Pediatric physiologically based pharmacokinetic models mediated by drug-metabolizing enzymes: Predictions, applications, and challenges

Juan Wu , Yawen Yuan , Haonan Li , Junyu Li , Cheng Jin

Intelligent Pharmacy ›› 2026, Vol. 4 ›› Issue (3) : 217 -225.

PDF (1014KB)
Intelligent Pharmacy ›› 2026, Vol. 4 ›› Issue (3) :217 -225. DOI: 10.1016/j.ipha.2026.04.001
Review article
Pediatric physiologically based pharmacokinetic models mediated by drug-metabolizing enzymes: Predictions, applications, and challenges
Author information +
History +
PDF (1014KB)

Abstract

Children, as a special population, exhibit significant age-specific characteristics and inter-individual variability in drug metabolism due to the influences of physiological development and genetic factors, bringing challenges to the efficacy and safety of clinical medication. Traditional pharmacokinetic studies are difficult to conduct in children and are constrained by numerous ethical limitations. Physiologically based pharmacokinetic (PBPK) modeling, as a mechanistic modeling tool, can integrate pediatric developmental physiological parameters, enzyme kinetic characteristics, and genetic polymorphism data. It has gradually become an important tool for predicting pediatric drug metabolism and optimizing dosing regimens. This review summarizes the key enzymatic characteristics of pediatric drug metabolism and the technological pathways and current application status of PBPK models in assessing pediatric drug metabolism. It discusses the major challenges, such as uncertainties in parameter extrapolation and the lack of pediatric-specific data. It further elaborates on the practical value of PBPK models in optimizing dosing regimens, supporting pediatric clinical trial design, and aiding regulatory decision-making. Finally, it proposes future directions for development by integrating artificial intelligence and multi-omics data. This review aims to provide a theoretical basis and practical reference for precision dosing in children and to promote the broader application of PBPK models in the pediatric pharmaceutical field.

Graphical abstract

Keywords

Physiologically based pharmacokinetic model Pediatrics / Drug-metabolizing enzymes / Genetic polymorphism / Precision dosing / Dose prediction

Cite this article

Download citation ▾
Juan Wu, Yawen Yuan, Haonan Li, Junyu Li, Cheng Jin. Pediatric physiologically based pharmacokinetic models mediated by drug-metabolizing enzymes: Predictions, applications, and challenges. Intelligent Pharmacy, 2026, 4 (3) : 217-225 DOI:10.1016/j.ipha.2026.04.001

登录浏览全文

4963

注册一个新账户 忘记密码

1 Introduction

In the pediatric population, drug metabolism is a dynamic process fundamentally shaped by the forces of both physiological development and genetic inheritance, leading to significant variability in therapeutic outcomes. The liver, as the core organ of drug metabolism, features an enzyme system whose ontogeny displays significant isoform specificity and non-linear dynamic characteristics 1–4. The content of the cytochrome P450 (CYP450) enzyme family reaches only 50% of adult levels by 6–12 months after birth, and the capacity for phase II conjugation reactions is also significantly delayed.5 Concurrently, neonatal hepatic blood flow is relatively low and may not reach adult levels until infancy or early school age. These characteristics result in unique pharmacokinetic profiles for hepatically metabolized drugs across different developmental stages in children.

Genetic polymorphism represents another critical factor influencing pediatric drug metabolism. By modulating CYP450 enzyme activity, it classifies individuals into ultra-rapid metabolizers (UMs), extensive metabolizers (EMs), intermediate metabolizers (IMs), and poor metabolizers (PMs), leading to significant variability in drug metabolic rates. This variability can potentially result in accumulation toxicity, therapeutic failure, or excessive generation of toxic metabolites.6 The immature development of the hepatic enzyme system in children confers a weaker compensatory capacity for fluctuations in enzyme activity, making the genetic determinism of metabolic phenotypes particularly pronounced. Therefore, accurate prediction of drug metabolism is essential to optimize dosing, enable personalized medicine, and reduce adverse effects in children.

The PBPK model is a mechanism-driven modeling approach grounded in physiological, anatomical, and physicochemical principles. By designating each tissue or organ as a distinct compartment, it quantifies hemodynamics, tissue characteristics, and drug-enzyme interactions, thereby enabling dynamic simulation of the entire absorption, distribution, metabolism, and excretion (ADME) process of drugs.7,8 In the context of predicting pediatric drug metabolism, PBPK models can integrate developmental physiological parameters, enzyme kinetic data, and genetic information to accurately simulate metabolic differences from neonates to adolescents, offering dose optimization strategies for diverse patient populations and identifying potential safety risks. Currently, PBPK models have been widely applied throughout the entire process of pediatric drug development.9,10

This review aims to systematically summarize the core technologies and clinical applications of PBPK models in assessing pediatric drug metabolism. It will focus on elucidating how these models achieve accurate predictions by integrating pediatric developmental physiological parameters with genetic polymorphism data. The review will also explore current challenges in model application, such as uncertainties in parameter extrapolation and the lack of pediatric-specific data. Furthermore, it will discuss the practical value of PBPK models in optimizing individualized dosing regimens, improving pediatric clinical trial design, and supporting regulatory decision-making, ultimately providing a theoretical basis and practical reference for precision dosing research in children.

2 Drug-metabolizing enzymes in pediatric drug metabolism

2.1 CYP450 enzymes

2.1.1 CYP1A2

CYP1A2 accounts for approximately 13% of the total human hepatic CYP450 protein content and is primarily expressed in the liver, with distribution also observed in tissues such as the intestine.11 This enzyme is involved in the metabolic conversion of various clinically commonly used drugs, including theophylline and caffeine.12,13

The expression of CYP1A2 exhibits significant age-dependent characteristics. During the fetal period, CYP1A2 is virtually absent. By postnatal day 7, its expression reaches approximately 5% of adult levels. Expression becomes active between 1 and 3 months of age, attains 50% of adult levels by 1 year, and approaches adult levels by approximately 3 years of age.14

The CYP1A2 gene is highly polymorphic. The most extensively studied functional variants includes CYP1A2*1C, *1D, *1E, and *1F.15 Additionally, the -3113A > G polymorphism has a distribution frequency of approximately 10% in the Chinese population and is associated with decreased enzyme activity.16

The regulation of enzyme activity by different genotypes is complex in children. A study involving 104 Korean preterm infants found that CYP1A2 activity was significantly correlated with theophylline clearance rate (P < 0.001), but was primarily influenced by postnatal age and postmenstrual age, with no significant differences observed in the effects of CYP1A2*1C and *1F genotypes on enzyme activity.17 Another study in preterm infants similarly confirmed that these genotypes did not significantly affect theophylline clearance.18 This suggests that developmental factors dominate CYP1A2 regulation during the neonatal period, whereas the influence of genetic polymorphisms gradually emerges in older children. A study in Serbian pediatric patients with epilepsy demonstrated that children carrying the CYP1A2 -163A/A genotype exhibited significantly increased clearance of carbamazepine.19 Berel et al.20 reported two pediatric patients receiving clozapine therapy who presented with persistently low plasma drug concentrations. Both were confirmed to be CYP1A2 UMs. This issue was subsequently corrected by the adjunctive administration of fluvoxamine, a potent CYP1A2 inhibitor. Furthermore, as CYP1A2 serves as the primary metabolic enzyme for atypical antipsychotics such as olanzapine and clozapine, its ontogeny and genetic polymorphisms may significantly impact pediatric pharmacokinetics. However, relevant research remains limited.21 Currently, drug labeling information pertaining to CYP1A2-related considerations is also scarce.22

2.1.2 CYP2C9

CYP2C9 constitutes approximately 20% of total hepatic CYP protein content and mediates the metabolism of over 100 drugs.23,24 Its sub-strates include phenytoin, warfarin, tolbutamide, and various non-steroidal anti-inflammatory drugs (NSAIDs), and it is also involved in the activation of prodrugs such as cyclophosphamide and ifosfamide. Given that many of these drugs possess narrow therapeutic windows, the ontogeny and genetic polymorphisms of CYP2C9 have critical implications for medication safety in pediatric populations.

The expression of CYP2C9 exhibits marked age dependence. During early gestation, expression levels are only 1%–2% of mature values. By mid-to-late gestation, they increase to approximately 27%, with sub-stantial interindividual variability.25,26 Postnatally, CYP2C9 expression increases significantly. Between the neonatal period and 5 months of age, 51% of samples have achieved mature expression levels, yet considerable variability remains. Notably, CYP2C9 activity in children aged 3–10 years significantly exceeds adult levels and does not decline to adult values until the end of puberty.25

The most clinically significant alleles of CYP2C9 are the wild-type(CYP2C9*1) and two major variants: the R144C mutation (CYP2C9*2) and the I359L mutation (CYP2C9*3), with the latter being the primary cause of reduced enzyme activity. Based on metabolic phenotype, individuals carrying CYP2C9*1/*1、*1/*2 and *2/*2 are classified as EMs. Those with *1/*3、*2/*3 are classified as IMs, and *3/*3 as PMs.26

The impact of CYP2C9 genetic polymorphisms on pediatric pharmacotherapy has been substantiated in multiple clinical studies. A meta-analysis encompassing 8 clinical studies involving 507 pediatric patients demonstrated that children carrying CYP2C9*1/*2、*1/*3, or at least one variant allele (*2 or *3) required 15% to 41% lower warfarin maintenance doses compared to wild-type carriers (P < 0.05).27 Another meta-analysis comprising 6 studies with 847 pediatric epilepsy patients revealed a significant association between CYP2C9 polymorphisms and valproic acid plasma concentrations (P = 0.02). The impact of the *1/*3 phenotype was more pronounced in children than in mixed-age populations (P = 0.04).28 A study involving 89 Indian children with epilepsy found that serum phenytoin levels were significantly higher in the CYP2C9*3 genotype group compared to the wild-type group(P = 0.009), although no significant difference in adverse reaction incidence was observed (P = 0.597).29

As a key metabolic enzyme for numerous drugs with narrow therapeutic windows, CYP2C9 exhibits both developmental overshoot in activity and genetic polymorphisms, collectively contributing to the complexity of pediatric pharmacotherapy. However, dose optimization studies specifically targeting children at different developmental stages and with distinct genotypes remain relatively limited. Prospective clinical investigations are urgently needed to provide more robust evidence-based guidance.

2.1.3 CYP2C19

CYP2C19 is primarily localized in hepatic microsomes and shares 92% sequence homology with CYP2C9. It predominantly mediates hydroxylation reactions of drugs. Common substrates include proton pump inhibitors (e.g., omeprazole, lansoprazole), tricyclic antidepressants (e.g., imipramine, amitriptyline), selective serotonin reuptake inhibitors(e.g., fluoxetine and sertraline), benzodiazepines (e.g., diazepam), phenobarbital, phenytoin, voriconazole, and selegiline.30

CYP2C19 protein and catalytic activity are detectable as early as the first trimester of gestation, reaching approximately 12%–15% of adult levels by 8 weeks of gestation31 and remaining relatively stable throughout the fetal period. Postnatally, the expression pattern of CYP2C19 shows no abrupt changes.32 Activity gradually increases between 0 and 5 months of age. From 5 months to 10 years of age, sub-stantial interindividual variability is observed, which may be attributable to genetic factors. During the period of 10 to 18 years of age, expression continues to increase, approaching adult levels. Some studies have also reported that CYP2C19 activity in children aged 2–8 years exceeds that observed in adults.33

CYP2C19 exhibits highly genetic polymorphisms. Among the most common variant alleles are CYP2C19*2 (G681A in exon 5) and CYP2C19*3 (G636A in exon 4), both of which are loss-of-function alleles resulting in reduced enzyme activity. Based on genotype, individuals can be classified as EMs (*1/*1), IMs (*1/*2 or *1/*3), and PMs (*2/*2,*3/*3). Additionally, CYP2C19*17 has been demonstrated to be associated with increased enzyme activity.34

CYP2C19 genetic polymorphisms significantly influence the disposition of various drugs in pediatric populations. A pharmacokinetic study conducted in 40 children demonstrated that pantoprazole AUC was significantly lower in individuals carrying the CYP2C19*17/*17 genotype compared to those with *1/*1 and *1/*17 genotypes, with significant associations observed between genotype and dose-normalized AUC (P < 0.0001) as well as apparent elimination rate constant (P = 0.0012).35 Another clinical study involving 207 pediatric patients with Helicobacter pylori infection reported that although the eradication rate was slightly higher in PMs (96.4%) than in EMs (78.3%) and IMs (83.3%), the differences did not reach statistical significance.36 Regarding adverse effects, a retrospective analysis of 259 children undergoing allogeneic hematopoietic stem cell transplantation revealed that those carrying CYP2C19*2 mutations had a significantly increased risk of hemorrhagic cystitis (P = 0.021), potentially associated with altered cyclophosphamide metabolism leading to elevated urinary acrolein exposure.37 Multiple studies have also confirmed that CYP2C19 genotype is a key determinant of voriconazole plasma concentrations in children, with PMs exhibiting significantly higher trough concentrations and requiring lower maintenance doses to achieve therapeutic targets.38,39 A retrospective study in Japanese patients with epilepsy(n = 1901) demonstrated that CYP2C19 genotype significantly influenced lacosamide plasma concentrations and treatment persistence, with three-year treatment retention rates being significantly higher in non-PMs than in PMs (P < 0.05), suggesting that poor metabolizer status may compromise long-term medication adherence due to excessively high drug concentrations.40

2.1.4 CYP2D6

Although CYP2D6 constitutes less than 2% of the total hepatic CYP enzyme content, it is involved in the metabolism of approximately 25% of commonly used clinical drugs, primarily exerting its effects through first-pass metabolism in the liver.41 This enzyme is also expressed at low levels in the kidney, intestine, lung, placenta, and brain. Substrates metabolized by CYP2D6 include dextromethorphan, tricyclic antidepressants (e.g., imipramine, nortriptyline), selective serotonin reuptake inhibitors (e.g., fluoxetine, paroxetine), non-tricyclic antidepressants (e.g., atomoxetine, venlafaxine), antipsychotics (e.g., olanzapine, risperidone, haloperidol), and β-blockers (e.g., metoprolol, propranolol).42

Studies have demonstrated that CYP2D6 enzyme activity stabilizes after 42 weeks of postmenstrual age. Consequently, genetic variation serves as the primary determinant of CYP2D6-mediated substrate disposition capacity in infants, children, and adolescents.43

CYP2D6 exhibits over 100 allelic variants.44 Based on functional activity, these can be categorized into normal-function groups (e.g., CYP2D6*1 and *2), reduced-function groups (e.g., CYP2D6*9, *10, *17, and *41), and non-functional groups (e.g., CYP2D6*3, *4, *5, and*6).45,46 Corresponding phenotypes include UMs, EMs, IMs, and PMs.47

CYP2D6 genetic polymorphism is a key genetic determinant of interindividual variability in drug metabolism. A prospective observational study involving 63 psychiatric patients aged 10–18 years demonstrated that EMs exhibited better therapeutic responses to anti-psychotic medications, with significant improvements in clinical symptoms and social functioning, whereas IMs showed poorer efficacy, delayed treatment response, and a higher prevalence of abnormal brain imaging findings.48 In a study of aripiprazole for the treatment of tic disorders in children, significant differences in metabolic ratios and clearance rates were observed among UM, EM, and IM groups, with UMs showing significantly lower clinical response rates compared to EM and IM patients.49 Another study involving 56 adolescents with psychiatric disorders receiving risperidone, aripiprazole, or olanzapine reported that CYP2D6 genotype influenced treatment efficacy and quality of life. EMs demonstrated significantly greater reductions in Positive and Negative Syndrome Scale (PANSS) scores at 3, 6, and 12 months of treatment compared to IMs (P < 0.05), PMs exhibited the highest PANSS scores, and EMs had the lowest incidence of adverse reactions.50 Furthermore, a pharmacokinetic model developed using intravenous administration data in neonates revealed that CYP2D6 genotype significantly affected the clearance of tramadol and its active metabolite, underscoring the necessity of genotype-guided dosing.51

2.1.5 CYP3A

CYP3A represents one of the most critical subfamilies of drugmetabolizing enzymes in humans, involved in the metabolism of approximately 50% of drugs. CYP3A4 and CYP3A5 are the primary functional isoforms. CYP3A enzymes are highly expressed in the liver and small intestine, with additional distribution in tissues such as the kidney, lung, brain, duodenum, and lymphocytes.52

CYP3A4 is the predominant (approximately 30%) CYP enzyme expressed in the liver after birth. Significant interindividual variability has been observed in the hepatic expression and functional activity of CYP3A4. Its expression is extremely low during the fetal and early neonatal periods, then transcriptional activation occurs within the first week after birth. Protein levels then increase gradually over the first 6 months and approach adult levels by around 4 years of age.53 The clearance of CYP3A4 substrates appears to follow an age-dependent pattern. Activity may be higher in toddlers and children than in adults, showing an inverted U-shaped trajectory.54,55

The CYP3A4 gene exhibits polymorphisms, but coding region variants have low allele frequencies. Their impact on enzyme expression or function appears limited. These variants may partially explain interindividual differences in drug clearance, but they may not the major cause.56 However, many drugs can induce (e.g., carbamazepine, phenytoin, rifampicin) or inhibit (e.g., erythromycin, itraconazole, cyclosporine) CYP3A4 activity. Developmental changes in enzyme function may amplify the clinical effects of these drug interactions, especially in children.

CYP3A5 is the major extrahepatic CYP3A isoform. Its activity varies markedly among individuals but shows no clear age-related pattern. The substrate profile of CYP3A5 overlaps with that of CYP3A4, with minor differences in kinetic properties. Genetic polymorphisms of CYP3A5 are closely linked to enzyme activity. However, its overall contribution to drug metabolism remains to be further clarified.57

In clinical studies, one investigation in children with nephrotic range proteinuria showed that co-administration of diltiazem increased the dose-adjusted trough concentration of tacrolimus by approximately 2-fold. The increase was more pronounced in CYP3A5*3 expressers (1.8-fold) than in non-expressers (1.3-fold).58 Another retrospective study involving 135 pediatric patients found that age under 4 years and co-administration of enzyme inducers significantly reduced the concentration - dose ratio of perampanel. In contrast, co-administration of valproic acid or carrying the CYP3A5*3/*3 genotype significantly increased this ratio.59 Among 60 pediatric heart transplant recipients, children carrying CYP3A4*22 or meeting the CYP3A PM phenotype(CYP3A5*3/*3 and CYP3A4*22) required significantly lower doses to achieve target tacrolimus trough concentrations and exhibited higher dose-adjusted trough concentrations.60

2.2 UDP-glucuronosyltransferase

UDP-glucuronosyltransferase (UGT) catalyzes glucuronidation, a major phase II metabolic pathway in drug disposition. Based on amino acid sequence homology, the UGT superfamily is primarily divided into two families, UGT1 and UGT2. Members of the UGT1A subfamily play a key role in drug metabolism and the clearance of endogenous substances.

2.2.1 UGT1A1

UGT1A1 encodes the UGT enzyme responsible for bilirubin glucuronidation. This isoform is primarily expressed in the liver. It catalyzes the glucuronidation of bilirubin and other hydrophobic endogenous and exogenous compounds. As a specific substrate of UGT1A1, the metabolic clearance efficiency of bilirubin directly reflects the in vivo functional status of this enzyme.61 The UGT1A1 gene is highly polymorphic. Among these variants, UGT1A1*6 and UGT1A1*28 are the two most common loss-of-function alleles in Asian populations. Both mutations can lead to a significant reduction (up to 70%) in enzyme activity.62

Several clinical studies have investigated the impact of UGT1A1 polymorphisms on drug metabolism and treatment safety in children. In a study of 44 pediatric oncology patients receiving irinotecan, no significant association was found between UGT1A1*6 or UGT1A1*28 genotypes and the incidence of severe (grade III-IV) toxicity. However, carrying a mutant allele (UGT1A1*6 or UGT1A1*28) appeared to be linked to lower overall survival rates. Further analysis revealed that children with the UGT1A1*6 mutation had a significantly higher risk of grade I–IV irinotecan-related diarrhea (P = 0.043) and anemia(P = 0.002). However, this conclusion requires validation in large-scale prospective studies.63 In antiepileptic therapy, a population pharmacokinetic study analyzed 236 topiramate plasma concentrations from 181 children with epilepsy. The results identified UGT1A1 rs4148324 as a significant covariate affecting topiramate clearance.64 Another cohort study of 163 children with epilepsy further confirmed that UGT1A1 rs4148324 was significantly associated with topiramate plasma concentrations. Patients with the TG or GG genotypes exhibited significantly higher standardized topiramate concentrations than those with the wild-type TT genotype. This phenomenon may be explained by the high linkage disequilibrium between rs4148324 and rs34983651(UGT1A1*28). These two variants may synergistically affect UGT1A1 transcriptional activity and metabolic capacity, thereby altering the glucuronidation clearance of topiramate.65

2.2.2 UGT1A6

UGT1A6 is another important member of the UGT1A family. It primarily catalyzes the glucuronidation of phenolic compounds in the liver. It is also expressed to varying degrees in extrahepatic tissues, including the kidney, stomach, small intestine, colon, and lung.61 UGT1A6 conjugates numerous drugs and xenobiotics. It also exhibits high conjugation efficiency for serotonin, an important neurotransmitter.66

Valproic acid is a first-line treatment for childhood epilepsy, and its in vivo clearance is primarily achieved through UGT-mediated glucuronidation. A study of 98 epilepsy patients receiving valproic acid monotherapy showed that patients with double heterozygous mutations at the T19G, A541G, and A552C loci of the UGT1A6 gene required higher valproic acid doses than those with wild-type or single heterozygous mutations (P = 0.010). Dose-adjusted plasma concentrations were significantly lower than those in single heterozygous mutation patients (P = 0.027).67 However, other studies reported no significant association between UGT1A6 polymorphisms (T19G, A541G, A552C) and valproic acid plasma concentrations or clinical outcomes in children.68,69 Therefore, larger, multicenter, prospective studies will help clarify the clinical value of UGT1A6 in personalized pediatric therapy.

3 Application of PBPK models in pediatric drug development and precision dosing

3.1 Development of pediatric PBPK models

PBPK modeling offers distinct theoretical advantages and broad application prospects in pediatric drug research compared to traditional population pharmacokinetic approaches. Its core strength lies in integrating key physiological characteristics (e.g., developmental stage, disease status, and genetic factors) into a mechanistic framework through organ-specific parameterization. This enables mechanistic insights into drug disposition in special populations, particularly in children. Consequently, relatively abundant data from adults or healthy subjects can be extrapolated to pediatric populations with limited clinical data, thereby supporting initial pediatric dosing regimen design.70 For pediatric populations, the key value of PBPK modeling is its capacity to incorporate individualization equations based on physiological developmental trajectories. These equations integrate developmental changes, including organ volume and blood perfusion, renal function maturation, and age-dependent expression of drug-metabolizing enzymes and transporters, into the prediction framework. Unlike conventional models relying on statistical fitting, PBPK modeling provides more detailed mechanistic insights. This facilitates a deeper understanding of how age-related factors influence the ADME of drugs, leading to more reliable predictions of pediatric drug safety and efficacy. Following decades of methodological development and validation, pediatric PBPK modeling has become an essential tool in drug development and clinical individualized therapy.

The general workflow for developing pediatric PBPK models is illustrated in Fig. 1. To ensure reliable model structure and accurate parameter estimation, a widely recommended modeling strategy involves first establishing a well-validated adult PBPK model before constructing a pediatric version. Adult model development helps identify key pharmacokinetic processes of the drug in vivo (e.g., major clearance pathways, tissue distribution characteristics) and determine drugspecific parameters independent of age (e.g., lipophilicity, plasma protein binding, membrane permeability). Subsequently, the adult model is extrapolated to pediatric populations by substituting age-specific physiological parameters (e.g., organ volume, tissue blood perfusion rate, plasma protein concentrations, and glomerular filtration rate).

It should be noted that pediatric PBPK modeling still faces data gaps for several key parameters. Although anatomical and physiological parameters such as organ volume and blood flow are well-characterized across age groups, systematic and quantitative age-related data on the expression and activity of drug-metabolizing enzymes and transporters during childhood development remain limited. These information gaps directly impact the predictive accuracy of models for critical parameters such as drug clearance. However, if age-related expression patterns of metabolic enzymes and transporters can be obtained through literature integration or in vitro experiments, and model predictions are adequately validated in independent clinical datasets, pediatric PBPK models can still be considered suitable for their intended use. Such models can support initial pediatric dose selection and optimization of clinical trial protocols.71

3.2 PBPK modeling in characterizing pediatric drug metabolism

PBPK modeling leverages in vitro data to enable in vitro-in vivo extrapolation (IVIVE), providing an important approach for predicting drug clearance in pediatric populations. For example, based on in vitro activity assays using recombinant drug-metabolizing enzymes, drug clearance per unit of enzyme protein can be calculated. This value can then be extrapolated to whole-liver clearance for different age groups by incorporating age-specific liver weight and target enzyme expression per gram of liver tissue. This approach is particularly valuable for innovative drugs lacking pediatric clinical data, offering critical references for first-in-pediatric dose design.72 Complementarily, when adult in vivo pharmacokinetic data are available, adult clearance can be scaled to pediatric populations using physiological knowledge of age-related differences in metabolic enzyme expression. This can further improve prediction reliability. If data from in vitro experiments (e.g., recombinant enzymes, human liver S9 fractions, human liver microsomes, or human hepatocytes) have been used to estimate hepatic clearance in adult models, these data can also be extrapolated to pediatric models by accounting for reported differences in enzyme expression and activity. Currently, for drug metabolism mediated by enzymes such as CYP1A2, CYP2C8, CYP2C9, CYP2C19, CYP2D6, and CYP3A4, ontogeny models have been further refined through validation against clinically observed clearance values. These models may enable accurate predictions in children as young as 1 month of age.73

Parvez et al.74 measured the expression levels of CYP enzymes involved in metronidazole metabolism using liver microsomes from 115 children and 35 adults. Results showed that CYP2A6, CYP2E1, and CYP3A4 expression increased with age, while CYP3A7 exhibited a postnatal decline. CYP2A6 and CYP3A5 expression were significantly associated with rs4803381 and rs776746 genotypes, respectively. A PBPK model developed based on the proteomics data successfully characterized the in vivo disposition of metronidazole and its active metabolite. The metabolite-to-parent drug ratio increased with age, and CYP2A6 expression showed a strong correlation with metabolite formation (r2 = 0.75). Notably, CYP3A7 contributed approximately 75% to metronidazole clearance in neonates, suggesting that variability in CYP2A6 and CYP3A7 expression may lead to interindividual differences in metronidazole pharmacokinetics among young children. The large-scale developmental dataset of hepatic CYP enzymes established in this study provides a generalizable methodological framework for pediatric PBPK modeling of similar drugs. Yao et al.75 systematically collected demographic and physiological parameters from children aged 0-18 years in the Chinese population. They established a virtual Chinese pediatric population model using the Simcyp platform. The model was validated using probe drugs with different elimination pathways, such as theophylline (CYP1A2 metabolism), fentanyl (CYP3A4 metabolism), and ceftazidime and vancomycin (renal clearance). The predicted peak concentration, area under the curve (AUC), and clearance values were all within 0.5–2.0-fold of the observed values. These results confirm that the model adequately characterizes the pharmacokinetics of drugs metabolized by CYP1A2 and CYP3A4 or eliminated renally in Chinese children. Southall et al.76 further integrated physiological parameters specific to Chinese children to develop a PBPK model. They selected drugs with different metabolic pathways (fentanyl, midazolam, ruxolitinib, efavirenz, and theophylline) as well as renally cleared drugs for validation. The results showed that 79% of the predicted pharmacokinetic parameters fell within 0.8–1.25-fold of the observed values, and 100% fell within 0.67–1.5-fold, demonstrating good predictive accuracy of the model. Johnson et al.77 developed a Japanese pediatric PBPK model by integrating Japanese children's physiological parameters into a Northern European pediatric model. The model was validated using 14 drugs with various clearance pathways. Results showed that 71% of predicted pharmacokinetic parameters fell within 0.8–1.25-fold of observed values, and 90% fell within 0.67–1.5-fold.

These studies collectively demonstrate the critical role of pediatric PBPK modeling in integrating race-specific physiological parameters, quantitatively characterizing drug exposure profiles across different metabolic pathways in diverse pediatric populations, and supporting pediatric dose extrapolation and regimen optimization. However, the application of PBPK models in young children, particularly neonates and preterm infants, still faces numerous challenges. Insufficient data on the maturation of certain metabolic enzymes lead to difficulties in model predictions. Additionally, system-specific parameters in this population can change rapidly over short periods, making PBPK model predictions of measured drug concentrations less accurate compared to those in older children and adults.78 Nevertheless, with growing understanding of developmental physiological parameters in preterm infants and neonates, related research is providing a more solid theoretical basis for drug metabolism mechanisms in this age group. This is driving further optimization and application of PBPK models in pediatric drug development.

Michelet et al.79 developed a PBPK model spanning from preterm infants to adults using propofol as a model drug for CYP3A4. This study integrated CYP3A4 ontogeny data and successfully predicted drug clearance and concentration-time profiles across different age groups. The findings validated the reliability of PBPK models based on metabolic enzyme developmental characteristics for cross-age extrapolation. Kovar et al.80 established a parent-metabolite PBPK model for buprenorphine covering adults, children (4.6–7.5 years), and preterm infants(gestational age 27–34 weeks). The model integrated developmental data for metabolic enzymes including CYP3A4 and UGT2B7. It demonstrated excellent predictive performance and provided important references for the safe use of opioids in preterm infants. Van Hoogdalem et al.81 further employed PBPK modeling to investigate the impact of maturity on buprenorphine pharmacokinetics in neonates with neonatal opioid withdrawal syndrome. The study identified biliary clearance, buccal mucosal absorption, and CYP3A4 expression levels as key factors influencing pharmacokinetic variability in this population.

3.3 Application of pediatric PBPK models

3.3.1 Pediatric dose prediction and precision dosing strategies

By integrating ontogeny functions and genotype information of metabolic enzymes, PBPK models enable precise dose prediction and adjustment. For drugs with known metabolic enzyme polymorphisms, PBPK models can recommend individualized starting doses or adjustment regimens by incorporating pediatric-specific factors such as age, body weight, and genotype. These models can also integrate patient-specific factors, including organ function (e.g., hepatic or renal impairment), body composition (e.g., BMI or body fat percentage), and disease status. Furthermore, given the frequent need for combination medication in pediatric patients, PBPK models can predict changes in drug exposure with different genotypes when enzyme inhibitors or inducers are co-administered, thereby identifying potential high-risk subpopulations. This provides precision dosing adjustments for specific pediatric populations not included in clinical trials during drug development.

Soliman et al.82 successfully developed and validated PBPK models for omeprazole and its major metabolites in both adults and children. Simulations of a 0.7 mg/kg dosing regimen demonstrated that the duration with intragastric pH > 4 within a 24-h period was significantly prolonged in infants under 1 year of age compared to older children and adults, suggesting an increased risk of infection and impaired calcium absorption in this age group. Drug-drug interaction (DDI) simulations further revealed that complete inhibition of CYP3A4 (and CYP3A7 during infancy) resulted in an approximately 4-fold increase in omeprazole exposure in children under 2 years, whereas no significant change was observed in adults. These findings indicate that infants and young children are particularly susceptible to CYP3A inhibitor-mediated DDI. In a pediatric PBPK modeling study of posaconazole,83 the investigators incorporated age-specific anatomical and physiological parameters along with ontogeny data for the UGT1A4 to extrapolate to the pediatric population. Simulations revealed that the conventional oral suspension was unlikely to achieve effective exposure levels in children under 7 years of age due to solubility limitations. For the intravenous formulation, the optimal therapeutic dose was determined to be 11–13 mg/kg once daily, while the prophylactic dose was 8–9 mg/kg once daily. For the delayed-release oral suspension, the optimal therapeutic dose ranged from 12 to 14 mg/kg once daily, with a prophylactic dose of 8.5–10 mg/kg once daily. Eltanami et al.84 developed and validated PBPK models for the commonly used combination regimen of clobazam(CLB) and stiripentol (STP) in children with Dravet syndrome. The models were used to evaluate the impact of age and CYP2C19 genotype on DDI. CLB is metabolized primarily by CYP3A4 to form its active metabolite, N-desmethylclobazam (N-CLB), which is subsequently cleared via CYP2C19. STP inhibits multiple CYP enzymes. The investigators first constructed a model based on adult data to optimize the inhibitory parameters of STP on CYP2C19, and then extrapolated to children aged 2 years and older successfully by incorporating CYP enzyme ontogeny profiles. DDI risk assessment based on this model revealed that STP increased N-CLB exposure by 7- to 8-fold in CYP2C19 EMs, whereas no increase or even a decrease was observed in PMs. Further extrapolation to infants aged 6 months to 2 years predicted that exposures of CLB, N-CLB, and STP were comparable to those in older children and remained within therapeutic windows, supporting the safe use of STP in combination with CLB in infants older than 6 months.

3.3.2 Optimal design of pediatric clinical trials

PBPK modeling plays an important role in pediatric clinical trial design. By generating virtual subpopulations and simulating various scenarios, it enables the design of more efficient and targeted studies, thereby reducing trial costs and improving drug development success rates. Extrapolating data from adults or older children to infants and young children represents a typical application scenario for PBPK models. By integrating key physiological factors such as enzyme activity differences and organ function status, PBPK models can simulate drug metabolism processes across different pediatric age groups. This provides important references for first-in-pediatric dosing regimen design and reduces the need for large-scale trials in this vulnerable population.

Jo et al.85 developed a PBPK model for dapagliflozin to support pediatric dosing regimen design. The model was initially constructed and validated using in vitro data and adult clinical parameters, then refined using data from adolescent patients with type 2 diabetes mellitus aged 11–17 years. Given that UGT1A9 ontogeny maturity is critical for predicting drug exposure in young children, the study compared the impact of two different UGT1A9 ontogeny models on dose recommendations. Based on the optimized model, the study predicted exposure levels following monotherapy in children and adolescents aged 1 month to 18 years. This was used to identify pediatric regimens matching the exposure of a single 10 mg adult dose, thereby guiding first-dose selection in young children lacking clinical data. Traver et al.86 developed a PBPK model for liglitazone to guide pediatric starting dose selection. The investigators first constructed and validated an adult PBPK model using in vitro data along with single ascending dose and multiple ascending dose data from Phase I studies. This model characterized the drug's elimination pathways, including CYP2C8 metabolism (45%), CYP3A4 meta-bolism (24%), and biliary clearance (19.5%). When extrapolating to the pediatric population, CYP2C8 and CYP3A4 ontogeny profiles were integrated into the model. Pediatric starting doses were predicted based on the target steady-state exposure of the 150 mg once-daily oral dose in adults. Preliminary clinical data from five pediatric patients aged 5-12 years with cerebral adrenoleukodystrophy showed that observed drug concentrations fell within the 95% confidence intervals of model predictions. Bayesian predicted clearance values were consistent with PBPK model predictions, confirming the predictive performance of the model.

3.3.3 Regulatory submissions and considerations for review

Regulatory submissions utilizing PBPK modeling to predict pediatric pharmacokinetics are playing an increasingly important role in elucidating how metabolism influences drug disposition, safety, and efficacy. Regulatory agencies including the US FDA and European EMA have progressively accepted PBPK modeling as a supportive tool in pediatric drug development, particularly in the absence of clinical data. However, it is worth noting that current new drug applications submitted to the FDA remain predominantly focused on DDIs. Submissions involving models for susceptible subpopulations at different life stages (e.g., children), organ dysfunction, and genetic polymorphisms still account for a relatively small proportion. Among new drugs approved by the FDA between 2020 and 2024, only 2.6% of submissions employed PBPK modeling for pediatric dose prediction, highlighting ongoing challenges related to data scarcity and insufficient model validation in this field. During regulatory review, the FDA places particular emphasis on whether models adequately account for physiological changes associated with pediatric growth and development, such as enzyme expression and organ function maturation. The agency also stresses that models must possess a reliable in vitro-to-clinical extrapolation chain and validation data specific to targeted age groups.87

For example, in the regulatory submission for ALYFTREK, the applicant developed PBPK models to predict drug exposure in children aged 6 to 11 years. These models aimed to evaluate CYP3A-mediated DDIs, establish pediatric dosing regimens, and predict potential risks to CYP2C9 substrates. Regulatory review concluded that while adult DDI predictions and dosing regimens were successfully validated, pediatric extrapolation proved inadequate, and the CYP2C9-mediated interaction risk remained incompletely characterized.88 In another case, VORANIGO, the applicant developed PBPK models to predict DDIs for VORANIGO in two contexts, as a victim drug during co-administration with the CYP1A2 inhibitor fluvoxamine or inducers rifampin/phenytoin, and as a perpetrator drug affecting substrates of CYP3A4, CYP2C8, CYP2B6 enzymes and BCRP/P-gp transporters. Regulatory assessment concluded that the model was deemed adequate to support labeling for VORANIGO as a victim drug, accurately predicting a 10-fold increase in exposure when co-administered with the strong CYP1A2 inhibitor fluvoxamine and an approximately 40% decrease with moderate inducers (rifampin, phenytoin). Conversely, the model was considered inadequate for predicting VORANIGO as a perpetrator of CYP3A4 and CYP2C8 induction, leading to a conservative label statement that DDI potential cannot be excluded, thereby alerting clinicians to potential interaction risks.89 These cases demonstrate that when reviewing pediatric PBPK models, the FDA carefully evaluates mechanistic completeness, reliability of parameter sources, and the ability to characterize age-related physiological changes. The agency emphasizes that limitations identified in pediatric model applications must be clearly stated in product labeling.

4 Challenges and future perspectives

Despite the promising applications of PBPK models incorporating drug-metabolizing enzymes in pediatric populations, numerous challenges currently limit their predictive performance and clinical translation. These challenges are primarily manifested in the following aspects.

First, the acquisition of pediatric physiological parameters is difficult, resulting in incomplete foundational data. Knowledge gaps persist regarding the ontogeny of drug-metabolizing enzymes and transporters in children, particularly in neonates and infants. Limited data exist on protein expression levels and activity of certain metabolic enzymes in preterm infants, full-term neonates, and infants. System-specific parameters such as organ volumes and tissue perfusion rates across different age groups remain incomplete. Additionally, substantial interethnic variability in the distribution of metabolic enzyme genetic polymorphisms necessitates more detailed data to develop ethnically representative models.

Second, the scarcity of pediatric clinical trial data impedes model validation. Due to ethical and operational constraints, blood sample collection in children is limited, resulting in insufficient pharmacokinetic data. Sparse sampling designs often fail to comprehensively evaluate model predictive accuracy across different age subgroups and dosing regimens. This limitation undermines confidence in predicted results among regulatory agencies and clinicians, restricting model application in critical decision-making.

Third, the complex influence of non-genetic factors increases modeling difficulty. Beyond genetic factors, non-genetic variables including nutritional status, dietary components, and concomitant medications collectively influence pediatric drug metabolism, further complicating PBPK model development.

With advances in methodological approaches and deepening multi-disciplinary collaboration, pediatric PBPK models are expected to overcome current limitations and play a central role in pediatric drug development and clinical pharmacotherapy. The integration of artificial intelligence and machine learning techniques will create new opportunities for pediatric PBPK modeling. These approaches can optimize model parameter identification, improving the efficiency and accuracy of parameter estimation. Big data analytics enables metabolic phenotype classification to identify metabolic capacity in children with different genotypes, while deep learning methods can capture complex nonlinear relationships to enhance model prediction accuracy for interindividual variability. Together, these techniques facilitate the construction of hybrid modeling frameworks with greater predictive power.

The rapid development of omics technologies provides rich information for understanding inter-individual differences in drug meta-bolism. Integrating absolute quantitative data of metabolic enzymes from proteomic assays into PBPK models enhances the ability to predict age-related changes in drug clearance. Future systematic integration of multi-omics data, including genomics, transcriptomics, proteomics, and metabolomics, holds promise for constructing refined models of enzyme and transporter expression. This would reveal the complete regulatory cascade from genotype to phenotype, providing a mechanistic foundation for personalized dosing.

Clinical decision support systems based on PBPK modeling are gradually transitioning from research tools to clinical applications. Future development of clinical decision systems integrated with electronic health records could enable dynamic dose adjustment by incorporating patient genotype, real-time physiological monitoring data, and medication history. Such systems would directly translate model predictive capabilities into clinical practice, advancing the implementation of precision medicine in pediatric populations.

Promoting the standardization and shared platform development of pediatric PBPK models represents an important direction for sustainable advancement in this field. Establishing public databases of pediatric physiological parameters, model code repositories, and validation benchmarks can enhance model transparency and reusability. Concurrently, collaborative efforts among academia, industry, and regulatory agencies may lead to unified guidelines for model development and validation, providing technical specifications for regulatory acceptance of pediatric PBPK models. International collaborative research programs also contribute to integrating data resources from pediatric populations across different ethnicities, facilitating the construction of more representative population models.

5 Conclusions

PBPK models incorporating drug-metabolizing enzymes provide powerful mechanistic tools for understanding drug metabolism characteristics in pediatric populations. These models effectively integrate ontogeny data of metabolic enzymes, genetic polymorphism information, and physiological parameter variations to support decision-making across the entire continuum from early drug development to clinical pharmacotherapy. Compared with conventional empirical dose conversion approaches, PBPK modeling enables more accurate characterization of drug disposition patterns across pediatric age groups from neonates to adolescents, providing scientific rationale for first-in-pediatric dose selection, clinical trial design, and precision dosing guidance.

Despite current challenges, including critical data gaps, difficulties in model validation, and complex drug interactions, pediatric PBPK models are poised to play increasingly important roles in precision medicine practice. This progress is driven by the accumulating developmental data on metabolic enzymes and transporters, deep integration of artificial intelligence technologies, and systematic incorporation of multi-omics data. Moving forward, continued advancement in standardized platform development and multidisciplinary collaboration will progressively overcome existing limitations. Ultimately, PBPK modeling will make substantial contributions to improving the safety, efficacy, and accessibility of medications for pediatric populations.

References

[1]

Matalová P, Urbánek K, Anzenbacher P, et al. Specific features of pharmacokinetics in children. Drug Metab Rev. 2016;48(1):70–79.

[2]

Lacroix D, Sonnier M, Moncion A, et al. Expression of CYP3A in the human liver–evidence that the shift between CYP3A7 and CYP3A4 occurs immediately after birth. Eur J Biochem. 1997 Jul 15;247(2):625–634.

[3]

Koukouritaki SB, Manro JR, Marsh SA, et al. Developmental expression of human hepatic CYP2C9 and CYP2C19. J Pharmacol Exp Therapeut. 2004 Mar;308(3):965–974.

[4]

Ratanasavanh D, Beaune P, Morel F, et al. Intralobular distribution and quantitation of cytochrome P-450 enzymes in human liver as a function of age. Hepatology. 1991 Jun;13(6):1142–1151.

[5]

Maruo Y, Iwai M, Mori A, et al. Polymorphism of UDP-glucuronosyltransferase and drug metabolism. Curr Drug Metabol. 2005 Apr;6(2):91–99.

[6]

Zhou SF, Liu JP, Chowbay B. Polymorphism of human cytochrome P450 enzymes and its clinical impact. Drug Metab Rev. 2009;41(2):89–295.

[7]

Jones H, Rowland-Yeo K. Basic concepts in physiologically based pharmacokinetic modeling in drug discovery and development. CPT Pharmacometrics Syst Pharmacol.2013 Aug 14;2(8):e63.

[8]

Jorga K, Chavanne C, Frey N, et al. Bottom-up meets top-down: complementary physiologically based pharmacokinetic and population pharmacokinetic modeling for regulatory approval of a dosing algorithm of valganciclovir in very young children. Clin Pharmacol Ther. 2016 Dec;100(6):761–769.

[9]

Green FG, Park K, Burckart GJ. Methods used for pediatric dose selection in drug development programs submitted to the US FDA 2012-2020. J Clin Pharmacol. 2021 Jun;61(Suppl 1):S28–S35.

[10]

Leong R, Vieira MLT, Zhao P, et al. Regulatory experience with physiologically based pharmacokinetic modeling for pediatric drug trials. Clin Pharmacol Ther. 2012 May;91(5):926–931.

[11]

Shimada T, Yamazaki H, Mimura M, et al. Interindividual variations in human liver cytochrome P450 enzymes involved in the oxidation of drugs, carcinogens, and toxic chemicals. J Pharmacol Exp Therapeut. 1994 Jul;270(1):414–423.

[12]

Xiong S, Li L. The effect of CYP1A2 gene polymorphism on the metabolism of theophylline. Exp Ther Med. 2018 Jan;15(1):109–114.

[13]

Barreto G, Grecco B, Merola P, et al. Novel insights on caffeine supplementation, CYP1A2 genotype, physiological responses and exercise performance. Eur J Appl Physiol. 2021 Mar;121(3):749–769.

[14]

Tateishi T, Asoh M, Yamaguchi A, et al. Developmental changes in urinary elimination of theophylline and its metabolites in pediatric patients. Pediatr Res.1999;45:66–70.

[15]

Chida M, Yokoi T, Fukui T, et al. Detection of three genetic polymorphisms in the 5’- flanking region and intron 1 of human CYP1A2 in the Japanese population. Jpn J Cancer Res. 1999;90:899–902.

[16]

Chen X, Wang L, Zhi L, et al. The G-113A polymorphism in CYP1A2 affects the caffeine metabolic ratio in a Chinese population. Clin Pharmacol Ther. 2005 Sep;78(3):249–259.

[17]

Sohn JA, Kim HS, Oh J, et al. Prediction of serum theophylline concentrations and cytochrome P450 1A2 activity by analyzing urinary metabolites in preterm infants. Br J Clin Pharmacol. 2017 Jun;83(6):1279–1286.

[18]

Kim SE, Kim BH, Lee S, et al. Population pharmacokinetics of theophylline in premature Korean infants. Ther Drug Monit. 2013 Jun;35(3):338–344.

[19]

Djordjevic N, Milovanovic DD, Radovanovic M, et al. CYP1A2 genotype affects carbamazepine pharmacokinetics in children with epilepsy. Eur J Clin Pharmacol.2016 Apr;72(4):439–445.

[20]

Berel C, Mossé U, Wils J, et al. Interest of fluvoxamine as an Add-On to clozapine in children with severe psychiatric disorder according to CYP polymorphisms: experience from a case series. Front Psychiatr. 2021 Jun 21;12:669446.

[21]

Merino D, Fernandez A, Gérard AO, et al. Adverse drug reactions of olanzapine, clozapine and loxapine in children and youth: a systematic pharmacogenetic review. Pharmaceuticals. 2022 Jun 14;15(6):749.

[22]

FDA. Table of pharmacogenomic biomarkers in drug labeling [EB/OL] [2026-02-23].2024-06.

[23]

Miners JO, Birket DJ. Cytochrome P4502C9: an enzyme of major importance in human drug metabolism. Br J Clin Pharmacol. 1998;45:525–538.

[24]

Rettie AE, Jones JP. Clinical and toxicological relevance of CYP2C9: drug-drug interactions and pharmacogenetics. Annu Rev Pharmacol Toxicol. 2005;45:477–494.

[25]

Blake MJ, Castro L, Leeder JS, et al. Ontogeny of drug metabolizing enzymes in the neonate. Semin Fetal Neonatal Med. 2005 Apr;10(2):123–138.

[26]

Jetter A, Kinzig-Schippers M, Skott A, et al. Cytochrome P450 2C9 phenotyping using low-dose tolbutamide. Eur J Clin Pharmacol. 2004 May;60(3):165–171.

[27]

Zhang J, Tian L, Huang J, et al. Cytochrome P450 2C9 gene polymorphism and warfarin maintenance dosage in pediatric patients: a systematic review and meta-analysis. Cardiovasc Ther. 2017 Feb;35(1):26–32.

[28]

Fang H, Wang X, Hou K, et al. The association of adjusted plasma valproic acid concentration with CYP2C9 gene polymorphism in patients with epilepsy: a systematic review and meta-analysis. Ann Transl Med. 2021 May;9(10):846.

[29]

Chaudhary N, Kabra M, Gulati S, et al. Frequencies of CYP2C9 polymorphisms in North Indian population and their association with drug levels in children on phenytoin monotherapy. BMC Pediatr. 2016 May 14;16:66.

[30]

Gardiner SJ, Begg EJ. Pharmacogenetics, drug-metabolizing enzymes, and clinical practice. Pharmacol Rev. 2006 Sep;58(3):521–590.

[31]

Koukouritaki SB, Manro JR, Marsh SA, et al. Developmental expression of human hepatic CYP2C9 and CYP2C19. J Pharmacol Exp Therapeut. 2004 Mar;308(3):965–974.

[32]

Knebel W, Tammara B, Udata C, et al. Population pharmacokinetic modeling of pantoprazole in pediatric patients from birth to 16 years. J Clin Pharmacol. 2011 Mar;51(3):333–345.

[33]

Yanni SB, Annaert PP, Augustijns P, et al. In vitro hepatic metabolism explains higher clearance of voriconazole in children versus adults: role of CYP2C19 and flavin-containing monooxygenase3. Drug Metab Dispos. 2010 Jan;38(1):25–31.

[34]

Sim SC, Risinger C, Dahl ML, et al. A common novel CYP2C19 gene variant causes ultrarapid drug metabolism relevant for the drug response to proton pump inhibitors and antidepressants. Clin Pharmacol Ther. 2006 Jan;79(1):103–113.

[35]

Kearns GL, Leeder JS, Gaedigk A. Impact of the CYP2C19*17 allele on the pharmacokinetics of omeprazole and pantoprazole in children: evidence for a differential effect. Drug Metab Dispos. 2010 Jun;38(6):894–897.

[36]

Thuy LLT, Nguyen LT, Vu HA, et al. Effect of MDR1 C3435T and CYP2C19 genetic polymorphisms on the outcome of Helicobacter pylori eradication treatment in children with gastritis and peptic ulcer, Vietnam. BMC Pediatr. 2024 Jul 19;24(1):464.

[37]

Zhang W, Ji Q, Wang Q, et al. Association of CYP2C19 single nucleotide polymorphism with hemorrhagic cystitis after pediatric allogeneic hematopoietic stem cell transplantation. Clin Transplant. 2026 Feb;40(2):e70474.

[38]

Chen X, Xiao Y, Li H, et al. Therapeutic drug monitoring and CYP2C19 genotyping guide the application of voriconazole in children. Transl Pediatr. 2022 Aug;11(8):1311–1322.

[39]

Tilen R, Paioni P, Goetschi AN, et al. Pharmacogenetic analysis of voriconazole treatment in children. Pharmaceutics. 2022 Jun 17;14(6):1289.

[40]

Yamamoto Y, Shiratani Y, Nishida T, et al. Lacosamide pharmacokinetics and retention in Japanese patients with epilepsy: a retrospective study on the influence of age, comedications, and cytochrome P450 2C19 polymorphism. Ther Drug Monit.2025 Jun 1;47(3):407–412.

[41]

Ingelman-Sundberg M, Sim SC, Gomez A, et al. Influence of cytochrome P450 polymorphisms on drug therapies: pharmacogenetic, pharmacoepigenetic, and clinical aspects. Pharmacol Ther. 2007 Dec;116(3):496–526.

[42]

van den Anker JN. Developmental pharmacology. Dev Disabil Res Rev. 2010;16(3):233–238.

[43]

Allegaert K, van Schaik RH, Vermeersch S, et al. Postmenstrual age and CYP2D6 polymorphisms determine tramadol O-demethylation in critically ill neonates and infants. Pediatr Res. 2008 Jun;63(6):674–679.

[44]

Pratt VM, Zehnbauer B, Wilson JA, et al. Characterization of 107 genomic DNA reference materials for CYP2D6, CYP2C19, CYP2C9, VKORC1, and UGT1A1: a GeT-RM and association for molecular pathology collaborative project. J Mol Diagn. 2010 Nov;12(6):835–846.

[45]

Crews KR, Gaedigk A, Dunnenberger HM, et al. Clinical pharmacogenetics implementation consortium guidelines for cytochrome P450 2D6 genotype and codeine therapy:2014 update. Clin Pharmacol Ther. 2014 Apr;95(4):376–382.

[46]

Gaedigk A, Simon SD, Pearce RE, et al. The CYP2D6 activity score: translating genotype information into a qualitative measure of phenotype. Clin Pharmacol Ther.2008 Feb;83(2):234–242.

[47]

Desmeules J. The AmpliChip CYP450 test: cytochrome P450 2D6 genotype assessment and phenotype prediction. Pharmacogenomics J. 2009 Feb;9(1):34–41.

[48]

Cojocaru A, Braha A, Andreescu NI, et al. Correlation of neuroimaging biomarkers and pharmacogenetic profiles in optimizing personalized therapy in children and adolescents with psychotic disorders. Neurol Int. 2025 Aug 14;17(8):128.

[49]

Xin Y, Gao L, Li S, et al. Effect of CYP2D6 and ABCB1 polymorphisms on pharmacokinetics and efficacy of aripiprazole in pediatric tic disorders. BMC Pediatr.2025 Jul 2;25(1):495.

[50]

Cojocaru A, Braha A, Jeleriu R, et al. The implications of cytochrome P450 2D6/CYP2D6 polymorphism in the therapeutic response of atypical antipsychotics in adolescents with Psychosis-A prospective study. Biomedicines. 2024 Feb 22;12(3):494.

[51]

Healy P, Allegaert K, Della Pasqua O. Evaluation of the effect of CYP2D6 and OCT1 polymorphisms on the pharmacokinetics of tramadol: implications for clinical safety and dose rationale in paediatric chronic pain. Br J Clin Pharmacol. 2025 Feb;91(2):283–296.

[52]

Zhou SF. Drugs behave as substrates, inhibitors, and inducers of human cytochrome P450 3A4. Curr Drug Metabol. 2008 May;9(4):310–322.

[53]

Lacroix D, Sonnier M, Moncion A, et al. Expression of CYP3A in the human liverevidence that the shift between CYP3A7 and CYP3A4 occurs immediately after birth. Eur J Biochem. 1997 Jul 15;247(2):625–634.

[54]

Hines RN. Ontogeny of human hepatic cytochromes P450. J Biochem Mol Toxicol.2007;21(4):169–175.

[55]

Kerr BM, Thummel KE, Wurden CJ, et al. Human liver carbamazepine metabolism. Role of CYP3A4 and CYP2C8 in 10,11-epoxide formation. Biochem Pharmacol. 1994 Jun 1;47(11):1969–1979.

[56]

Lamba JK, Lin YS, Thummel K, et al. Common allelic variants of cytochrome P4503A4 and their prevalence in different populations. Pharmacogenetics. 2002 Mar;12(2):121–132.

[57]

Stevens JC, Hines RN, Gu C, et al. Developmental expression of the major human hepatic CYP3A enzymes. J Pharmacol Exp Therapeut. 2003 Nov;307(2):573–582.

[58]

Yang Q, Wang Y, Wang X, et al. Drug-drug interaction between diltiazem and tacrolimus in relation to CYP3A5 genotype status in Chinese pediatric patients with nephrotic range proteinuria: a retrospective study. Front Pharmacol. 2024 Sep 3;15:1463595.

[59]

Wang H, Wang J, Lin B, et al. Effect of age, comedications, and CYP3A4/5 polymorphisms on perampanel exposure in Chinese pediatric patients with epilepsy. J Clin Pharmacol. 2024 Jun;64(6):737–743.

[60]

Gijsen VM, van Schaik RH, Elens L, et al. CYP3A4*22 and CYP3A combined genotypes both correlate with tacrolimus disposition in pediatric heart transplant recipients. Pharmacogenomics. 2013 Jul;14(9):1027–1036.

[61]

Ito M, Yamamoto K, Maruo Y, et al. Effect of a conserved mutation in uridine diphosphate glucuronosyltransferase 1A1 and 1A6 on glucuronidation of a metabolite of flutamide. Eur J Clin Pharmacol. 2002 Apr;58(1):11–14.

[62]

Kweekel D, Guchelaar H, Gelderblom H. Clinical and pharmacogenetic factors associated with irinotecan toxicity. Cancer Treat Rev. 2008;34(7):656–669.

[63]

Zhu X, Zhu J, Sun F, et al. Influence of UGT1A1 *6/*28 polymorphisms on irinotecan-related toxicity and survival in pediatric patients with relapsed/refractory solid tumors treated with the VIT regimen. Pharmgenomics Pers Med. 2021 Mar 23;14:369–377.

[64]

Wei S, Li X, Zhang Q, et al. Population pharmacokinetics of topiramate in Chinese children with epilepsy. Eur J Clin Pharmacol. 2023 Oct;79(10):1401–1415.

[65]

Wei S, Li X, Wu H, et al. UGT1A polymorphism rs4148324 associated with topiramate plasma concentration to dose ratio in children with epilepsy. Seizure.2024 Mar;116:107–112.

[66]

King CD, Rios GR, Assouline JA, et al. Expression of UDP-glucuronosyltransferases(UGTs) 2B7 and 1A6 in the human brain and identification of 5-hydroxytryptamine as a substrate. Arch Biochem Biophys. 1999 May 1;365(1):156–162.

[67]

Guo Y, Hu C, He X, et al. Effects of UGT1A6, UGT2B7, and CYP2C9 genotypes on plasma concentrations of valproic acid in Chinese children with epilepsy. Drug Metabol Pharmacokinet. 2012;27(5):536–542.

[68]

Nandith PB, Adiga U, Shenoy V, et al. UGT1A6 and UGT2B7 gene polymorphism and its effect in pediatric epileptic patients on sodium valproate monotherapy. Indian J Pediatr. 2021 Aug;88(8):764–770.

[69]

Banawalikar N, Adiga S, Adiga U, et al. Association of UGT1A6 gene polymorphism with clinical outcome in pediatric epileptic patients on sodium valproate monotherapy. Braz J Med Biol Res. 2021 Jun 14;54(9):e11097.

[70]

Yellepeddi V, Rower J, Liu X, et al. State-of-the-Art review on physiologically based pharmacokinetic modeling in pediatric drug development. Clin Pharmacokinet. 2019 Jan;58(1):1–13.

[71]

Jaroch K, Jaroch A, Bojko B. Cell cultures in drug discovery and development: the need of reliable in vitro-in vivo extrapolation for pharmacodynamics and pharmacokinetics assessment. J Pharm Biomed Anal. 2018 Jan 5;147:297–312.

[72]

Verscheijden LFM, Koenderink JB, Johnson TN, et al. Physiologically-based pharmacokinetic models for children: starting to reach maturation? Pharmacol Ther.2020 Jul;211:107541.

[73]

Templeton IE, Jones NS, Musib L. Pediatric dose selection and utility of PBPK in determining dose. AAPS J. 2018 Feb 13;20(2):31.

[74]

Parvez MM, Thakur A, Mehrotra A, et al. Age-dependent abundance of CYP450 enzymes involved in metronidazole metabolism: application to pediatric PBPK modeling. Clin Pharmacol Ther. 2024 Oct;116(4):1090–1099.

[75]

Yao X, Liu X, Tu S, et al. Development of a virtual Chinese pediatric population physiological model targeting specific metabolism and kidney elimination pathways. Front Pharmacol. 2021 May 11;12:648697.

[76]

Southall RL, Dinh J, Pan X, et al. Development, verification, and application of a Chinese pediatric physiologically based pharmacokinetic model: emphasis on CYP metabolism and renal elimination. AAPS J. 2025 Nov 10;28(1):17.

[77]

Johnson TN, Abduljalil K, Pan X, et al. Development and verification of a Japanese pediatric physiologically based pharmacokinetic model with emphasis on drugs eliminated by cytochrome P450 or renal excretion. J Clin Pharmacol. 2023 Oct;63(10):1156–1168.

[78]

Abduljalil K, Jamei M, Johnson TN. Fetal physiologically based pharmacokinetic models: systems information on the growth and composition of fetal organs. Clin Pharmacokinet. 2019 Feb;58(2):235–262.

[79]

Michelet R, Van Bocxlaer J, Allegaert K, et al. The use of PBPK modeling across the pediatric age range using propofol as a case. J Pharmacokinet Pharmacodyn J Pharmacokinet Pharmacodyn. 2018 Dec;45(6):765–785.

[80]

Kovar L, Schrapel C, Selzer D, et al. Physiologically-based pharmacokinetic (PBPK) modeling of buprenorphine in adults, children and preterm neonates. Pharmaceutics.2020 Jun 23;12(6):578.

[81]

van Hoogdalem MW, Johnson TN, McPhail BT, et al. Physiologically-based pharmacokinetic modeling to investigate the effect of maturation on buprenorphine pharmacokinetics in newborns with neonatal opioid withdrawal syndrome. Clin Pharmacol Ther. 2022 Feb;111(2):496–508.

[82]

Soliman A, Rodriguez-Vera L, Alarcia-Lacalle A, et al. Leveraging omeprazole PBPK/PD modeling to inform drug-drug interactions and specific recommendations for pediatric labeling. Pharmaceutics. 2025 Mar 14;17(3):373.

[83]

Malik P, Mian P. Physiologically based pharmacokinetic modeling to refine dosing of posaconazole in young children. Clin Ther. 2025 Apr;47(4):261–270.

[84]

Eltanameli B, Al Sahlawi S, Cristofoletti R. Physiologically based pharmacokinetic modeling of clobazam and stiripentol Co-Therapy in Dravet syndrome. J Personalized Med. 2025 Nov 11;15(11):549.

[85]

Jo H, Pilla Reddy V, Parkinson J, et al. Model-informed pediatric dose selection for dapagliflozin by incorporating developmental changes. CPT Pharmacometrics Syst Pharmacol. 2021 Feb;10(2):108–118.

[86]

Traver E, Rodríguez-Pascau L, Meya U, et al. Clinical pharmacokinetics of leriglitazone and a translational approach using PBPK modeling to guide the selection of the starting dose in children. CPT Pharmacometrics Syst Pharmacol. 2024 Jun;13(6):982–993.

[87]

Li Y, Sun H, Zhang Z. The evolution and future directions of PBPK modeling in FDA regulatory review. Pharmaceutics. 2025 Oct 31;17(11):1413.

[88]

ALYFTREK-INTEGRATED review. Available online: . Accessed February 5, 2026.

[89]

VORANIGO-MULTI-DISCIPLINE review. Available online: February 5 2026.

Rights & permissions

2026 The Authors. Publishing services by Elsevier B.V. on behalf of Higher Education Press and KeAi Communications Co. Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

PDF (1014KB)

0

Accesses

0

Citation

Detail

Sections
Recommended

/

〈 〉