Population pharmacokinetic-pharmacodynamic analysis of elafibranor and metabolite GFT1007 to support exposure-response characterization and dose selection in patients with primary biliary cholangitis
Jurij Aguiar Zdovc (1), Karl Brendel (2), Qing Xi Ooi (1), Maddlie Bardol (1), Marion Dehez (2)
(1) Pharmetheus AB, Uppsala, Sweden, (2) Ipsen, Les Ulis, France
Objectives: Elafibranor is a dual peroxisome proliferator-activated receptor α/δ agonist. It is an orally administered, liver-targeted drug candidate developed for the treatment of primary biliary cholangitis (PBC), which is a rare, chronic cholestatic liver disease. In June 2023, the Phase-3 readout for elafibranor in patients with PBC became available[1]. A pharmacokinetic-pharmacodynamic (PKPD) analysis was performed based on a previous PK analysis. The overall aim was to provide dose justification and inform on the efficacy of elafibranor and its equipotent main active metabolite GFT1007 in patients with PBC. The efficacy endpoints of primary interest were alkaline phosphatase (ALP) and total bilirubin (TBIL). The specific objectives were to describe the relationship between elafibranor and GFT1007 exposure and responses in ALP and TBIL and to evaluate the impact of relevant covariates.
Methods: The data for this analysis originated from a Phase II study (Study GFT505B-216-1) and a Phase III study (Study GFT505B-319-1), where patients with PBC received oral doses of elafibranor (80 mg/day or 120 mg/day) or placebo for up to 12 weeks and 52 weeks, respectively. The sequential approach using individual PK parameters from a previous PK analysis was used for the development of the PKPD models. The sum of the areas under the concentration-time curve during a dosing interval at steady state of elafibranor and GFT1007 (AUCτ,ss sum) as predicted by the final PK models for elafibranor and GFT1007 was used as a driver for drug effects. In the PKPD analysis of ALP and TBIL, separate models for ALP and TBIL were first developed, followed by joint modelling of these endpoints, also considering the correlations in the inter-individual variability (IIV) on selected parameters. The forest plots were generated based on the final ALP model and the final TBIL model to illustrate the covariate effects. The final joint ALP-TBIL model was used to perform simulations of ALP and TBIL. The relative change from baseline ALP, relative change from baseline TBIL, normalized ALP (ALP/upper limit of normal value (ULN)), and normalized TBIL (TBIL/ULN) were derived and compared to the threshold defining the primary endpoint of ALP level of < 1.67 times the ULN, a decrease of ALP from baseline of at least 15%, and TBIL ≤ ULN at week 52.
The modeling analyses were performed using NONMEM version 7.5.0 and simulations were performed using the mrgsolve package in R.
Results: Data from 206 adult patients with PBC were included and a total of 1892 and 1693 plasma concentrations of ALP and TBIL were evaluated, respectively. An indirect response PD model with AUCτ,ss sum as the PK driver inhibiting the zero-order production rate constant (kin) of ALP as described by a maximum effect (Emax) model, was used to describe the ALP dynamics and the delay in the placebo and drug effects. The same model structure was also used for TBIL but with a linear model instead of the Emax model to describe the drug effect. The final joint ALP-TBIL model consisted of the structure of the final ALP model, the structure of the final TBIL model, and the correlations in IIV between the baseline of ALP and TBIL. Based on the forest plot, patients with mild-to-moderate hepatic impairment had approximately 30% higher baseline ALP concentration compared to patients with no hepatic impairment. Patients with baseline TBIL of 5.00 umol/L and 16.4 umol/L were associated with 3% lower and 10% higher ALP at baseline compared to a reference PBC patient with baseline TBIL of 8.4 umol/L. Patients with baseline liver stiffness of 4.91 kPa and 15.5 kPa had 15% lower and 20% higher TBIL at baseline compared to a reference PBC patient with baseline liver stiffness of 8.10 kPa. Based on the PKPD simulations, the 80 mg/day regimen showed a clear effect and improvement in the ALP and TBIL response compared to placebo. Saturation in the ALP response was observed for AUCτ,ss sum of approximately > 30 umol·h/L, and the majority of the patients with PBC administered 80 mg/day (∼60%) had AUCτ,ss sum above 30 umol·h/L. Most of the simulated patients receiving the 80 mg/day regimen were predicted to achieve the primary endpoints' target: 53% achieved ALP < 1.67 times the ULN, 91% achieved a relative ALP decrease from baseline of at least 15%, and 97% had TBIL ≤ ULN at week 52.
Conclusions: The proposed dose of 80 mg/day was considered efficacious in the treatment of PBC in adult patients.
References:
[1] Kowdley K, Bowlus C, Levy C, et al (2024) Efficacy and Safety of Elafibranor in Primary Biliary Cholangitis. N Engl J Med 390:795-805. doi:10.1056/NEJMoa2306185
*Analysis and publication sponsor: Ipsen; included studies sponsor: GENFIT
