Bayesian adaptive design for clinical trials in Duchenne muscular dystrophy

Abstract

Bayesian adaptive design is proposed for a clinical trial in Duchenne muscular dystrophy. The trial was designed to demonstrate treatment efficacy on an ambulatory-based clinical endpoint and to identify early success on a biomarker (dystrophin protein levels) that can serve as a basis for accelerated approval in the United States. The trial incorporates placebo augmentation using placebo data from past clinical trials. A thorough simulation study was conducted to understand the operating characteristics of the trial. This trial design was selected for the US FDA Complex Innovative Trial Design Pilot Meeting Program and the experience in that program is summarized. K E Y W O R D Sadaptive design, Bayesian, clinical trial, clinical trial simulation, interim analysis INTRODUCTIONDuchenne muscular dystrophy (DMD) is an X-linked genetic disorder caused by mutations in the dystrophin-encoding DMD gene 1,2 and is the most common type of childhood muscular dystrophy. [3][4][5] As it affects only one in 5000 males born worldwide, it is a rare disease as defined by the US Food and Drug Administration (FDA; affects fewer than 200 000 people) and European Medicines Agency (EMA; affects not more than 5 in 10 000 persons). Progressive muscle degeneration in patients with DMD eventually leads to loss of ambulation in most by 10 to 12 years of age. [6][7][8] With supportive care, such as corticosteroid use, ventilatory assistance, and scoliosis surgery, life expectancy in DMD patients may be extended to approximately 30 years of age. 9 Effective therapies are needed to slow the course of this devastating disease. The development of therapeutics to treat DMD has proven challenging. To date, no therapy has demonstrated a convincing benefit on a clinical endpoint which in DMD clinical trials are typically measures of ambulatory function, such as the timed 6-minute walk test (6MWT) or the North Star Ambulatory Assessment (NSAA 10 ). The high level of variability in ambulation-based endpoints has complicated the interpretation of several trials in DMD. 11,12 In 2016, EXONDYS 51®, an antisense oligonucleotide (ASO) intended to target human dystrophin premessenger ribonucleic acid in order to induce exon 51 skipping and dystrophin protein restoration, was approved by the FDA under the accelerated approval pathway using a surrogate endpoint that was considered reasonably likely to predict clinical benefit. The basis for the accelerated approval was the observed increase in dystrophin levels in clinical trial patients. Recently, other ASOs (VYONDYS 53™, VILTEPSO®) for exon 53 skipping were granted accelerated approval on similar grounds. Confirmation of clinical benefit for these drugs in the postmarketing setting is a condition of continued US FDA approval. While the EMA also has an accelerated approval mechanism for medicines intended to treat serious orStephen L. Lake and Melanie A. Quintana contributed equally to this work.

Adaptive Platform Trials to Transform Amyotrophic Lateral Sclerosis Therapy Development

“…The primary Bayesian shared parameter model combines function and mortality (death or use of permanent assisted ventilation) by introducing two components with a common shared treatment effect across them. The functional component of the analysis model is a Bayesian repeated measures model of ALSFRS‐R at baseline through 24 weeks (measured every 4 weeks) for all participants who have not experienced mortality and measures the slowing in the rate of progression (common across all time points) on the ALSFRS‐R in treated participants relative to the shared controls for survivors similar to that used in other neurological diseases 34–36 . The mortality component of the analysis model is an exponential proportional hazards model of mortality through 24 weeks and measures the mortality hazard ratio in treated participants relative to control.…”

“…The functional component of the analysis model is a Bayesian repeated measures model of ALSFRS-R at baseline through 24 weeks (measured every 4 weeks) for all participants who have not experienced mortality and measures the slowing in the rate of progression (common across all time points) on the ALSFRS-R in treated participants relative to the shared controls for survivors similar to that used in other neurological diseases. [34][35][36] The mortality component of the analysis model is an exponential proportional hazards model of mortality through 24 weeks and measures the mortality hazard ratio in treated participants relative to control. A shared treatment effect parameter between mortality and function is introduced as a common slowing in the time of progression of the disease.…”

Application of Bayesian methods to accelerate rare disease drug development: scopes and hurdles

“…Pre-trial simulations must be performed to assess trial characteristics under a variety of real-life settings to decide if incorporating adaptive components is advantageous. Additional examples of Bayesian interim analyses in rare disease are the Focal Cerebral Arteriopathy Steroid (FOCAS) trial [ 86 ] and the study of Suvodirsen in ambulatory patients with Duchenne Muscular Dystrophy (DYSTANCE 51) [ 87 ]. There are also methods available for a Bayesian response adaptive design that allocates more participants to the best performing arm, has higher power than a traditional fixed design, and has small bias and mean squared error of the treatment effect estimates [ 88 ].…”

Regulatory Framework for Drug Development in Rare Diseases

“…One example of a clinical trial selected for this program is an adaptive design for clinical trials in Duchenne muscular dystrophy. 43 Patient-Focused Drug Development Given the small number of available subjects to enroll in studies of rare diseases it is important for study conduct that barriers to patient participation be removed and that their concerns and worries are addressed. Clinical development programs that are patient focused are more likely to be of more interest to potential participants and investigators, as well as payers.…”

“…The pilot program was initiated in 2019 and will continue through 2022. One example of a clinical trial selected for this program is an adaptive design for clinical trials in Duchenne muscular dystrophy 43 …”