Background Mavrilimumab, a fully human monoclonal antibody targeting the granulocyte-macrophage colony-stimulating factor receptor (GM–CSFR) α, is in Phase IIb investigation for the treatment of rheumatoid arthritis (RA).
Objectives We characterized the exposure–efficacy relationship of mavrilimumab with data from two Phase II randomized controlled trials of patients with moderate to severe RA, via population modeling and clinical simulations.
Methods In a Phase IIa study, adult patients with active RA received 7 subcutaneous administrations of mavrilimumab (10, 30, 50 or 100 mg every other week [eow]). Mavrilimumab pharmacokinetics and efficacy (DAS28–CRP, ACR20/50) were modeled using a population approach. Clinical simulations were performed to facilitate the design of a Phase IIb trial, in which RA patients received placebo or mavrilimumab (30, 100, or 150 mg eow) for 24 weeks. We developed a joint probability model to describe the risk of voluntary dropouts at each scheduled visit of the Phase IIb study. The exposure–response relationship for mavrilimumab was characterized by clinical simulations via the PK-efficacy dropout model.
Results Mavrilimumab and placebo treatment effects observed in the Phase IIa study were adequately described by mechanistic PK-efficacy model. Although mavrilimumab 150 mg eow was not evaluated in the Phase IIa study, improved efficacy was observed at this dosage in Phase IIb, as predicted by a priori clinical simulations. The hazard rate of patient dropout prior to next scheduled visits was significantly greater for non-responders, and substantially greater for lower dosage groups at Week 12, when patient rollover to an open-label extension study was permitted. Clinical simulations indicated that the maximum efficacy was reached at the 150-mg eow dosage.
Conclusions Efficacy outcomes of the Phase IIb study confirmed the rational selection of mavrilimumab 150 mg eow as the best dosage, through modeling of Phase IIa data. Joint modeling of placebo effect, treatment response, and informative dropout greatly facilitates interpretation of Phase IIb results, and will assist in the design of late-stage clinical trials for mavrilimumab.
Disclosure of Interest D. Jin Shareholder of: AstraZeneca, Employee of: MedImmune, C.-Y. Wu Shareholder of: AstraZeneca, Employee of: MedImmune, L. Roskos Shareholder of: AstraZeneca, Employee of: MedImmune, A. Godwood Shareholder of: AstraZeneca, Employee of: MedImmune, D. Close Shareholder of: AstraZeneca, Employee of: MedImmune, B. Wang Shareholder of: AstraZeneca, Employee of: MedImmune