Cytochrome P450s (CYPs) 2C subfamily (eg, 2C8, 2C9, and 2C19) and phase II enzymes such as uridine 5'-diphospho-glucuronosyltransferases (UGTs) are increasingly relevant in drug development and key targets for enzymatic induction. However, for these enzymes, weak induction signals in standard in vitro tools, such as sandwich-cultured human hepatocytes, challenge drug-drug interaction (DDI) risk assessment. This study evaluated an all-human hepatocyte coculture system (TruVivo) as a more sensitive model for CYP2Cs and UGT1A1 induction. After treatment of cells with rifampicin, carbamazepine, and phenytoin, we demonstrated robust mRNA and activity-fold-induction exceeding or meeting the 2-fold threshold in the coculture system, allowing for estimation of CYP2Cs and UGT1A1 induction parameters (IndC50, Indmax), unlike sandwich culture. Using TruVivo IndC50, Indmax of these precipitants in physiologically based pharmacokinetic (PBPK) modeling resulted in high predictive accuracy. In rifampicin studies, using TruVivo mRNA-derived data from the most sensitive donor 1 and average parameters across donors, was essential to properly predict in vivo DDI, particularly for object drugs mainly metabolized by CYP2Cs and UGT1A1, or with moderate to low CYP3A4 contribution (fm ≤ 0.5) in multipathway metabolism. For object drugs metabolized by CYP3A4 beyond 2Cs and UGTs, carbamazepine and phenytoin PBPK predictions highlighted the applicability of TruVivo uncalibrated CYP3A4 data for accurate assessment, whereas parameters calibrated against rifampicin showed a conservative trend in estimating DDI. Overall, the all-human coculture system, paired with PBPK, offers a breakthrough for CYP2Cs and UGT1A1 preclinical DDI induction risk assessment. SIGNIFICANCE STATEMENT: Cryopreserved human hepatocytes in sandwich culture show limited sensitivity toward cytochrome P450s 2C and uridine 5' -diphospho-glucuronosyltransferases induction, challenging in vitro-in vivo translation of the drug-drug interaction risk. This study confirms that TruVivo is a more sensitive in vitro model. By using physiologically based pharmacokinetic modeling, we investigated the impact of the measured induction parameters on predictive accuracy, showing TruVivo as a useful tool for cytochrome P450s 2C and uridine 5' -diphospho-glucuronosyltransferases risk assessment.

BACKGROUND:Voriconazole (VRC), a triazole antifungal agent, is recommended as primary treatment for invasive aspergillosis. Due to its unpredictable dose-exposure relationship, therapeutic drug monitoring (TDM) is recommended. Despite TDM, the probability of achieving therapeutic targets is only approximately 50%. This highlights the need for new dosing strategies. Calculating the metabolic ratio (MR) of NOX, the principal N-oxide metabolite of voriconazole primarily formed by the genetically polymorphic CYP2C19 enzyme, to VRC concentrations could serve as a surrogate for phenotype information. This is particularly relevant since the CYP2C19 genotype is often unavailable in acute clinical settings. Moreover, genotype may be less informative during inflammation due to a phenomenon known as 'phenoconversion', where inflammation masks genotype effects through downregulation of CYP450 enzymes. However, before integrating NOX in population PK models and/or as an adjunct in voriconazole TDM flowcharts, it is important to understand the factors impacting NOX exposure, and hence the MR. The objective of this narrative review is therefore to summarise recent literature on patient- and drug-related covariates impacting NOX exposure. METHODS:A comprehensive search was conducted in PubMed and Embase for articles published from 2003 to 2025 focusing on factors impacting NOX or the MR. RESULTS/DISCUSSION:Twenty-one articles were included, predominantly involving adult patients from an Asian or white ethnicity. Studies were mainly set up as pharmacokinetic (PK) studies, in which associations between NOX or MR were assessed in uni- or multivariate analyses. Key covariates identified were as follows: (1) voriconazole dose (6/11 studies)-higher doses increasing NOX, although this relationship was complicated by saturated metabolism; (2) C-reactive protein (CRP) (8/9)-higher inflammation levels correlated with reduced metabolic capacity due to CYP450 enzyme downregulation; (3) age (7/12)-children exhibited higher MR, probably driven by higher clearance due to increased CYP2C19 and FMO3 expression, while older adults (> 60 years) showed decreased MR due to a decline in CYP2C19 activity; (4) CYP2C19 genotype (10/13)-poor metabolizers had lower MR and lower NOX, supporting the utility of genotyping, if available with acceptable turnaround time. A CRP threshold of 100 mg/L has been found in some studies and may serve as a crucial indicator for clinicians in optimising dosage adjustments. Although hepatic impairment is recognised as a significant covariate for voriconazole, its influence on NOX or the MR was found limited in this review. This is likely due to the exclusion of patients with hepatic impairment from most studies, which resulted in lower interindividual variation in liver function test results. While renal function does not affect voriconazole exposure, it may impact NOX or MR, as some studies suggest a higher renal clearance of NOX compared to voriconazole. However, only one of the included studies suggested that impaired renal function resulted in a higher MR, indicating NOX accumulation. Further research is however needed. CONCLUSION:VRC dose, CRP, age and CYP2C19 genotype are the primary covariates affecting NOX or MR, and renal and liver function should also be considered in future population pharmacokinetic models and TDM protocols.

AIMS:Amoxicillin, a widely used β-lactam antibiotic, requires improved pharmacokinetic characterization during breastfeeding. This study used a population pharmacokinetic (PopPK) approach to model amoxicillin concentrations in breast milk, identify variability sources and estimate infant exposure, applying worst-case scenarios. METHODS:Breastfeeding mothers receiving amoxicillin for at least 2 days were enrolled. At steady state, three (nearly) paired blood and milk samples were collected between two doses: 15-30 min, 1-2 h and 3-4 h post-dose. Samples were analysed using LC-MS/MS. PopPK modelling was performed with Monolix. Milk-to-plasma (M/P) concentrations ratios and relative infant dose (RID) were evaluated. RID was calculated relative to both adult and paediatric dosing regimens. Monte Carlo simulations estimated infant exposure at high maternal doses (6 g/day). RESULTS:Twenty-five mother-infant pairs were included. All plasma and milk samples (n = 75/matrix) contained quantifiable amoxicillin. A two-compartment model described the data, with drug transfer from plasma to milk and elimination from milk. Simulated M/P concentrations ratios were 4.1%-5.4%, while observed values ranged from 4.5% to 9.6%. RID based on adult or paediatric dosages remained <0.4%, regardless of measured or simulated concentrations. No adverse effects were reported in breastfed infants. CONCLUSIONS:These findings support the low transfer of amoxicillin into breast milk and align with the absence of adverse effects in breastfed infants, reinforcing its safety during lactation.

Aim: Pharmacogenetic variant frequencies vary substantially across ancestral populations, yet African populations remain understudied. This systematic review aimed to characterize the distribution of pharmacogenetic variants among Ethiopians. Methods: Following the PRISMA guidelines, a literature search was performed across PubMed, Web of Science, Embase, Scopus, and Cochrane Central Register of Controlled Trials. Random-effects meta-analyses were performed to estimate pooled allele, genotype, and genotype-predicted phenotype frequencies. Results: A total of 73 studies were included, involving Ethiopian participants from diverse study populations. Overall, 37 pharmacogenes were investigated, predominantly drug-metabolizing enzymes. Pooling and meta-analyses revealed substantial frequencies of clinically relevant decreased-, increased- and no-function alleles. The no-function CYP3A5*3 allele and decreased-function CYP2B6*6 occurred at a frequency of 62% and 31%, respectively. Decreased-function CYP2C9 alleles (*2 and *3) were observed at lower frequencies (<5%). For CYP2D6, both decreased-function alleles (*41: 21%; *17: 12%), and gene duplications (*1xN and/or *2xN: 14%) were common. The frequencies of no-function CYP2C19*2 and increased-function CYP2C19*17 alleles were 13% and 18%, respectively. The no-function SLCO1B1 c.521T>C variant occurred at a frequency of 19%. Ethnicity substantially contributed to the observed heterogeneity in CYP3A5, CYP3A4, CYP2C9, CYP2C19 and TPMT allele frequencies. Conclusion: These findings provide comprehensive evidence on the distribution and clinical relevance of pharmacogenetic variants in Ethiopians, providing an important foundation for future pharmacogenomics research and implementation. The findings also underscore the need for greater inclusion of African populations in precision medicine research to ensure that genomic medicine is informed by population-specific evidence.

Introduction:Ongoing maternal and clinical hesitancy in breastfeeding-related shared decision-making is driven by limited safety data on maternal pharmacotherapy. Theoretical exposure to maternal enalapril and its active metabolite enalaprilat in breastfed infants has previously been reported in two studies of eight mother-infant pairs. However, actual infant plasma concentrations remain uncharacterized. Methods:A 30-year-old white woman started enalapril (5 mg, 1x/day) for IgA nephropathy at 11 weeks postpartum while exclusively breastfeeding. On day 101 postpartum, 25 days after therapy start, she collected six steady-state milk samples over 24 h, along with two maternal and one infant blood sample, used to calculate milk-to-plasma (M/P) ratio and estimated infant exposure. Samples were analyzed using liquid chromatography with tandem mass spectrometry. Maternal and infant health information was concurrently collected via structured questionnaires. Results:Low levels of enalapril (0.01-1.22 ng/mL) and enalaprilat (0.32-0.77 ng/mL) were measured in human milk. Using 200 and 150 mL/kg/day milk intake, estimated daily infant dosage was 78.32 ng/kg/day and 58.82 ng/kg for enalapril and 124.45 ng/kg/day and 93.34 ng/kg/day for enalaprilat, corresponding to a relative infant dose (RID) of 0.097% and 0.073% for enalapril. Infant enalapril and enalaprilat plasma concentrations were below the lower limit of quantification. Discussion:These data support evidence of minimal transfer of enalapril and enalaprilat into human milk, suggesting low risk to breastfed infants, and may help address uncertainties in current clinical guidelines. This case report provides detailed maternal pharmacokinetic data for both compounds in human milk, alongside estimated infant exposure at 3 months postpartum.

Introduction:The aim of the study was to develop a physiologically based pharmacokinetic (PBPK) model of tafenoquine (TQ), a new drug for the radical cure of Plasmodium vivax malaria and assess drug-drug interactions with substrates of the transporters OCT2 and MATE1. Methods:A whole-body PBPK model was built using PK-Sim® to simulate 100 adults aged 19-53 years weighing 51-108 kg. Hepatic clearance was incorporated into the model. Literature data on the oral administration of TQ were used to verify the model. The qualified model was used to simulate co-administration with metformin (MTF), an OCT2 and MATE1 substrate, and assess drug-drug interactions. Results and Discussion:The PBPK model presented a good descriptive and predictive performance for plasma concentration-time profiles. Quantitative measures of the model performance fell within the 2-fold acceptance criteria when predictions were compared to observed values. Co-administration of TQ with MTF presented no alteration of MTF's exposure. Model simulations suggest that there is no significant interaction between TQ and substrates of OCT2/MATE1, such as MTF.

Iodine is an essential element of human milk. However, high therapeutic maternal doses of potassium iodide (KI) during breastfeeding and its effect on the infant have been insufficiently studied. Therefore, human milk samples were collected within the UmbrelLACT study to determine human milk iodine concentration (HMIC) and estimate infant exposure during and after high maternal KI intake. One mother used KI 3 × 50 mg/day as preparation for thyroidectomy at 5 months postpartum. Human milk samples were collected opportunistically during and within 2 weeks after KI discontinuation. Furthermore, milk samples were collected over 24-h periods at 3 weeks post-discontinuation and monthly from 1 month up to 5 months after discontinuation of KI intake. Breastfeeding was temporarily interrupted during maternal treatment, with partial resumption of breastfeeding 1 week after KI discontinuation, complemented with infant formula upon the participant’s decision. HMIC were analysed using inductively coupled plasma tandem mass spectrometry (ICP-MS/MS) after basic extraction with tetra-methyl ammonium hydroxide (TMAH) and infant exposure was estimated. Median HMIC were 70.8 (46.6–80.8) µg/mL during and 0.064 (0.016–0.227) µg/mL from day six after maternal treatment stop onwards. After discontinuation of KI intake, HMIC decreased quickly and remained within reference range (0.1–0.2 µg/mL). Using the infant requirement of 15 µg/kg/day, relative infant recommended dose (RID recommended ) during and after KI intake was 70805% and 98%, respectively. The infant’s thyroid values were within normal ranges at 3 months after discontinuation of maternal treatment and no adverse events were reported for the infant. Post-discontinuation HMIC are in line with mean concentrations in Nordic countries (68–90 ng/mL). High therapeutic doses of KI should not be combined with breastfeeding, but based on these results (exclusive) breastfeeding could be resumed from 3 weeks after finalisation of maternal treatment. Further research is warranted to determine exposure and safety for breastfed infants. Registration UmbrelLACT study: NCT06042803 ( https://clinicaltrials.gov/study/NCT06042803?cond=umbrellact&rank=1 ).

ABSTRACT Reliable physiologically based pharmacokinetic (PBPK) modeling depends on tissue‐specific expression profiles that reflect protein activities governing drug disposition. In PK‐Sim, existing expression databases rely on transcriptomics data. However, mRNA levels often exhibit limited correlation with protein abundance, frequently necessitating empirical expression modification to align bottom‐up simulations with clinical observations. To address this limitation, we developed ProteinDB as a proteomics‐based expression database for PK‐Sim, using proteomics data primarily from PaxDb v6.0. Raw proteomics data were mapped to gene identifiers and standardized into absolute concentrations (μmol/L tissue) before integration into PK‐Sim. Cross‐platform comparisons were performed for hepatic protein abundance across PBPK platforms, while cross‐omics comparisons were made of relative tissue distributions with transcriptomics‐based PK‐Sim databases. The performance of ProteinDB in PBPK modeling was evaluated using the probe substrates midazolam, digoxin, rifampicin, and tizanidine, with associated drug–drug interactions. Cross‐platform comparisons showed strong agreement for most hepatic enzymes and transporters, while revealing divergences for proteins with greater inter‐individual variability, lower abundance, or limited evidence base. Cross‐omics analyses demonstrated tissue‐dependent discrepancies between transcript‐ and protein‐based expression patterns, with higher consistency observed for kidney and small intestine, particularly with the RT‐PCR and Bgee databases. For PBPK modeling, ProteinDB showed consistently comparable or superior predictive performance for systemic exposure and other clinical endpoints compared with transcriptomics‐based baseline and empirically modified library profiles. By providing a direct physiological basis for system parameterization, ProteinDB offers a robust alternative to current transcriptomics PK‐Sim databases and reduces the reliance on empirical expression modification, thus improving the reliability of prospective PBPK modeling.

Aminoglycoside dosing in suspected neonatal sepsis remains difficult due to highly variable pharmacokinetics driven by marked physiological diversity, from extremely preterm to term neonates, and further complicated by acute kidney injury, perinatal asphyxia, and concomitant interventions. We developed multiscale medical digital twins combining a physiologically-based pharmacokinetic model with an eco-evolutionary pharmacodynamic module capturing drug-modulated bacterial growth and resistance. Glomerular filtration rate is continuously updated using a long short-term memory neural network trained on real-world data. Calibrated on 1634 neonates, the framework enables in silico optimization of full-course antibiotic therapy through real and virtual cohorts, balancing efficacy and safety while accounting for resistance-driven changes in the minimum inhibitory concentration (MIC). Nonlinear optimal control achieved bacteriostatic exposure across all digital-twin neonates, with safety preserved in most cases at higher MICs. Model predictive control further reduced bacterial rebound during late therapy. This framework supports evolution-aware precision dosing of renally cleared antibiotics in vulnerable neonatal populations.

Serum creatinine (Scr) centile values were recently described in a cohort of 1136 (near)-term neonates that underwent therapeutic hypothermia (TH) because of moderate to severe hypoxic-ischemic encephalopathy. Recent methodological progress enables conversion of these Scr centiles to estimated glomerular filtration rate (eGFR) values. Scr centiles in the TH dataset during the first 10 days of life were converted to eGFR values, using the Schwartz formula, with the Smeets k-value (0.31) and fixed body length (50 cm) to generate postnatal reference eGFR values, centiles, and an equation for median eGFRs. These findings were compared to published eGFR data in term controls. A polynomial function was estimated: eGFR(mL/min∙ 1.73m^2)=9.1667+7.1173 -0.3439x^2,(x=days) for eGFR in TH neonates. The median eGFR increases 2- to threefold over the first week (day 1: 16.1; day 2: 19.4; day 7: 41.2 mL/min∙1.73 m2), while the polynomial function does not fully reflect the interindividual variability in eGFR values (intra-day variability is also 2- to threefold). Patterns in acute kidney injury (AKI) TH cases differ significantly from non-AKI TH cases. Based on pooling of published eGFR data, this was compared to a function in healthy term neonates: eGFR(mL/min∙ 1.73 m^2)=14.2167+6.7644-0.3901x^2(x=days) (day 1: 20; day 2: 26; day 7: 42 mL/min/1.73 m2). Based on a pooled dataset in TH cases, we converted Scr centiles to eGFR centiles. Based on median values, this resulted in a polynomial function in TH cases, compared to healthy term neonates. This eGFR function enables precision pharmacotherapy for GFR-cleared drugs in this vulnerable population.

Hypoxic-ischemic encephalopathy (HIE) often results in multi-organ damage. Liver injury following HIE is typically characterized by time-dependent altered patterns of alanine aminotransferase (ALT), aspartate aminotransferase (AST) or total bilirubin (TB) concentrations. This study aims to describe age-dependent patterns of liver injury biomarkers during and after therapeutic hypothermia (TH). Liver injury biomarkers (ALT, AST, TB) over the first 10 days of postnatal age (PNA) from six cohorts, including 428 neonates with moderate-to-severe HIE treated with TH were pooled. Statistical modelling with linear mixed models and quantile regression was applied to assess trends and correlations with HIE severity, gestational age, and birth weight. ALT and AST concentrations were highest on PNA day 1 (median [IQR]: 37.5 [18.75–85.50] U/L), and 154 [79–345 U/L], respectively) and declined thereafter. ALT and AST values were both associated with HIE severity (p < 0.001). The AST/ALT (De Ritis) ratio remained > 3.0 throughout PNA days 1–10. Conclusion: We characterize age-dependent reference values of liver-injury biomarkers in HIE neonates treated with TH. Along with PNA, HIE severity has a strong association with liver biomarkers, while the De Ritis ratio reflects ischemic hepatic injury throughout PNA (day 1–10). These age-dependent reference values can be used to assess intra- and interpatient variability, or to explore other causes of hepatic toxicity like drug-induced liver injury, preferably following external validation.

IntroductionKnowledge regarding lorazepam use during breastfeeding is limited and so are quantitative data on lorazepam transfer into human milk. Lorazepam transfer to human milk was characterized in a single case and pharmacovigilance databases were explored.MethodsA clinical lactation study was conducted in a breastfeeding mother who used lorazepam as needed for anxiety symptoms (single 1 mg oral dose, 6.5 months postpartum). Serial human milk (n = 6) and maternal plasma (n = 2) samples were collected over 24 h. Concentration-time data were used to calculate plasma and milk AUC0–24h (linear trapezoidal interpolation), milk-to-plasma (M/P) ratios, daily and relative infant dose (DID, RID). Because of the a priori pragmatic design (limited sampling), published plasma concentration–time profiles following oral dose were searched, digitized and dose-normalized to derive a plasma AUC0–24h. Pharmacovigilance data from the U.S. Food and Drug Administration Adverse Event Reporting System (FAERS) database and Eudravigilance were analyzed for “lorazepam related exposure via human milk”.ResultsMilk concentrations ranged from 0.65 to 2.10 ng/mL, with highest levels at 1.5 h. The milk AUC0–24h was 25.36 ng*h/mL. Literature-derived plasma AUC0–24h was 158.94 ng*h/mL resulting in an AUC-based M/P ratio of 0.16, calculated DID was 218 ng/kg/day. Assuming milk intakes of 150 and 200 mL/kg/day, estimated DID values were 176.26 and 235.02 ng/kg/day respectively, corresponding to RID values of 1.35% and 1.80%, increasing up to 3.85% upper-bound estimates. The RIDtherapeutic was 0.44%, increasing to 0.81% in upper-bound estimates. FAERS identified 25 reports of lorazepam exposure via human milk, Eudravigilance one, with mainly central nervous system (CNS) related symptoms. In all these cases, lorazepam was part of CNS polypharmacy, while no lorazepam monotherapy related events were retrieved.ConclusionThese findings suggest low infant exposure following occasional single, low-dose lorazepam use during breastfeeding. While interpretation should remain cautious given the single-case design, the single low dose, limited plasma sampling, and milk concentrations below the validated lower limit of quantification, the available pharmacokinetic data suggest low estimated infant exposure following occasional low-dose lorazepam use. Pharmacovigilance findings provided contextual information but should not be interpreted as confirmatory evidence of safety.

In vitro to in vivo extrapolation (IVIVE) methods for hepatic clearance (CLH) prediction often underpredict, partly due to reliance on mathematical liver disposition models such as the well-stirred model (WSM) or parallel tube model (PTM). The ex vivo isolated perfused rat liver (IPRL) model bridges in vitro and in vivo data, providing mechanistic insights into the predictive accuracy of IVIVE models. This study evaluates the IPRL model across a diverse selection of 16 compounds, and benchmarks results against in vitro and in vivo data to verify the predictive performance of the WSM and PTM. Results demonstrate that both the IPRL and in vivo clearance conflict with assumptions of the WSM (AAFE = 2.85) or PTM (AAFE = 1.74), which consider the liver outlet concentration as a driver for the hepatic elimination rate. However, except for terfenadine, IPRL clearance predictions were within two-fold (AAFE = 1.59) of in vivo clearance when the liver inlet concentration was utilized to calculate the CLH. When employing the WSM or PTM for in vitro to ex vivo extrapolation, underpredictions were observed for compounds with high plasma protein binding and subject to sinusoidal hepatic uptake, reflecting model oversimplification compared to in vivo dynamics. Our findings experimentally challenge the theoretical assumptions underlying the use of the WSM and PTM in IVIVE methods. Unique insights from the IPRL model point to the next steps needed to advance IVIVE: refining current liver disposition models through enhanced and next-generation in vitro assays, capturing dynamic in vivo disposition mechanisms, and exploring complementary models.

BACKGROUND AND PURPOSE:Therapeutic hypothermia (TH) decreases morbidity and mortality in neonates with perinatal asphyxia (PA) and moderate-to-severe hypoxic-ischaemic encephalopathy. Midazolam is an antiepileptic and substrate of CYP3A4. Drug metabolism during PA/TH remains partially characterized, and disentangling effects of PA and TH is difficult, making precision dosing challenging. The I-PREDICT project aims to develop a framework for drug disposition in PA/TH neonates, using physiologically based pharmacokinetic (PBPK) modelling. The predictive performance of the model was explored for midazolam and its active metabolites, 1'-hydroxymidazolam and 1'-hydroxymidazolam-O-glucuronide, in neonates with PA/TH. EXPERIMENTAL APPROACH:The model was established stepwise, starting with healthy adults, and extrapolated to paediatric (1 month-18 years) and neonatal (0-27 days, non-asphyxiated) populations. Finally, PA/TH-related physiological and metabolic changes were incorporated. For each population, performance was verified using clinical data. KEY RESULTS:Model predictive performance for midazolam and 1'-hydroxymidazolam was acceptable in adults and children, with AFE and AAFE values ranging from 0.87 to 1.22 and 1.15 to 1.23, respectively. The neonatal PA/TH PBPK model captured midazolam pharmacokinetics, with 71% of the observations within the 10th-90th percentiles and 46% within the 25th-75th percentiles of the prediction. Incorporating PA/TH physiology improved the neonatal PA/TH model by 6%-43% (within the 10th-90th percentiles), depending on the compound. CONCLUSIONS AND IMPLICATIONS:This PA/TH-adjusted neonatal PBPK model may support future midazolam dose optimization through mechanistic understanding and provides a framework for other drugs in neonates with PA/TH. Future knowledge on the impact of TH on protein binding and renal clearance of 1'-hydroxymidazolam-O-glucuronide would further improve the model.

Background Clinical lactation studies are valuable to guide pharmacotherapy during breastfeeding, but are not easy to conduct. This paper reports on challenges and mitigation strategies while conducting the UmbrelLACT study, a prospective observational lactation study.Methods The UmbrelLACT study includes breastfeeding women taking a relevant medicine (eg, absence of safety evidence during lactation) and after feasibility verification (eg, access to bioanalytical assay). Participants collect human milk samples at home over 24 hours. Optionally, maternal and infant blood samples are collected, together with self-reported questionnaires on clinical maternal and child variables. Medicine concentrations and estimated infant exposure (eg, daily and relative infant dose) are determined.Results Regarding informed consent, study objectives should be carefully phrased to prevent pharmacotherapy avoidance due to a lack of safety information. On sample collection, the study team should provide all necessary materials if not available at home, including an electric breast pump. Due to the milk expressions over 24 hours, the infant might miss soothing moments. This can be mitigated by giving the breast after expressing the milk for study purposes, creating alternative soothing moments. Sample handling might be complicated by additional steps needed for certain compounds. A bioanalysis method should be available or developed for the milk matrix. Finding the assay might be challenging, but it can be facilitated by a network of specialised labs. In the UmbrelLACT study, pharmacokinetic values are collected fragmentarily. This results in the need for data pooling with data from the literature. In contrast, at-home sampling and multidisciplinary collaboration are clear strengths.Conclusions It is feasible to tailor the approach to each participant and compound, to minimise the burden on the patient and their family. While there are limitations related to the pragmatic design, the opportunities of this lactation study outweigh the challenges to result in clinically significant scientific observations.Trial registration number NCT06042803.

Developed at Bayer Technology Services, PK-Sim and MoBi transitioned into the Open Systems Pharmacology (OSP) Suite, released as free open-source software in 2017. An active community with stakeholders from academia, industries, and regulators contributes to the continuous improvement of open-source model-informed drug development (MIDD). This perspective summarizes the latest advancements presented at the second OSP Community Conference (OSP-CC) hosted from 29 to 30th of September 2025 at Sanofi Paris, that gathered over 100 attendees from more than 40 institutions.

Model-informed precision dosing (MIPD) of tacrolimus in renal allograft recipients, evaluated in silico, demonstrated improved (time to and) probability of target concentration attainment and smaller deviations from target range. Using simulated tacrolimus concentration-time profiles, a study of 200 patients was predicted to have sufficient power to demonstrate superior performance of MIPD for these exposure end points compared with physician-based dosing. A fully automated tacrolimus MIPD application integrated in the electronic patient file was tested in 293 de novo recipients in the first 14 days after transplantation in a prospective randomized controlled clinical validation study. More patients dosed with the MIPD application reached the primary study end point of three in-target tacrolimus pre-dose trough concentrations by Day 8, compared with physician-dosed patients: 52.2% (95% CI: 45.3-59.6) vs. 35.7% (95% CI: 27.4-45.6); HR: 1.64 (95% CI: 1.10-2.43) (P = 0.015). The mean fraction of samples per patient in target during the complete study period was higher in the MIPD arm: 0.38 ± 0.14 compared with the physician-dosed arm: 0.28 ± 0.14 (P < 0.001). The mean distance from target window was significantly lower in MIPD-treated patients: 0.022 (95% CI: 0.019-0.024) vs. 0.040 (95% CI: 0.036-0.044) (P < 0.001). On 19 occasions (< 1%), MIPD tacrolimus dose suggestions were actively overruled by physicians. A fully automated MIPD application for tacrolimus in de novo renal recipients, integrated in the electronic patient file, demonstrated superior performance in achieving tacrolimus exposure targets with limited active overruling of MIPD dose executions by physicians. Automated MIPD can be tested in larger trials to evaluate the impact of dosing decision support on clinical outcomes after renal transplantation.

Amoxicillin is frequently prescribed to peripartum women as prophylaxis and treatment for infection. Despite frequent usage of amoxicillin in lactating women, quantitative evidence regarding its transfer between systemic circulation and breastmilk remains limited. However, such data is fundamentally essential for assessing the safety risk of medicine usage during breastfeeding. This research performed population pharmacokinetic (PopPK) modeling and simulation to characterize the pharmacokinetics of amoxicillin in plasma and breastmilk in lactating Göttingen Minipigs, a physiologically relevant animal model to humans that enables studying medication transfer during lactation. Demographic characteristics and amoxicillin concentrations in plasma and breastmilk following daily intramuscular doses (7 mg/kg) in 9 minipigs from two studies were included in the PopPK analysis. The milk-to-plasma (M/P) ratio, daily infant dosage (DID), and relative infant dose (RID) were calculated based on the simulated steady-state exposure. The final model consisted of a two-compartment distribution with zero-order absorption and linear elimination. The central compartment was deemed as the plasma compartment with time-varying volume of distribution, and linked to the breastmilk compartment via bidirectional first-order transfer. No covariate was included. The developed PopPK model well described amoxicillin in plasma and breastmilk in the lactating Göttingen Minipigs. The predicted M/P ratio, DID, and RID of amoxicillin following the maximal doses were approximately 0.25, 0.11 mg·kg−1·day−1, and 1.6