Determinants of clinical response to levosimendan in the REFALS phase 3 study in people with ALS
Merit Cudkowicz1, Angela Genge2, Nicholas Maragakis3, Susanne Petri4, Leonard van den Berg5, Valtteri Aho6, Chris Garratt6, Toni Sarapohja6, Ammar Al-Chalabi7
1Massachusetts General Hospital, 2Mcgill University, 3Johns Hopkins University School of Medicine, 4Medizinische Hochschule Hannover, 5University Medical Center Utrecht, 6Orion Pharma, 7King'S College London

To evaluate whether baseline characteristics determined response to levosimendan in people with amyotrophic lateral sclerosis (ALS) in the REFALS phase 3 trial (NCT03505021).


Treatments for people with ALS may not work equally for all patients and it may be possible to identify a population more likely to derive a relevant treatment effect. The REFALS study explored the clinical effects of oral levosimendan during prolonged treatment in people with ALS.

The REFALS study enrolled 496 adult participants with ALS and some degree of respiratory dysfunction (sitting SVC 60-90%) at 99 clinical sites in 14 countries in EU, North America and Australia. Participants were randomized (2:1 allocation) to oral levosimendan 1-2mg daily or placebo for 48 weeks in a double-blind, parallel-group design. The primary endpoint was supine SVC at 12 weeks and the key secondary endpoint was combined assessment of ALSFRS-R function and survival (CAFS) through 48 weeks. The analysis plan pre-specified a number of baseline characteristics potentially relevant to the outcome of the study. These characteristics were evaluated in pre-specified sub-group analyses of supine SVC and ALSFRS-R, based on which further sub-groups were determined and analyzed post-hoc.

492 participants with ALS, mean 26 months from symptom onset, were treated with levosimendan or placebo. Treatment groups were balanced at baseline: spinal onset 81%, mean sitting SVC 76.5%, ALSFRS-R 36.0 points. 83% and 17.5% patients were taking riluzole and edaravone, respectively. Neither the primary  nor the key secondary endpoint showed significant difference between oral levosimendan treatment and placebo. The results of the individual pre-specified and post-hoc sub-group analyses of supine SVC and ALSFRS-R will be presented in detail.

Although the REFALS study of levosimendan was not positive, subgroup analyses may be of value in identifying patient groups with a stronger treatment effect and inform future trial design.