S1748 Mirikizumab Crohn's Disease VIVID-1 Study Placebo Response Differences Between the United States Labeling and Primary Manuscript Due to Methodological Differences in Handling Participants Randomized to Placebo

Introduction: Different methods for handling placebo (PBO) response in a treat-through design study, where patients receiving PBO are switched to active treatment after induction, can lead to differences in reported PBO response rates. Mirikizumab (MIRI) effect size has been reported differently in publications compared with the United States Package Insert (USPI)2. We discuss these different methodologies and their impact on determining MIRI’s effect size. Methods: For VIVID-1 publications, inclusion criteria for the primary analysis set (PAS) population were a baseline unweighted daily average frequency of loose/watery stool ≥4 and/or unweighted daily average abdominal pain ≥2 and a centrally read Simple Endoscopic Score for Crohn’s Disease (SES-CD) score ≥7 or ≥4 for patients with isolated ileal disease. Week (W) 12 clinical response was determined by Patient Reported Outcome (PRO). Nonresponder imputation (NRI) was used. PBO patients who did not achieve PRO clinical response at W12 were switched to MIRI and analyzed as NRs for all W52 efficacy outcomes (PBO/PBO W12 Observation Carried Forward method). In the USPI, PBO patients switched to MIRI at W12 were analyzed as PBO patients with their observed W52 results used for all W52 efficacy endpoints despite having received MIRI for 40 weeks (PBO/PBO + PBO/MIRI Observed method). To assess the impact of the 2 methods on MIRI effect size, we examined efficacy in the USPI population using both methods. Results: For the endpoints included, W52 PBO rates were 5%-17% greater for the observed method compared with the observation carried forward method, thereby changing MIRI effect size by the same amount. Across the 4 W52 endpoints, a range of 48%-69% of the W52 PBO response rate for the observed method was driven by patients with 40 weeks treatment on MIRI. Conclusion: PBO rate differences between the MIRI USPI and published VIVID-1 analyses are due to different methodologic approaches to calculating treatment response. Published greater differences in the treatment effect between MIRI and PBO over 52 weeks included MIRI-treated patients but as NRs (40 weeks of MIRI treatment) in the PBO response rate, whereas these patients were included with observed data in the method used for the USPI.