A novel adaptive design strategy increases the efficiency of clinical trials in secondary progressive multiple sclerosis

Jeremy Chataway,Susan Todd,Nigel Stallard,Tim Friede,Elsa Valdés-Márquez,David H Miller,Nicholas Parsons,Richard Nicholas

doi:10.1177/1352458510382129

Abstract

Background:Adaptive seamless designs (ASDs) have been proposed to test multiple candidate compounds using an interim decision point which allows potentially effective therapies to be taken into the next design stage and to be assessed using a phase III outcome. Objective:To determine whether ASDs are feasible in secondary progressive multiple sclerosis (SPMS) and to compare them with conventional trial designs. Methods:We develop an innovative adaptive trial design for SPMS, which builds on recent developments in statistical methodology. A literature search and individual clinical datasets were used to inform a framework to run simulations to evaluate the proposed design. Results:ASDs are feasible in SPMS with MRI informing an interim decision point and Expanded Disability Status Scale (EDSS) as the final disability endpoint. Furthermore ASDs are more efficient than conventional designs with sample size savings of up to 40%. Sample sizes of 1000–1250 patients are sufficient to test up to four experimental treatments. Controlled recruitment is important to realize the full benefits of ASDs. Conclusions:Although more complex in design, ASDs have the potential to be more efficient and more powerful than conventional designs.

Highlights

Despite the success of increasingly sophisticated compounds in modulating the early phases of multiple sclerosis (MS), both delaying the conversion of clinically isolated syndromes to the full disease state[1] and reducing relapse frequency,[2] the pivotal problem of altering an established gradient of progression, as the disease evolves from an inflammatory to an axonal destructive state,[3] remains
To apply Adaptive seamless designs (ASDs) in MS, two components are required: first, methodology is needed to develop a design for the trial; and, second, a framework is necessary to facilitate the conduct of simulation studies that evaluate how the trial might run in practice
Individual patient data were obtained from seven neuroscience centres: longitudinal natural history studies (n 1⁄4 6) and randomized controlled trials (RCTs) control groups (n 1⁄4 1), all of which reported Expanded Disability Status Scale (EDSS), confirming again that this is clearly the most common endpoint used and would be appropriate for an ASD

Summary

Introduction

Despite the success of increasingly sophisticated compounds in modulating the early phases of multiple sclerosis (MS), both delaying the conversion of clinically isolated syndromes to the full disease state[1] and reducing relapse frequency,[2] the pivotal problem of altering an established gradient of progression (primary or secondary), as the disease evolves from an inflammatory to an axonal destructive state,[3] remains. A confirmatory phase III analysis can be conducted at the end of the study using combined information from the first and second stages Such an approach is more efficient than the traditional phase II followed by phase III approach.[6] Previously, such designs have largely been described in settings where the interim treatment selection is based on the final outcome, which is unrealistic in progressive MS. Adaptive seamless designs (ASDs) have been proposed to test multiple candidate compounds using an interim decision point which allows potentially effective therapies to be taken into the design stage and to be assessed using a phase III outcome. Conclusions: more complex in design, ASDs have the potential to be more efficient and more powerful than conventional designs

Objectives

Methods

Results

Conclusion