PK model was employed to investigate the drug exposure differences across various age groups and to simulate scenarios of potential overdose. Results: Sixty-three children were enrolled, and 24 of them received a cetirizine dose exceeding the recommended dosage. A PPK model, based on published literature, served as the basis of our analysis, with adjustment made to estimate certain parameters. The final model evaluation and validation indicated accurate predictive performance and robust parameter estimation. Simulations conducted for the label-dose among age 1-12 indicated median maximum concentration at steady state (Cmax,ss) of 7year old children could be the highest. The model was also used to predict the off-label dose scenarios and overdose patient to support the clinical decision. There were no adverse drug reactions in either group. Conclusion: This study provides evidence-based and model-based exploration for optimizing cetirizine usage in Chinese pediatric patients. The cetirizine PPK model showed accurate predictive performance and could be utilized to simulate individual patient exposure in real-world clinical scenarios.

Intrapulmonary and Systemic Pharmacokinetics of Colistin Following Nebulization of Low-Dose Colistimethate Sodium in Patients with Ventilator-Associated Pneumonia Caused by Carbapenem-Resistant Acinetobacter baumannii.

Colistimethate sodium (CMS) nebulization is associated with reduced systemic toxicity compared to intravenous injection, with potentially enhanced clinical efficacy. This study aimed to assess the pharmacokinetic (PK) properties of colistin during low-dose CMS nebulization in patients with ventilator-associated pneumonia (VAP) caused by carbapenem-resistant Acinetobacter baumannii. A nonlinear mixed-effects modeling approach was applied to develop population PK models for colistin in both epithelial lining fluid (ELF) and plasma. Twenty patients participated, and 80 ELF and 100 plasma samples were used for model development. Median colistin concentrations measured in ELF were 614-fold, 408-fold, and 250-fold higher than in plasma at 1, 3, and 5 h, respectively. Time courses in both ELF and plasma were best described by a one-compartment model with a Weibull absorption process. When the final model was simulated, the maximum free concentration and area under the free colistin concentration-time curve at steady state over 24 h in the plasma were approximately 1/90 and 1/50 of the corresponding values in ELF at steady state, respectively. For an A. baumannii MIC of 1 mg/L, inhaling 75 mg of CMS at 6 h intervals was deemed appropriate, with dose adjustments needed for MICs exceeding 2 mg/L. Using a nebulizer for CMS resulted in a notably higher exposure of colistin in the ELF than plasma, indicating the potential of nebulization to reduce systemic toxicity while effectively treating VAP.

Integrating Full Bayesian Inference and Student's t-Distribution Method for Enhanced Outlier Handling in Caffeine Population Pharmacokinetics: Assessing Drug-Drug Interactions with Enasidenib in Relapsed or Refractory AML and MDS Patients.

As the first-in-class, selective, and potent inhibitor of the isocitrate dehydrogenase-2 (IDH2) mutant protein, enasidenib was approved by the US Food and Drug Administration (FDA) in 2017 for the treatment of adult patients with relapsed or refractory acute myeloid leukemia (AML) with an IDH2 mutation. Known for its interactions with various cytochrome P450 (CYP)enzymes and transporters in vitro, a clinical pharmacokinetics (PK) trial was initiated to assess the impact of multiple doses of enasidenib on the single-dose PK of sensitive probe substrates of several cytochrome P450 enzymes and transporters. In this study, a population pharmacokinetic analysis approach was employed to address challenges posed by high, nonzero baseline caffeine concentrations. Moreover, we integrated full Bayesian inference into this approach innovatively for a more detailed understanding of parameter uncertainty and greater modeling flexibility, alongside Student's t-distribution for robust error modeling in handling the abnormal outlier caffeine concentration data observed in this trial. Our analyses demonstrated that multiple doses of enasidenib altered caffeine clearance to a clinically meaningful extent, as evidenced by an approximate 8-fold decrease. This finding led to a specific recommendation in the package insert to avoid the concurrent use of certain CYP1A2 substrates with enasidenib, unless directed otherwise in the prescribing information. Furthermore, this research underlines the technical benefits of integrating full Bayesian inference and incorporating Student's t-distribution for residual error modeling in the PK field.

Pharmacokinetics of propofol in severely obese surgical patients.

Existing PK models of propofol include sparse data from very obese patients. The aim of this study was to develop a PK model based on standardised surgical conditions and spanning from normal-weight up to, and including, a high number of very obese patients. Adult patients scheduled for laparoscopic cholecystectomy or bariatric surgery were studied. Anaesthesia was induced with propofol 2 mg/kg adjusted body weight over 2 min followed by 6 mg/kg/h adjusted body weight over 30 min. For the remainder of the operation anaesthesia was maintained with sevoflurane. Remifentanil was dosed according to clinical need. Eight arterial samples were drawn in a randomised block sampling regimen over a span of 24 h. Time-concentration data were analysed by population PK modelling using non-linear mixed-effects modelling. Four hundred and seventy four serum propofol concentrations were collected from 69 patients aged 19-60 years with a BMI 21.6-67.3 kg/m2. Twenty one patients had a BMI above 50 kg/m2. A 3-compartment PK model was produced wherein three different body weight descriptors and sex were included as covariates in the final model. Total body weight was found to be a covariate for clearance and Q3; lean body weight for V1, V2 and Q2; predicted normal weight for V3 and sex for V1. The fixed allometric exponent of 0.75 applied to all clearance parameters improved the performance of the model. Accuracy and precision were 1.4% and 21.7% respectively in post-hoc performance evaluation. We have developed a new PK model of propofol that is suitable for all adult weight classes. Specifically, it is based on data from an unprecedented number of individuals with very high BMI.

A Randomized, Double-Blind, Placebo-Controlled Trial Evaluating the Effect of Topical Urea for Secondary Prophylaxis of Hand Foot Skin Reaction in Renal Cell Cancer Patients on Sunitinib Therapy

IntroductionHand foot skin reaction (HFSR) is a common dose-limiting adverse effect of multi kinase inhibitors (MKI) whose mechanism is not fully understood, and the prophylaxis is inadequate. ObjectiveIn this pilot study, a double-blind, randomized placebo-controlled trial was conducted to evaluate the effect of topical urea in secondary prevention of sunitinib-induced HFSR in renal cell cancer patients. MethodsOut of 55 screened patients, 14 were randomized to receive topical urea or placebo for four weeks. The association of HFSR with drug levels of sunitinib and its metabolite (n-desethyl sunitinib), genetic polymorphism of VEGFR2 gene, quality of life (QOL) and biochemical markers was also assessed. ResultsThe results showed that urea-based cream was not superior to placebo (P = .075). There was no change in the QOL in both the groups. Single nucleotide polymorphism was checked for two nucleotides rs1870377 and rs2305948 located in VEGFR2 gene on chromosome 4. SNP (variant T > A) at rs1870377 was associated with appearance of new HFSR as compared to the wild type, although the association was not statistically significant (OR 0.714). There was no statistically significant difference between mean plasma levels of sunitinib and N-desethyl sunitinib in urea arm as compared to placebo arm as compared to placebo. The best fit population pharmacokinetic model for sunitinib was one compartment model with first order absorption and linear elimination. The median (IQR) of population parameters calculated from the population pharmacokinetics model for Ka, V and Cl was 0.22 (0.21-0.24) h-1, 4.4 (4.09-4.47) L, 0.049 (0.042-0.12) L/hr, respectively. ConclusionThe study suggested that the urea-based cream was not superior to placebo in decreasing the appearance of new HFSR in renal cancer patients receiving 4:2 regimen of sunitinib.

Population pharmacokinetics of everolimus in renal transplant recipients receiving long-term multiple immunosuppressive therapy

Everolimus is used for immunosuppression after renal transplantation. This study aimed to develop a population pharmacokinetic (PopPK) model of everolimus using therapeutic drug monitoring (TDM) data of patients under long-term multiple immunosuppressive therapy, including tacrolimus. To develop the model, 185 renal transplant recipients with 3358 everolimus blood concentrations during a median postoperative period of 35.3 months were included. The PopPK model is described as a one-compartment model with first-order absorption. The population mean of apparent clearance is 8.92 L/h (relative standard error = 3.6%), and this negatively correlated with the dose-normalized concentration (C/D) of tacrolimus and hematocrit value, and positively correlated with a daily dose of everolimus (i.e. TDM effect). The usefulness of dose adjustment using the final popPK model was assessed by a simulation study. The ratio of the first trough measurement within the therapeutic range of 3–8 ng/mL increased from 69.8% in the original dose to 87.9% in the individual dose calculated by the final PopPK model. The tacrolimus C/D ratio before initiating everolimus therapy and the hematocrit value were useful to estimate the initial dose of everolimus and can improve the safety and effectiveness of immunosuppressive therapy involving everolimus.

A Deep Learning-Based Approach for Prediction of Vancomycin Treatment Monitoring: Retrospective Study Among Patients With Critical Illness.

Vancomycin pharmacokinetics are highly variable in patients with critical illnesses, and clinicians commonly use population pharmacokinetic (PPK) models based on a Bayesian approach to dose. However, these models are population-dependent, may only sometimes meet the needs of individual patients, and are only used by experienced clinicians as a reference for making treatment decisions. To assist real-world clinicians, we developed a deep learning-based decision-making system that predicts vancomycin therapeutic drug monitoring (TDM) levels in patients in intensive care unit. This study aimed to establish joint multilayer perceptron (JointMLP), a new deep-learning model for predicting vancomycin TDM levels, and compare its performance with the PPK models, extreme gradient boosting (XGBoost), and TabNet. We used a 977-case data set split into training and testing groups in a 9:1 ratio. We performed external validation of the model using 1429 cases from Kangwon National University Hospital and 2394 cases from the Medical Information Mart for Intensive Care-IV (MIMIC-IV). In addition, we performed 10-fold cross-validation on the internal training data set and calculated the 95% CIs using the metric. Finally, we evaluated the generalization ability of the JointMLP model using the MIMIC-IV data set. Our JointMLP model outperformed other models in predicting vancomycin TDM levels in internal and external data sets. Compared to PPK, the JointMLP model improved predictive power by up to 31% (mean absolute error [MAE] 6.68 vs 5.11) on the internal data set and 81% (MAE 11.87 vs 6.56) on the external data set. In addition, the JointMLP model significantly outperforms XGBoost and TabNet, with a 13% (MAE 5.75 vs 5.11) and 14% (MAE 5.85 vs 5.11) improvement in predictive accuracy on the inner data set, respectively. On both the internal and external data sets, our JointMLP model performed well compared to XGBoost and TabNet, achieving prediction accuracy improvements of 34% and 14%, respectively. Additionally, our JointMLP model showed higher robustness to outlier data than the other models, as evidenced by its higher root mean squared error performance across all data sets. The mean errors and variances of the JointMLP model were close to zero and smaller than those of the PPK model in internal and external data sets. Our JointMLP approach can help optimize treatment outcomes in patients with critical illnesses in an intensive care unit setting, reducing side effects associated with suboptimal vancomycin administration. These include increased risk of bacterial resistance, extended hospital stays, and increased health care costs. In addition, the superior performance of our model compared to existing models highlights its potential to help real-world clinicians.

Population pharmacokinetics of vancomycin in term neonates with perinatal asphyxia treated with therapeutic hypothermia.

Little is known about the population pharmacokinetics (PPK) of vancomycin in neonates with perinatal asphyxia treated with therapeutic hypothermia (TH). We aimed to describe the PPK of vancomycin and propose an initial dosing regimen for the first 48 h of treatment with pharmacokinetic/pharmacodynamic target attainment. Neonates with perinatal asphyxia treated with TH were included from birth until Day 6 in a multicentre prospective cohort study. A vancomycin PPK model was constructed using nonlinear mixed-effects modelling. The model was used to evaluate published dosing guidelines with regard to pharmacokinetic/pharmacodynamic target attainment. The area under the curve/minimal inhibitory concentration ratio of 400-600 mg*h/L was used as target range. Sixteen patients received vancomycin (median gestational age: 41 [range: 38-42] weeks, postnatal age: 4.4 [2.5-5.5] days, birth weight: 3.5 [2.3-4.7] kg), and 112 vancomycin plasma concentrations were available. Most samples (79%) were collected during the rewarming and normothermic phase, as vancomycin was rarely initiated during the hypothermic phase due to its nonempirical use. An allometrically scaled 1-compartment model showed the best fit. Vancomycin clearance was 0.17 L/h, lower than literature values for term neonates of 3.5kg without perinatal asphyxia (range: 0.20-0.32 L/h). Volume of distribution was similar. Published dosing regimens led to overexposure within 24 h of treatment. A loading dose of 10mg/kg followed by 24 mg/kg/day in 4 doses resulted in target attainment. Results of this study suggest that vancomycin clearance is reduced in term neonates with perinatal asphyxia treated with TH. Lower dosing regimens should be considered followed by model-informed precision dosing.

Implications of Immunogenicity Testing for Therapeutic Monoclonal Antibodies: A Quantitative Pharmacology Framework

The interpretation of immunogenicity results for a mAb product and prediction of its clinical consequences remain difficult, despite enormous advances in methodologies and efforts toward the best practice for consistent data generation and reporting. To this end, the contribution from the clinical pharmacology discipline has been largely limited to comparing descriptively the pharmacokinetic (PK) profiles by antidrug antibodies (ADA) status or testing the significance of ADA as a covariate in a population PK setting, similar to the practice for small-molecule drugs in investigating the effect of an intrinsic/extrinsic factor on the drug disposition. There is a need for a mAb disposition framework that captures the dynamics of ADA formation and drug’s interactions with the ADA and target as parts of the drug distribution and elimination. Here we describe such a framework and examine it against the PK, ADA, and clinical response data from a phase 3 trial in patients treated with adalimumab. The proposed framework offered a generalized understanding of how the dose, target affinity, and drug/ADA analyte forms affects the manifestation of ADA response with regard to its detections and alterations of drug disposition and effectiveness. Furthermore, as an example, its utility for dose considerations was demonstrated through predicting for late-stage trials of a PCSK9 inhibitor in terms of development in ADA incidence and titers, and consequences on the drug disposition, interaction with target, and downstream lowering effect on LDL-C.Graphical

Pharmacokinetics and exposure-safety relationship of ciprofol for sedation in mechanically ventilated patients in the intensive care unit.

Ciprofol (HSK3486) is a newly developed, highly selective γ-aminobutyric acid-A (GABAA) receptor potentiator that is recently approved for a new indication of sedation for patients in the intensive care unit (ICU) in China. This analysis aimed to characterize the population pharmacokinetics (PopPKs) of ciprofol and evaluate the relationship of exposure with hypotension in mechanically ventilated patients in the ICU. A total of 462 subjects with 3918 concentration measurements from two clinical trials of mechanically ventilated patients in the ICU, four clinical trials of elective surgical patients, and six clinical trials of healthy subjects were used in the PopPK analysis. Exposure-safety relationship for hypotension was evaluated based on the data gathered from 112 subjects in two clinical trials of mechanically ventilated patients in the ICU. Ciprofol pharmacokinetics (PKs) was adequately described by a three-compartment linear disposition model with first-order elimination. Body weight, age, sex, blood sampling site (vein vs. arterial), study design (long-term infusion vs. short-term infusion), and patient population (ICU vs. non-ICU) were identified as statistically significant covariates on the PKs of ciprofol. Within the exposure range of the mechanically ventilated ICU patient population, no meaningful association was observed between ciprofol exposure and the incidence of hypotension. These results support the dosing regimen currently used in mechanically ventilated patients in the ICU.