Developing Interim Outcome Measures for Multi-Arm Multistage Clinical Trials in Multiple Sclerosis

Abstract

Background and Objectives Multi-arm multistage (MAMS) clinical trials use interim outcome stage analyses to test multiple candidate treatments, determining which should proceed to the next trial stage. The reduced time, cost, and participant numbers, compared with current separated phases 2 and 3 clinical trials, could expedite the development of treatments in neurodegenerative diseases, such as progressive multiple sclerosis (PMS). MAMS trial interim outcomes must reflect the final stage outcome and be able to detect treatment effects with high power at interim stages. Patient-reported outcomes (PROMs) are cost-effective and highly accessible, making them promising MAMS trial interim outcome measures. We determined whether the MS Quality of Life 54 Physical Health Composite, MS Impact Scale 29 version 2 physical health subscore (MSIS-29v2 PH), MS Walking Scale 12 version 2 (MSWS-12v2), or their composite could act as interim outcome measures in MAMS PMS trials. Methods Using a pooled analysis of 11,052 longitudinal visits from 3 PMS trials, we assessed the relationship of PROM change at early time points with 6-month confirmed disability progression (CDP) in the Expanded Disability Status Scale (EDSS) up to 36 months through Cox proportional hazards. To minimize PROM ceiling effects, we generated an MSIS-29v2 PH and MSWS-12v2 composite, prioritizing greatest change. We calculated the required sample size when these PROMs were used as interim measures. Results We analyzed 1,927 of the 2,046 available participants (mean age 54 [SD 7.2] years and 66% female). EDSS 6-month CDP risk was significantly associated with each 1-standard deviation 12-month score increase for the MSIS-29v2 PH (hazard ratio [HR] 1.20 [1.08–1.33]), MSWS-12v2 (HR 1.26 [1.13–1.40]), and composite (HR 1.29 [1.16–1.44]; all p < 0.0001). In simulations, using the composite, n = 125/arm can detect a 50% treatment effect on mean 12-month change with power ≥90% and 200/arm with a power ≥95%. Discussion Changes in the MSIS-29v2 PH, MSWS-12v2, and composite scores associated with ≥6-month EDSS CDP up to 36 months. The composite has sufficient statistical power to detect treatment effects at 12 months. These PROMs are well placed to play a role in interim analysis in MAMS trials in PMS.