P234: The ICH E9(R1) Estimand Framework Implemented in a Phase III Equivalence RCT Conducted during COVID-19 Pandemic - A Case Study
Poster Presenter
Marian Mitroiu
Associate Director Biostatistics
Biogen Switzerland
Objectives
Implement ICH E9(R1) estimand framework with intercurrent events and strategies for addressing intercurrent events (related or not to COVID-19 pandemic) to define different estimands for clinical efficacy evaluation of a Phase III study following recommendations from different Regulatory Agencies.
Method
A multiregional, double-blind Phase III equivalence RCT of a proposed biosimilar BAT1806/BIIB800 vs reference tocilizumab was conducted partially during the COVID-19 pandemic (Dec 2018 – Jan 2021). ICH E9(R1) estimand framework was implemented for efficacy evaluation.
Results
As a consequence of COVID-19 pandemic and consequent lockdowns, intercurrent events related or not related to COVID-19 were identified and incorporated in the efficacy evaluations. The implementation of the estimand framework resulted in three different estimand constructs, primary and secondary, depending on agreement/recommendations from each Regulatory Agency (EMA/FDA/NMPA).
Primary estimand for EMA assessed the treatment effect at Week 12 where the clinical question was: “What is the treatment effect had no subject discontinued the treatment, nor missed a study treatment infusion, for any reason, nor needed rescue medication within protocolled window (hypothetical), and death considered a non-response (composite)?”. To establish equivalence a two-sided 95% confidence interval (CI; a=0·025) and a margin of (-14.5%, +14.5%) was applied.
Primary estimand for FDA/ NMPA assessed the treatment effect at Week 24 where the clinical question of interest was: “What is the treatment effect regardless of any treatment discontinuation or missed study treatment infusion, for any reason, regardless of any rescue medication need within protocolled window (treatment policy), and death considered a non-response (composite)?”. A two-sided 90% CI (a=0·05) and equivalence margins of (-12.0%, +15.0%) and for NMPA, a two-sided 95% CI (a=0·025) and equivalence margins of (-13.6%, +13.6%) was applied.
Secondary estimands for Regulatory Agencies assessed the treatment effect at Week 12/Week 24 where the clinical question was: “What is the treatment effect regardless of any treatment discontinuation or missed study treatment infusion not related to COVID-19 (treatment policy), and had no subject discontinued treatment or missed a study treatment infusion related to COVID-19 (hypothetical), and regardless of rescue medication need within protocolled window (treatment policy), and death considered a non-response (composite)?”.
Conclusion
The ICH E9(R1) estimand framework was implemented for a phase III following guidance and methodological considerations for trials affected by COVID-19 pandemic. The E9(R1) guideline was helpful to define treatment effects using estimand attributes, especially the strategies for addressing intercurrent events.
Equivalence was demonstrated for both primary and secondary estimands with corresponding CIs contained within the predefined equivalence margins.
Authors: Marian Mitroiu, Hans Ebbers, Yinbo Zhou, Xiaolei Yang, QingFeng Dong, Mourad F. Rezk, Janet Addison
