Background and ObjectiveTo study whether dimethyl fumarate is superior to placebo in decreasing CSF concentrations of neurofilament light chain (NFL) in patients with primary progressive MS (PPMS).MethodsIn the double-blind, placebo-controlled phase 2 study dimethyl FUMArate treatment in Progressive Multiple Sclerosis (FUMAPMS), patients with PPMS were randomly assigned to treatment with 240 mg dimethyl fumarate or placebo in a 1:1 ratio for 48 weeks. The primary endpoint was change in concentration of NFL in the CSF. Secondary endpoints included other CSF biomarkers and clinical and MRI measures. Efficacy was evaluated for the full data set by multiple imputations to account for missing data. Safety was assessed for the full data set.ResultsFifty-four patients (mean age 54.9 years [SD 6.1], median Expanded Disability Status Scale 4.0 [nterquartile range 4.0–6.0], disease duration 14.1 [SD 9.4], and 21 [39%] female) were randomized to either placebo (n = 27) or dimethyl fumarate (n = 27) therapy. At screening CSF concentrations, adjusted for age and sex, of NFL, myelin basic protein (MBP), soluble CD27, chitinase 3-like 1, and B-cell maturation antigen were higher than in a group of symptomatic controls. Twenty-six patients (96%) in the dimethyl fumarate group and 24 patients (89%) in the placebo group completed the randomized phase. Mean change in CSF concentrations of NFL did not differ between groups (mean difference 99 ng/L; 95% CI −292 to 491 ng/L). MBP in CSF decreased in the treatment group (−182 ng/L, 95% CI −323 to −41 ng/L compared with placebo). The difference observed in the multiple imputation data set was not significant in a per protocol analysis. This was nominally significant in the multiple imputation data set but not in the per protocol analysis This was not found in the per protocol analysis Other secondary and tertiary outcomes were not affected. Various infections, lymphopenia, flushing, and gastrointestinal side effects were more frequent in the dimethyl fumarate group. Serious adverse events were similar between groups.DiscussionDimethyl fumarate treatment for 48 weeks had no effect on any of the investigated efficacy measures in patients with PPMS. We did not observe adverse events not anticipated for dimethyl fumarate treatment.Trial Registration InformationClinicaltrials.gov identifier NCT02959658.Classification of EvidenceThis study provides Class I evidence that for patients with PPMS, dimethyl fumarate treatment has no effect on CSF NFL levels compared with placebo treatment.