Francesco Bellanti 1, Marta Neve 1, Italo Poggesi 1, Sara Cazzaniga 2, Paolo Bettica 2
1 Certara (Radnor, USA), 2 Italfarmaco S.p.A. (Milan, Italy)
Introduction: Givinostat is an oral histone deacetylase inhibitor indicated for the treatment of Duchenne muscular Dystrophy (DMD) in patients aged ≥6 years. Optimizing pediatric dosing while ensuring efficacy and safety required a model-informed approach due to ethical and logistical trial constraints common to rare diseases. A robust pharmacometric (PMx) strategy was implemented to overcome these challenges and support givinostat approval.
Objectives:
• To characterize the pharmacokinetic (PK) of givinostat and to quantify givinostat pharmacokinetic-pharmacodynamic (PK-PD) relationship with platelet levels using a model-based approach.
• To optimize dosing recommendations in DMD patients aged ≥6 years leveraging the developed PMx framework.
Methods: PMx analyses included data from six clinical studies (including one phase 2 and the pivotal phase 3 study EPIDYS in DMD patients). 3325 PK observations were available in 242 subjects and 3241 PD observations (platelet counts) were available in 169 subjects. A non-linear mixed effects approach in NONMEM was used for the population PK and PK-PD analysis. Model development was performed utilizing first-order conditional estimation method with interaction (FOCEI). The developed models were leveraged to simulate givinostat PK and PD in a virtual pediatric population > 2 years of age and > 10 kg (n=1000/weight group) based on age and weight distributions from the NHANES database [1]. Simulations accounted for inter-individual variability (IIV) and were performed in R using the package rxode2 (v 2.1.2). Simulated platelet counts were assessed biweekly; if predicted values were <150 × 10⁹/L, dose reductions were applied.
Results: A 2-compartment model with first order input with lag and first order elimination best described the PK of givinostat. IIV was incorporated on CL/F, V2/F and bioavailability, and residual variability was addressed using a combined error model. The model included the effect of body weight on CL/F, with an allometric coefficient of 0.326 (95% CI 0.221-0.431). A Friberg-like model [2] was used to describe the time course of platelet counts. An inhibitory linear drug model included on the proliferation rate (Kprol) parameter was used to describe the effect of givinostat on the platelet levels. Both models were suitable to simulate givinostat PK and PK-PD in pediatric DMD patients. The results of the simulations allowed simplifying the 9-weight band dosing regimen originally adopted in the EPIDYS study to a 4-weight band regimen. In the revised regimen, geometric mean [CV%] simulated givinostat exposures were comparable in the 10–20 kg and 20–40 kg groups (area under the curve [AUC]0-12: 380 [33.6] and 429 [33.4] hr•ng/mL) but slightly higher in the 40–60 kg and >60 kg groups (AUC0-12 504 [34.5] and 534 [33.6] hr•ng/mL) following the first administration. At steady-state, following final dose modifications, simulated exposures were comparable across all weight groups (AUC0-12: 363 [32.0], 375 [32.0], 399 [32.8], and 415 [31.1] hr•ng/mL for 10–20 kg, 20–40 kg, 40–60 kg, and >60 kg groups, respectively). Dose reduction was required in 41.5% of virtual patients; most occurred within the first 3 months.
Conclusions: The results from the pivotal Phase 3 study EPIDYS showed that the devised dosing instructions allowed to safely manage platelet decrease, with >95% of patients completing the study without clinical manifestations. The developed PMx framework allowed to successfully revise the regimen and to ensure consistent exposures across weight groups while substantially simplifying dosing and lowering the risk of thrombocytopenia-related reductions. Overall, these findings demonstrate that model informed analyses enabled rational simplification of the givinostat dosing regimen without compromising efficacy or safety.
References:
[1] https://wwwn.cdc.gov/nchs/nhanes/
[2] Friberg LE, Henningsson A, Maas H, Nguyen L and Karlsson MO. Model of chemotheraphy-induced myelosuppresion with parameters consistency across drugs. J. Clin. Oncol 20:4713:4721, 2002.
Reference: PAGE 34 (2026) Abstr 12041 [www.page-meeting.org/?abstract=12041]
Poster: Drug/Disease Modelling - Paediatrics