Emilie Schindler* (1), Anna Largajolli* (1), Petra Jauslin (1), Tomohisa Saito (2), Luca Loprete (3), Nathalie Wagner (3), Vincent Duval (1). *Equal contribution.
(1) Certara Strategic Consulting, (2) Chugai Pharmaceutical, (3) Nestlé Skin Health, Galderma Research and Development
Objectives: To characterize the time-course and exposure-response relationship of the investigator’s global assessment (IGA) score in subjects with moderate-to-severe atopic dermatitis (AD) treated with subcutaneous nemolizumab, a humanized antibody against interleukin-31 receptor A.
Methods: A total of 3,818 IGA scores collected from 486 AD patients were available from two randomized placebo-controlled clinical studies: a phase 2a safety and efficacy study (CIM003JG: N = 261, dose range: 0.1-2 mg/kg Q4W) [1] and a phase 2b dose-range finding study (RD.03.SPR.114322: N = 225, dose groups: 20 mg loading dose (LD) + 10 mg , 60 mg LD + 30 mg or 90 mg Q4W) [2]. The 6-point IGA scale from CIM003JG (ranging 0-5) was adapted into the 5-point IGA scale from RD.03.SPR.114322 (ranging 0-4) by lumping scores of 4 and 5 into one “severe” category. In addition, the proportion of observed scores of 0 was small (1.8%) and therefore lumped together with scores of 1 in the analysis. Longitudinal nemolizumab serum concentrations were described by a one-compartment distribution population pharmacokinetic (popPK) model with linear elimination, firstorder absorption with lag time and dose effect on bioavailability. Inter-individual variability (IIV) was included in clearance (CL), volume of distribution (V) and the absorption rate constant. CL and V were correlated at an individual level. Concentrations were log-transformed for analysis. Hence, the residual error was additive in the log domain. Covariate effects of serum albumin on CL and body weight on CL and V were identified. A previously developed continuoustime Markov model (CTMM) [3] was extended to describe longitudinal IGA scores. The CTMM assimilated the probability of each IGA score to a compartment amount and modelled the ascending (toward worse scores) and descending (toward better scores) transitions between compartments. Linear and non-linear drug effects driven by model-predicted daily concentrations and placebo effects were evaluated on the rate constants of ascending transitions (λasc) and descending transitions (λdesc). The effect of covariates (age, body weight, sex, and baseline IGA) on model parameters was assessed. A minimal CTMM approach was also evaluated [4]. All models were implemented in NONMEM 7.3 [5].
Results: Longitudinal IGA data were best described by a 4-compartment CTMM with a stimulating drug effect common to all λdesc. A simplified IIV structure using one term common to all λasc and one term common to all λdesc improved numerical stability without significantly worsening model fit. A statistically significant linear effect of time (i.e., placebo effect) on λasc was identified in study RD.03.SPR.114322, but not in study CIM003JG. Baseline IGA score of 4 or 5 was found to be a significant covariate for all λasc. All model parameters, except for the drug concentration producing half of maximal effect (64% RSE), were estimated with acceptable precision (≤31% RSE and ≤23% RSE for fixed and random effects, respectively). Visual predictive checks showed an adequate predictive performance of the model for all scores and dose levels up to 24 weeks of treatment. The alternative minimal CTMM could well describe the time-courses of most IGA scores but failed to describe the placebo data. It was therefore not retained.
Conclusions: The developed CTMM provides a good description of the IGA data in AD patients treated with a wide range of nemolizumab doses. The IGA model, combined with models for two other pharmacodynamic endpoints – the eczema area and severity index (EASI) and weekly average peak pruritus numeric rating scale (PP-NRS), both described in a separate abstract – will further be used to support the development program of nemolizumab.
References:
[1] https://clinicaltrials.gov/ct2/show/NCT01986933.
[2] https://clinicaltrials.gov/ct2/show/NCT03100344.
[3] Saito T. et al. WcoP (2016) Abstract 265. https://go-wcop.org/2016/application-of-continuous-time-marcov-exposure-response-model-to-siga-nemolizumab-case.
[4] Schindler E., Karlsson M.O. AAPS J (2017).
[5] Beal SL et al. NONMEM 7.3.0 users guides. (1989–2013). Hanover: ICON Development Solutions.
Reference: PAGE 28 (2019) Abstr 9178 [www.page-meeting.org/?abstract=9178]
Poster: Drug/Disease Modelling - Other Topics