Impact of covariate model building methods on their clinical relevance evaluation in population pharmacokinetic analyses: comparison of the full model, stepwise covariate model (SCM) and SCM+ approaches.

Covariate analysis in population pharmacokinetics is key for adjusting doses for patients. The main objective of this work was to compare the adequacy of various modeling approaches on covariate clinical relevance decision-making. The full model, stepwise covariate model (SCM) and SCM+ PsN algorithms were compared in a clinical trial simulation of a 383-patient population pharmacokinetic study mixing rich and sparse designs. A one-compartment model with first-order absorption was used. A base model including a body weight effect on CL/F and V/F and a covariate model including 4 additional covariates-parameters relationships were simulated. As for forest plots, ratios between covariates at a specific value and that of a typical individual were calculated with their 90% confidence interval (CI90) using standard errors. Covariates on CL, V and KA were considered relevant if their CI90 fell completely outside the reference area [0.8-1.2]. All approaches provided unbiased covariate ratio estimates. For covariates with a simulated effect, the 3 approaches correctly identify their clinical relevance. However, significant covariates were missed in up to 15% of cases with SCM/SCM+. For covariate with no simulated effects, the full model mainly identified them as non-relevant or with insufficient information while SCM/SCM+ mainly did not select them. SCM/SCM+ assume that non-selected covariates are non-relevant when it could be due to insufficient information, whereas the full model does not make this assumption and is faster. This study must be extended to other methods and completed by a more complex high-dimensional simulation framework.

Prediction of Individual Disease Progression Including Parameter Uncertainty in Rare Neurodegenerative Diseases: The Example of Autosomal-Recessive Spastic Ataxia Charlevoix Saguenay (ARSACS).

The aim of this study was to develop a model to predict individual subject disease trajectories including parameter uncertainty and accounting for missing data in rare neurological diseases, showcased by the ultra-rare disease Autosomal-Recessive Spastic Ataxia Charlevoix Saguenay (ARSACS). We modelled the change in SARA (Scale for Assessment and Rating of Ataxia) score versus Time Since Onset of symptoms using non-linear mixed effect models for a population of 173 patients with ARSACS included in the prospective real-world multicenter Autosomal Recessive Cerebellar Ataxia (ARCA) registry. We used the Multivariate Imputation Chained Equation (MICE) algorithm to impute missing covariates, and a covariate selection procedure with a pooled p-value to account for the multiply imputed data sets. We then investigated the impact of covariates and population parameter uncertainty on the prediction of the individual trajectories up to 5 years after their last visit. A four-parameter logistic function was selected. Men were estimated to have a 25% lower SARA score at disease onset and a moderately higher maximum SARA score, and time to progression (T50) was estimated to be 35% lower in patients with age of onset over 15 years. The population disease progression rate started slowly at 0.1 points per year peaking to a maximum of 0.8 points per year (at 36.8 years since onset of symptoms). The prediction intervals for SARA scores 5 years after the last visit were large (median 7.4 points, Q1-Q3: 6.4-8.5); their size was mostly driven by individual parameter uncertainty and individual disease progression rate at that time.

Cardiac Adverse Events and Remdesivir in Hospitalized Patients with Coronavirus Disease 2019 (COVID-19): A Post Hoc Safety Analysis of the Randomized DisCoVeRy Trial.

BACKGROUND: We aimed to evaluate the cardiac adverse events (AEs) in hospitalized patients with Coronavirus Disease 2019 (COVID-19) receiving remdesivir plus standard of care (SoC) compared to SoC alone (control), as an association was noted in some cohort studies and disproportionality analyses of safety databases. METHODS: This post-hoc safety analysis is based on data from the multicenter, randomized, open-label, controlled DisCoVeRy trial in hospitalized patients with COVID-19 (NCT04315948). Any first AE occurring between randomization and day 29 in the modified intention-to-treat (mITT) population randomized to either remdesivir or control group was considered. Analysis was performed using Kaplan-Meier survival curves and Kaplan-Meier estimates were calculated for event rates. RESULTS: Cardiac AEs were reported in 46 (11.2%) of 410 and 48 (11.3%) of 423 patients in the mITT population (n = 833) enrolled in the remdesivir and control groups, respectively. The difference between both groups was not significant (HR 1.0, 95% CI 0.7-1.5, p = 0.98), even when evaluating serious and non-serious cardiac AEs separately. The majority of reports in both groups were of arrhythmic nature (remdesivir, 84.8%; control, 83.3%) and were associated with a favorable outcome. There was no significant difference between remdesivir and control groups in the occurrence of different cardiac AE subclasses, including arrhythmic events (HR 1.1, 95% CI: 0.7-1.7, p = 0.68). CONCLUSIONS: Remdesivir treatment was not associated with an increased risk of cardiac AEs, whether serious or not, and regardless of AE severity, compared to control, in patients hospitalized with moderate or severe COVID-19. This is consistent with the results of other randomized controlled trials and meta-analyses.

Distributive randomization: a pragmatic fractional factorial design to screen or evaluate multiple simultaneous interventions in a clinical trial.

BACKGROUND: In some medical indications, numerous interventions have a weak presumption of efficacy, but a good track record or presumption of safety. This makes it feasible to evaluate them simultaneously. This study evaluates a pragmatic fractional factorial trial design that randomly allocates a pre-specified number of interventions to each participant, and statistically tests main intervention effects. We compare it to factorial trials, parallel-arm trials and multiple head-to-head trials, and derive some good practices for its design and analysis. METHODS: We simulated various scenarios involving 4 to 20 candidate interventions among which 2 to 8 could be simultaneously allocated. A binary outcome was assumed. One or two interventions were assumed effective, with various interactions (positive, negative, none). Efficient combinatorics algorithms were created. Sample sizes and power were obtained by simulations in which the statistical test was either difference of proportions or multivariate logistic regression Wald test with or without interaction terms for adjustment, with Bonferroni multiplicity-adjusted alpha risk for both. Native R code is provided without need for compiling or packages. RESULTS: Distributive trials reduce sample sizes 2- to sevenfold compared to parallel arm trials, and increase them 1- to twofold compared to factorial trials, mostly when fewer allocations than for the factorial design are possible. An unexpectedly effective intervention causes small decreases in power (< 10%) if its effect is additive, but large decreases (possibly down to 0) if not, as for factorial designs. These large decreases are prevented by using interaction terms to adjust the analysis, but these additional estimands have a sample size cost and are better pre-specified. The issue can also be managed by adding a true control arm without any intervention. CONCLUSION: Distributive randomization is a viable design for mass parallel evaluation of interventions in constrained trial populations. It should be introduced first in clinical settings where many undercharacterized interventions are potentially available, such as disease prevention strategies, digital behavioral interventions, dietary supplements for chronic conditions, or emerging diseases. Pre-trial simulations are recommended, for which tools are provided.

Perturbation and resilience of the gut microbiome up to 3 months after ß-lactams exposure in healthy volunteers suggest an important role of microbial ß-lactamases.

BACKGROUND: Antibiotics notoriously perturb the gut microbiota. We treated healthy volunteers either with cefotaxime or ceftriaxone for 3 days, and collected in each subject 12 faecal samples up to day 90. Using untargeted and targeted phenotypic and genotypic approaches, we studied the changes in the bacterial, phage and fungal components of the microbiota as well as the metabolome and the ß-lactamase activity of the stools. This allowed assessing their degrees of perturbation and resilience. RESULTS: While only two subjects had detectable concentrations of antibiotics in their faeces, suggesting important antibiotic degradation in the gut, the intravenous treatment perturbed very significantly the bacterial and phage microbiota, as well as the composition of the metabolome. In contrast, treatment impact was relatively low on the fungal microbiota. At the end of the surveillance period, we found evidence of resilience across the gut system since most components returned to a state like the initial one, even if the structure of the bacterial microbiota changed and the dynamics of the different components over time were rarely correlated. The observed richness of the antibiotic resistance genes repertoire was significantly reduced up to day 30, while a significant increase in the relative abundance of ß-lactamase encoding genes was observed up to day 10, consistent with a concomitant increase in the ß-lactamase activity of the microbiota. The level of ß-lactamase activity at baseline was positively associated with the resilience of the metabolome content of the stools. CONCLUSIONS: In healthy adults, antibiotics perturb many components of the microbiota, which return close to the baseline state within 30 days. These data suggest an important role of endogenous ß-lactamase-producing anaerobes in protecting the functions of the microbiota by de-activating the antibiotics reaching the colon. Video Abstract.

Extending the code in the open-source saemix package to fit joint models of longitudinal and time-to-event data.

BACKGROUND AND OBJECTIVE: Joint modeling of longitudinal and time-to-event data has gained attention over recent years with extensive developments including nonlinear models for longitudinal outcomes and flexible time-to-event models for survival outcomes, possibly involving competing risks. However, in popular software such as R, the function used to describe the biomarker dynamic is mainly linear in the parameters, and the survival submodel relies on pre-implemented functions (exponential, Weibull, ...). The objective of this work is to extend the code from the saemix package (version 3.1 on CRAN) to fit parametric joint models where longitudinal submodels are not necessary linear in their parameters, with full user control over the model function. METHODS: We used the saemix package, designed to fit nonlinear mixed-effects models (NLMEM) through the Stochastic Approximation Expectation Maximization (SAEM) algorithm, and extended the main functions to joint model estimation. To compute standard errors (SE) of parameter estimates, we implemented a recently developed stochastic algorithm. A simulation study was proposed to assess (i) the performances of parameter estimation, (ii) the SE computation and (iii) the type I error when testing independence between the two submodels. Four joint models were considered in the simulation study, combining a linear or nonlinear mixed-effects model for the longitudinal submodel, with a single terminal event or a competing risk model. RESULTS: For all simulation scenarios, parameters were precisely and accurately estimated with low bias and uncertainty. For complex joint models (with NLMEM), increasing the number of chains of the algorithm was necessary to reduce bias, but earlier censoring in the competing risk scenario still challenged the estimation. The empirical SE of parameters obtained over all simulations were very close to those computed with the stochastic algorithm. For more complex joint models (involving NLMEM), some estimates of random effects variances had higher uncertainty and their SE were moderately under-estimated. Finally, type I error was controlled for each joint model. CONCLUSIONS: saemix is a flexible open-source package and we adapted it to fit complex parametric joint models that may not be estimated using standard tools. Code and examples to help users get started are freely available on Github.

Multivariate joint model under competing risks to predict death of hospitalized patients for SARS-CoV-2 infection.

During the coronavirus disease 2019 (COVID-19) pandemic, several clinical prognostic scores have been proposed and evaluated in hospitalized patients, relying on variables available at admission. However, capturing data collected from the longitudinal follow-up of patients during hospitalization may improve prediction accuracy of a clinical outcome. To answer this question, 327 patients diagnosed with COVID-19 and hospitalized in an academic French hospital between January and July 2020 are included in the analysis. Up to 59 biomarkers were measured from the patient admission to the time to death or discharge from hospital. We consider a joint model with multiple linear or nonlinear mixed-effects models for biomarkers evolution, and a competing risks model involving subdistribution hazard functions for the risks of death and discharge. The links are modeled by shared random effects, and the selection of the biomarkers is mainly based on the significance of the link between the longitudinal and survival parts. Three biomarkers are retained: the blood neutrophil counts, the arterial pH, and the C-reactive protein. The predictive performances of the model are evaluated with the time-dependent area under the curve (AUC) for different landmark and horizon times, and compared with those obtained from a baseline model that considers only information available at admission. The joint modeling approach helps to improve predictions when sufficient information is available. For landmark 6 days and horizon of 30 days, we obtain AUC [95% CI] 0.73 [0.65, 0.81] and 0.81 [0.73, 0.89] for the baseline and joint model, respectively (p = 0.04). Statistical inference is validated through a simulation study.

Early administration of tecovirimat shortens the time to mpox clearance in a model of human infection.

Despite use of tecovirimat since the beginning of the 2022 outbreak, few data have been published on its antiviral effect in humans. We here predict tecovirimat efficacy using a unique set of data in nonhuman primates (NHPs) and humans. We analyzed tecovirimat antiviral activity on viral kinetics in NHP to characterize its concentration-effect relationship in vivo. Next, we used a pharmacological model developed in healthy volunteers to project its antiviral efficacy in humans. Finally, a viral dynamic model was applied to characterize mpox kinetics in skin lesions from 54 untreated patients, and we used this modeling framework to predict the impact of tecovirimat on viral clearance in skin lesions. At human-recommended doses, tecovirimat could inhibit viral replication from infected cells by more than 90% after 3 to 5 days of drug administration and achieved over 97% efficacy at drug steady state. With an estimated mpox within-host basic reproduction number, R0, equal to 5.6, tecovirimat could therefore shorten the time to viral clearance if given before viral peak. We predicted that initiating treatment at symptom onset, which on average occurred 2 days before viral peak, could reduce the time to viral clearance by about 6 days. Immediate postexposure prophylaxis could not only reduce time to clearance but also lower peak viral load by more than 1.0 log10 copies/mL and shorten the duration of positive viral culture by about 7 to 10 days. These findings support the early administration of tecovirimat against mpox infection, ideally starting from the infection day as a postexposure prophylaxis.

MOSAIC: A cohort study of human mpox virus disease.

Background: Human mpox is a viral disease caused by an Orthopoxvirus, human mpox virus (hMPXV), typically causing fever and a rash. Mpox has historically been endemic to parts of Central and West Africa, with small numbers of imported cases reported elsewhere, but starting May 2022 an unprecedented global outbreak caused by clade IIb hMPXV was reported outside traditionally endemic countries. This prompted the initiation of MOSAIC, a cohort study implemented in Europe and Asia that aims to describe clinical and virologic outcomes of PCR-confirmed hMPXV disease, including those who receive antiviral therapy. The focus of this article, however, is on describing the study protocol itself with implementation process and operational challenges. Methods: MOSAIC recruits participants of any age with laboratory-confirmed mpox disease who provide informed consent. Participants enrol in the cohort for a total of six months. Blood, lesion and throat samples are collected at several timepoints from the day of diagnosis or the first day of treatment (Day 1) until Day 28 for PCR detection of hMPXV. Clinical data are collected by clinicians and participants (via a self-completion questionnaire) for six months to characterize the signs and symptoms associated with the illness, as well as short- and more long-term outcomes. Discussion: The design and prompt implementation of clinical research response is key in addressing emerging outbreaks. MOSAIC began enrolment within two months of the start of the international mpox epidemic. Enrolment has been stopped and the last follow-up visits are expected in January 2024. ICTRP registration: EU CT number: 2022-501132-42-00 (22/06/2022).

Association between SARS-CoV-2 viral kinetics and clinical score evolution in hospitalized patients.

The role of antiviral treatment in coronavirus disease 2019 hospitalized patients is controversial. To address this question, we analyzed simultaneously nasopharyngeal viral load and the National Early Warning Score 2 (NEWS-2) using an effect compartment model to relate viral dynamics and the evolution of clinical severity. The model is applied to 664 hospitalized patients included in the DisCoVeRy trial (NCT04315948; EudraCT 2020-000936-23) randomly assigned to either standard of care (SoC) or SoC + remdesivir. Then we use the model to simulate the impact of antiviral treatments on the time to clinical improvement, defined by a NEWS-2 score lower than 3 (in patients with NEWS-2 <7 at hospitalization) or 5 (in patients with NEWS-2 ≥7 at hospitalization), distinguishing between patients with low or high viral load at hospitalization. The model can fit well the different observed patients trajectories, showing that clinical evolution is associated with viral dynamics, albeit with large interindividual variability. Remdesivir antiviral activity was 22% and 78% in patients with low or high viral loads, respectively, which is not sufficient to generate a meaningful effect on NEWS-2. However, simulations predicted that antiviral activity greater than 99% could reduce by 2 days the time to clinical improvement in patients with high viral load, irrespective of the NEWS-2 score at hospitalization, whereas no meaningful effect was predicted in patients with low viral loads. Our results demonstrate that time to clinical improvement is associated with time to viral clearance and that highly effective antiviral drugs could hasten clinical improvement in hospitalized patients with high viral loads.

Robust and Adaptive Two-stage Designs in Nonlinear Mixed Effect Models.

To get informative studies for nonlinear mixed effect models (NLMEM), design optimization can be performed based on Fisher Information Matrix (FIM) using the D-criterion. Its computation requires knowledge about models and parameters, which are often prior guesses. Thus, adaptive designs composed of several stages may be used. Robust approach can also be used to account for various candidate models. In the estimation step of a given stage, model selection (MS) or model averaging (MA) can be performed. In this work we propose a new two-stage adaptive design strategy, based on the robust expected FIM and MA over several candidate models. The methodology is applied to a clinical trial simulation in ophthalmology to optimize doses and time measurements. A set of dose-response candidate models is defined, and one-stage designs are compared to two-stage 50/50 designs (i.e., each stage performed with half of the available subjects), using either local optimal design or robust design, and performing analysis with one model, MS or MA. Performing a two-stage design with MS at the interim analysis can correct the choice of a wrong model for designing the first stage. Overall, starting from a robust design (1- or 2-stage) is valuable and leads to reasonable bias and precision. The proposed robust adaptive design strategy is a new tool to design longitudinal studies that could be used in different therapeutic areas.

Management of pharmacovigilance during the COVID-19 pandemic crisis by the safety department of an academic sponsor: Lessons learnt and challenges from the EU DisCoVeRy clinical trial.

The current COVID-19 pandemic was an exceptional health situation, including for drug use. As there was no known effective drug for COVID-19 at the beginning of the pandemic, different drug candidates were proposed. In this article, we present the challenges for an academic Safety Department to manage the global safety of a European trial during the pandemic. The National Institute for Health and Medical Research (Inserm) conducted a European multicenter, open-label, randomized, controlled trial involving three repurposed and one-in development drugs (lopinavir/ritonavir, IFN-ß1a, hydroxychloroquine, and remdesivir) in adults hospitalized with COVID-19. From 25 March 2020 to 29 May 2020, the Inserm Safety Department had to manage 585 Serious Adverse Events (SAEs) initial notification and 396 follow-up reports. The Inserm Safety Department's staff was mobilized to manage these SAEs and to report Expedited safety reports to the competent authorities within the legal timeframes. More than 500 queries were sent to the investigators due to a lack of or incoherent information on SAE forms. At the same time, the investigators were overwhelmed by the management of patients suffering from COVID-19 infection. These particular conditions of missing data and lack of accurate description of adverse events made evaluation of the SAEs very difficult, particularly the assessment of the causal role of each investigational medicinal product. In parallel, working difficulties were accentuated by the national lockdown, frequent IT tool dysfunctions, delayed implementation of monitoring and the absence of automatic alerts for SAE form modification. Although COVID-19 is a confounding factor per se, the delay in and quality of SAE form completion and the real-time medical analysis by the Inserm Safety Department were major issues in the quick identification of potential safety signals. To conduct a high-quality clinical trial and ensure patient safety, all stakeholders must take their roles and responsibilities.

Ensuring quality control in a COVID-19 clinical trial during the pandemic: The experience of the Inserm C20-15 DisCoVeRy study.

SETTING: Health measures taken during the pandemic deeply modified the clinical research practices. At the same time, the demand for the results of the COVID-19 trials was urgent. Thus, the objective of this article is to share Inserm's experience in ensuring quality control in clinical trials in this challenging context. OBJECTIVES: DisCoVeRy is a phase III randomized study that aimed at evaluating the safety and efficacy of 4 therapeutic strategies in hospitalized COVID-19 adult patients. Between March, 22nd 2020 and January, 20th 2021, 1309 patients were included. In order to guarantee the best quality of data, the Sponsor had to adapt to the current sanitary measures and to their impact on clinical research activity, notably by adapting Monitoring Plan objectives, involving the research departments of the participating hospitals and a network of clinical research assistants (CRAs). RESULTS: Overall, 97 CRAs were involved and performed 909 monitoring visits. The monitoring of 100% of critical data for all patients included in the analysis was achieved, and despite of the pandemic context, a conform consent was recovered for more than 99% of patients. Results of the study were published in May and September 2021. DISCUSSION/CONCLUSION: The main monitoring objective was met thanks to the mobilization of considerable personnel resources, within a very tight time frame and external hurdles. There is a need for further reflection to adapt the lessons learned from this experience to the context of routine practice and to improve the response of French academic research during a future epidemic.

Evaluation of model-based bioequivalence approach for single sample pharmacokinetic studies.

In a traditional pharmacokinetic (PK) bioequivalence (BE) study, a two-way crossover study is conducted, PK parameters (namely the area under the time-concentration curve [AUC] and the maximal concentration [ C max ]) are obtained by noncompartmental analysis (NCA), and the BE analysis is performed using the two one-sided test (TOST) method. For ophthalmic drugs, however, only one sample of aqueous humor, in one eye, per eye can be obtained in each patient, which precludes the traditional BE analysis. To circumvent this issue, the U.S. Food and Drug Administration (FDA) has proposed an approach coupling NCA with either parametric or nonparametric bootstrap (NCA bootstrap). The model-based TOST (MB-TOST) has previously been proposed and evaluated successfully for various settings of sparse PK BE studies. In this paper, we evaluate, via simulations, MB-TOST in the specific setting of single sample PK BE study and compare its performance to NCA bootstrap. We performed BE study simulations using a published PK model and parameter values and evaluated multiple scenarios, including study design (parallel or crossover), sampling times (5 or 10 spread across the dosing interval), and geometric mean ratio (of 0.8, 0.9, 1, and 1.25). Using the simulated structural PK model, MB-TOST performed similarly to NCA bootstrap for AUC. For C max , the latter tended to be conservative and less powerful. Our research suggests that MB-TOST may be considered as an alternative BE approach for single sample PK studies, provided that the PK model is correctly specified and the test drug has the same structural model as the reference drug.

Pharmacokinetic Model of Tenofovir and Emtricitabine and Their Intracellular Metabolites in Patients in the ANRS 134-COPHAR 3 Trial Using Dose Records.

Tenofovir (TFV) and emtricitabine (FTC) are part of the recommended highly active antiretroviral therapy (ART). Both molecules show a large interindividual pharmacokinetic (PK) variability. Here, we modeled the concentrations of plasma TFV and FTC and their intracellular metabolites (TFV diphosphate [TFV-DP] and FTC triphosphate [FTC-TP]) collected after 4 and 24 weeks of treatment in 34 patients from the ANRS 134-COPHAR 3 trial. These patients received daily (QD) atazanavir (300 mg), ritonavir (100 mg), and a fixed-dose combination of coformulated TFV disoproxil fumarate (300 mg) and FTC (200 mg). Dosing history was collected using a medication event monitoring system. A three-compartment model with absorption delay (Tlag) was selected to describe the PK of, respectively, TFV/TFV-DP and FTC/FTC-TP. TFV and FTC apparent clearances, 114 L/h (relative standard error [RSE] = 8%) and 18.1 L/h (RSE = 5%), respectively, were found to decrease with age. However, no significant association was found with the polymorphisms ABCC2 rs717620, ABCC4 rs1751034, and ABCB1 rs1045642. The model allows prediction of TFV-DP and FTC-TP concentrations at steady state with alternative regimens.

Characteristics and outcomes of an international cohort of 600 000 hospitalized patients with COVID-19.

BACKGROUND: We describe demographic features, treatments and clinical outcomes in the International Severe Acute Respiratory and emerging Infection Consortium (ISARIC) COVID-19 cohort, one of the world's largest international, standardized data sets concerning hospitalized patients. METHODS: The data set analysed includes COVID-19 patients hospitalized between January 2020 and January 2022 in 52 countries. We investigated how symptoms on admission, co-morbidities, risk factors and treatments varied by age, sex and other characteristics. We used Cox regression models to investigate associations between demographics, symptoms, co-morbidities and other factors with risk of death, admission to an intensive care unit (ICU) and invasive mechanical ventilation (IMV). RESULTS: Data were available for 689 572 patients with laboratory-confirmed (91.1%) or clinically diagnosed (8.9%) SARS-CoV-2 infection from 52 countries. Age [adjusted hazard ratio per 10 years 1.49 (95% CI 1.48, 1.49)] and male sex [1.23 (1.21, 1.24)] were associated with a higher risk of death. Rates of admission to an ICU and use of IMV increased with age up to age 60 years then dropped. Symptoms, co-morbidities and treatments varied by age and had varied associations with clinical outcomes. The case-fatality ratio varied by country partly due to differences in the clinical characteristics of recruited patients and was on average 21.5%. CONCLUSIONS: Age was the strongest determinant of risk of death, with a ∼30-fold difference between the oldest and youngest groups; each of the co-morbidities included was associated with up to an almost 2-fold increase in risk. Smoking and obesity were also associated with a higher risk of death. The size of our international database and the standardized data collection method make this study a comprehensive international description of COVID-19 clinical features. Our findings may inform strategies that involve prioritization of patients hospitalized with COVID-19 who have a higher risk of death.