The study was a phase 2, double-blind, randomized, placebo-controlled, multiple-dose investigation of reldesemtiv in 2 sequential ascending dose cohorts of patients with SMA (ClinicalTrials.gov identifier NCT02644668), conducted from December 2015 through May 2018 at 18 centers in Canada and the USA. The study received approval from institutional review boards before commencing, and was conducted in compliance with good clinical practice and the Declaration of Helsinki; written informed consent was obtained from patients ≥ 18 years of age, and parental permission and child assent were obtained for those < 18 years of age. Patients were randomized via an interactive web response system to receive reldesemtiv 150 mg bid or placebo (2:1; cohort 1), or reldesemtiv 450 mg bid or placebo (2:1; cohort 2), stratified by ambulatory status. The study drug was constituted with water at the site; the suspension was given to the patient for oral administration (9 mL) bid, approximately every 12 h after a ≥ 3-h fast, with a 1-h fast following dosing, for a total of 8 weeks.
Patients eligible for enrollment had genetically confirmed diagnosis of type II, III, or IV SMA and were ≥ 12 years of age. Ambulatory patients, after independently achieving a standing position, were required to complete at least 1 lap (≥ 50 m) in the 6-minute walk test (6MWT) without assistance. Non-ambulatory patients needed to be able to tolerate an upright supported sitting position continuously for 3 h, and required a wheelchair for mobility needs, though they may have been able to stand or walk less than 50 m without assistance in 6 min. Patients were also required to have forced vital capacity (FVC) > 20% predicted, Hammersmith Functional Motor Scale Expanded (HFMSE) score ≥ 10 and ≤ 54, and contracture of elbow flexion ≤ 90 degrees. Patients must have been able to swallow an oral suspension and were expected to be able to continue to do so for the duration of the study. Prior or concomitant treatment with nusinersen was not allowed.
Outcome measures were assessed at screening, day 1, at the end of weeks 1, 2, 4, and 8, and at follow-up (4 weeks after the last dose) for all except timed up and go (TUG) and 6MWT, which were assessed at screening, day 1, at the end of weeks 4 and 8, and at follow-up. The Spinal Muscular Atrophy-Health Index (SMA-HI) was completed on day 1 and at the end of week 8 in cohort 2 only. All clinical measures were standardized with certified clinical evaluators who underwent appropriate training to ensure consistency of measures across all sites and visits throughout the study.
Pulmonary function was assessed via FVC, as well as maximum inspiratory pressure (MIP) and maximum expiratory pressure (MEP), measured using calibrated spirometers by Micro Direct, Inc., Lewiston, ME (FVC: MicroLab Spirometer MK8; MIP and MEP: MicroRPM). Isometric muscle strength was assessed bilaterally using a make test, in which the examiner holds the dynamometer in a set position as the patient pushes against it, for 3 muscle groups (elbow flexion, knee flexion, and shoulder abduction) with a MicroFET2 handheld dynamometer (HHD, Hoggan Scientific, LLC, Salt Lake City, UT). The maximum muscle strength of 2 measurements was reported as percent change from baseline, with imputed muscle strength set to missing if the baseline value was 0. The mega-score, a composite score for strength across the 3 muscle groups, was calculated as the mean of transformed muscle strength scores. The HFMSE was used to evaluate functional mobility with a scale score ranging from 0 to 66, with higher scores reflecting better function . The revised upper limb module (RULM) was designed to be used in conjunction with the HFMSE; it tests upper limb function based on reachable workspaces from upper (shoulder), middle (elbow), and distal (wrist and hand) regions of the upper limb with a range of scoring from 0 to 44, with higher scores reflecting greater function .
The 6MWT, performed in ambulatory patients, measured the distance a patient walked in 6 min (6MWD) in order to assess functional endurance capacity and mobility; fatigue was measured by the difference in distance walked between the first and last minutes . To assess mobility, balance, and walking ability, the TUG, performed in ambulatory patients, measured the time for a patient to rise from a chair, traverse 3 m, turn around, return to the chair, and sit down . Patient-reported SMA burden was measured using the SMA-HI questionnaire, with higher scores reflecting greater burden of disease; this was performed only in cohort 2 given the timing of when the questionnaire was fully developed . Safety was assessed by monitoring AEs, coded using the Medical Dictionary for Regulatory Activities (MedDRA), version 18.0, clinical laboratory findings, and electrocardiogram intervals.
For PK endpoints, blood samples for determination of plasma concentrations of reldesemtiv were collected prior to dosing on day 1 and at the ends of weeks 1, 2, 4, and 8. Concentrations were determined by a validated solid-phase extraction method using high-performance liquid chromatography followed by tandem mass spectrometric detection.
The safety population consisted of all patients who received ≥ 1 dose of study drug, whereas the PD population included patients in the safety population with ≥ 1 post-baseline endpoint assessment, and the PK population included those in the safety population with ≥ 1 evaluable PK level. Baseline characteristics and safety data were summarized descriptively overall and by treatment, dose level, and ambulatory status in the safety population.
For dose-response PD effects, the change from baseline in a continuous endpoint was analyzed by repeated measures mixed effect models that accounted for within-patient correlation with an unstructured covariance matrix or compound symmetry when an unstructured covariance matrix could not converge. The covariates included dose level, visit, interaction between dose level and visit, ambulatory status, and the baseline value of the variable being analyzed (except for muscle strength mega-score, for which the score itself represents a percent change from baseline). Ambulatory status was removed from the model for subgroup analyses by ambulatory status. Important baseline characteristics (including age, gender, and age of disease onset) were also examined and included in the model.
The slope of change from baseline for continuous endpoints was analyzed by a mixed effect model with no intercept, and included dose level, days from the first dose of study treatment as a random covariate, interaction of dose level-by-days from the first dose of study treatment, ambulatory status, and baseline value of the variable being analyzed.
Maximum observed plasma concentration (Cmax), predose plasma concentration (Ctrough), and area under the plasma concentration-time curve from 0 to 12 h (AUC0–12) were calculated based on the plasma concentrations of reldesemtiv using noncompartmental PK methods and were summarized descriptively. For concentration-response PK analyses, the change in continuous endpoints from baseline was analyzed by a mixed effect model that accounted for within-patient correlation with an unstructured covariance matrix. The covariates included PK parameter, visit, interaction between visit and the PK parameter, ambulatory status, and the baseline value of the variable being analyzed.
For model-based PD analyses, least squares (LS) mean, difference of LS means between reldesemtiv and placebo (combined from the 2 cohorts), their standard errors (SEs) and 95% confidence intervals (CIs), and 2-sided p values were determined. Multiplicity was not addressed in this hypothesis-generating study; all p values of statistical significance are nominal.