Jessica Wojciechowski 1, Kevin Tianxiang Han 1, Howard Kallender 1, Vivek S. Purohit 1
1 Incyte Corporation (Wilmington, USA)
Objectives: Ruxolitinib (selective Janus kinase [JAK]1/JAK2 inhibitor) is available as a topical cream for the treatment of inflammatory skin diseases such as atopic dermatitis (AD) and vitiligo. The cream formulation was designed to minimize the plasma concentrations of ruxolitinib and thus avoid potential systemic pharmacologic effects. Model-based approaches enable an integrated evaluation of dose-concentration relationships and account for differences in the applied dose (dependent on affected body surface area [BSA; cm²] and active ingredient application rate [mg/cm²]) while quantifying covariate effects. A mixed-effects modeling analysis describing the ruxolitinib dose–concentration relationship was conducted to (1) quantify the relationship between the applied dose of ruxolitinib cream and the resulting plasma concentrations of ruxolitinib in adult patients with AD and vitiligo, (2) predict the probability of ruxolitinib plasma concentrations exceeding the threshold of 281 nM (whole blood ruxolitinib half-maximal inhibitory concentration [IC50] for JAK2 inhibition) considered clinically relevant for hematologic side effects, and (3) explore the association between plasma ruxolitinib concentrations and key hematologic laboratory parameters (hemoglobin, neutrophil count, and platelet count).
Methods: Mixed-effect modeling of the relationship between the average ruxolitinib dose per application of cream versus repeat steady-state plasma concentrations (CSS) was performed using pooled data from adult patients who applied 0.75% or 1.5% ruxolitinib cream twice daily (BID) continuously to the same areas identified at baseline. Data were included from 4 studies in AD (phase 2 and phase 3), 2 phase 3 studies in vitiligo, and 1 maximum-use trial (MUsT) in AD. Linear and saturable relationships were evaluated as structural models where slope and maximum concentration parameters, respectively, dictated increase in CSS with respect to the average ruxolitinib dose per application; these parameters are related to the inverse of apparent clearance. Log-normally distributed inter-individual variability (IIV) was incorporated on the slope or maximum concentration parameters, and effect of body weight was incorporated a priori (referenced to 70 kg) with a fixed exponent of –0.75. A combined proportional and additive residual error model was evaluated. Base model selection was guided by changes in the minimum objective function value, standard goodness-of-fit diagnostics, and visual predictive checks. Covariates evaluated on slope/maximum concentration parameters included age, sex, race, indication, disease severity, and geographic region. Covariate inclusion in the final model was guided by improvement in goodness-of-fit diagnostics. Parameter estimation was performed using NONMEM®. Stochastic simulations using the final model evaluating the probability of observing CSS above the clinically relevant threshold concentration of 281 nM were performed for candidate scenarios determined by covariates included in the final model. Exploratory graphical analysis was performed on the relationships between clinical laboratory test results of select hematology parameters and the CSS of ruxolitinib among adults who applied ruxolitinib cream BID continuously.
Results: The relationship between CSS and average ruxolitinib dose per application was best described by a saturable concentration model. Relative to reference patients with mild AD and after accounting for differences in applied ruxolitinib doses, patients with vitiligo had 72% higher CSS, while patients with AD in Europe or with baseline moderate to severe disease experienced 41% and 26% higher CSS, respectively. While these covariates were associated with higher CSS, the probability of observing concentrations >281 nM was approximately ≤5% across all scenarios. On average, concentrations across all target populations maintained a ≥6-fold margin relative to 281 nM. The evaluation of hematologic parameters by ruxolitinib plasma concentrations showed an absence of plasma concentration–dependent effects across CSS quartiles across all 3 hematologic parameters and demonstrated maintenance of values within reference ranges. No clinically relevant changes in hematologic parameters were observed in the few patients with CSS >281 nM.
Conclusions: Ruxolitinib cream (1.5%) BID (the approved strength) produced low plasma concentrations in adults with AD with CSS substantially below levels (≥6-fold margins to 281 nM) considered clinically relevant for JAK2-mediated hematologic effects. CSS was adequately described by a saturable relationship. Indication, Europe region, disease severity, and body weight were associated with differences in CSS. Ruxolitinib CSS following topical application did not adversely affect key hematologic parameters across the range of observed plasma concentrations. The results of this pooled analysis are consistent with the well-tolerated safety profile of 1.5% ruxolitinib cream for the treatment of AD and vitiligo.
Reference: PAGE 34 (2026) Abstr 12044 [www.page-meeting.org/?abstract=12044]
Poster: Drug/Disease Modelling - Other Topics