Dose confirmation based on population pharmacokinetics within LUMINESCE (NCT04963270); a Phase III, randomized, double-blind, placebo-controlled study evaluating satralizumab in patients with generalized myasthenia gravis
Hanna Silber Baumann (1), Sebastien Jolivet (1), Leonid Gibiansky (2), Gian-Andrea Thanei (3), Prune Schlewitz (3), Xiujing Kou (4), Petranka Krumova (5), Jill Smith (6), Nicolas Frey (1), Hajime Ito (7), Siân Lennon-Chrimes (6)
(1) Roche Pharma Research and Early development, Pharmaceutical Sciences, Roche Innovation Center Basel; (2) QuantPharm LLC, North Potomac, US; (3) Pharma Development Data Sciences, Hoffmann-La Roche, Switzerland; (4) Roche Holding Ltd, China; (5) Pharma Development Neuroscience, Hoffmann-La Roche, Switzerland; (6) Roche Products Ltd, Welwyn, UK; (7) Chugai Pharmaceutical Co., Ltd., Tokyo, Japan
Objectives:
Satralizumab is a humanized IgG2 monoclonal recycling IL-6 receptor antagonist antibody. It has been approved for treatment of neuromyelitis optica spectrum disorder (NMOSD) and is now being evaluated in new indications where IL-6 is associated with disease activity. Generalized Myasthenia Gravis (gMG) is a rare condition that causes muscle weakness. Based on satralizumab mechanism of action and disease understanding of gMG, together with efficacy in NMOSD, it was decided to go directly into Phase III. The Phase III dose was uncertain due to unexplained large differences in linear clearance (CL) between healthy volunteers and patients. A standalone PK study or a PK run-in to the Phase III study were envisaged but were considered too time-consuming and inefficient use of patient data. The alternative was to include a seamless dose confirmation step within the Phase III based on population pharmacokinetic (PK) by enabling an independent data monitoring committee (iDMC) to make a recommendation to either keep or change the dose without unblinding the sponsor or requiring a protocol amendment.
Methods:
A Phase III study design was developed which included an early interim analysis (IA) of PK data. The sponsor remained blinded, thereby protecting the integrity of this pivotal study. A decision framework for dose confirmation was developed defining criteria based on predicted exposure and receptor occupancy (RO) at steady state using NMOSD as a reference.
The disposition of satralizumab in NMOSD was well described [1] and included Michaelis-Menten elimination, which is considered to reflect the binding of satralizumab to its targets, soluble and membrane-bound IL-6R. Healthy volunteers (HV) had a 96% higher linear CL than NMOSD patients. This difference could not be explained by differences in baseline sIL-6R concentrations or any other investigated factor.
In NMOSD a flat dose of 120 mg every 4 weeks (Q4W) is approved. In the new indications a bodyweight tiered dosing regimen is planned with the aim to achieve a higher RO in heavier patients. The initial doses to be used in gMG were selected based on simulations with the population PK model assuming similarity to the NMOSD population. Alternative (higher) doses were selected assuming similarity to the HV population.
The study design needed to estimate CL adequately in subjects with gMG was investigated through simulations. It was deemed sufficient to estimate model parameters with the precision outlined in the FDA guidance [2] for pediatric studies. The population PK model was used to investigate the dependence of expected precision of parameter estimates on design parameters using a simulation-estimation procedure.
Results:
A bodyweight tiered dosing regimen with a dose of 120 mg or 180 mg for patients with a bodyweight below or above 100 kg, respectively (120 mg≤100 kg>180 mg), was selected assuming PK properties as in NMOSD. In the case of PK properties as in HV, the selected dose would be 180 mg≤100 kg>240 mg.
Evaluation of the study design showed that with a sample size of N=15 subjects and a minimum of 3 samples per subjects over 8 weeks, a sufficient precision of clearance could be achieved about 90% of the time, and the estimate of population clearance was within 80%-125% of true clearance 84% of the time.
A decision framework was developed which focused on identifying the dose option (initial or higher) that is most likely to result in exposure and RO close to the reference (NMOSD population receiving the 120 mg≤100 kg>180 mg dosing regimen). The collected patient data in gMG was analyzed using the population PK model and the CL was estimated. Simulations were performed with both possible dose options and the similarity to the reference was computed. Following the analysis, the iDMC recommended that the study continue with the initial doses. The result (while still blinded) indicates that the PK in the gMG population is more similar to the NMOSD population than to the HV population.
Conclusions:
The interim analysis was completed successfully with a recommendation by the iDMC to remain with the initial dose option, saving approximately 18 months on the development journey, compared to a standalone study. In addition, the study design allow efficient use of patient data since the patients included in the IA can also contribute to the primary analysis. The same approach will be applied to Phase III studies in two additional indications with satralizumab.
References:
[1] PAGE 29 (2021) Abstr 9665 [www.page-meeting.org/?abstract=9665]
[2] General clinical pharmacology considerations for pediatric studies for drugs and biological products. Draft FDA guidance for industry, December 2014.