Safety and efficacy of oral levosimendan in people with amyotrophic lateral sclerosis (the REFALS study): a randomised, double-blind, placebo-controlled phase 3 trial

Background: There is an urgent unmet need for new therapies in amyotrophic lateral sclerosis. In a clinical study with healthy volunteers, levosimendan, a calcium sensitiser, was shown to improve neuromechanical efficiency and contractile function of the human diaphragm. We aimed to evaluate the safety and efficacy of oral levosimendan in people with amyotrophic lateral sclerosis, with a focus on respiratory function.

Methods: The REFALS study is a randomised, double-blind, placebo-controlled phase 3 trial at 99 amyotrophic lateral sclerosis specialist centres in 14 countries worldwide. People with amyotrophic lateral sclerosis were eligible for participation if they were at least 18 years of age and had a sitting slow vital capacity (SVC) of 60-90% predicted. Participants were randomly assigned (2:1) by interactive web-response system to receive either levosimendan or placebo. The capsules for oral administration were identical in appearance to maintain blinding of participants and investigators. The primary endpoint was the change from baseline in supine SVC at 12 weeks, assessed as the percentage of predicted normal sitting SVC. The key secondary endpoint was the combined assessment of function and survival (CAFS) up to 48 weeks. Analyses were done in the intention-to-treat population, comprising all participants who were randomly assigned. This trial is registered at ClinicalTrials.gov (NCT03505021) and has been completed. An extension study (REFALS-ES; NCT03948178) has also been completed, but will be reported separately.

Findings: Between June 21, 2018, and June 28, 2019, 871 people were screened for the study, of whom 496 were randomly assigned either levosimendan (n=329) or placebo (n=167). Participants were followed up between June 27, 2018 and June 26, 2020, for a median duration of 50·1 (IQR 37·5-51·1) weeks. The median duration of treatment was 47·9 (IQR 26·4-48·1) weeks. Change from baseline in supine SVC at 12 weeks was -6·73% with levosimendan and -6·99% with placebo, with no significant difference between the treatments (estimated treatment difference 0·26%, 95% CI -2·03 to 2·55, p=0·83). Similarly, at week 48, CAFS did not differ between treatment groups (least squares mean change from baseline 10·69, 95% CI -15·74 to 37·12; nominal p value=0·43). The most frequent adverse events were increased heart rate (106 [33%] of 326 receiving levosimendan vs 12 [7%] of 166 receiving placebo), fall (85 [26%] vs 48 [29%]), headache (93 [29%] vs 36 [22%]), and dyspnoea (59 [18%] vs 32 [19%]). 33 (10%) participants allocated levosimendan and 20 (12%) assigned placebo died during the trial, mainly due to respiratory failure or progression of amyotrophic lateral sclerosis.

Interpretation: Levosimendan was not superior to placebo in maintaining respiratory function in a broad population with amyotrophic lateral sclerosis. Although levosimendan was generally well tolerated, increased heart rate and headache occurred more frequently with levosimendan than with placebo. The possibility of a clinically relevant subgroup of responsive individuals requires further evaluation.

Funding: Orion Corporation.