Recommended reading in population pharmacokinetic pharmacodynamics

Population pharmacokinetics and pharmacogenetics analyses of imatinib in Chinese patients with chronic myeloid leukemia in a real-world situation

BACKGROUND

Imatinib is presently the first-line choice for the treatment of chronic myeloid leukemia. However, there are limited real-world data on Chinese patients to support individualized medicine. This work aims to characterize population pharmacokinetics in Chinese patients with chronic myeloid leukemia, investigate the effects of several covariates on imatinib exposure, and provide support for personalized medicine and dose reduction.

METHODS

A total of 230 patients with chronic myeloid leukemia were enrolled, and 424 steady-state concentration measurements were taken to perform the population pharmacokinetic analysis and Monte Carlo simulations with Phoenix NLME software. The effects of the demographic, biological, and pharmacogenetic (ten SNP corresponding to CYP3A4, CYP3A5, ABCB1, ABCG2, SCL22A1 and POR) covariates on clearance were evaluated.

RESULTS

A one-compartmental model best-described imatinib pharmacokinetics. The hemoglobin and the estimated glomerular filtration rate (< 85 mL⋅min-1⋅1.73 m2) were associated with imatinib clearance. The genetic polymorphisms related to pharmacokinetics were not found to have a significant effect on the clearance of imatinib. The final model estimates of parameters are: ka (h-1) = 0.329; Vd/F (L) = 270; CL/F (L⋅h-1) = 7.60.

CONCLUSIONS

Key covariates in the study population accounting for variability in imatinib exposure are hemoglobin and the estimated glomerular filtration rate. There is some need for caution when treating patients with moderate-to-severe renal impairment and significant hemoglobin changes.

Non-compartmental and population pharmacokinetic analysis of dapsone in healthy NIGERIANS: A pilot study

Dapsone is employed for both non-dermatological and dermatological indications but with non-existent population pharmacokinetics (popPK) data in Nigerians. This study was therefore designed to develop a popPK model in Nigerians. Non-compartmental analysis and nonlinear mixed effects modelling were utilized for data analysis. Eleven participants administered 50 mg dapsone tablet were included in the analysis. Derived pharmacokinetic parameters were: Cmax  = 1.16 ± 0.32 μg/mL, Tmax  = 3.77 ± 2.40 h, and t1/2z  = 30.23 ± 11.76 h. PopPK model parameter estimates with inter-individual variability were Tlag = 0.40 h (10.0%, fixed); ka  = 1.78 h-1 (75.9%); V/F = 89.25 L (21.6%); and Cl/F = 1.32 Lh-1 (27.7%). Sex was significantly associated with Cl/F, and body weight with V/F. Best popPK model was one-compartment with lag time, and first-order absorption and elimination. Sex and body weight significantly influenced the clearance and distribution volume of dapsone respectively.

The Utility of Mixed Effects Models in the Evaluation of Complex Genomic Traits In Vitro

In pharmacogenomic studies, the use of human liver microsomes as a model system to evaluate the impact of complex genomic traits (i.e., linkage-disequilibrium patterns, coding, and non-coding variation, etc.) on efficiency of drug metabolism is challenging. To accurately predict the true effect size of genomic traits requires large richly sampled datasets representative of the study population. Moreover, the acquisition of this data can be labor-intensive if the study design or bioanalytical methods are not high throughput, and it is potentially unfeasible if the abundance of sample needed for experiments is limited. To overcome these challenges, we developed a novel strategic approach using non-linear mixed effects models (NLME) to determine enzyme kinetic parameters for individual liver specimens using sparse data. This method can facilitate evaluation of the impact that complex genomic traits have on the metabolism of xenobiotics in vitro when tissue and other resources are limited. In addition to facilitating the accrual of data, it allows for rigorous testing of covariates as sources of kinetic parameter variability. In this in silico study, we present a practical application of such an approach using previously published in vitro cytochrome P450 (CYP) 2D6 data and explore the impact of sparse sampling, and experimental error on known kinetic parameter estimates of CYP2D6 mediated formation of 4-hydroxy-atomoxetine in human liver microsomes. SIGNIFICANCE STATEMENT: This study presents a novel non-linear mixed effects model (NLME)-based framework for evaluating the impact of complex genomic traits on saturable processes described by a Michaelis-Menten kinetics in vitro using sparse data. The utility of this approach extends beyond gene variant associations, including determination of covariate effects on in vitro kinetic parameters and reduced demand for precious experimental material.

A mathematical model of the within-host kinetics of SARS-CoV-2 neutralizing antibodies following COVID-19 vaccination

Compelling evidence continues to build to support the idea that SARS-CoV-2 Neutralizing Antibody (NAb) levels in an individual can serve as an important indicator of the strength of protective immunity against infection. It is not well understood why NAb levels in some individuals remain high over time, while in others levels decline rapidly. In this work, we present a two-state mathematical model of within-host NAb dynamics in response to vaccination. By fitting only four host-specific parameters, the model is able to capture individual-specific NAb levels over time as measured by the AditxtScore™ for NAbs. The model can serve as a foundation for predicting NAb levels in the long-term, understanding connections between NAb levels, protective immunity, and breakthrough infections, and potentially guiding decisions about whether and when a booster vaccination may be warranted.

Fitting Pharmacodynamic Data to the Emax Model to Assess the Inherent Potency of Topical Corticosteroids

The widespread use of topical corticosteroids (TCs) in dermatotherapy requires a consideration of their potency and benefit/risk ratios. Although there are a variety of topical corticosteroid products (TCPs) available on the market and their potencies are ranked using different classification systems, to our knowledge, no classification system to rank the inherent potencies of TC active pharmaceutical ingredients (APIs) currently exists. Most of the published classification systems for TCPs are based on randomized clinical comparative studies and/or vasoconstrictor assay (VCA) data. The objective was to apply the US FDA's VCA to classify the inherent potencies of several TCs using standardized doses to make appropriate comparisons of the relevant APIs in solutions of the same molar concentrations. Six TC APIs were assessed for their relative potencies using healthy human participants. The E_max model was used to fit skin blanching data following application of the respective TCs, and the parameters, E_max and ED₅₀, were derived. E_max values were used as the metric to assess potency. Statistical analyses of the data revealed that the inherent potencies of fluticasone propionate, mometasone furoate, and hydrocortisone butyrate were similar. However, there was no significant difference between hydrocortisone butyrate and clobetasol propionate, while there was a significant difference between clobetasol propionate, fluticasone propionate, and mometasone furoate. Hence, the potency of hydrocortisone butyrate appears to overlap two potency classes. Furthermore, the potencies of betamethasone valerate and methylprednisolone aceponate were similar but lower than those of all of the other APIs. The application of the VCA to classify inherent potency provides a reliable method to establish a classification system for TCs. Inherent potency assessment of TCs provides information that will be useful when choosing an appropriate TC for the development of a TCP for a specific clinical indication.

Application of Emax model to assess the potency of topical corticosteroid products

The objective of this study was to compare the potencies of two topical corticosteroid products (TCPs) using the E_max model to fit the skin blanching responses obtained from the US FDA's vasoconstrictor assay (VCA) and to illustrate the influence of formulation on potency. The potencies of two marketed TCPs, Dermovate® cream containing clobetasol propionate (CP) and Elocon® cream containing mometasone furoate (MF), were assessed using healthy human subjects. In order to investigate the influence of formulation and associated vehicle properties, the creams were compared with their respective topical corticosteroids (TCs) from a previously published study wherein the inherent potencies of those TCs were assessed using a validated VCA method. Whereas the inherent potency of MF (E_max  = -94.45 ± 0.21) was found to be greater than CP (E_max  = -58.80 ± 15.65), when formulated as creams, the TCP containing CP had a higher potency (E_max  = -86.15 ± 0.17) than that containing MF (E_max  = -42.61 ± 26.04). This reversal of potency may be attributed to the effect of formulation factors. The comparison of the potencies of TCPs with inherent potencies of their corresponding TCs confirmed the influence of formulation parameters on the potency of those products.

Integration of advanced methods and models to study drug absorption and related processes: An UNGAP perspective

This collection of contributions from the European Network on Understanding Gastrointestinal Absorption-related Processes (UNGAP) community assembly aims to provide information on some of the current and newer methods employed to study the behaviour of medicines. It is the product of interactions in the immediate pre-Covid period when UNGAP members were able to meet and set up workshops and to discuss progress across the disciplines. UNGAP activities are divided into work packages that cover special treatment populations, absorption processes in different regions of the gut, the development of advanced formulations and the integration of food and pharmaceutical scientists in the food-drug interface. This involves both new and established technical approaches in which we have attempted to define best practice and highlight areas where further research is needed. Over the last months we have been able to reflect on some of the key innovative approaches which we were tasked with mapping, including theoretical, in silico, in vitro, in vivo and ex vivo, preclinical and clinical approaches. This is the product of some of us in a snapshot of where UNGAP has travelled and what aspects of innovative technologies are important. It is not a comprehensive review of all methods used in research to study drug dissolution and absorption, but provides an ample panorama of current and advanced methods generally and potentially useful in this area. This collection starts from a consideration of advances in a priori approaches: an understanding of the molecular properties of the compound to predict biological characteristics relevant to absorption. The next four sections discuss a major activity in the UNGAP initiative, the pursuit of more representative conditions to study lumenal dissolution of drug formulations developed independently by academic teams. They are important because they illustrate examples of in vitro simulation systems that have begun to provide a useful understanding of formulation behaviour in the upper GI tract for industry. The Leuven team highlights the importance of the physiology of the digestive tract, as they describe the relevance of gastric and intestinal fluids on the behaviour of drugs along the tract. This provides the introduction to microdosing as an early tool to study drug disposition. Microdosing in oncology is starting to use gamma-emitting tracers, which provides a link through SPECT to the next section on nuclear medicine. The last two papers link the modelling approaches used by the pharmaceutical industry, in silico to Pop-PK linking to Darwich and Aarons, who provide discussion on pharmacometric modelling, completing the loop of molecule to man.

Machine learning in pharmacometrics: Opportunities and challenges

The explosive growth in medical devices, imaging and diagnostics, computing, and communication and information technologies in drug development and healthcare has created an ever-expanding data landscape that the pharmacometrics (PMX) research community must now traverse. The tools of machine learning (ML) have emerged as a powerful computational approach in other data-rich disciplines but its effective utilization in the pharmaceutical sciences and PMX modelling is in its infancy. ML-based methods can complement PMX modelling by enabling the information in diverse sources of big data, e.g. population-based public databases and disease-specific clinical registries, to be harnessed because they are capable of efficiently identifying salient variables associated with outcomes and delineating their interdependencies. ML algorithms are computationally efficient, have strong predictive capabilities and can enable learning in the big data setting. ML algorithms can be viewed as providing a computational bridge from big data to complement PMX modelling. This review provides an overview of the strengths and weaknesses of ML approaches vis-à-vis population methods, assesses current research into ML applications in the pharmaceutical sciences and provides perspective for potential opportunities and strategies for the successful integration and utilization of ML in PMX.

The parameter Houlihan: A solution to high-throughput identifiability indeterminacy for brutally ill-posed problems

One way to interject knowledge into clinically impactful forecasting is to use data assimilation, a nonlinear regression that projects data onto a mechanistic physiologic model, instead of a set of functions, such as neural networks. Such regressions have an advantage of being useful with particularly sparse, non-stationary clinical data. However, physiological models are often nonlinear and can have many parameters, leading to potential problems with parameter identifiability, or the ability to find a unique set of parameters that minimize forecasting error. The identifiability problems can be minimized or eliminated by reducing the number of parameters estimated, but reducing the number of estimated parameters also reduces the flexibility of the model and hence increases forecasting error. We propose a method, the parameter Houlihan, that combines traditional machine learning techniques with data assimilation, to select the right set of model parameters to minimize forecasting error while reducing identifiability problems. The method worked well: the data assimilation-based glucose forecasts and estimates for our cohort using the Houlihan-selected parameter sets generally also minimize forecasting errors compared to other parameter selection methods such as by-hand parameter selection. Nevertheless, the forecast with the lowest forecast error does not always accurately represent physiology, but further advancements of the algorithm provide a path for improving physiologic fidelity as well. Our hope is that this methodology represents a first step toward combining machine learning with data assimilation and provides a lower-threshold entry point for using data assimilation with clinical data by helping select the right parameters to estimate.

Mechanistic machine learning: how data assimilation leverages physiologic knowledge using Bayesian inference to forecast the future, infer the present, and phenotype

We introduce data assimilation as a computational method that uses machine learning to combine data with human knowledge in the form of mechanistic models in order to forecast future states, to impute missing data from the past by smoothing, and to infer measurable and unmeasurable quantities that represent clinically and scientifically important phenotypes. We demonstrate the advantages it affords in the context of type 2 diabetes by showing how data assimilation can be used to forecast future glucose values, to impute previously missing glucose values, and to infer type 2 diabetes phenotypes. At the heart of data assimilation is the mechanistic model, here an endocrine model. Such models can vary in complexity, contain testable hypotheses about important mechanics that govern the system (eg, nutrition's effect on glucose), and, as such, constrain the model space, allowing for accurate estimation using very little data.

Factors Influencing Norvancomycin Concentration in Plasma and Cerebrospinal Fluid in Patients After Craniotomy and Dosing Guideline: A Population Approach

PURPOSE

Antibacterial spectrum and activity of norvancomycin are comparable with vancomycin, and it has been widely used in China. Norvancomycin can penetrate into the cerebrospinal fluid (CSF) through the damaged blood-brain barrier in patients after craniotomy. Because higher inter-individual variability was observed, we aimed to identify factors related to drug concentration to guide clinicians with norvancomycin dosing.

METHODS

After craniotomy, patients with an indwelling catheter in the operational area/ventricle were intravenously administered norvancomycin. Venous blood and CSF specimens were collected at a scheduled time for measuring drug concentrations. Blood and CSF data were fitted simultaneously with the use of the nonlinear fixed-effects modeling method to develop the population pharmacokinetic model. Covariate analysis was applied to select candidate factors associated with pharmacokinetic parameters. A model-based simulation was performed to find optimized regimens for different subgroups of patients.

FINDINGS

A 3-compartmental model (central, peripheral, and CSF compartments) with 2 elimination pathways (drug elimination from the kidney and CSF outflow) was developed to characterize the in vivo process of norvancomycin. The covariate analysis identified that weight and drainage amount were strongly associated with the central volume and the drug clearance from CSF, respectively. Goodness-of-fit and model validation suggested that the proposed model was acceptable. A dosage regimen table was created for specific patient populations with different weights and drainage amounts to facilitate clinical application.

IMPLICATIONS

We identified 2 clinical markers associated with plasma and CSF concentrations. The proposed simulation may be useful to clinicians for norvancomycin dosing in this specific population with normal kidney function.

Understanding and applying pharmacometric modelling and simulation in clinical practice and research

Understanding the dose–concentration–effect relationship is a fundamental component of clinical pharmacology. Interpreting data arising from observations of this relationship requires the use of mathematical models; i.e. pharmacokinetic (PK) models to describe the relationship between dose and concentration and pharmacodynamic (PD) models describing the relationship between concentration and effect. Drug development requires several iterations of pharmacometric model‐informed learning and confirming. This includes modelling to understand the dose–response in preclinical studies, deriving a safe dose for first‐in‐man, and the overall analysis of Phase I/II data to optimise the dose for safety and efficacy in Phase III pivotal trials. However, drug development is not the boundary at which PKPD understanding and application stops. PKPD concepts will be useful to anyone involved in the prescribing and administration of medicines for purposes such as determining off‐label dosing in special populations, individualising dosing based on a measured biomarker (personalised medicine) and in determining whether lack of efficacy or unexpected toxicity maybe solved by adjusting the dose rather than the drug. In clinical investigator‐led study design, PKPD can be used to ensure the optimal dose is used, and crucially to define the expected effect size, thereby ensuring power calculations are based on sound prior information. In the clinical setting the most likely people to hold sufficient expertise to advise on PKPD matters will be the pharmacists and clinical pharmacologists. This paper reviews fundamental PKPD principles and provides some real‐world examples of PKPD use in clinical practice and applied clinical research.

Utility of a Bayesian Mathematical Model to Predict the Impact of Immunogenicity on Pharmacokinetics of Therapeutic Proteins

The impact of an anti-drug antibody (ADA) response on pharmacokinetic (PK) of a therapeutic protein (TP) requires an in-depth understanding of both PK parameters and ADA characteristics. The ADA and PK bioanalytical assays have technical limitations due to high circulating levels of TP and ADA, respectively, hence, significantly hindering the interpretation of this assessment. The goal of this study was to develop a population-based modeling and simulation approach that can identify a more relevant PK parameter associated with ADA-mediated clearance. The concentration-time data from a single dose PK study using five monoclonal antibodies were modeled using a non-compartmental analysis (NCA), one-compartmental, and two-compartmental Michaelis-Menten kinetic model (MMK). A novel PK parameter termed change in clearance time of the TP (α) derived from the MMK model could predict variations in α much earlier than the time points when ADA could be bioanalytically detectable. The model could also identify subjects that might have been potentially identified as false negative due to interference of TP with ADA detection. While NCA and one-compartment models can estimate loss of exposures, and changes in clearance, the two-compartment model provides this additional ability to predict that loss of exposure by means of α. Modeling data from this study showed that the two-compartment model along with the conventional modeling approaches can help predict the impact of ADA response in the absence of relevant ADA data.

Nonlinear mixed-effects modelling for single cell estimation: when, why, and how to use it

Background

Studies of cell-to-cell variation have in recent years grown in interest, due to improved bioanalytical techniques which facilitates determination of small changes with high uncertainty. Like much high-quality data, single-cell data is best analysed using a systems biology approach. The most common systems biology approach to single-cell data is the standard two-stage (STS) approach. In STS, data from each cell is analysed in a separate sub-problem, meaning that only data from the same cell is used to calculate the parameter values within that cell. Because only parts of the data are considered, problems with parameter unidentifiability are exaggerated in STS. In contrast, a related approach to data analysis has been developed for the studies of patient-to-patient variations. This approach, called nonlinear mixed-effects modelling (NLME), makes use of all data, when estimating the patient-specific parameters. NLME would therefore be advantageous compared to STS also for the study of cell-to-cell variation. However, no such systematic evaluation of the two approaches exists.

Results

Herein, such a systematic comparison between STS and NLME has been performed. Different examples, both linear and nonlinear, and both simulated and real experimental data, have been examined. With informative data, there is no significant difference in the results for either parameter or noise estimation. However, when data becomes uninformative, NLME is significantly superior to STS. These results hold independently of whether the loss of information is due to a low signal-to-noise ratio, too few data points, or a bad input signal. The improvement is shown to come from both the consideration of a joint likelihood (JLH) function, describing all parameters and data, and from an a priori postulated form of the population parameters. Finally, we provide a small tutorial that shows how to use NLME for single-cell analysis, using the free and user-friendly software Monolix.

Conclusions

When considering uninformative single-cell data, NLME yields more accurate parameter and noise estimates, compared to more traditional approaches, such as STS and JLH.

Electronic supplementary material

The online version of this article (doi:10.1186/s12918-015-0203-x) contains supplementary material, which is available to authorized users.

Systems Pharmacology of the NGF Signaling Through p75 and TrkA Receptors

The nerve growth factor (NGF) pathway has been shown to play a key role in pain treatment. Recently, a systems pharmacology model has been proposed that can aid in the identification and validation of drug targets in the NGF pathway. However, this model did not include the role of the p75 receptor, which modulates the signaling of NGF through the tropomyosin receptor kinase A (TrkA). The precise mechanism of the interaction between these two receptors has not been completely elucidated, and we therefore adopted a systems pharmacology modeling approach to gain understanding of the effect of p75 on the dynamics of NGF signal transduction. Specifically, models were developed for the so-called heterodimer and for the ligand-passing hypotheses. We used the model to compare the effect of inhibition of NGF and TrkA and its implication for drug discovery and development for pain treatment.

Pharmacokinetic-pharmacodynamic modeling of the effect of varenicline on nicotine craving in adult smokers

INTRODUCTION

Varenicline has been shown to significantly reduce craving and several aspects of smoking reinforcement in clinical trials, compared with placebo. This is the first report describing the concentration-effect relationship of varenicline on relief of craving.

METHODS

The pharmacokinetics (PK) and pharmacodynamics (PD) of a single 2mg dose of varenicline were investigated in 40 smokers in a randomized, crossover study comparing the effect of varenicline with placebo on ameliorating abstinence-and cue-induced craving and withdrawal symptoms. Subjects were asked to complete self-reported questionnaires (Smoking Urges Scale and Minnesota Nicotine Withdrawal Scale [MNWS]) and blood samples were simultaneously collected for measurement of varenicline concentrations. Only the data from the 4-hr postdose abstinence period (just prior to the cue session) were analyzed. Data were described by a 2-compartment PK model and a linear PD model with first-order onset/offset rate constants describing the placebo response "kinetics." Response was described as the net effect of the baseline, placebo, and drug responses.

RESULTS

Varenicline significantly decreased mean craving score when compared with placebo and the magnitude of this response was related to varenicline concentration. The time-course and magnitude of both placebo and varenicline craving response were characterized by a large degree of unexplained variability. Simulations were used to illustrate the expected craving response over time and its associated random variability after chronic dosing.

CONCLUSIONS

Craving reduction is associated with increased varenicline concentrations. The relatively rapid onset of this effect within 4 hr postdose suggests that, smokers may experience some craving relief after acute administration of varenicline.

Pharmacokinetic/pharmacodynamic modelling approaches in paediatric infectious diseases and immunology

Pharmacokinetic/pharmacodynamic (PKPD) modelling is used to describe and quantify dose-concentration-effect relationships. Within paediatric studies in infectious diseases and immunology these methods are often applied to developing guidance on appropriate dosing. In this paper, an introduction to the field of PKPD modelling is given, followed by a review of the PKPD studies that have been undertaken in paediatric infectious diseases and immunology. The main focus is on identifying the methodological approaches used to define the PKPD relationship in these studies. The major findings were that most studies of infectious diseases have developed a PK model and then used simulations to define a dose recommendation based on a pre-defined PD target, which may have been defined in adults or in vitro. For immunological studies much of the modelling has focused on either PK or PD, and since multiple drugs are usually used, delineating the relative contributions of each is challenging. The use of dynamical modelling of in vitro antibacterial studies, and paediatric HIV mechanistic PD models linked with the PK of all drugs, are emerging methods that should enhance PKPD-based recommendations in the future.

Population pharmacokinetics of inhaled tobramycin powder in cystic fibrosis patients

Tobramycin powder for inhalation (TOBI Podhaler or TIP) is approved for the treatment of Pseudomonas aeruginosa airway infection in patients with cystic fibrosis (CF). A population pharmacokinetic model for tobramycin inhalation powder (TIP) in CF patients was developed to characterize the effect of covariates including body mass index (BMI) and lung function (forced expiratory volume in 1 s as percent of the predicted value (FEV₁% predicted) at baseline) on the serum exposure parameters.

A two-compartment model with first-order elimination and first-order absorption was developed. Across a range of baseline demographic values in the study population, the predicted mean values for the maximum (C_max) and trough (C_trough) plasma concentrations at steady state were at least 7.5 and 5-fold lower, respectively, than the recommended thresholds for tobramycin toxicity (12 µg/ml for C_max and 2 µg/ml for C_trough). This model adequately described the tobramycin serum concentration–time course in CF patients following inhalation of TIP. The results indicate that no BMI- or FEV₁-based dose adjustment is needed for use of TIP in CF patients.

Non linear mixed effects analysis in PET PK-receptor occupancy studies

The characterisation of a pharmacokinetic-receptor occupancy (PK-RO) relationship derived from a PET study is typically modelled in a conventional non-linear least squares (NLLS) framework. In the present work, we explore the application of a non-linear mixed effects approach (NLME) and compare this with NLLS estimation (using both naive pooled data and two-stage approaches) in the context of a direct PK-RO relationship described by an Emax model, using simulated data sets. Target and reference tissue time-activity curves were simulated using a two-tissue compartmental model and an arterial plasma input function for a typical PET study (12 subjects in 3 dose groups with 3 scans each). A range of different PET scenarios was considered to evaluate the impact of between-subject variability and reference region availability. The PET outcome measures derived from the simulations were then used to estimate the parameters of the PK-RO model. The performance of the two approaches was compared in terms of parameters estimates (square mean error SME, root mean square error RMSE) and prediction of the exposure-occupancy relationship. In general, both NLME and NLLS estimation methods provided unbiassed and precise population estimates for the Emax model parameters, although a slight bias was observed for the individual-NLLS method due to a few outliers. The increased value of NLME over NLLS was most notable in the estimation of the between-subject variability (BSV), especially in the case of a more complex PK-RO model when no reference region was available (maximum SME and RMSE values related to BSV of EC₅₀ of 27.6% and 86.5% from NLME versus 264.6% and 689.5% from NLLS). Overall, the NLME approach provided a more robust estimation and produced less-biassed estimates of the population means and variances than either the NLLS approach for the simulations considered.

Clinical trials in neonatal sepsis

Antibiotic licensing studies remain a problem in neonates. The classical adult clinical syndrome-based licensing studies do not apply to neonates, where sepsis is the most common infection. The main obstacle to conducting neonatal antibiotic trials is a lack of consensus on the definition of neonatal sepsis itself and the selection of appropriate endpoints. This article describes the difficulties of the clinical and laboratory definitions of neonatal sepsis and reviews the varying designs of previous neonatal sepsis trials. The optimal design of future trials of new antibiotics will need to be based on pharmacokinetic/pharmacodynamic parameters, combined with adequately powered clinical studies to determine safety and efficacy.

Estimating insulin sensitivity from glucose levels only: Use of a non-linear mixed effects approach and maximum a posteriori (MAP) estimation

Insulin Sensitivity is an important parameter for the management of Diabetes. It can be derived for a particular patient using data derived from some glucose challenge tests using measured glucose and insulin levels at various times. Whilst a useful approach, deriving insulin sensitivities to inform insulin dosing in other settings such as Intensive Care Units can be more challenging - especially as insulin levels have to be assayed in a laboratory, not at the bedside. This paper investigates an approach to measure insulin sensitivity from glucose levels only. Estimates of mean and between individual parameter variances are used to derive conditional estimates of insulin sensitivity. The method is demonstrated to perform reasonably well, with conditional estimates comparing well with estimates derived from insulin data as well.

Quantifying short-term dynamics of Parkinson's disease using self-reported symptom data from an Internet social network

Background

Parkinson’s disease (PD) is an incurable neurological disease with approximately 0.3% prevalence. The hallmark symptom is gradual movement deterioration. Current scientific consensus about disease progression holds that symptoms will worsen smoothly over time unless treated. Accurate information about symptom dynamics is of critical importance to patients, caregivers, and the scientific community for the design of new treatments, clinical decision making, and individual disease management. Long-term studies characterize the typical time course of the disease as an early linear progression gradually reaching a plateau in later stages. However, symptom dynamics over durations of days to weeks remains unquantified. Currently, there is a scarcity of objective clinical information about symptom dynamics at intervals shorter than 3 months stretching over several years, but Internet-based patient self-report platforms may change this.

Objective

To assess the clinical value of online self-reported PD symptom data recorded by users of the health-focused Internet social research platform PatientsLikeMe (PLM), in which patients quantify their symptoms on a regular basis on a subset of the Unified Parkinson’s Disease Ratings Scale (UPDRS). By analyzing this data, we aim for a scientific window on the nature of symptom dynamics for assessment intervals shorter than 3 months over durations of several years.

Methods

Online self-reported data was validated against the gold standard Parkinson’s Disease Data and Organizing Center (PD-DOC) database, containing clinical symptom data at intervals greater than 3 months. The data were compared visually using quantile-quantile plots, and numerically using the Kolmogorov-Smirnov test. By using a simple piecewise linear trend estimation algorithm, the PLM data was smoothed to separate random fluctuations from continuous symptom dynamics. Subtracting the trends from the original data revealed random fluctuations in symptom severity. The average magnitude of fluctuations versus time since diagnosis was modeled by using a gamma generalized linear model.

Results

Distributions of ages at diagnosis and UPDRS in the PLM and PD-DOC databases were broadly consistent. The PLM patients were systematically younger than the PD-DOC patients and showed increased symptom severity in the PD off state. The average fluctuation in symptoms (UPDRS Parts I and II) was 2.6 points at the time of diagnosis, rising to 5.9 points 16 years after diagnosis. This fluctuation exceeds the estimated minimal and moderate clinically important differences, respectively. Not all patients conformed to the current clinical picture of gradual, smooth changes: many patients had regimes where symptom severity varied in an unpredictable manner, or underwent large rapid changes in an otherwise more stable progression.

Conclusions

This information about short-term PD symptom dynamics contributes new scientific understanding about the disease progression, currently very costly to obtain without self-administered Internet-based reporting. This understanding should have implications for the optimization of clinical trials into new treatments and for the choice of treatment decision timescales.

Interpreting population pharmacokinetic-pharmacodynamic analyses - a clinical viewpoint

The population analysis approach is an important tool for clinical pharmacology in aiding the dose individualization of medicines. However, due to their statistical complexity the clinical utility of population analyses is often overlooked. One of the key reasons to conduct a population analysis is to investigate the potential benefits of individualization of drug dosing based on patient characteristics (termed covariate identification). The purpose of this review is to provide a tool to interpret and extract information from publications that describe population analysis. The target audience is those readers who are aware of population analyses but have not conducted the technical aspects of an analysis themselves. Initially we introduce the general framework of population analysis and work through a simple example with visual plots. We then follow-up with specific details on how to interpret population analyses for the purpose of identifying covariates and how to interpret their likely importance for dose individualization.

Understanding the time course of pharmacological effect: a PKPD approach

The key concepts that underpin the choice of drug and dosing regimen are an understanding of the drugs' effectiveness, the potential for adverse effects, and the expected time course over which both desired and adverse effects are likely to occur. Research in clinical pharmacology should therefore address three fundamental questions: (1) What is the magnitude of drug effects (beneficial or adverse) from a given dose? (2) How quickly will any given effects occur? (3) How long will these effects last? Under steady-state conditions, only the magnitude of drug effects can be examined. This requires researchers to consider non-steady-state conditions, which require more complex models and an understanding of the mechanisms that drive the time course of drug effect. The aim of this review is to provide a conceptual framework for understanding the time course of drug effects using pharmacokinetic-pharmacodynamic models. Key examples will illustrate how this can inform the optimal use of drugs in the clinic.

Population pharmacokinetics of therapeutic monoclonal antibodies

A growing number of population pharmacokinetic analyses of therapeutic monoclonal antibodies (mAbs) have been published in the scientific literature. The aims of this article are to summarize the findings from these studies and to relate the findings to the general pharmacokinetic and structural characteristics of therapeutic mAbs. A two-compartment model was used in the majority of the population analyses to describe the disposition of the mAb. Population estimates of the volumes of distribution in the central (V(1)) and peripheral (V(2)) compartments were typically small, with median (range) values of 3.1 (2.4-5.5) L and 2.8 (1.3-6.8) L, respectively. The estimated between-subject variability in the V(1) was usually moderate, with a median (range) coefficient of variation (CV) of 26% (12-84%). Between-subject variability in other distribution-related parameters such as the V(2) and intercompartmental clearance were often not estimated. Although the pharmacokinetic models used most frequently in the population analyses were models with linear clearance, other models with nonlinear, or parallel linear and nonlinear clearance pathways were also applied, as many therapeutic mAbs are eliminated via saturable target-mediated mechanisms. Population estimates of the maximum elimination rate (V(max)) and the mAb concentration at which elimination was at half maximum for Michaelis-Menten-type elimination pathways varied considerably among the different therapeutic mAbs. However, estimates of the total clearance (CL) of mAbs with linear clearance characteristics and of the clearance of mAbs via the linear clearance pathway (CL(L)) with parallel linear and nonlinear clearance were quite similar for the different mAbs and typically ranged from 0.2 to 0.5 L/day, which is relatively close to the estimated clearance of endogenous IgG of 0.21 L/day. The between-subject variability in the V(max), CL and CL(L) was moderate to high, with estimated CVs ranging from 15% to 65%. Measures of body size were the covariates most commonly identified as influencing the pharmacokinetics of therapeutic mAbs. In summary, many features of the population pharmacokinetics of currently used therapeutic mAbs are similar, despite differences in their pharmacological targets and studied patient populations.

Measuring the glomerular filtration rate in obese individuals without overt kidney disease

BACKGROUND

Identifying methods to accurately measure the glomerular filtration rate (GFR) in obese individuals without kidney overt kidney disease is necessary to understanding the pathophysiology and natural history of obesity-related kidney disease.

METHODS

Using a cross-sectional design, iohexol clearance and disposition was measured, an optimal sampling schedule was identified, and the reliability of GFR-estimating methods was described in 29 obese individuals with normal serum creatinine levels. Iohexol disposition was measured using population pharmacokinetics. The agreement with GFR-estimating equations was assessed by intraclass coefficients.

RESULTS

Mean age was 44 ± 10 years, body mass index 45 ± 10, creatinine 0.7 ± 0.2 mg/dl (62 ± 18 μmol/l) , and cystatin C 0.83 ± 0.18 mg/dl (8.3 ± 1.8 mg/l). Iohexol disposition fit a two-compartment model and 5 sampling windows were identified over a 4-hour period to optimize model accuracy and minimize blood draws. Precision was not compromised with this sampling design. Neither creatinine nor cystatin C were linearly correlated with the measured GFR though cystatin C was independent of body composition. Agreement was fair to poor between the measured GFR and GFR-estimating equations.

CONCLUSION

This study offers a rigorous method to study obesity-related kidney disease and improve upon suboptimal GFR-estimating methods.

Pharmacokinetics of clindamycin in pregnant women in the peripartum period

The study presented here was performed to determine the pharmacokinetics of intravenously administered clindamycin in pregnant women. Seven pregnant women treated with clindamycin were recruited. Maternal blood and arterial and venous umbilical cord blood samples were obtained. Maternal clindamycin concentrations were analyzed by nonlinear mixed-effects modeling with the NONMEM program. The data were best described by a linear three-compartment model. The clearance and the volume of distribution at steady state were 10.0 liters/h and 6.32 x 10(3) liters, respectively. Monte Carlo simulations were performed to determine the area under the concentration curve (AUC) for the free (unbound) drug (f) in maternal serum for 24 h divided by the MIC (fAUC(0-24)/MIC). At a MIC of 0.5 mg/liter, which is the EUCAST breakpoint, the attainment at the lower 95% confidence interval (CI) was 24.6 if the level of protein binding was 65%, and this value concurred well with the target value of 27. However, for higher degrees of protein binding, as has been described in the literature, the attainment was lower, down to 10.2 for a protein binding level of 85% (lower 95% CI). The concentrations in umbilical cord blood were lower than those in maternal blood. The concentration-time profiles in maternal serum indicate that the level of exposure to clindamycin may be too low in these patients. Together with the lower concentrations in umbilical cord blood, this finding suggests that the current dosing regimen may not be adequate to protect all neonates from group B streptococcal disease.

The influence of labour on the pharmacokinetics of intravenously administered amoxicillin in pregnant women

AIMS

Many physiological changes take place during pregnancy and labour. These might change the pharmacokinetics of amoxicillin, necessitating adjustment of the dose for prevention of neonatal infections. We investigated the influence of labour on the pharmacokinetics of amoxicillin.

METHODS

Pregnant women before and during labour were recruited and treated with amoxicillin intravenously. A postpartum dose was offered. Blood samples were obtained and amoxicillin concentrations were determined using high-pressure liquid chromatography. The pharmacokinetics were characterized by nonlinear mixed-effects modelling using NONMEM.

RESULTS

The pharmacokinetics of amoxicillin in 34 patients was best described by a three-compartment model. Moderate interindividual variability was identified in CL, central and peripheral volumes of distribution. The volume of distribution (V) increased with an increasing amount of oedema. Labour influenced the parameter estimate of peripheral volume of distribution (V(2)). V(2) was decreased during labour, and even more in the immediate postpartum period. For all patients the population estimates (mean +/- SE) for CL and V were 21.1 +/- 4.1 l h(-1) (CL), 8.7 +/- 6.6 l (V(1)), 11.8 +/- 7.7 l (V(2)) and 20.5 +/- 15.4 l (V(3)) respectively.

CONCLUSIONS

The peripheral distribution volume of amoxicillin in pregnant women during labour and immediately postpartum is decreased. However, these changes are not clinically relevant and do not warrant deviations from the recommended dosing regimen for amoxicillin during labour in healthy pregnant patients.

Pharmacokinetics of amoxicillin in maternal, umbilical cord, and neonatal sera

The pharmacokinetics of amoxicillin were studied in umbilical cord and neonatal sera relative to maternal concentrations in prevention of neonatal group B streptococcus infection. The subjects were 44 pregnant women receiving amoxicillin as 1 or 2 g as an intravenous infusion. To measure the concentrations, blood samples were obtained from the mother, the arterial and venous umbilical cord, and the neonate. The pharmacokinetics were characterized by a five-compartment model by using nonlinear mixed-effects (population) modeling. The population estimates for the clearance, central volume of distribution, and the two peripheral maternal volumes of distribution were 19.7 +/- 0.99 liters/h, 6.40 +/- 0.61 liters, and 5.88 +/- 0.83 liters (mean +/- standard error), respectively. The volume of distribution of the venous umbilical cord and the neonatal volume of distribution were 3.40 liters and 11.9 liters, respectively. The pharmacokinetic parameter estimates were used to simulate the concentration-time profiles in maternal, venous umbilical cord, and neonatal sera. The peak concentration in the venous umbilical cord serum was 18% of the maternal peak concentration. It was reached 3.3 min after the maternal peak concentration. The concentration-time profile in neonatal serum was determined by the profile in venous umbilical cord serum, which in turn depended on the profile in maternal serum. Furthermore, the simulated concentrations in maternal, venous umbilical cord, and neonatal sera exceeded the MIC for group B streptococcus for more than 90% of the 4-h dosing interval. In a first approximation, the 2-g infusion to the mother appears to be adequate for the prevention of group B streptococcal disease. However, to investigate the efficacy of the prophylaxis, further studies of the interindividual variability in pharmacokinetics are indicated.

Amoxicillin pharmacokinetics in pregnant women with preterm premature rupture of the membranes

OBJECTIVE

This study was undertaken to study the pharmacokinetics of intravenously administered amoxicillin in pregnant women with preterm premature rupture of the membranes (PPROM).

STUDY DESIGN

Healthy women with PPROM were recruited and treated with amoxicillin (2 g initially and 1 g subsequently). Blood samples were obtained from the opposite arm and concentrations determined with the use of high-pressure liquid chromatography. Nonlinear mixed-effects modeling was performed in nonlinear mixed effect (population) modeling.

RESULTS

The pharmacokinetics of 17 patients was described by a 3-compartment model. Clearance and volume of distribution at steady state were 22.8 L/h and 21.4 L/h, respectively, similar to values in nonpregnant individuals. There was little variability between patients. No relationship was observed between values of individual pharmacokinetic parameters and various covariates.

CONCLUSION

The pharmacokinetics of amoxicillin in pregnant patients with PPROM similar to nonpregnant individuals. Given the small interindividual variability in pharmacokinetics, no dose adjustments are required to account for differences between subjects under normal circumstances.

Integrated pharmacokinetics and pharmacodynamics in drug development

Drug development is a complex, lengthy and expensive process. Pharmaceutical companies and regulatory authorities have recognised that the drug development process needs optimisation for efficiency in view of the return on investments. Pharmacokinetics and pharmacodynamics are the two main principles determining the relationship between dose and response. This article provides an update on integrated approaches towards drug development by linking pharmacokinetics, pharmacodynamics and disease aspects into mathematical models. Gradually, a transition is taking place from a rather empirical approach towards a modelling- and simulation-based approach to drug development. The main learning phases should be phases 0, I and II, whereas phase III studies should merely have a confirmatory purpose. In model-based drug development, mechanism-based mathematical models, which are iteratively refined along the path of development, incorporate the accumulating knowledge of the investigational drug, the disease and their mutual interference in different subsets of the target population. These models facilitate the design of the next study and improve the probability of achieving the projected efficacy and safety endpoints. In this article, several theoretical and practical aspects of an integrated approach towards drug development are discussed, together with some case studies from different therapeutic areas illustrating the application of pharmacokinetic/pharmacodynamic disease models at different stages of drug development.

Overview of model-building strategies in population PK/PD analyses: 2002-2004 literature survey

AIMS

A descriptive survey of published population pharmacokinetic and/or pharmacodynamic (PK/PD) analyses from 2002 to 2004 was conducted and an evaluation made of how model building was performed and reported.

METHODS

We selected 324 articles in Pubmed using defined keywords. A data abstraction form (DAF) was then built comprising two parts: general characteristics including article identification, context of the analysis, description of clinical studies from which the data arose, and model building, including description of the processes of modelling. The papers were examined by two readers, who extracted the relevant information and transmitted it directly to a MySQL database, from which descriptive statistical analysis was performed.

RESULTS

Most published papers concerned patients with severe pathology and therapeutic classes suffering from narrow therapeutic index and/or high PK/PD variability. Most of the time, modelling was performed for descriptive purposes, with rich rather than sparse data and using NONMEM software. PK and PD models were rarely complex (one or two compartments for PK; E(max) for PD models). Covariate testing was frequently performed and essentially based on the likelihood ratio test. Based on a minimal list of items that should systematically be found in a population PK-PD analysis, it was found that only 39% and 8.5% of the PK and PD analyses, respectively, published from 2002 to 2004 provided sufficient detail to support the model-building methodology.

CONCLUSIONS

This survey allowed an efficient description of recent published population analyses, but also revealed deficiencies in reporting information on model building.