Physiologically-Based Pharmacokinetics and Empirical Pharmacodynamic Modeling for Pediatric Henagliflozin Dosing: Clinical Insights for Chinese Patients

Abstract

Objective: This study aimed to present a quantitative modeling and simulation approach for oral henagliflozin, a selective sodium-glucose cotransporter 2 (SGLT2) inhibitor primarily metabolized by uridine diphosphate-glucuronosyltransferase (UGT) enzymes. Methods: A physiologically-based pharmacokinetic (PBPK) model for henagliflozin was developed using in vitro metabolism and clinical pharmacokinetic (PK) data, with validation across multiple contexts, including healthy adults, and hepatic impairment populations. Additionally, empirical pharmacodynamic (PD) modeling was employed to optimize pediatric dosing based on exposure-response relationships for urinary glucose excretion (UGE). Predicting henagliflozin exposure in pediatric patients poses challenges due to UGT enzyme ontogeny and the scarcity of clinical PK data in younger age groups. Using twofold acceptance criteria, model-predicted and observed drug exposures and PK parameters (area under the curve and peak concentration) were compared in diverse scenarios, including monotherapy in healthy adults (single/multiple doses), hepatic impairment, and extrapolation to pediatric age groups. Results: The PBPK model accurately captured observed exposures within a twofold range in both adults and adolescents, supporting the model's predictive utility. The verified PBPK and empirical PD models informed dosing recommendations in pediatric populations aged 1 month to 18 years, achieving henagliflozin exposures comparable to those in adult patients receiving a 5-10 mg dose. Conclusion: This study shows that PBPK and PD modeling effectively guide pediatric dosing of henagliflozin, reducing trial reliance and supporting real-world validation.

Keywords: Chinese population; henagliflozin; pediatric dosing; pharmacodynamic (PD) modeling; physiologically-based pharmacokinetic (PBPK) modeling.

Figure 1

Structures for selected SGLT2 inhibitors.

Figure 2

The workflow of PBPK model in adult and pediatric populations for henagliflozin.

Figure 3

Predicted vs. observed plasma concentrations and goodness of hit for henagliflozin at different doses. (A) Predicted vs. observed plasma concentrations for different single doses of henagliflozin ranging from 2.5 to 200 mg, with log–log scaling. (B) Predicted vs. observed plasma concentrations for different multiple doses of henagliflozin ranging from 1.25 to 100 mg, with log–log scaling. (C) Goodness of fit plot for henagliflozin single-dose predictions. (D) Goodness of fit plot for henagliflozin multiple-dose predictions. The solid line represents the 1.25-fold error, and the dashed lines represent the twofold error boundaries, for evaluating prediction performance.

Figure 4

Predicted vs. observed plasma concentrations and goodness of fit for henagliflozin under fasted state at various doses. (A) Predicted vs. observed plasma concentrations for a 5 mg dose, fasted state. (C) Predicted vs. observed concentrations for a 10 mg dose, fasted state, with repeated dosing. (E) Predicted vs. observed concentrations for a 25 mg dose, fasted state, with repeated dosing. Shaded areas represent 5%–95% prediction intervals, solid lines show model predictions, and symbols indicate observed data points from different datasets. (B), (D), and (F) present GOF plots for 5, 10, and 25 mg doses, respectively, in the fasted state. Solid lines represent the 1.25-fold error, while dashed lines represent the twofold error boundaries, providing an assessment of model prediction accuracy.

Figure 5

Predicted vs. observed plasma concentrations and goodness of fit for henagliflozin under fed and fasting dtates in T2DM Patients. (A, B) show predicted vs. observed plasma concentrations for 5 and 10 mg doses in the fed state, respectively, while (C–E) display 5, 10, and 20 mg doses under fasting conditions in T2DM patients. The shaded areas represent 5%–95% prediction intervals, with solid lines indicating model predictions and symbols representing observed data points. (F, G) illustrate the GOF plots for the fed and fasting T2DM states, respectively, with solid lines representing the 1.25-fold error and dashed lines indicating the twofold error boundaries for performance assessment.

Figure 6

Boxplots of AUC, C_max, and UGE across different age groups with adult reference lines. (A) AUC values across age groups from neonates to adults, with adult 5 and 10 mg reference ranges indicated by solid and dashed lines, respectively. (B) C_max values across age groups, with the solid line representing the adult 5 mg reference range and the dashed line indicating the adult 10 mg range. (C) UGE calculated from AUC values, with reference lines as in (A). (D) UGE calculated from C_max values, with reference lines as in (B).