Evaluation of ChatGPT and Gemini large language models for pharmacometrics with NONMEM

To assess ChatGPT 4.0 (ChatGPT) and Gemini Ultra 1.0 (Gemini) large language models on NONMEM coding tasks relevant to pharmacometrics and clinical pharmacology. ChatGPT and Gemini were assessed on tasks mimicking real-world applications of NONMEM. The tasks ranged from providing a curriculum for learning NONMEM, an overview of NONMEM code structure to generating code. Prompts in lay language to elicit NONMEM code for a linear pharmacokinetic (PK) model with oral administration and a more complex model with two parallel first-order absorption mechanisms were investigated. Reproducibility and the impact of "temperature" hyperparameter settings were assessed. The code was reviewed by two NONMEM experts. ChatGPT and Gemini provided NONMEM curriculum structures combining foundational knowledge with advanced concepts (e.g., covariate modeling and Bayesian approaches) and practical skills including NONMEM code structure and syntax. ChatGPT provided an informative summary of the NONMEM control stream structure and outlined the key NONMEM Translator (NM-TRAN) records needed. ChatGPT and Gemini were able to generate code blocks for the NONMEM control stream from the lay language prompts for the two coding tasks. The control streams contained focal structural and syntax errors that required revision before they could be executed without errors and warnings. The code output from ChatGPT and Gemini was not reproducible, and varying the temperature hyperparameter did not reduce the errors and omissions substantively. Large language models may be useful in pharmacometrics for efficiently generating an initial coding template for modeling projects. However, the output can contain errors and omissions that require correction.

PmWebSpec: An Application to Create and Manage CDISC-Compliant Pharmacometric Analysis Dataset Specifications

A well-documented pharmacometric (PMx) analysis dataset specification ensures consistency in derivations of the variables, naming conventions, traceability to the source data, and reproducibility of the analysis dataset. Lack of standards in creating the dataset specification can lead to poor quality analysis datasets, negatively impacting the quality of the PMx analysis. Standardization of the dataset specification within an individual organization helps address some of these inconsistencies. The recent introduction of the Clinical Data Interchange Standards Consortium (CDISC) Analysis Data Model (ADaM) Population Pharmacokinetic (popPK) Implementation Guide (IG) further promotes industry-wide standards by providing guidelines for the basic data structure of popPK analysis datasets. However, manual implementation of the standards can be labor intensive and error-prone. Hence, there is still a need to automate the implementation of these standards. In this paper, we present PmWebSpec, an easily deployable web-based application to facilitate the creation and management of CDISC-compliant PMx analysis dataset specifications. We describe the application of this tool through examples and highlight its key features including pre-populated dataset specifications, built-in checks to enforce standards, and generation of an electronic Common Technical Document (eCTD)-compliant data definition file. The application increases efficiency, quality and semi-automates PMx analysis dataset, and specification creation and has been well accepted by pharmacometricians and programmers internally. The success of this application suggests its potential for broader usage across the PMx community.

Translational PK-PD/TD modeling of antitumor effects and peripheral neuropathy in gemcitabine and nab-paclitaxel chemotherapy from xenograft mice to patients for optimal dose and schedule

PURPOSE

Gemcitabine and nab-paclitaxel (GnP) treatment, the standard first-line chemotherapy for unresectable pancreatic cancer, often causes peripheral neuropathy (PN). To develop alternative dosing strategies to avoid severe PN, understanding the relationship between pharmacokinetics (PK) and pharmacodynamics/toxicodynamics (PD/TD) is necessary. We established a PK-PD/TD model of GnP treatment to develop an optimal dose schedule.

METHODS

A mouse xenograft model of human pancreatic cancer was generated to measure drug concentrations in the plasma and tumor, antitumor effects, and PN after GnP treatment. The Simeoni tumor growth inhibition model with tumor concentrations and empirical indirect response models were used for the PD and TD models, respectively. Clinical outcomes were predicted with reported population estimates of PK parameters in cancer patients.

RESULTS

The PK-PD/TD model simultaneously described the observed tumor volume and paw withdrawal frequency in the von Frey test. For the standard GnP regimen, the model predicted clinical overall response (75.1%), which was overestimated compared to that in a recent phase II study (42.1%) but lower than the observed disease control rate (96.5%). Model simulation showed that dose reduction to less than 40% GnP dose was not effective; a change of dose schedule from every week for 3 weeks to every 2 weeks was a more favorable approach than dose reduction to 60% every week.

CONCLUSION

The PK-PD/TD model-based translational approach provides a guide for optimal dose determination to avoid severe PN while maintaining antitumor effects during GnP chemotherapy. Further research is needed to enhance its applicability and potential for combination chemotherapy regimens.

Learning pharmacometric covariate model structures with symbolic regression networks

Efficiently finding covariate model structures that minimize the need for random effects to describe pharmacological data is challenging. The standard approach focuses on identification of relevant covariates, and present methodology lacks tools for automatic identification of covariate model structures. Although neural networks could potentially be used to approximate covariate-parameter relationships, such approximations are not human-readable and come at the risk of poor generalizability due to high model complexity.In the present study, a novel methodology for the simultaneous selection of covariate model structure and optimization of its parameters is proposed. It is based on symbolic regression, posed as an optimization problem with a smooth loss function. This enables training of the model through back-propagation using efficient gradient computations.Feasibility and effectiveness are demonstrated by application to a clinical pharmacokinetic data set for propofol, containing infusion and blood sample time series from 1031 individuals. The resulting model is compared to a published state-of-the-art model for the same data set. Our methodology finds a covariate model structure and corresponding parameter values with a slightly better fit, while relying on notably fewer covariates than the state-of-the-art model. Unlike contemporary practice, finding the covariate model structure is achieved without an iterative procedure involving manual interactions.

Low-dimensional neural ODEs and their application in pharmacokinetics

Machine Learning (ML) is a fast-evolving field, integrated in many of today's scientific disciplines. With the recent development of neural ordinary differential equations (NODEs), ML provides a new tool to model dynamical systems in the field of pharmacology and pharmacometrics, such as pharmacokinetics (PK) or pharmacodynamics. The novel and conceptionally different approach of NODEs compared to classical PK modeling creates challenges but also provides opportunities for its application. In this manuscript, we introduce the functionality of NODEs and develop specific low-dimensional NODE structures based on PK principles. We discuss two challenges of NODEs, overfitting and extrapolation to unseen data, and provide practical solutions to these problems. We illustrate concept and application of our proposed low-dimensional NODE approach with several PK modeling examples, including multi-compartmental, target-mediated drug disposition, and delayed absorption behavior. In all investigated scenarios, the NODEs were able to describe the data well and simulate data for new subjects within the observed dosing range. Finally, we briefly demonstrate how NODEs can be combined with mechanistic models. This research work enhances understanding of how NODEs can be applied in PK analyses and illustrates the potential for NODEs in the field of pharmacology and pharmacometrics.

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.

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

OBJECTIVE

Predictions 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.

METHODS

A 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.

RESULTS

Considering 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.

CONCLUSION

Given 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.

Warfarin-A natural anticoagulant: A review of research trends for precision medication

BACKGROUND

Warfarin is a widely prescribed anticoagulant in the clinic. It has a more considerable individual variability, and many factors affect its variability. Mathematical models can quantify the quantitative impact of these factors on individual variability.

PURPOSE

The aim is to comprehensively analyze the advanced warfarin dosing algorithm based on pharmacometrics and machine learning models of personalized warfarin dosage.

METHODS

A bibliometric analysis of the literature retrieved from PubMed and Scopus was performed using VOSviewer. The relevant literature that reported the precise dosage of warfarin calculation was retrieved from the database. The multiple linear regression (MLR) algorithm was excluded because a recent systematic review that mainly reviewed this algorithm has been reported. The following terms of quantitative systems pharmacology, mechanistic model, physiologically based pharmacokinetic model, artificial intelligence, machine learning, pharmacokinetic, pharmacodynamic, pharmacokinetics, pharmacodynamics, and warfarin were added as MeSH Terms or appearing in Title/Abstract into query box of PubMed, then humans and English as filter were added to retrieve the literature.

RESULTS

Bibliometric analysis revealed important co-occuring MeShH and index keywords. Further, the United States, China, and the United Kingdom were among the top countries contributing in this domain. Some studies have established personalized warfarin dosage models using pharmacometrics and machine learning-based algorithms. There were 54 related studies, including 14 pharmacometric models, 31 artificial intelligence models, and 9 model evaluations. Each model has its advantages and disadvantages. The pharmacometric model contains biological or pharmacological mechanisms in structure. The process of pharmacometric model development is very time- and labor-intensive. Machine learning is a purely data-driven approach; its parameters are more mathematical and have less biological interpretation. However, it is faster, more efficient, and less time-consuming. Most published models of machine learning algorithms were established based on cross-sectional data sourced from the database.

CONCLUSION

Future research on personalized warfarin medication should focus on combining the advantages of machine learning and pharmacometrics algorithms to establish a more robust warfarin dosage algorithm. Randomized controlled trials should be performed to evaluate the established algorithm of warfarin dosage. Moreover, a more user-friendly and accessible warfarin precision medicine platform should be developed.

Training the next generation of pharmacometric modelers: a multisector perspective

The current demand for pharmacometricians outmatches the supply provided by academic institutions and considerable investments are made to develop the competencies of these scientists on-the-job. Even with the observed increase in academic programs related to pharmacometrics, this need is unlikely to change in the foreseeable future, as the demand and scope of pharmacometrics applications keep expanding. Further, the field of pharmacometrics is changing. The field largely started when Lewis Sheiner and Stuart Beal published their seminal papers on population pharmacokinetics in the late 1970's and early 1980's and has continued to grow in impact and use since its inception. Physiological-based pharmacokinetics and systems pharmacology have grown rapidly in scope and impact in the last decade and machine learning is just on the horizon. While all these methodologies are categorized as pharmacometrics, no one person can be an expert in everything. So how do you train future pharmacometricians? Leading experts in academia, industry, contract research organizations, clinical medicine, and regulatory gave their opinions on how to best train future pharmacometricians. Their opinions were collected and synthesized to create some general recommendations.

Comparison of Two Methods for Determining Item Characteristic Functions and Latent Variable Time-Course for Pharmacometric Item Response Models

There are examples in the literature demonstrating different approaches to defining the item characteristic functions (ICF) and characterizing the latent variable time-course within a pharmacometrics item response theory (IRT) framework. One such method estimates both the ICF and latent variable time-course simultaneously, and another method establishes the ICF first then models the latent variable directly. To date, a direct comparison of the "simultaneous" and "sequential" methodologies described in this work has not yet been systematically investigated. Item parameters from a graded response IRT model developed from Parkinson's Progression Marker Initiative (PPMI) study data were used as simulation parameters. Each method was evaluated under the following conditions: (i) with and without drug effect and (ii) slow progression rate with smaller sample size and rapid progression rate with larger sample size. Overall, the methods performed similarly, with low bias and good precision for key parameters and hypothesis testing for drug effect. The ICF parameters were well determined when the model was correctly specified, with an increase in precision in the scenario with rapid progression. In terms of drug effect, both methods had large estimation bias for the slow progression rate; however, this bias can be considered small relative to overall progression rate. Both methods demonstrated type 1 error control and similar discrimination between model with and without drug effect. The simultaneous method was slightly more precise than the sequential method while the sequential method was more robust towards longitudinal model misspecification and offers practical advantages in model building.

Clinical antiviral efficacy of favipiravir in early COVID-19 (PLATCOV): an open-label, randomised, controlled, adaptive platform trial

In early symptomatic COVID-19 treatment, high dose oral favipiravir did not accelerate viral clearance.

BACKGROUND

Favipiravir, an anti-influenza drug, has in vitro antiviral activity against SARS-CoV-2. Clinical trial evidence to date is inconclusive. Favipiravir has been recommended for the treatment of COVID-19 in some countries.

METHODS

In a multicentre open-label, randomised, controlled, adaptive platform trial, low-risk adult patients with early symptomatic COVID-19 were randomised to one of ten treatment arms including high dose oral favipiravir (3.6g on day 0 followed by 1.6g daily to complete 7 days treatment) or no study drug. The primary outcome was the rate of viral clearance (derived under a linear mixed-effects model from the daily log10 viral densities in standardised duplicate oropharyngeal swab eluates taken daily over 8 days [18 swabs per patient]), assessed in a modified intention-to-treat population (mITT). The safety population included all patients who received at least one dose of the allocated intervention. This ongoing adaptive platform trial was registered at ClinicalTrials.gov (NCT05041907) on 13/09/2021.

RESULTS

In the final analysis, the mITT population contained data from 114 patients randomised to favipiravir and 126 patients randomised concurrently to no study drug. Under the linear mixed-effects model fitted to all oropharyngeal viral density estimates in the first 8 days from randomisation (4,318 swabs), there was no difference in the rate of viral clearance between patients given favipiravir and patients receiving no study drug; a -1% (95% credible interval: -14 to 14%) difference. High dose favipiravir was well-tolerated.

INTERPRETATION

Favipiravir does not accelerate viral clearance in early symptomatic COVID-19. The viral clearance rate estimated from quantitative measurements of oropharyngeal eluate viral densities assesses the antiviral efficacy of drugs in vivo with comparatively few studied patients.

The MILK study: Investigating intergenerational transmission of low-calorie sweeteners in breast milk

Introduction

Forty-four percent of lactating women in the United States consume beverages containing low calorie sweeteners (LCS), and the presence of LCS in the food supply has continued to increase in recent years. While LCS are approved by the United States Food and Drug Administration (FDA) and are believed to be safe for human consumption, intergenerational LCS transmission and the health impacts of early life LCS exposure are severely understudied.

Methods and analysis

In a tightly controlled, single site, prospective interventional study, mothers' plasma and breast milk, and infants' plasma will be collected from 40 mother-infant dyads over the course of 72 h, with rich sampling following maternal ingestion of a LCS sweetened beverage containing sucralose and acesulfame potassium (ace-K). Concentration-time data will be used to build maternal and infant pharmacokinetic models for future simulations and analysis.

Conclusion

This study aims to measure LCS concentrations in breast milk, maternal plasma, and infant plasma, to gain insight into infant exposure and inform recommendations for LCS consumption during breastfeeding.

Early Identification of Patients at Risk of Cabazitaxel-induced Severe Neutropenia

BACKGROUND

Cabazitaxel frequently causes severe neutropenia. A higher cabazitaxel systemic exposure is related to a lower nadir absolute neutrophil count (ANC).

OBJECTIVE

To describe the effect of cabazitaxel systemic exposure on ANC by a population pharmacokinetic/pharmacodynamic (POP-PK/PD) model, and to identify patients at risk of severe neutropenia early in their treatment course using a PK threshold.

DESIGN, SETTING, AND PARTICIPANTS

Data from five clinical studies were pooled to develop a POP-PK/PD model using NONMEM, linking both patient characteristics and cabazitaxel systemic exposure directly to ANC.

OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS

A PK threshold, predictive of severe neutropenia (grade ≥3), was determined using a receiver operating characteristic curve.

RESULTS AND LIMITATIONS

Ninety-six patients were included with a total of 1726 PK samples and 1081 ANCs. The POP-PK/PD model described both cabazitaxel PK and ANC accurately. A cabazitaxel plasma concentration of >4.96 ng/ml at 6 h after the start of infusion was found to be predictive of severe neutropenia, with a sensitivity of 76% and a specificity of 65%.

CONCLUSIONS

Early cabazitaxel plasma levels are predictive of severe neutropenia. Implementation of the proposed PK threshold results in early identification of almost 76% of all severe neutropenias. If prospectively validated, patients at risk could benefit from prophylactic administration of granulocyte colony stimulating factors, preventing severe neutropenia in an early phase of treatment. Implementation of this threshold permits a less restricted use of the 25 mg/m2 dose, potentially increasing the therapeutic benefit.

PATIENT SUMMARY

Treatment with cabazitaxel chemotherapy often causes neutropenia, leading to susceptibility to infections, which might be life threatening. We found that a systemic cabazitaxel concentration above 4.96 ng/ml 6 h after the start of infusion is predictive of the occurrence of severe neutropenia. Measurement of systemic cabazitaxel levels provides clinicians with the opportunity to prophylactically stimulate neutrophil growth.

Generation and application of avatars in pharmacometric modelling

Simulations from population models have critical applications in drug discovery and development. Avatars or digital twins, defined as individual simulations matching clinical criteria of interest compared to observations from a real subject within a predefined margin of accuracy, may be a better option for simulations performed to inform future drug development stages in cases where an adequate model is not achievable. The aim of this work was to (1) investigate methods for generating avatars with pharmacometric models, and (2) explore the properties of the generated avatars to assess the impact of the different selection settings on the number of avatars per subject, their closeness to the individual observations, and the properties of the selected samples subset from the theoretical model parameters probability density function. Avatars were generated using different combinations of nature and number of clinical criteria, accuracy of agreement, and/or number of simulations for two examples models previously published (hemato-toxicity and integrated glucose-insulin model). The avatar distribution could be used to assess the appropriateness of the models assumed parameter distribution. Similarly it could be used to assess the models ability to properly describe the trajectories of the observations. Avatars can give nuanced information regarding the ability of a model to simulate data similar to the observations both at the population and at the individual level. Further potential applications for avatars may be as a diagnostic tool, an alternative to simulations with insurance to replicate key clinical features, and as an individual measure of model fit.

Population pharmacokinetic analysis of lopinavir in HIV negative individuals exposed to SARS-CoV-2: a COPEP (COronavirus Post-Exposure Prophylaxis) sub-study

BACKGROUND

Lopinavir/ritonavir (LPV/r) is a drug traditionally used for the treatment of HIV that has been repurposed as a potential post-exposure prophylaxis agent against COVID-19 in the COronavirus Post-Exposure Prophylaxis (COPEP) study. The present analysis aims to evaluate LPV levels in individuals exposed to SARS-CoV-2 versus people living with HIV (PLWH) by developing a population pharmacokinetic (popPK) model, while characterizing external and patient-related factors that might affect LPV exposure along with dose-response association.

METHODS

We built a popPK model on 105 LPV concentrations measured in 105 HIV-negative COPEP individuals exposed to SARS-CoV-2, complemented with 170 LPV concentrations from 119 PLWH followed in our routine therapeutic drug-monitoring programme. Published LPV popPK models developed in PLWH and in COVID-19 patients were retrieved and validated in our study population by mean prediction error (MPE) and root mean square error (RMSE). The association between LPV model-predicted residual concentrations (Cmin) and the appearance of the COVID-19 infection in the COPEP participants was investigated.

RESULTS

A one-compartment model with linear absorption and elimination best described LPV concentrations in both our analysis and in the majority of the identified studies. Globally, similar PK parameters were found in all PK models, and provided close MPEs (from -19.4% to 8.0%, with a RMSE of 3.4% to 49.5%). No statistically significant association between Cmin and the occurrence of a COVID-19 infection could be detected.

CONCLUSION

Our analysis indicated that LPV circulating concentrations were similar between COPEP participants and PLWH, and that published popPK models described our data in a comparable way.

Plasma and intraperitoneal pharmacokinetics of ceftazidime/avibactam in peritoneal dialysis patients

OBJECTIVES

Peritonitis is a serious complication in patients undergoing automated peritoneal dialysis (APD) that increases morbidity and frequently disqualifies patients from the peritoneal dialysis programme. Ceftazidime/avibactam (CAZ/AVI) is a potential treatment option for APD patients with peritonitis caused by resistant Gram-negative bacteria, but limited data exist on systemic and target-site pharmacokinetics (PK) in patients undergoing APD. This study set out to investigate the PK of CAZ/AVI in plasma and peritoneal dialysate (PDS) of patients undergoing APD.

METHODS

A prospective, open-label PK study was conducted on eight patients undergoing APD. CAZ/AVI was administered as a single intravenous dose of 2 g/0.5 g over 120 minutes. APD cycles were initiated 15 hours after the study drug administration. Dense PDS and plasma sampling was performed for 24 hours after the start of administration. PK parameters were analysed with population PK modelling. Probability of target attainment (PTA) was simulated for different CAZ/AVI doses.

RESULTS

PK profiles of both drugs in plasma and PDS were similar, indicating that the two drugs are well suited for a fixed-dose combination. A two-compartment model best described the PK of both drugs. A single dose of 2 g/0.5 g CAZ/AVI led to concentrations that far exceeded the PK/PD targets of both drugs. In the Monte Carlo simulations, even the lowest dose (750/190 mg CAZ/AVI) achieved a PTA of >90% for MICs up to 8 mg/L (The European Committee on Antimicrobial Susceptibility Testing epidemiological cut-off value for Pseudomonas aeruginosa) in plasma and PDS.

DISCUSSION

On the basis of PTA simulations, a dose of 750/190 mg CAZ/AVI would be sufficient to treat plasma and peritoneal fluid infections in patients undergoing APD.

Population pharmacokinetic modelling of fibrinogen in patients with congenital or acquired-chronic or acute-hypofibrinogenaemia

AIMS

Fibrinogen is the key substrate for coagulation. Fibrinogen pharmacokinetics (PK) after single doses of fibrinogen concentrate (FC), using modelling approaches, has only been evaluated in congenital afibrinogenaemic patients. The aims of this study are to characterize the fibrinogen PK in patients with acquired-chronic (cirrhosis) or acute-hypofibrinogenaemia (critical haemorrhage), showing endogenous production. Influencing factors of differences on the fibrinogen PK between subpopulations will be identified.

METHODS

A total of 428 time-concentration values from 132 patients were recorded. Eighty-two out of 428 values were from 41 cirrhotic patients administered with placebo and 90 out of 428 were from 45 cirrhotic patients that were given FC, 161 out of 428 values were from 14 afibrinogenaemic patients and 95 out of 428 values were from 32 severe acute trauma haemorrhagic patients. A turnover model that accounted for endogenous production and exogenous dose was fitted using NONMEM74. The production rate (Ksyn), distribution volume (V), plasma clearance (CL) and concentration yielding to 50% of maximal fibrinogen production (EC50) were estimated.

RESULTS

Fibrinogen disposition was described by a one-compartment model with CL and V values of 0.0456 L·h-1 and 4.34 L·70 kg-1 , respectively. Body weight was statistically significant in V. Three different Ksyn values were identified that increased from 0.00439 g·h-1 (afibrinogenaemia), to 0.0768 g·h-1 (cirrhotics) and 0.1160 g·h-1 (acute severe trauma). EC50 was of 0.460 g·L-1 .

CONCLUSIONS

This model will be key as a support tool for dose calculation to achieve specified target fibrinogen concentrations, in each of the studied populations.

International society of Pharmacometrics Mentorship Program (IMP): feedback survey from the first cohort of mentor-mentee pairs

The International Society of Pharmacometrics (ISoP) Mentorship Program (IMP) aims to help professionals at all career stages to transition into the pharmacometrics field, move to a different role/area within pharmacometrics, or expand their skillsets. The program connects mentees at various stages of their careers with mentors based on established criteria for mentor-mentee matching. Pairing mentees with appropriate mentors ensures strong alignment between mentees' interests and mentors' expertise as this is critical to the success and continuation of the relationship between the mentor and mentee. Once mentors and mentees are connected, they are strongly encouraged to meet at least once per month for an hour. The mentor and mentee have the freedom to tailor their sessions to their liking, including frequency, duration, and topics they choose to focus on. Mentees are encouraged to clearly define their goals to help direct their mentor-mentee relationship and conversations. Mentees and mentors alike are given the opportunity to provide feedback about the program to the ISoP Education Committee through surveys and testimonials. Due to the program's infancy, structured guidelines for mentor-mentee sessions are still being developed and instituted using the program evaluation described in this paper.

External validation of a tumor growth inhibition-overall survival model in non-small-cell lung cancer based on atezolizumab studies using alectinib data

BACKGROUND

A modeling framework was previously developed to simulate overall survival (OS) using tumor growth inhibition (TGI) data from six randomized phase 2/3 atezolizumab monotherapy or combination studies in non-small-cell lung cancer (NSCLC). We aimed to externally validate this framework to simulate OS in patients with treatment-naive advanced anaplastic lymphoma kinase (ALK)-positive NSCLC in the alectinib ALEX study.

METHODS

TGI metrics were estimated from a biexponential model using longitudinal tumor size data from a Phase 3 study evaluating alectinib compared with crizotinib in patients with treatment-naive ALK-positive advanced NSCLC. Baseline prognostic factors and TGI metric estimates were used to predict OS.

RESULTS

286 patients were evaluable (at least baseline and one post-baseline tumor size measurements) out of 303 (94%) followed for up to 5 years (cut-off: 29 November 2019). The tumor growth rate estimate and baseline prognostic factors (inflammatory status, tumor burden, Eastern Cooperative Oncology Group performance status, race, line of therapy, and sex) were used to simulate OS in ALEX study. Observed survival distributions for alectinib and crizotinib were within model 95% prediction intervals (PI) for approximately 2 years. Predicted hazard ratio (HR) between alectinib and crizotinib was in agreement with the observed HR (predicted HR 0.612, 95% PI 0.480-0.770 vs. 0.625 observed HR).

CONCLUSION

The TGI-OS model based on unselected or PD-L1 selected NSCLC patients included in atezolizumab trials is externally validated to predict treatment effect (HR) in a biomarker-selected (ALK-positive) population included in alectinib ALEX trial suggesting that TGI-OS models may be treatment independent.

mvLognCorrEst: an R package for sampling from multivariate lognormal distributions and estimating correlations from uncomplete correlation matrix

BACKGROUND AND OBJECTIVE

Pharmacometrics (PMX) is a quantitative discipline which supports decision-making processes in all stages of drug development. PMX leverages Modeling and Simulations (M&S), which represents a powerful tool to characterize and predict the behavior and the effect of a drug. M&S-based methods, such as Sensitivity Analysis (SA) and Global Sensitivity Analysis (GSA), are gaining interest in PMX as they allow the evaluation of model-informed inference quality. Simulations should be correctly designed to obtain reliable results. Neglecting correlations between model parameters can significantly alter the results of simulations. However, the introduction of a correlation structure between model parameters can cause some issues. Sampling from a multivariate lognormal distribution, which is the typically distribution assumed for PMX model parameters, is not straightforward when a correlation structure is introduced. Indeed, correlations need to respect some constraints which depend by the CVs (i.e., coefficients of variation) of lognormal variables. In addition, when correlation matrices have some unspecified values, they should be properly fixed preserving the positive semi-definiteness of the correlation structure. In this paper, we present mvLognCorrEst, an R package developed to address these issues.

METHODS

The proposed sampling strategy was based on reconducting the extraction from the multivariate lognormal distribution of interest to the underlying Normal distribution. However, with high lognormal CVs, a positive semi-definite Normal covariance matrix cannot be obtained due to the violation of some theoretical constraints. In these cases, the Normal covariance matrix was approximated to its nearest positive definite matrix using Frobenius norm as matrix distance. For the estimation of unknown correlations terms, the graph theory was used to represent the correlation structure as weighed undirected graph. Plausible value ranges for the unspecified correlations were derived considering the paths between variables. Then, their estimation was performed by solving a constrained optimization problem.

RESULTS

Package functions are presented and applied on a real case study, that is the GSA of a PMX model that has been recently developed to support preclinical oncological studies.

CONCLUSIONS

mvLognCorrEst package is an R tool to support simulation-based analysis for which sampling from multivariate lognormal distributions with correlated variables and/or estimation of partially defined correlation matrix are required.

A novel approach for pharmacological substantiation of safety signals using plasma concentrations of medication and administrative/healthcare databases: a case study using Danish registries for an FDA warning on lamotrigine

PHARMACOM-EPI is a novel framework to predict plasma concentrations of drugs at the time of occurrence of clinical outcomes. In early 2021, the U.S. Food and Drug Administration (FDA) issued a warning on the antiseizure drug lamotrigine claiming that it has the potential to increase the risk of arrhythmias and related sudden cardiac death due to a pharmacological sodium channel-blocking effect. We hypothesized that the risk of arrhythmias and related death is due to toxicity. We used the PHARMACOM-EPI framework to investigate the relationship between lamotrigine's plasma concentrations and the risk of death in older patients using real-world data. Danish nationwide administrative and healthcare registers were used as data sources and individuals aged 65 years or older during the period 1996 - 2018 were included in the study. According to the PHARMACOM-EPI framework, plasma concentrations of lamotrigine were predicted at the time of death and patients were categorized into non-toxic and toxic groups based on the therapeutic range of lamotrigine (3-15mg/L). Over 1 year of treatment, the incidence rate ratio (IRR) of all-cause mortality was calculated between the propensities score matched toxic and non-toxic groups. In total, 7286 individuals were diagnosed with epilepsy and were exposed to lamotrigine, 432 of which had at least one plasma concentration measurement The pharmacometric model by Chavez et al. was used to predict lamotrigine's plasma concentrations considering the lowest absolute percentage error among identified models (14.25%, 95% CI: 11.68-16.23). The majority of lamotrigine associated deaths were cardiovascular-related and occurred among individuals with plasma concentrations in the toxic range. The IRR of mortality between the toxic group and non-toxic group was 3.37 [95% CI: 1.44-8.32] and the cumulative incidence of all-cause mortality exponentially increased in the toxic range. Application of our novel framework PHARMACOM-EPI provided strong evidence to support our hypothesis that the increased risk of all-cause and cardiovascular death was associated with a toxic plasma concentration level of lamotrigine among older lamotrigine users.

Toward model-informed precision dosing for tamoxifen: A population-pharmacokinetic model with a continuous CYP2D6 activity scale

BACKGROUND

Tamoxifen is important in the adjuvant treatment of breast cancer. A plasma concentration of the active metabolite endoxifen of > 16 nM is associated with a lower risk of breast cancer-recurrence. Since inter-individual variability is high and > 20 % of patients do not reach endoxifen levels > 16 nM with the standard dose tamoxifen, therapeutic drug monitoring is advised. However, ideally, the correct tamoxifen dose should be known prior to start of therapy. Our aim is to develop a population pharmacokinetic (POP-PK) model incorporating a continuous CYP2D6 activity scale to support model informed precision dosing (MIPD) of tamoxifen to determine the optimal tamoxifen starting dose.

METHODS

Data from eight different clinical studies were pooled (539 patients, 3661 samples) and used to develop a POP-PK model. In this model, CYP2D6 activity per allele was estimated on a continuous scale. After inclusion of covariates, the model was subsequently validated using an independent external dataset (378 patients). Thereafter, dosing cut-off values for MIPD were determined.

RESULTS

A joint tamoxifen/endoxifen POP-PK model was developed describing the endoxifen formation rate. Using a continuous CYP2D6 activity scale, variability in predicting endoxifen levels was decreased by 37 % compared to using standard CYP2D6 genotype predicted phenotyping. After external validation and determination of dosing cut-off points, MIPD could reduce the proportion of patients with subtherapeutic endoxifen levels at from 22.1 % toward 4.8 %.

CONCLUSION

Implementing MIPD from the start of tamoxifen treatment with this POP-PK model can reduce the proportion of patients with subtherapeutic endoxifen levels at steady-state to less than 5 %.

Optimal dosing of gliclazide - a model-based approach

Gliclazide was approved as a treatment for type 2 diabetes in an era before model-based drug development and consequently the recommended doses were not optimized with modern methods. To investigate various dosing regimens of gliclazide, we used publicly available data to characterise the dose-response relationship using pharmacometric models. A literature search identified 21 published gliclazide pharmacokinetic (PK) studies with full profiles. These were digitized and a PK model was developed for immediate- (IR) and modified-release (MR) formulations. Data from a gliclazide dose-ranging study of postprandial glucose were used to characterise the concentration-response relationship, using the integrated glucose-insulin model. Simulations from the full model showed that the maximum effect was 44% of the patients achieving HbA1c<7 % with 11% experiencing glucose<3 mmol/L and the most sensitive patients (i.e., 5% most extreme) experiencing 35 min of hypoglycaemia. Simulations revealed that the recommended IR dose (320 mg) was appropriate with no efficacy gain with increased dose. However, the recommended dose for the MR formulation may be increased to 270 mg, with more patients achieving HbA1c goals (i.e., HbA1c<7%) without a hypoglycaemic risk higher than the resulting risk from the recommended IR dose.

Use of Clinical Trial Simulations to Compare the Performance of Different Approaches for Population Analyses of Pediatric Pharmacokinetic Data

Adequate characterization of the pharmacokinetics of a drug in pediatrics is mainstay to pediatric development programs and is critical for accurate dose selection in pediatrics. Analysis approaches can impact estimation and characterization of pediatric pharmacokinetic parameters. Analyses were conducted to compare performance of different approaches for analysis of pediatric pharmacokinetic data in the presence of extensive data from adult studies. Simulated clinical trial datasets were generated encompassing different scenarios which might be encountered in pediatric drug development. For each scenario, 250 clinical trials were simulated and analyzed using each of the following approaches: 1) estimating pediatric parameters using only pediatric data, 2) fixing specific parameters to adult estimates and estimating the remaining pediatric parameters using only pediatric data, 3) estimating pediatric parameters using adult parameters as informative Bayesian priors, 4) estimating pediatric parameters using combined adult and pediatric datasets with exponents for weight and clearance estimated using adult and pediatric data 5) estimating pediatric parameters using combined adult and pediatric datasets with exponents for weight and clearance estimated using pediatric data only. Each analysis approach was evaluated for its success in estimation of true pediatric pharmacokinetic parameter values. Results demonstrated that analyzing pediatric data using a Bayesian approach generally performed best and had the lowest probability of significant bias in the estimated pediatric pharmacokinetic parameters amongst different scenarios evaluated. This clinical trial simulation framework can be used to inform the optimal approach for analyses of pediatric data for other pediatric drug development program scenarios beyond the cases evaluated in these analyses. This article is protected by copyright. All rights reserved.

The convergence of pharmacometrics and quantitative systems pharmacology in pharmaceutical research and development

Quantitative systems pharmacology (QSP) models are an important facet of pharmaceutical and clinical research as they combine mechanistic models of physiology in health and disease with pharmacokinetics/pharmacodynamics to predict systems-level effects. The quantitative clinical pharmacology toolbox has traditionally included both mechanistic modeling and population approaches, collectively called pharmacometrics, but the current landscape requires the optimization and use of multiple models together. Here, we explore several case studies in drug development that exemplify three approaches for using QSP and pharmacometrics models together - parallel synchronization, cross-informative use, and sequential integration. While these approaches are increasingly applied in drug development, achieving a true convergence of QSP and pharmacometrics that fully exploits their synergy will require new tools and methods that enable greater technical integration, in addition to nurturing scientists with diverse modeling expertise that enable cross-discipline strategy. Extensions of existing methods used in each approach as well as additional resources including machine learning models, data-at-scale, end-to-end computation platforms, and real-time analytics will enable this convergence.

Pharmacometric modeling of drug adverse effects: an application of mixture models in schizophrenia spectrum disorder patients treated with clozapine

Clozapine has superior efficacy to other antipsychotics yet is underutilized due to its adverse effects, such as neutropenia, weight gain, and tachycardia. The current investigation aimed to introduce a pharmacometric approach to simultaneously model drug adverse effects, with examples from schizophrenia spectrum patients receiving clozapine. The adverse drug effects were represented as a function of time by incorporating a mixture model to describe individual susceptibility to the adverse effects. Applications of the proposed method were presented by analyzing retrospective data from patients' medical records in Psychiatric Clinic, Penang General Hospital. Tachycardia, weight gain, and absolute neutrophils count (ANC) decrease were best described by an offset, a piecewise linear, and a transient surge function, respectively. 42.9% of the patients had all the adverse effects, including weight gain (0.01 kg/m² increase every week over a baseline of 24.7 kg/m² until stabilizing at 279 weeks), ANC decrease (20% decrease from 4540 cells/µL week 12-20.8), and tachycardia (14% constant increase over a baseline of 87.9 bpm for a clozapine maintenance dose of 450 mg daily). 32.5% of the patients had only tachycardia, while the remaining 24.6% had none of the adverse effects. A new pharmacometric approach was proposed to describe adverse drug effects with examples of clozapine-induced weight gain, ANC drop, and tachycardia. The current approach described the longitudinal time changes of continuous data while assessing patient susceptibility. Furthermore, the model revealed the possible co-existence of ANC drop and weight gain; thus, neutrophil monitoring might predict future changes in body weight.

Computing optimal drug dosing with OptiDose: implementation in NONMEM

Determining a drug dosing recommendation with a PKPD model can be a laborious and complex task. Recently, an optimal dosing algorithm (OptiDose) was developed to compute the optimal doses for any pharmacometrics/PKPD model for a given dosing scenario. In the present work, we reformulate the underlying optimal control problem and elaborate how to solve it with standard commands in the software NONMEM. To demonstrate the potential of the OptiDose implementation in NONMEM, four relevant but substantially different optimal dosing tasks are solved. In addition, the impact of different dosing scenarios as well as the choice of the therapeutic goal on the computed optimal doses are discussed.

A population pharmacokinetic model of remdesivir and its major metabolites based on published mean values from healthy subjects

Remdesivir is a direct-acting anti-viral agent. It was originally evaluated against filoviruses. However, during the COVID-19 pandemic, it was investigated due to its anti-viral activities against (SARS-CoV-2) virus. Therefore remdesivir received conditional approval for treatment of patients with severe coronavirus disease. Yet, its pharmacokinetic properties are inadequately understood. This report describes the population pharmacokinetics of remdesivir and its two plasma-detectable metabolites (GS-704277 and GS-441524) in healthy volunteers. The data was extracted from published phase I single escalating and multiple i.v remdesivir dose studies conducted by the manufacturer. The model was developed by standard methods using non-linear mixed effect modeling. Also, a series of simulations were carried out to test suggested clinical doses. The model describes the distribution of remdesivir and each of its metabolites by respective two compartments with sequential metabolism between moieties, and elimination from central compartments. As individual data were not available, only inter-cohort variability could be assessed. The estimated point estimates for central (and peripheral) volumes of distribution for remdesivir, GS-704277, and GS-441524 were 4.89 L (46.5 L), 96.4 L (8.64 L), and 26.2 L (66.2 L), respectively. The estimated elimination clearances of remdesivir, GS704277, and GS-441524 reached 18.1 L/h, 36.9 L/h, and 4.74 L/h, respectively. The developed model described the data well. Simulations of clinically approved doses showed that GS-441524 concentrations in plasma exceeded the reported EC₅₀ values during the complete duration of treatment. Nonetheless, further studies are needed to explore the pharmacokinetics of remdesivir and its relationship to clinical efficacy, and the present model may serve as a useful starting point for additional evaluations.

Pharmacometrics: The Already-Present Future of Precision Pharmacology

The use of mathematical modeling to represent, analyze, make predictions or providing information on data obtained in drug research and development has made pharmacometrics an area of great prominence and importance. The main purpose of pharmacometrics is to provide information relevant to the search for efficacy and safety improvements in pharmacotherapy. Regulatory agencies have adopted pharmacometrics analysis to justify their regulatory decisions, making those decisions more efficient. Demand for specialists trained in the field is therefore growing. In this review, we describe the meaning, history, and development of pharmacometrics, analyzing the challenges faced in the training of professionals. Examples of applications in current use, perspectives for the future, and the importance of pharmacometrics for the development and growth of precision pharmacology are also presented.

Generative models for age, race/ethnicity, and disease state dependence of physiological determinants of drug dosing

Dosing requires consideration of diverse patient-specific factors affecting drug pharmacokinetics and pharmacodynamics. The available pharmacometric methods have limited capacity for modeling the inter-relationships and patterns of variability among physiological determinants of drug dosing (PDODD). To investigate whether generative adversarial networks (GANs) can learn a generative model from real-world data that recapitulates PDODD distributions. A GAN architecture was developed for modeling a PDODD panel comprised of: age, sex, race/ethnicity, body weight, body surface area, total body fat, lean body weight, albumin concentration, glomerular filtration rate (EGFR), urine flow rate, urinary albumin-to-creatinine ratio, alanine aminotransferase to alkaline phosphatase R-value, total bilirubin, active hepatitis B infection status, active hepatitis C infection status, red blood cell, white blood cell, and platelet counts. The panel variables were derived from National Health and Nutrition Examination Survey (NHANES) data sets. The dependence of GAN-generated PDODD on age, race, and active hepatitis infections was assessed. The continuous PDODD biomarkers had diverse non-normal univariate distributions and bivariate trend patterns. The univariate distributions of PDODD biomarkers from GAN simulations satisfactorily approximated those in test data. The joint distribution of the continuous variables was visualized using three 2-dimensional projection methods; for all three methods, the points from the GAN simulation random variate vectors were well dispersed amongst the test data. The age dependence trend patterns in GAN data were similar to those in test data. The histograms for R-values and EGFR from GAN simulations overlapped extensively with test data histograms for the Hispanic, White, African American, and Other race/ethnicity groups. The GAN-simulated data also mirrored the R-values and EGFR changes in active hepatitis C and hepatitis B infection. GANs are a promising approach for simulating the age, race/ethnicity and disease state dependencies of PDODD.

Comparison of Dose Adjustment Strategies for Obesity in High-dose Cyclophosphamide Among Adult Hematopoietic Cell Transplantation Recipients: Pharmacokinetic Analysis

Cyclophosphamide (CY) is an alkylating agent widely used in the field of oncology and hematopoietic cell transplantation (HCT). It is recommended to use an adjusted body weight with an adjustment factor of 0.25 (ABW25) for dosing of CY in obese patients undergoing HCT. However, evidence based on the pharmacokinetics (PK) of CY to support this recommendation is lacking. We aimed to identify a dosing strategy of CY that achieves equivalent exposures among obese and nonobese patients. The present study is a secondary analysis of a previously conducted observational PK study of phosphoramide mustard (PM), the final cytotoxic metabolite of CY. Data were collected from 85 adults with hematologic malignancy who received a single infusion of CY 50 mg/kg, fludarabine, ± anti-thymocyte globulin, and a single fraction of total body irradiation as HCT conditioning therapy. A previously developed population PK model in these patients was used for simulations. Using individualized PK parameters from that analysis, simulations were performed to assess cumulative exposures of PM (i.e., area-under-the-curve [AUC]) resulting from 8 different dosing strategies according to various measures of body size: (1) "mg/kg" by total body weight (TBW); (2) "mg/kg" by ideal body weight (IBW); (3) "mg/kg" by fat free mass; (4) "mg/m²" by body surface area (BSA); (5) "mg/kg" by TBW combined with ABW25 (TBW-ABW25); (6) "mg/kg" by IBW combined with ABW25 (IBW-ABW25); (7) "mg/kg" by TBW combined with ABW by adjustment factor of 0.50 (TBW-ABW50); and (8) "mg" by fixed-dose. We defined equivalent exposure as the effect of obesity on PM AUC within ±20% from the PM AUC in the nonobese group, where obesity is defined based on TBW/IBW ratio (i.e., nonobese, <1.2; mildly obese, 1.2-1.5; and moderately/severely obese, >1.5). Primary and secondary outcomes were PM AUC_0-8hours and PM AUC_0-infinity, respectively. In the 85 patients, with the median age of 63 years (range 21-75), 46% were classified as mildly and 25% were moderately/severely obese based on the TBW/IBW ratio. Negative correlations (i.e., higher the extent of obesity, lower the PM AUC) were shown when dosing simulations were based on IBW, TBW-ABW25, and fixed dosing (P < .05). Positive correlations were shown when dosing was simulated by TBW (P < .05). None of the 8 dosing strategies attained equivalent PM AUC_0-8hours between patients with versus without obesity, whereas dosing by BSA and TBW-ABW50 attained equivalent PM AUC_0-infinity (P < .05). Our study predicted that the recommended ABW25 dose adjustment may result in lower exposure of CY therapy in obese patients than in nonobese. A CY dosing strategy that would result in similar PM concentrations between obese and nonobese was not identified for early exposure (i.e., PM AUC_0-8hours). The data suggest though that CY dosing based on "mg/m²" by BSA or "mg/kg" by TBW-ABW50 would result in similar total exposure (i.e., PM AUC_0-infinity) and may minimize exposure differences in obese and nonobese patients.

Deciphering longitudinal optical-density measurements to guide clinical dosing regimen design: A model-based approach

BACKGROUND

Model-based analysis of longitudinal optical density measurements from a bacterial suspension exposed to antibiotics has been proposed as a potentially efficient and effective method for extracting useful information to improve the individualized design of treatments for bacterial infections. To that end, the authors developed in previous work a mathematical modeling framework that can use such measurements for design of effective dosing regimens.

OBJECTIVES

Here we further explore ways to extract information from longitudinal optical density measurements to predict bactericidal efficacy of clinically relevant antibiotic exposures.

METHODS

Longitudinal optical density measurements were collected in an automated instrument where Acinetobacter baumannii, ATCC BAA747, was exposed to ceftazidime concentrations of 1, 4, 16, 64, and 256 mg/L and to ceftazidime/amikacin concentrations of 1/0.5, 4/2, 16/8, 64/32, and 256/128 (mg/L)/(mg/L) over 20 h. Calibrated conversion of measurements produced total (both live and dead) bacterial cell concentration data (CFU/mL equivalent) over time. Model-based data analysis predicted the bactericidal efficacy of ceftazidime and of ceftazidime/amikacin (at ratio 2:1) for periodic injection every 8 h and subsequent exponential decline with half-life of 2.5 h. Predictions were experimentally tested in an in vitro hollow-fiber infection model, using peak concentrations of 60 and 150 mg/L for injected ceftazidime and of 40/20 (mg/L)/(mg/L) for injected ceftazidime/amikacin.

RESULTS

Model-based analysis predicted low (<62%) confidence in clinically relevant suppression of the bacterial population by periodic injections of ceftazidime alone, even at high peak concentrations. Conversely, analysis predicted high (>95%) confidence in bacterial suppression by periodic injections of ceftazidime/amikacin combinations at a wide range of peak concentrations ratioed at 2:1. Both predictions were experimentally confirmed in an in vitro hollow fiber infection model, where ceftazidime was periodically injected at peak concentrations 60 and 150 mg/L (with predicted suppression confidence 38% and 59%, respectively) and a combination of ceftazidime/amikacin was periodically injected at peak concentrations 40/20 (mg/L)/(mg/L) (with predicted suppression confidence 98%).

CONCLUSIONS

The paper highlights the potential of clinicians using the proposed mathematical framework to determine the utility of different antibiotics to suppress a patient-specific isolate. Additional studies will be needed to consolidate and expand the utility of the proposed method.

In Silico Prediction of Andrographolide Dosage Regimens for COVID-19 Treatment

Andrographolide (APE) has been used for COVID-19 treatment in various clinical settings in South-East Asia due to its benefits on reduction of viral clearance and prevention of disease progression. However, the limitation of APE clinical use is the high incidence of adverse events. The objective of this study was to find the optimal dosage regimens of APE for COVID-19 treatment. The whole-body physiologically-based pharmacokinetic (PBPK) models were constructed using data from the published articles and validated against clinical observations. The inhibitory effect of APE was determined for the potency of drug efficacy. For prevention of pneumonia, multiple oral doses such as 120[Formula: see text]mg for three doses, followed by 60[Formula: see text]mg three times daily for 4 consecutive days, or 200[Formula: see text]mg intravenous infusion at the rate of 20 mg/h once daily is advised in patients with mild COVID-19. For prevention of pneumonia and reduction of viral clearance time, the recommended dosage regimen is 500[Formula: see text]mg intravenous infusion at the rate of 25[Formula: see text]mg/h once daily in patients with mild-to-moderate COVID-19. One hundred virtual populations (50 males and 50 females) were simulated for oral and intravenous infusion formulations of APE. The eligible PBPK/PD models successfully predicted optimal dosage regimens and formulations of APE for prevention of disease progression and/or reduction of viral clearance time. Additionally, APE should be co-administered with other antiviral drugs to enhance therapeutic efficacy for COVID-19 treatment.

Anti-cancer drug combinations approved by US FDA from 2011 to 2021: main design features of clinical trials and role of pharmacokinetics

During the last decade, the treatment for many cancer indications has evolved due to intensive clinical research into anti-tumor agents' combination. In most instances, new combination treatments consist of an add-on to the standard of care (SOC), which then demonstrate a substantial gain in efficacy and no detrimental effect in tolerability. In the era of targeted therapies, for which maximum tolerated dose (MTD)-based dosing strategies are no longer applicable, early stage studies exploring new combinations are often conducted in the population of interest, expediting the collection of preliminary safety data, to be promptly expanded to collect preliminary efficacy data. Nevertheless, rule-based dose-finding studies are still a prevailing approach for early stage cancer, especially for chemotherapy (CT)-containing combinations. Pharmacokinetic (PK) assessments play a key role throughout the clinical development of drug combinations, informing potential PK interactions. But most importantly, they allow the development of innovative exposure-response (E-R) models aimed at exploring the contribution of each agent to the overall effect of the combination therapy. This review identifies 81 new drug combinations approved by the United States Food and Drug Administration (FDA) for hemato-oncology during the 2011-2021 period and summarizes the main design features of clinical trials and the role of PK assessments.

Population kinetics of homoarginine and optimized supplementation for cardiovascular risk reduction

Homoarginine is an endogenous amino acid whose levels are reduced in patients with renal, cardio- and cerebrovascular disease. Moreover, low homoarginine concentrations independently predict morbidity and mortality in these patients. Besides endogenous synthesis, homoarginine is also a constituent of the human diet. The objective of the present study was to analyze the kinetics of orally supplemented homoarginine in human plasma by means of a pharmacometric approach. We developed a pharmacometric model to evaluate different dosing regimens, especially the regimen of 125 mg once weekly, based on a previous clinical study (n = 20). The model was adapted to account for differences in baseline homoarginine plasma concentrations between healthy and diseased individuals. A novel dosing regimen of 25 mg once daily led to higher attainment of homoarginine reference concentrations using clinical trial simulations. With 25 mg/day, the trough concentration of only 6% of the older and 3.8% of the younger population was predicted to be below the target concentration of 2.0–4.1 µmol/L. In synopsis, the new dosing regimen recapitulates the kinetics of homoarginine in healthy individuals optimally.

Amikacin pharmacokinetics in elderly patients with severe infections

OBJECTIVE

The aim of this study was to characterize the population pharmacokinetics of amikacin in elderly patients by means of nonlinear mixed effects modelling and to propose initial dosing schemes to optimize therapy based on PK/PD targets.

METHOD

A total of 137 elderly patients from 65 to 94 years receiving intravenous amikacin and routine therapeutic drug monitoring at Hospital Universitario Severo Ochoa were included. Concentration-time data and clinical information were retrospectively collected; initial doses of amikacin ranged from 5.7 to 22.5 mg/kg/day and each patient provided between 1 and 10 samples.

RESULTS

Amikacin pharmacokinetics were best described by a two-compartment open model; creatinine clearance (CrCL) was related to drug clearance (2.75 L/h/80 mL/min) and it was augmented 28% when non-steroidal anti-inflammatory drugs were concomitantly administered. Body mass index (BMI) influenced the central volume of distribution (17.4 L/25 kg/m²). Relative absolute prediction error was reduced from 33.2% (base model) to 17.9% (final model) when predictive performance was evaluated with a different group of elderly patients. A nomogram for initial amikacin dosage was developed and evaluated based on stochastic simulations considering final model to achieve PK/PD targets (Cmax/MIC>10 and AUC/MIC>75) and to avoid toxic threshold (Cmin<2.5 mg/L).

CONCLUSION

Initial dosing approach for amikacin was designed for elderly patients based on nonlinear mixed effects modeling to maximize the probability to attain efficacy and safety targets considering individual BMI and CrCL.

Knowledge dissemination and central indexing of resources in pharmacometrics: an ISOP education working group initiative

Pharmacometrics is a constantly evolving field that plays a major role in decision making in drug development and clinical monitoring. Scientists in Pharmacometrics, especially in their early phases of career, are often faced with the challenge of identifying adequate resources for self-training and education. Hence, the ISoP Education Committee through its working group dedicated to Central Indexing and knowledge Dissemination has built a database of worldwide educational programs and most common references in Pharmacometrics.

Therapeutic Drug Monitoring as a Tool for Therapy Optimization

The use of pharmacotherapy for improving healthcare in society is increasing. A vast majority of patients have either received subtherapeutic treatment (which could result from low pharmacokinetic) or experienced adverse effects due to the toxic levels of the drug. The medicines used to treat chronic conditions, such as epilepsy; cardiovascular diseases; and oncological, neurological, and psychiatric disorders, require routine monitoring. New targeted therapies suggest an individualized treatment that can slowly move practitioners away from the concept of a one-size-fits-all-fixed-dosing approach. Therapeutic drug use can be monitored based on pharmacokinetic, pharmacodynamic, and pharmacometric methods. Based on the experiences of therapeutic drug monitoring of various agents across the globe, we can look ahead to the possible developments of therapeutic drug monitoring in India.

Assessment of drug-drug interaction and optimization in capecitabine and irinotecan combination regimen using a physiologically based pharmacokinetic model

Capecitabine and irinotecan (CPT-11) combination regimen (XELIRI) is used for colorectal cancer treatment. Capecitabine is metabolized to 5-fluorouracil (5-FU) by three enzymes, including carboxylesterase (CES). CES can also convert CPT-11 to 7-ethyl-10-hydroxycamptotecin (SN-38). CES is involved in the metabolic activation of both capecitabine and CPT-11, and it is possible that drug-drug interactions occur in XELIRI. Here, a physiologically based pharmacokinetic (PBPK) model was developed to evaluate drug-drug interactions. Capecitabine (180 mg/kg) and CPT-11 (180 mg/m²) were administered to rats, and blood (250 μL) was collected from the jugular vein nine times after administration. Metabolic enzyme activities and K_i values were calculated through in vitro experiments. The plasma concentration of 5-FU in XELIRI was significantly decreased compared to capecitabine monotherapy, and metabolism of capecitabine by CES was inhibited by CPT-11. A PBPK model was developed based on the in vivo and in vitro results. Furthermore, a PBPK model-based simulation was performed with the capecitabin dose ranging from 0 to 1000mol/kg in XELIRI, and it was found that an approximately 1.7-fold dosage of capecitabine was required in XELIRI for comparable 5-FU exposure with capecitabine monotherapy. PBPK model-based simulation will contribute to the optimization of colorectal cancer chemotherapy using XELIRI.

Population pharmacokinetic model selection assisted by machine learning

A fit-for-purpose structural and statistical model is the first major requirement in population pharmacometric model development. In this manuscript we discuss how this complex and computationally intensive task could benefit from supervised machine learning algorithms. We compared the classical pharmacometric approach with two machine learning methods, genetic algorithm and neural networks, in different scenarios based on simulated pharmacokinetic data. Genetic algorithm performance was assessed using a fitness function based on log-likelihood, whilst neural networks were trained using mean square error or binary cross-entropy loss. Machine learning provided a selection based only on statistical rules and achieved accurate selection. The minimization process of genetic algorithm was successful at allowing the algorithm to select plausible models. Neural network classification tasks achieved the most accurate results. Neural network regression tasks were less precise than neural network classification and genetic algorithm methods. The computational gain obtained by using machine learning was substantial, especially in the case of neural networks. We demonstrated that machine learning methods can greatly increase the efficiency of pharmacokinetic population model selection in case of large datasets or complex models requiring long run-times. Our results suggest that machine learning approaches can achieve a first fast selection of models which can be followed by more conventional pharmacometric approaches.

Machine learning-guided, big data-enabled, biomarker-based systems pharmacology: modeling the stochasticity of natural history and disease progression

The incidence of systemic and metabolic co-morbidities increases with aging. The purpose was to investigate a novel paradigm for modeling the orchestrated changes in many disease-related biomarkers that occur during aging. A hybrid strategy that integrates machine learning and stochastic modeling was evaluated for modeling the long-term dynamics of biomarker systems. Bayesian networks (BN) were used to identify quantitative systems pharmacology (QSP)-like models for the inter-dependencies for three disease-related datasets of metabolic (MB), metabolic with leptin (MB-L), and cardiovascular (CVB) biomarkers from the NHANES database. Biomarker dynamics were modeled using discrete stochastic vector autoregression (VAR) equations. BN were used to derive the topological order and connectivity of a data driven QSP model structure for inter-dependence of biomarkers across the lifespan. The strength and directionality of the connections in the QSP models were evaluated using bootstrapping. VAR models based on QSP model structures from BN were assessed for modeling biomarker system dynamics. BN-restricted VAR models of order 1 were identified as parsimonious and effective for characterizing biomarker system dynamics in the MB, MB-L and CVB datasets. Simulation of annual and triennial data for each biomarker provided good fits and predictions of the training and test datasets, respectively. The novel strategy harnesses machine learning to construct QSP model structures for inter-dependence of biomarkers. Stochastic modeling with the QSP models was effective for predicting the age-varying dynamics of disease-relevant biomarkers over the lifespan.

Wide size dispersion and use of body composition and maturation improves the reliability of allometric exponent estimates

To evaluate study designs and the influence of dispersion of body size, body composition and maturation of clearance or reliable estimation of allometric exponents. Non-linear mixed effects modeling and parametric bootstrap were employed to assess how the study sample size, number of observations per subject, between subject variability (BSV) and dispersion of size distribution affected estimation bias and uncertainty of allometric exponents. The role of covariate model misspecification was investigated using a large data set ranging from neonates to adults. A decrease in study sample size, number of observations per subject, an increase in BSV and a decrease in dispersion of size distribution, increased the uncertainty of allometric exponent estimates. Studies conducted only in adults with drugs exhibiting normal (30%) BSV in clearance may need to include at least 1000 subjects to be able to distinguish between allometric exponents of 2/3 and 1. Nevertheless, studies including both children and adults can distinguish these exponents with only 100 subjects. A marked bias of 45% (95%CI 41-49%) in the estimate of the allometric exponent of clearance was obtained when maturation and body composition were ignored in infants. A wide dispersion of body size (e.g. infants, children and adults) is required to reliably estimate allometric exponents. Ignoring differences in body composition and maturation of clearance may bias the exponent for clearance. Therefore, pharmacometricians should avoid estimating allometric exponent parameters without suitable designs and covariate models. Instead, they are encouraged to rely on the well-developed theory and evidence that clearance and volume parameters in humans scale with theory-based exponents.

Population Pharmacokinetic Modelling to Support the Evaluation of Preclinical Pharmacokinetic Experiments with Lorlatinib

The effect of transporters and enzymes on drug pharmacokinetics is increasingly evaluated using genetically modified animals that have these proteins either knocked-out or their human orthologues transgenically expressed. Analysis of pharmacokinetic data obtained in such experiments is typically performed using non-compartmental analysis (NCA), which has limitations such as not being able to identify the PK parameter that is affected by the genetic modification of the enzymes or transporters and the requirement of intense and homogeneous sampling of all subjects. Here we used a compartmental population pharmacokinetic modeling approach using PK data from a series of genetically modified mouse experiments with lorlatinib to extend the results and conclusions from previously reported NCA analyses. A compartmental population pharmacokinetic model was built and physiologically plausible covariates were evaluated for the different mouse strains. With the model, similar effects of the strains on the area under the concentration-time curve (AUC) from 0 to 8 hours were found as for the NCA. Additionally, the differences in AUC between the strains were explained by specific effects on clearance and bioavailability for the strain with human expressing CYP3A4. Finally, effects of multidrug efflux transporters ATP-binding cassette (ABC) sub-family B member 1 (ABCB1) and G member 2 (ABCG2) on brain efflux were quantified. Use of compartmental population PK modeling yielded additional insight into the role of drug-metabolizing enzymes and drug transporters in mouse experiments compared to the NCA. Furthermore, these models allowed analysis of heterogeneous pooled datasets and the sparse organ concentration data in contrast to classical NCA analyses.

Population pharmacokinetic analysis of entrectinib in pediatric and adult patients with advanced/metastatic solid tumors: support of new drug application submission

PURPOSE

Entrectinib (ROZLYTREK^®) is a CNS-active, potent, and selective inhibitor of ROS1, TRK A/B/C, and ALK kinase activity. It was recently approved for the treatment of ROS1-positive non-small cell lung cancer and NTRK gene fusion-positive solid tumors. The main objective of this analysis was to characterize the pharmacokinetics (PK) of entrectinib and its main active metabolite, M5.

METHODS

A total of 276 cancer patients receiving oral entrectinib were included in the analysis. A model-based population approach was used to characterize the PK profiles of both entities using NONMEM^® 7.4. A joint model captures the PK of both entrectinib and M5. The effects of pH modifiers, formulation, weight, age, and sex on model parameters were assessed. Model performance was evaluated using visual predictive checks (VPCs).

RESULTS

The absorption of entrectinib was best described using a sequential zero- and first-order absorption model and the disposition with one-compartment model for each entity with linear elimination. Moderate-to-high between-patient variability was estimated in model parameters (from 30.8% for the apparent clearance of entrectinib to 122% for the first-order absorption rate constant). Theory-based allometric scaling using body weight on clearances and volumes and a 28% lower relative bioavailability of the F1 formulation in pediatric patients were retained in the model. The VPC confirmed the good predictive performance of the PopPK model.

CONCLUSIONS

A robust population PK model was built and qualified for entrectinib and M5, describing linear PK for both entities. This model was used to support the ROZLYTREK^® new drug application.

Evaluation of covariate effects using variance-based global sensitivity analysis in pharmacometrics

In pharmacometrics, understanding a covariate effect on an interested outcome is essential for assessing the importance of the covariate. Variance-based global sensitivity analysis (GSA) can simultaneously quantify contribution of each covariate effect to the variability for the interested outcome considering with random effects. The aim of this study was to apply GSA to pharmacometric models to assess covariate effects. Simulations were conducted with pharmacokinetic models to characterize the GSA for assessment of covariate effects and with an example of quantitative systems pharmacology (QSP) models to apply the GSA to a complex model. In the simulations, covariate and random variables were generated to simulate the outcomes using the models. Ratios of variance explained by each factor (each covariate and random effect) over the overall variance of the outcome were used as sensitivity indices. The sensitivity indices were consistent with the effect size of covariate. The sensitivity indices identified the importance of creatinine clearance on the pharmacokinetic exposure for a renally-excreted drug. These sensitivity indices could be applied to plasma concentrations over time (repeated measurable outcomes over time) as interested outcomes. Using the GSA, each contribution of all of the covariate effects could be efficiently identified even in the complex QSP model. Variance-based GSA can provide insight when considering the importance of covariate effects by simultaneously and quantitatively assessing all covariate and random effects on interested outcomes in pharmacometrics.

Linking categorical models for prediction of pleasantness score using individual predictions of sweetness and creaminess: An advancement of categorical modeling

The aim of this work was to develop and evaluate approaches of linked categorical models using individual predictions of probability. A model was developed using data from a study which assessed the perception of sweetness, creaminess, and pleasantness in dairy solutions containing variable concentrations of sugar and fat. Ordered categorical models were used to predict the individual sweetness and creaminess scores and these individual predictions were used as covariates in the model of pleasantness response. The model using individual predictions was compared to a previously developed model using the amount of fat and sugar as covariates driving pleasantness score. The model using the individual prediction of odds of sweetness and creaminess had a lower variability of pleasantness than the model using the content of sugar and fat in the test solutions, which indicates that the individual odds explain part of the variability in pleasantness. Additionally, simultaneous and sequential modeling approaches were compared for the linked categorical model. Parameter estimation was similar, but precision was better with sequential modeling approaches compared to the simultaneous modeling approach. The previous model characterizing the pleasantness response was improved by using individual predictions of sweetness and creaminess rather than the amount of fat and sugar in the solution. The application of this approach provides an advancement within categorical modeling showing how categorical models can be linked to enable the utilization of individual prediction. This approach is aligned with biology of taste sensory which is reflective of the individual perception of sweetness and creaminess, rather than the amount of fat and sugar in the solution.

Pharmacometric estimation methods for aggregate data, including data simulated from other pharmacometric models

Lack of data is an obvious limitation to what can be modelled. However, aggregate data in the form of means and possibly (co)variances, as well as previously published pharmacometric models, are often available. Being able to use all available data is desirable, and therefore this paper will outline several methods for using aggregate data as the basis of parameter estimation. The presented methods can be used for estimation of parameters from aggregate data, and as a computationally efficient alternative for the stochastic simulation and estimation procedure. They also allow for population PK/PD optimal design in the case when the data-generating model is different from the data-analytic model, a scenario for which no solutions have previously been available. Mathematical analysis and computational results confirm that the aggregate-data FO algorithm converges to the same estimates as the individual-data FO and yields near-identical standard errors when used in optimal design. The aggregate-data MC algorithm will asymptotically converge to the exactly correct parameter estimates if the data-generating model is the same as the data-analytic model. The performance of the aggregate-data methods were also compared to stochastic simulations and estimations (SSEs) when the data-generating model is different from the data-analytic model. The aggregate-data FO optimal design correctly predicted the sampling distributions of 200 models fitted to simulated datasets with the individual-data FO method.

Model-informed precision dosing to optimise immunosuppressive therapy in renal transplantation

Immunosuppressive therapy is pivotal for sustained allograft and patient survival after renal transplantation. However, optimally balanced immunosuppressive therapy is challenged by between-patient and within-patient pharmacokinetic (PK) variability. This could warrant the application of personalised dosing strategies to optimise individual patient outcomes. Pharmacometrics, the science that investigates the xenobiotic-biotic interplay using computer-aided mathematical modelling, provides options to describe and quantify this PK variability and enables identification of patient characteristics affecting immunosuppressant PK and treatment outcomes. Here, we review and critically appraise the available pharmacometric model-informed dosing solutions for the typical immunosuppressants in modern renal transplantation, to guide their initial and subsequent dosing.