Clinical Pharmacology & Therapeutics

Background: Mocravimod is an oral sphingosine-1-phosphate (S1P) receptor modulator. This Phase 1 multiple-ascending-dose study evaluated its safety, tolerability, pharmacokinetics (PK), and pharmacodynamics (PD) in healthy volunteers. Methods: In this double-blind, randomized, placebo-controlled, parallel-group trial, 60 healthy male volunteers were enrolled in five cohorts. Mocravimod was administered once daily at 0.3, 0.6, 1.2, or 3.0 mg for 14 days, or at 2.0 mg for 28 days. Safety assessments included adverse events (AEs), laboratory tests, vital signs, electrocardiography, and Holter monitoring. PK of mocravimod and its active metabolite, mocravimod-phosphate, and PD effects on absolute lymphocyte count (ALC) and leukocyte subsets were assessed. Results: Fifty-nine of 60 participants completed the study. One participant in the 3.0 mg cohort discontinued treatment because of asymptomatic transaminase elevation. No deaths or serious AEs occurred. AEs were mostly mild or moderate, transient, and showed no clear dose relationship. Mocravimod produced dose-dependent reductions in ALC from 0.6 mg onward, with maximum geometric mean reductions of 65%, 74%, 83%, and 77% at 0.6, 1.2, 2.0, and 3.0 mg, respectively. ALC values recovered to above the lower limit of normal during follow-up in all cohorts. Holter monitoring showed an initial placebo-corrected reduction in heart rate of approximately 10-15 beats/min at doses of 1.2-3.0 mg, which attenuated with continued dosing. One participant in the 3.0 mg cohort had a recurrent daytime second-degree atrioventricular block (Mobitz I/Wenckebach), reported as a mild non-dose-limiting AE. No QT prolongation was observed. Exposure to mocravimod and mocravimod-phosphate increased approximately dose-proportionally. Steady state was reached by Day 14 (Day 28 in the 2.0 mg cohort), accumulation was approximately five- to sevenfold, terminal half-lives were approximately 100-40 hours for both analytes, and parent-to-metabolite exposure ratios were close to 1. Conclusions: Once-daily mocravimod up to 3.0 mg for 14 days and 2.0 mg for 28 days was generally well tolerated and showed predictable S1P-modulator class effects on lymphocyte counts and heart rate, with PK properties supporting once-daily dosing and further clinical development.

Background: Cysteamine is the only disease-modifying therapy for nephropathic cystinosis and has shown promise in mitochondrial disorders, but its clinical utility is limited by poor tolerability due to high peak concentrations with existing formulations. TTI-0102 is a novel natural controlled-release cysteamine prodrug designed to provide sustained cysteamine exposure with improved tolerability. Methods: A multi-center, randomized, single-blind, placebo-controlled Phase 2 trial enrolled 9 patients with MELAS syndrome caused by mtDNA m.3243A>G mutation (>50% heteroplasmy) and moderate disease severity (NMDAS score 15-45). Patients received placebo (n=3) or TTI-0102 at 2.75 g/day for one week then 5.5 g/day (n=6, equivalent to 2.5 g/day cysteamine base). Pharmacokinetic parameters, safety, and pharmacodynamic biomarkers including pyruvate, taurine, pantothenic acid, tryptophan, GSH/GSSG, lactate, GDF-15, and FGF-21 were assessed. Clinical efficacy was evaluated using the Modified Fatigue Impact Scale (MFIS) and 12-minute walk test. Results: TTI-0102 demonstrated expected gastrointestinal side effects (nausea, vomiting, diarrhea) consistent with the cysteamine class, with dropout occurring in patients 50 kg receiving fixed 5.5 g/day dosing. Weight-based dosing at 60 {+/-} 5 mg/kg TTI-0102 (~26 mg/kg cysteamine base equivalent) achieved sustained 24-hour cysteamine exposure with half the daily dose and peak concentrations lower than expected by dose proportionality, compared to approved formulations (Procysbi: 56 mg/kg, peak 2.5 mg/L vs. TTI-0102: 26 mg/kg, peak ~2 mg/L). TTI-0102 significantly elevated pantothenic acid (plateauing at 2 weeks) and taurine levels, providing mitochondrial cofactor support and antioxidant effects. Statistically significant pharmacodynamic effects included increased plasma pyruvate (p=0.03) without lactate elevation, suggesting enhanced glycolytic flux, and decreased tryptophan (p<0.01), potentially reducing oxidative stress from neurotoxic kynurenine pathway metabolites. Interestingly, increase in plasma pyruvate and decrease in tryptophan were negligible at doses up to 40 mg/kg/day, optimal at 60 mg/kg/day, and slightly less at 65 mg/kg/day. GSH/GSSG measurements were confounded by sample stability issues. GDF-15, FGF-21, and 12-minute walk distance showed no treatment-related changes. Most notably, MFIS total scores demonstrated significant improvement in TTI-0102-treated patients at 60 mg/kg/day average dose compared to placebo (p=0.04). Polynomial regression revealed therapeutic onset at ~4 weeks, maximal benefit at ~12 weeks, and subsequent plateau. Conclusions: This Phase 2 trial provides proof-of-concept that TTI-0102 is safe and well-tolerated in MELAS patients while treated with less than 65 mg/kg/day, with efficacy signals in fatigue reduction, a cardinal symptom affecting 71-100% of mitochondrial disease patients. The drug tri-faceted mechanism through sustained cysteamine, taurine, and pantothenic acid delivery addresses oxidative stress, mitochondrial energy metabolism, and cofactor deficiency. Significant MFIS improvement coupled with favorable modulation of pyruvate and tryptophan supports advancing TTI-0102 to larger Phase 2b/3 trials in mitochondrial disease employing weight-based dosing (60 {+/-} 5 mg/kg), validated patient-reported outcomes, and minimum 12-week treatment duration. The same mechanism of cysteamine/cystine thiol-disulfide exchange in lysosomes that may benefit mitochondrial diseases also supports cystinosis treatment. An investigator-initiated study in cystinosis will evaluate whether once-daily TTI-0102 at 60 {+/-} 5 mg/kg can maintain therapeutic WBC cystine levels, potentially offering improved adherence and quality of life compared to current twice-daily or four-times-daily regimens, and this weight-adjusted dosing strategy and pharmacodynamic biomarkers identified in the MELAS study are going to be used to inform the design of the planned Phase 2 study in Leigh syndrome, another mitochondrial disorder, in collaboration with the Childrens Hospital of Philadelphia (CHOP), with particular attention to dose optimization and biomarker-based assessment of pharmacological activity. Acknowledgement: We are very thankful to the patients and the clinical teams of Radboud University Nijmegen Medical Centre (Netherlands) and Centre Hospitalier Universitaire d'Angers (France) for their participation in this operationally challenging study.

Background Classical pharmacokinetic-pharmacodynamic (PK/PD) theory models exposure-effect in two dimensions: magnitude and time. Rate-dependent toxicity has been documented across therapeutic domains but never formalised as a conserved biological constraint. Methods We developed the Human Adaptive Rate Limit (HARL) framework, formalising the maximum tolerable velocity as |dS/dt|_max = sigma_max / tau. We validated HARL across five domains using published trial data and a reanalysis of the longitudinal biomarker data from the 202-patient CAR-T cohort of Wei et al (2023). An 8-ODE quantitative systems pharmacology model guided biomarker selection. Early biomarker velocities (maximum positive slope within days 0-5) were computed for ferritin and D-dimer. Patients were classified as high-risk only if both velocities exceeded their thresholds (dual-velocity classifier). Thresholds were identified by grid-search optimisation of the Youden index and assessed by leave-one-out cross-validation. Findings A prospective crossover study (Kleinbloesem 1987, n=8) demonstrated that matched steady-state nifedipine concentrations produce divergent haemodynamic responses depending solely on rate of rise, anticipating the dose-related mortality signal subsequently reported across ~8350 patients with coronary heart disease (Furberg 1995), a meta-analysis that was itself debated. Convergent evidence spans haematology (CHOIR, 1432 patients, hazard ratio [HR] 1.34 [1.03-1.74] for aggressive Hb correction), radiation (dose-rate effectiveness factor [DDREF] 1.5-2.0), and infusion pharmacology. In the CAR-T cohort, high-risk classification (ferritin >232 ng/mL per day AND D-dimer >1.21 mg/L per day) predicted severe CRS with 100% sensitivity (~78% specificity) in safety rule-out mode and 91.1% sensitivity (93.6% specificity, AUC 0.95 [95% CI 0.91-0.98]) in Youden-optimised mode. Median kinetic lead time was 4 days (range 3-7) before clinical decompensation. Interpretation Biological tolerability is three-dimensional. HARL unifies rate-dependent toxicity across domains spanning minutes to weeks. MTDyn--specifying target level and allowable rate of change--should supplement conventional dose-response assessment.

ObjectiveConventional pharmacodynamic (PD) modeling workflows require manual model selection, repeated equation rewriting, and empirical parameter adjustment, resulting in limited automation, high cross-scenario migration costs, and insufficient reproducibility. This study aims to develop PD Union, a unified, automated, and interpretable framework for mechanistic PD modeling. MethodsPD Union is built upon a unified continuous dynamical skeleton that organizes absorption and systemic exposure module, the receptor module, the drug input module, the first delay module, the primary pharmacodynamic function module, the primary pharmacodynamic state module, the downstream pharmacodynamic state module, the second delay module, the feedback module, the circadian modulation module, the biophase module, the direct effect module, the disease state module, the second PD axis first delay module, the second PD axis primary pharmacodynamic function module, the second PD axis primary pharmacodynamic state module, the second PD axis downstream pharmacodynamic state module, the second PD axis second delay module, and the second PD axis feedback module. A machine learning-based structure identification module is incorporated to recognize drug input modes and mechanism labels from population PK/PD time series, followed by constrained population parameter optimization, forming an integrated pipeline of structure identification, candidate generation, and parameter fitting. ResultsValidation was conducted at two levels. In standardized synthetic benchmarking across 14 representative single-endpoint scenarios, the structure identification model achieved an output mode accuracy(NRMSE) of 0.7600 and macro-average F1 of 0.6307; parameter fitting yielded an NRMSE mean of 0.146 and median of 0.117. In the unified reconstruction validation based on 15 population pharmacokinetics/pharmacodynamics (PK/PD) literature data, the mean NRMSE of PDUnion model for PD was 0.261, and the median was 0.228. Among the 15 studies, 14 performed better than the models provided in the original literature. ConclusionsPD Union demonstrates that interpretable mechanistic modularization combined with machine learning-assisted structure identification is feasible for automated PD modeling. The framework provides an executable methodological foundation for unified, reproducible, and extensible mechanistic PD modeling, with potential applicability to multi-endpoint and complex disease-state modeling scenarios.

Background: This study aimed to evaluate real-world adverse event (AE) signals of EV to provide evidence-based guidance for its safe clinical application. Methods: Data from the FDA Adverse Event Reporting System (FAERS) database from the period of 2019 Q1-2025 Q3 were analyzed. Disproportionality analysis algorithms, including the reporting odds ratio (ROR), proportional reporting ratio (PRR), Bayesian confidence propagation neural network (BCPNN), and empirical Bayes geometric mean (EBGM), were utilized to mine safety signals.The time to onset (TTO) was evaluated using the Weibull distribution model. Results: Among 11,697,906 reports, 4,177 EV-treated patients experienced 14,511 AEs. The most common System Organ Classes (SOCs) were skin and subcutaneous tissue disorders (18.23%), general disorders and administration site conditions (13.17%).Multi-algorithm consensus identified 179 positive signals. Alongside known toxicities (rash, peripheral neuropathy, hyperglycemia), potential new signals emerged, including dysgeusia, atypical skin lesions, and myelosuppression. Median TTO was 14 days, with the Weibull {beta} of 0.736, confirming an "early failure" profile. Subgroup analysis revealed toxicity heterogeneity: patients aged [&ge;]65 and females exhibited stronger signals for fatal severe cutaneous adverse reactions, while patients aged < 65 and males showed higher susceptibility to neurological and metabolic toxicities. Conclusions: The real-world safety profile of EV confirms known toxicities, reveals new risks (e.g., dysgeusia), and shows toxicity concentrated in the first treatment cycle. Clinical practice requires proactive monitoring during the first two weeks using demographic-specific strategies: vigilance for fatal skin toxicity in elderly and female patients, and close follow-up of neurological and metabolic indicators in younger and male populations.

Abstract Introduction The sharing of individual patient data is essential for advancing pharmacometrics but is strictly limited by privacy regulations (e.g., GDPR). While synthetic data generation offers a legally compliant alternative, its structural impact on complex nonlinear mixed-effects (NLME) modelling remains largely unexplored. This study aimed to benchmark five generative artificial intelligence algorithms by evaluating the balance between data privacy and the preservation of structural PK properties and clinical dosing guidance. Material & methods A daptomycin two-compartment PopPK model was used to simulate a reference cohort of 500 patients. Five generative algorithms (Modified AVATAR, Gaussian Copula, Synthpop, TVAE, and CTGAN) produced 100 independent synthetic datasets each. A two-stage evaluation framework was applied: first, a statistical indistinguishability test based on logistic regression (AUC ROC) was used as a macroscopic pre-selection criterion to determine algorithm eligibility for NLME modelling and privacy risk assessment. Privacy risk was independently quantified using the Anonymeter framework (Singling Out and Linkability attacks). Eligible algorithms were further evaluated on PK parameter recovery bias and clinical dosing simulations. Results Deep learning architectures (TVAE, CTGAN) were excluded at the pre-selection stage due to both biologically implausible covariate generation and high macroscopic detectability (mean AUC ROC = 0.837 and 0.986, respectively). Synthpop, AVATAR, and Gaussian Copula all passed the indistinguishability threshold (AUC ROC = 0.475 +- 0.033, 0.490 +- 0.013, and 0.619 +- 0.031, respectively) and proceeded to NLME evaluation. However, attack-based privacy assessment revealed that Synthpop carried an unacceptable singling-out risk (0.035), disqualifying it from privacy-preserving data sharing. AVATAR and Gaussian Copula demonstrated acceptable privacy profiles (singling-out = 0.004 and 0.001; linkability = 0.010 and 0.003, respectively). At the structural level, Gaussian Copula injected stochastic noise inflating residual error (+157.0%) and V1; (+25.9%), blunting predicted Cmax and predisposing to empirical dose escalation and risk of toxicity. AVATAR acted aSs a smoothing filter, deflating V2; (-48.3%) and underestimating CL (-12.9%). Forward clinical simulations confirmed directionally opposed prediction errors: Gaussian Copula consistently underestimated Cmax across standard and renally impaired profiles (-14.5% and -16.0%, respectively), predisposing to empirical dose escalation, whereas AVATAR- and Synthpop-derived models overestimated Cmax and Cmin in the obese infected patient (+14.7% and +8.2%, respectively), compounding the accumulation risk already present in this profile. Conclusion While no generative algorithm currently offers a perfect solution, AVATAR and Gaussian Copula represent the most viable candidates, being the only methods to satisfy both macroscopic indistinguishability and attack-based privacy criteria. These findings highlight the necessity of a structured, two-stage validation framework and suggest that, when coupled with therapeutic drug monitoring, synthetic datasets could significantly enhance multicentre collaboration while maintaining strict regulatory compliance

Mocravimod (KRP203) is a selective sphingosine 1-phosphate (S1P) receptor modulator currently in development for patients with haematological malignancies undergoing allogenic haematopoietic cell transplantation (HCT). This first-in-human, randomised, double-blind, placebo-controlled, single ascending oral dose study evaluated the safety, tolerability, pharmacokinetics (PK), and pharmacodynamics (PD) of mocravimod in 136 healthy adult participants (EudraCT No. 2006-006814-13). Participants received single doses ranging from 0.01 to 40 mg or placebo, with a cohort dedicated to studying food-effect at 3 mg. Mocravimod demonstrated slow absorption (mean Tmax 6-11 hrs), extensive distribution, and a long terminal half-life (91-132 hrs). Exposure increased dose-proportionally for doses [&ge;]2 mg. The most common adverse events were headache, dizziness, and fatigue, all graded as mild or moderate; no serious adverse events or deaths occurred. Mocravimod-phosphate induced robust, dose-dependent reductions in lymphocyte counts, with significant decreases at doses [&ge;]2 mg and recovery to baseline observed in all but the highest dose groups. Cardiac effects included transient bradycardia and benign second-degree atrioventricular (AV) block at higher doses, without clinically significant arrhythmias. Food intake had minimal impact on PK. No clinically meaningful changes in pulmonary function or laboratory safety signals were detected. These results indicate that single oral doses of mocravimod up to 40 mg are safe and well tolerated in healthy adults, with predictable PK and expected PD effects. The findings support further clinical development of mocravimod as a targeted immunomodulator in settings such as allogeneic HCT for haematological malignancies.

BackgroundLarge language models (LLMs) are increasingly used in medical education and clinical decision-making, but their reliability in high-risk medication dosing remains unclear. Opioid rotation is a common task requiring precise calculations where errors may result in overdose or inadequate pain relief. MethodsThirteen LLMs were tested using an API-based framework to ensure independent queries across trials. First, fictional clinical scenarios were tested to simulate real-world clinical situations involving opioid rotation; to test the effects of changes in wording, scenarios were revised into 4 "vignettes" showing the same clinical situation. Next, opioid pairs were tested with a random-dose paradigm across a clinically-pertinent range (5-120 mg daily morphine equivalents). LLM outputs were compared with expected values derived from reference standards. Accuracy was assessed using predefined safety thresholds: tight accuracy (0.85-1.15x expected dose) and broad accuracy (0.6-1.7x). We tested models naively and with prompts augmented with reference tables and unit explanations. ResultsNaive models generally exhibited low tight-range accuracy across opioid pairs. For any given opioid pair, each model would consistently produce similar incorrect conversion ratios despite wide variability across opioid pairs and language models. Vignette wording changes accounted for 76% of within-scenario response variance. Reference-based prompt augmentation significantly improved performance, with over half of models achieving high proportions of conversions within tight accuracy for morphine-equivalent conversions. ConclusionsWhile commercial LLMs demonstrated variable accuracy in the native state, prompt augmentation significantly improved their performance.

BackgroundThe discontinuation of Fasiglifam (TAK-875), a GPR40/FFAR1 full agonist, during Phase 3 clinical trials due to hepatotoxicity led to widespread abandonment of GPR40 as a viable therapeutic target for type 2 diabetes mellitus (T2DM). However, mechanistic evidence suggests that Fasiglifams hepatotoxicity arises from mitochondrial liability driven by high lipophilicity (aLogP = 5.31), rather than from on-target GPR40 signaling. We hypothesized that target-level failure was incorrectly inferred from compound-level safety concerns, and that superior candidates exist within publicly available databases. MethodsWe queried ChEMBL Release 36 (28 GB SQLite, 74 tables) for all compounds with documented GPR40/FFAR1 activity (UniProt: O14842). Compounds were filtered by EC50 [&le;] 10 nM in nM units with standard relation "=". Drug-likeness was assessed using Lipinskis Rule of Five (Ro5), aLogP, molecular weight (MW), hydrogen bond donors/acceptors (HBD/HBA), and polar surface area (PSA). A parallel analysis of Therapeutic Target Database (TTD v10.1.01, 4,298 targets) provided clinical context. A real-world evidence (RWE) patient stratification framework was constructed using EMR data from tens of millions of patients with >10 years of longitudinal follow-up. ResultsOf 2,637 GPR40-active compounds in ChEMBL 36, 526 (19.9%) demonstrated EC50 < 100 nM and 102 (3.9%) demonstrated EC50 < 10 nM. Eight compounds met stringent drug-likeness criteria (Ro5 violations = 0, aLogP < 5.0, EC50 [&le;] 1 nM). The lead compound (CHEMBL4859651) exhibited EC50 = 0.04 nM (8.75-fold more potent than Fasiglifam), MW = 297 Da (43% lower), and aLogP = 4.30 (19% lower), with zero Ro5 violations. Mean MW of the eight candidates was 317 {+/-} 28 Da versus 524 Da for Fasiglifam. A parallel GCK analysis identified a protein-protein interaction target (CHEMBL3885579, GCK-GKRP interface) harboring 40 exclusive compounds as an orthogonal strategy for partial GCK activation. ConclusionsSystematic cheminformatic analysis reveals that compounds with substantially superior activity and drug-likeness profiles relative to Fasiglifam exist within ChEMBL 36. Fasiglifams hepatotoxicity is attributable to compound-specific physicochemical properties, not GPR40-mediated toxicity. RWE patient stratification may further mitigate hepatotoxicity risk for next-generation GPR40 agonists. These findings argue for systematic reappraisal of GPR40 as a viable therapeutic target for T2DM.

ObjectivesTo develop a population pharmacokinetic (PPK) model of polymyxin B (PMB) for intravenous (IV) and combined intravenous plus inhaled (IV+IH) administration in critically ill patients, and evaluate the association between the 24-h steady-state area under concentration-time curve to minimum inhibitory concentration ratio (AUCss,24h/MIC) and clinical outcomes. MethodsThis prospective cohort was conducted in the ICU of the Third Xiangya Hospital, Central South University (ethics R19048; ChiCTR1900028602). Adults with multidrug-resistant Gram-negative bacterial infections receiving PMB [&ge;]48 h were enrolled and assigned to IV or IV+IH groups. Serial plasma samples were analyzed by validated LC-MS/MS. The PPK model was developed with NONMEM(R). Clinical efficacy at end of treatment was blindly assessed. ResultsForty-three patients were enrolled (IV, n=22; IV+IH, n=21), with an overall clinical success rate of 66.7%. A two-compartment PPK model best described the data, with typical values of clearance (2.6 L/h), central volume (13.6 L), and peripheral volume (17.6 L). Clearance was influenced by creatinine clearance and total bile acids. In the overall cohort, neither AUCss,24h nor AUCss,24h/MIC differed significantly between clinical success and failure (p=0.591 and 0.143). In the IV group, AUCss,24h/MIC was significantly higher in responders (p=0.005) with an ROC-derived efficacy threshold of 94.37; AUCss,24h showed a non-significant trend (p=0.076). No exposure- response relationship was observed in the IV+IH group (p=0.398 and 0.495). ConclusionsPlasma AUCss,24h/MIC appears to be associated with clinical efficacy during IV monotherapy but not in IV+IH regimens, likely due to high pulmonary exposure. Plasma-based PK/PD targets should be applied cautiously when inhalation is added.

BackgroundPost-market pharmacovigilance relies predominantly on single-database disproportionality analysis of spontaneous adverse event reports, which lacks corroboration across independent evidence streams and cannot integrate randomised trial evidence. No publicly accessible platform has previously combined European national pharmacovigilance registries, the US FDA Adverse Event Reporting System (FAERS), and clinical trial meta-analyses into a unified, continuously scored signal detection framework. MethodsWe describe the Signal Consensus Index (SCI), a composite 0-100 pharmacovigilance signal score integrating disproportionality evidence from the Danish National Pharmacovigilance Database, the UK MHRA Yellow Card scheme, and FAERS, with DerSimonian-Laird meta-analytic risk ratios from ClinicalTrials.gov, across 6,905,874 drug-adverse event pairs. Each source contributes a continuous score derived from the lower bounds of three complementary disproportionality metrics (ROR, PRR, IC025) for spontaneous reporting sources, and from the pooled risk ratio lower confidence bound for clinical trials. The SCI is publicly accessible via the Adverse Event Atlas (aeatlas.com). We report reference set validation against the EU-ADR reference standard, a single-source comparison with discordance characterisation, temporal stability analysis across eight cumulative data windows (2015-2023), and a weight sensitivity analysis across seven pre-specified weighting schemes. ResultsThe SCI generated 129,176 Moderate-or-Strong signals (SCI [&ge;] 50, confidence [&ge;] 50) and 7,290 Strong signals (SCI [&ge;] 70, confidence [&ge;] 70). Reference set validation against 88 classifiable drug-event pairs (44 positive controls, 44 negative controls) yielded 18 true positives, 0 false positives, 44 true negatives, and 26 false negatives (sensitivity 40.9%, specificity 100.0%, PPV 100.0%, NPV 62.9%). Zero false positives were observed across all 44 classifiable negative controls, with five false negatives attributable to the confidence gate correctly suppressing single-source signals pending multi-source corroboration. Single-source comparison demonstrated that FAERS alone generated 1,438,246 disproportionality signals, of which 94.8% were not confirmed by the SCI, while 54,184 SCI-detected signals were absent from FAERS, of which 8.3% involved drugs absent from the US reporting system. Discordance analysis showed that 99.8% of Danish non-confirmation reflected data availability constraints. Temporal stability was high: 98.5% of pairs received identical classifications across all seven weight scenarios, and 57.0% of final Strong signals were already detectable as Moderate or Strong in the earliest data window (2015-2016). Strong classifications were stable across weight scenarios (94.0% of Strong observations remaining Strong). ConclusionsThe SCI provides a transparent, openly accessible framework for cross-source pharmacovigilance signal prioritisation with 100% specificity and PPV against an established reference standard and stable classifications across weighting schemes. Progressive signal emergence through the Moderate tier supports its use as an early detection layer. The platform is available at aeatlas.com.

Drug-induced seizures remain a major safety concern in drug development, yet human seizure liability is difficult to predict using conventional preclinical models. Here, we evaluated whether spontaneous calcium network activity in human induced pluripotent stem cell-derived CNS-3D Brain Organoids could predict clinically observed seizure risk across a pharmacokinetically anchored drug set. CNS-3D organoids contained neuronal and astrocytic populations, expressed neuroactive receptor and ion-channel gene programs that aligned with human cortical tissue, and exhibited reproducible spontaneous calcium oscillations across production batches. A retrospective drug panel of 66 small-molecule drugs was assembled from human clinical evidence, including 30 seizure-associated drugs and 36 comparator drugs without documented clinical seizure liability. Drugs were tested across concentration ranges anchored to reported clinical Cmax, and calcium time-series responses were integrated with chemical structure features using a machine-learning workflow. The final model predicted clinical seizure liability with an AUROC of 0.872, achieving 83.3% sensitivity and 88.9% specificity in drug-level cross-validation. Model scores also stratified seizure-associated drugs by clinical context and prevalence, suggesting that CNS-3D activity profiles capture clinically meaningful differences in seizure risk. Compared with published in vitro and preclinical seizure-liability models, CNS-3D organoid-based predictions showed improved balanced sensitivity and specificity. These findings support high-throughput calcium profiling in human CNS-3D organoids as a scalable, exposure-aware platform for predicting human seizure liability and contributing functional human data to neuro-safety assessment.

Population pharmacokinetic (popPK) models are commonly developed using ordinary differential equations (ODEs) to describe deterministic concentration-time profiles, with unexplained variability typically attributed to interindividual variability or residual error. When model misspecification is present, system-level deviations may be absorbed into these conventional variability terms, making the source and magnitude of model inadequacy difficult to assess quantitatively. Stochastic differential equations (SDEs) provide an alternative framework by introducing an explicit system-noise component into the structural model, allowing model-data mismatch to be evaluated more directly. However, historical implementation of SDE-based models in NONMEM has been technically challenging. The availability of the Fortran plug-in subroutine SDE.f90 substantially lowers this barrier and enables more practical implementation of SDE-based models in NONMEM. In this work, SDE-based nonlinear mixed-effects models were evaluated as a quantitative diagnostic framework for probing popPK model misspecification. The SDE.f90 implementation was first verified using simulated one-compartment intravenous bolus datasets with stochastic process noise. Additional simulation-estimation scenarios were then conducted under intentionally misspecified structural or stochastic assumptions, including time-varying elimination, compartmental misspecification, and residual error misspecification. Across these scenarios, the estimated system-noise parameter was generally sensitive to misspecification, with larger values usually associated with greater structural or stochastic mismatch. SDE-based modeling also helped partially separate system-level variability from residual variability and, in selected settings, supported localization of misspecification to specific model components, thereby helping guide model refinement. Overall, SDE-based popPK modeling is a useful addition to the pharmacometric diagnostic toolbox, with system-noise estimates best interpreted together with structural model evaluation, residual diagnostics, parameter behavior, and pharmacologic plausibility.

1.Pharmacokinetic studies in critically ill patients are often constrained by small sample sizes, limiting the strength and generalizability of conclusions drawn solely from observed data. Bayesian inference offers a powerful strategy to address this challenge by incorporating prior knowledge. In this study, we evaluated two model-based approaches for characterizing the population pharmacokinetics of ceftolozane and tazobactam in critically ill patients, comparing nonlinear mixed-effects modeling with Bayesian hierarchical analyses. The Bayesian methods incorporated literature-derived prior information. The data was collected from 13 critically ill patients receiving 3.0 g of ceftolozane combined with tazobactam (2:1) via intravenous infusion. Pharmacokinetic modeling was performed using NONMEM and Stan software with the Torsten extension. Model diagnostics and graphical analyses were conducted in RStudio with relevant packages. In the absence of prior information, a one-compartment model with a limited set of parameters describing inter-individual variability adequately characterized the pharmacokinetics of ceftolozane and tazobactam. When prior information was incorporated, a two-compartment model became feasible and yielded a characterization of parameter variability and correlations that was more consistent with published literature. The application of Bayesian inference ensured alignment with existing literature on ceftolozane and tazobactam pharmacokinetics and mitigated some systematic biases observed in the data-driven approaches. Moreover, the Bayesian approach enables direct decision-making by incorporating uncertainty into the analysis, as demonstrated by probability of target attainment analysis. Collectively, these results underscore the utility of Bayesian methods in pharmacokinetic modeling for critically ill patients, offering a robust framework for optimizing dosing strategies in data-limited settings.

BackgroundNon-adherence to lipid-lowering therapy (LLT) affects up to half of patients and contributes substantially to preventable cardiovascular morbidity and mortality. Existing measures, such as the proportion of days covered, provide cross-sectional summaries but fail to capture the dynamic patterns of adherence over time. Although group-based trajectory modelling identifies distinct longitudinal adherence patterns, no approach currently predicts trajectory membership prospectively while incorporating patient-reported barriers. We developed BRIDGE, a barrier-informed Bayesian model to predict adherence trajectories and identify their underlying drivers. MethodsBRIDGE incorporates patient-reported barriers as structured prior information within a Bayesian framework for adherence-trajectory prediction. The model was designed not only to estimate which patients are likely to follow different adherence trajectories, but also to generate clinically interpretable probability estimates that help explain why those trajectories may arise and what modifiable factors may be most relevant for intervention. ResultsBRIDGE achieved a macro AUROC of 0.809 (95% CI 0.806 to 0.813), comparable to random forest (0.815 (95% CI 0.812 to 0.819)) and XGBoost (0.821 (95% CI 0.818 to 0.824)), two widely used machine-learning benchmarks for structured clinical prediction. Calibration was superior to random forest (Brier score 0.530 vs 0.545; ), and performance was stable across six independent training runs (AUROC SD = 0.003). Incorporating barrier-informed priors improved accuracy by 3.5% and calibration by 5.5% compared to flat priors, showing that incorporation of patient-reported barriers added value beyond electronic medical record data alone. Four clinically distinct adherence trajectories were identified: gradual decline associated with treatment deprioritisation amid polypharmacy (10.4%), early discontinuation linked to asymptomatic risk dismissal (40.5%), rapid decline associated with intolerance (28.8%), and persistent adherence (20.2%). Counterfactual analysis identified trajectory-specific intervention levers. ConclusionsBRIDGE provides accurate and well-calibrated prediction of adherence trajectories while offering clinically actionable insights into their underlying drivers. By integrating patient-reported barriers with routine clinical data, the model supports targeted, mechanism-informed interventions at the point of prescribing to improve adherence to cardioprotective therapies. FundingMRFF CVD Mission Grant 2017451 Evidence before this studyWe searched PubMed and Scopus from database inception to December 2025 using the terms "medication adherence", "trajectory", "prediction model", "Bayesian", "lipid-lowering therapy", and "barriers", with no language restrictions. Group-based trajectory modelling has consistently identified three to five adherence patterns across cardiovascular cohorts; however, these applications have been descriptive rather than predictive. Machine-learning models for adherence prediction achieve moderate discrimination but treat adherence as a binary or continuous outcome, thereby overlooking the clinically meaningful heterogeneity captured by trajectory approaches. One prior study applied a Bayesian dynamic linear model to examine adherence-outcome associations, but it did not predict adherence trajectories or incorporate patient-reported barriers. To our knowledge, no published model integrates patient-reported barriers into trajectory prediction. Added value of this studyBRIDGE is, to our knowledge, the first model to incorporate patient-reported adherence barriers as hierarchical domain-informed priors within a Bayesian framework for trajectory prediction. Using 108 predictors derived from routine electronic medical records, the model achieves discrimination comparable to state-of-the-art machine-learning approaches while additionally providing uncertainty quantification, barrier-level interpretability, and counterfactual insights to inform intervention strategies. The identified trajectories differed not only in adherence level but also in switching behaviour, drug-class evolution, and medication burden, suggesting distinct underlying mechanisms of non-adherence that may require tailored clinical responses. Implications of all the available evidenceEach adherence trajectory implies a distinct intervention target: asymptomatic risk communication for early discontinuers (40.5% of patients), proactive tolerability management for rapid decliners, medication simplification for patients with gradual decline associated with polypharmacy, and maintenance support for persistent adherers. By integrating routinely collected clinical data with patient-reported barriers, BRIDGE can be deployed within existing primary care EMR infrastructure to generate actionable, trajectory and patient--specific recommendations at the point of prescribing, helping to bridge the gap between adherence measurement and targeted adherence management.

Background Plasma gelsolin (pGSN) is a non-immunosuppressive anti-inflammatory immunomodulator with demonstrated efficacy in animal models of acute lung injury. Its potential role in moderate-to-severe acute respiratory distress syndrome (ARDS) is currently under investigation. Methods We conducted a phase 1, randomized, double-blind, placebo-controlled study to evaluate the safety, tolerability, and pharmacokinetics of recombinant human pGSN (rhu-pGSN) following intravenous (IV) administration to healthy volunteers. Thirty-two participants were assigned to 4 sequentially ascending dose cohorts (6, 12, 18, 24 mg/kg of body weight) to receive five IV infusions of rhu-pGSN or saline placebo. Each cohort includes 8 subjects randomized 3:1 with rhu-pGSN or placebo. Doses were administered at 0 hours, 12 hours, 36 hours, 60 hours, and 84 hours. The primary outcome is the incidence and severity of clinical and laboratory AEs regardless of causality. Secondary outcomes include the pharmacokinetics of IV rhu-pGSN and the presence of anti-rhu-pGSN antibodies at Day 28. Results Overall, 10 subjects (41.7%) who received rhu-pGSN reported a total of 13 adverse events (AEs), and 1 subject (12.5%) who received placebo reported an AE. All AEs were mild or moderate. AEs in system organ classes that were reported by 2 or more subjects in either arm were skin and subcutaneous tissue disorders (12.5% rhu-pGSN; 0% placebo), gastrointestinal disorders (8.3% rhu-pGSN; 0% placebo), and nervous system disorders (12.5% rhu-pGSN; 12.5% placebo). No AEs by preferred term were reported by more than 1 subject in either arm. Three subjects (12.5%) experienced an AE assessed as related to study drug. No serious AEs occurred, and no AEs led to study discontinuation, dose interruption/reduction, or death. There were no apparent between-treatment differences in laboratory abnormalities, vital signs, or electrocardiogram findings. Conclusions Overall, in this study, IV rhu-pGSN (up to 24 mg/kg daily) appeared safe and well tolerated compared to placebo. The median half-life of rhu-pGSN exceeded 14 h across all dosing regimens, supporting once daily IV dosing in healthy subjects. Trial registration This study was registered with ClinicalTrials.gov on 2023-03-29 under the registration identifier NCT05789745.

Long-acting cabotegravir and rilpivirine combination (LA-CAB/RPV) is approved for HIV treatment whilst long-acting cabotegravir alone (LA-CAB) is approved for HIV prevention, both in adults. However, individuals who become pregnant might prefer to discontinue it due to lack of definitive data on safety. The aim of this study was to characterise the tail-phase maternal and fetal pharmacokinetics of LA-CAB/RPV following discontinuation at steady-state early in pregnancy. A virtual population of non-pregnant women (n = 100 per scenario) initiated intramuscular injections of LA-CAB/RPV at the approved dosage and continued maintenance dose (400/600 mg once monthly or 600/900 mg once every two months) until steady state. We simulated discontinuation at steady state after only one injection during pregnancy. Tail-phase pharmacokinetics of CAB and RPV from LA injections were characterised during gestation and until 6 months postpartum. Pharmacokinetic tails of LA-CAB/RPV were driven by the residual drug in the muscle depot which stabilised at steady state and reduced steadily upon dosing discontinuation. Upon discontinuation of the monthly dosing, predicted median (IQR) maternal plasma concentrations for LA-CAB were 415 (386-448) ng/mL at delivery and 125 (115-139) ng/mL 6 months postpartum. For LA RPV, these were 11.6 (11.0-12.6) ng/mL and 7.84 (7.30-8.49) ng/mL at delivery and 6 months postpartum, respectively. Pharmacokinetic tails of LA-CAB/RPV extend to several months postpartum, with levels falling below established minimum effective concentration in most women after gestation week 33. Potential strategies to minimise potential risks associated with LA-CAB/RPV discontinuation in this population are needed.

BackgroundAlkaptonuria (AKU) is a rare autosomal recessive metabolic disorder caused by deficiency of homogentisate 1,2-dioxygenase (HGD), resulting in systemic accumulation of homogentisic acid (HGA), ochronosis, and progressive multisystem disease. Although nitisinone (NTBC) lowers HGA levels, it does not correct the underlying genetic defect and induces hypertyrosinemia, highlighting the need for curative treatment approaches. We evaluated liver-directed adeno-associated virus (AAV)-mediated HGD gene therapy as a potential treatment for AKU. MethodsHgd-deficient (Hgd-/-) mice received liver-directed AAV2/8 vectors expressing codon-optimized human HGD under a liver-specific promoter. Reporter vectors were first used to assess hepatic biodistribution and transduction efficiency. Therapeutic efficacy was subsequently evaluated following AAV2/8-HGD administration (1 x 1012 vg/mouse). HGD expression was assessed by DNAscope, Western blotting, and RT-qPCR. Metabolic correction was determined using targeted LC-MS/MS and untargeted LC-HRMS metabolomics and compared with NTBC-treated Hgd-/- mice. ResultsReporter studies demonstrated liver-predominant transduction, with dose-dependent hepatocyte transduction reaching 89-93% at the highest dose. AAV2/8-HGD treatment produced robust hepatic HGD expression, with codon-optimized human HGD transcript levels approximately 33-fold higher than endogenous murine Hgd expression. Twelve weeks after treatment, plasma and urinary HGA levels were significantly reduced, with plasma HGA restored to near wild-type concentrations. Untargeted metabolomics further demonstrated marked reductions in HGA-derived phase I and II metabolites and revealed significant modulation of tricarboxylic acid cycle metabolism, consistent with partial restoration of metabolic homeostasis. Compared with NTBC-treated mice, AAV2/8-HGD achieved comparable plasma HGA reduction without elevation of upstream tyrosine pathway metabolites. ConclusionsLiver-directed AAV2/8-HGD gene therapy achieved substantial biochemical correction in Hgd-/- mice and restored metabolic flux without inducing hypertyrosinemia. These findings provide proof-of-concept supporting AAV-mediated HGD replacement as a promising long-term therapeutic strategy for AKU.

In the management of atrial fibrillation, the most frequently prescribed oral anticoagulant is apixaban, given at a fixed dose of 5mg BID. Apixaban is predominantly metabolized by cytochrome P4503A4 (CYP3A4) and is also a substrate for the drug efflux transporter P-glycoprotein (P-gp). In nearly 300,000 Medicare patients with AF receiving apixaban, we previously showed that concomitant therapy with drugs that inhibit both CYP3A4 and P-gp, specifically amiodarone or diltiazem, significantly increased serious bleeding that caused hospitalization and/or death. We hypothesized that this adverse effect was mediated by an increase in apixaban plasma concentrations caused by concomitant therapy that reduced drug elimination. Utilizing left-over samples obtained from clinically indicated blood draws that would typically be discarded, the Vanderbilt University Medical Center biobank BioVU contains >353,000 samples linked to de-identified electronic medical records (EMRs), with both DNA and plasma harvested. Of 35 samples drawn from patients taking apixaban 5mg BID, 5 were identified to be drawn from patients concomitantly taking drugs inhibiting both CYP3A4 and P-gp. Using a chromogenic anti-Xa assay, we found that plasma concentrations of apixaban were significantly higher (347{+/-}64 ng/mL; mean{+/-}SEM) for patients receiving concomitant CYP3A4/P-gp-inhibiting drugs compared to those not treated with these drugs (166{+/-}67 ng/mL; P=0.025, Mann Whitney). There were no differences between the 2 patient groups with respect to age, weight, or serum creatinine. The results of this pilot study provide preliminary data to support our hypothesis, and they demonstrate the practicality of obtaining pharmacokinetic data from a large cohort of plasma samples linked to deidentified EMRs. This approach could be used to define the role of apixaban levels in high-risk clinical scenarios and to better understand the relationship between drug levels and bleeding risk.

Prader-Willi syndrome (PWS) is a rare multisystemic disorder characterized by obesity, endocrine dysfunctions, and psychiatric comorbidities, which imply frequent use of psychotropic medications. They account for atypical responses to standard dosages of psychiatric drugs. Pharmacogenetics could be part of the reason for this situation, potentially offering a valuable tool for individualized treatment. This study analyzed allelic and phenotypic frequency distributions of five of the main cytochrome P450 enzymes (CYP2D6, CYP2B6, CYP2C19, CYP2C9, CYP3A4) involved in psychiatric drug metabolism in 47 patients with genetically confirmed diagnosis of PWS and compared them to reference frequencies in the general European population. Allelic frequency comparisons between the European reference population and the overall PWS cohort revealed a significant global difference for CYP2B6, with CYP2C19 and CYP2D6 showing trends toward significance. Although no global allelic differences remained significant after false discovery rate correction, post-hoc analyses consistently identified an enrichment of reduced- or non-functional alleles CYP2B619 and CYP2D610 in patients with PWS. Predicted metabolizer phenotype analyses showed a significant shift toward intermediate metabolizers of CYP3A4 in the PWS cohort, with corresponding depletion of normal metabolizers. Subgroup analyses indicated that allelic differences were more pronounced in maternal uniparental disomy and non-deletion subtypes, particularly for CYP2B6, although no significant differences were observed between PWS genetic subtypes. Overall, results imply potential differences in metabolizing activity in PWS patients, and subsequent implications in drug efficacy and tolerability. These results support the idea that pharmacogenetic testing may improve therapeutic decision-making in PWS for psychiatric treatment. Larger studies are needed to confirm these preliminary results.