Propensity Score Matched Comparative Effectiveness Analysis of Dimethyl Fumarate Relative to Interferon, Glatiramer Acetate, or Teriflunomide Treated RRMS Outpatients in the German NeuroTransData Registry

Objective: Comparative effectiveness analysis of patients initiating dimethyl fumarate (DMF) versus a pair-wise propensity-score-matched (PSM) cohort initiating interferons (IFN), glatiramer acetate (GA), or teriflunomide (TERI). Background: No randomized controlled trials have been conducted to compare efficacy outcomes between DMF versus IFN, TERI, or GA. NeuroTransData (NTD) is a German network of u003e130 practice-based neurologists running a MS registry including ~25,000 patients with RRMS. Design/Methods: Data were sourced from the NTD MS registry on 1 October 2016, including RRMS patients aged ≥18 years at therapy initiation with ≥1 relapse and EDSS assessment on-therapy. 1:1 PSM was used to match DMF to comparator cohorts. Primary outcome: time-to-first-relapse (TTFR). Secondary outcomes: annualized relapse rate (ARR) and time-to-treatment-discontinuation (TTD). Exploratory outcomes: time-to-3-months- and -6-months-confirmed-disability-progression (TTCDP3, TTCDP6). TTFR, TTD, and TTCDP3/TTCDP6 were analyzed using a Kaplan-Meier approach and Cox-marginal-regression-model, ARR using a GEE Poisson-regression-model. The clustered nature of the matched design was taken into account. Non-pairwise censoring was applied. Results: DMF patients were matched 1:1 to IFN (n=439), GA (n=535), and TERI (n=388); more than 75%, 61%, and 36% of the patients were DMD-naive, respectively. A statistically significant reduction was observed in TTFR for DMF vs IFN (Hazard Ratio [HR] 0.59; 95% CI 0.44, 0.79; P =0.0004), GA (HR 0.65; 95% CI 0.50, 0.84; P =0.001), and TERI (HR 0.53; 95% CI 0.38, 0.75; P =0.0003). A statistically significant reduction was observed in ARR for DMF vs IFN (Rate ratio [RR] 0.70; 95% CI 0.53, 0.94; P =0.017), GA (RR 0.76; 95% CI 0.59, 0.98; P =0.035), and TERI (RR 0.55; 95% CI 0.39, 0.77; P =0.001). There was no evidence for differences in TTCDP and TTD for any comparison. Conclusions: A PSM 1:1 analysis of DMF versus IFN, GA, and TERI, sourced from the NTD network, showed a statistically significant reduction in TTFR and ARR in favor of DMF. Study Supported by: Biogen Disclosure: Dr. Bergmann has received personal compensation for consulting, serving on a scientific advisory board, speaking, or other activities with NeuroTransData for project management, clinical studies, travel expenses from Novartis, Servier. Dr. Bergmann has received compensation for serving on the Board of Directors of NeuroTransData. Dr. Braune has received personal compensation for consulting, serving on a scientific advisory board, speaking, or other activities with Kassenarztliche Vereinigung Bayern and HMOs for patient care, from Biogen, Lilly, Medday, Merck, NeuroTransData, Novartis, Roche, Thieme Verlag for consultancy, project management, clinical studies and lectures. Dr. Braune has received compensation for serving on the Board of Directors of NeuroTransData. Dr. Grimm has received personal compensation for consulting, serving on a scientific advisory board, speaking, or other activities with employee of PwC, contracted to perform statistical projects for NeuroTransData. Dr. van Hovell has received personal compensation for consulting, serving on a scientific advisory board, speaking, or other activities with employee of PwC, contracted to perform statistical projects for NeuroTransData. Dr. Freudensprung has received personal compensation for consulting, serving on a scientific advisory board, speaking, or other activities with Employee of Biogen. Dr. Hyde has received personal compensation for consulting, serving on a scientific advisory board, speaking, or other activities with Employee of Biogen. Dr. Group has received research support from NTD members (including Dr. Braune and Dr. Bergmann) receive honorariums from many pharmaceutical companies for studies, consultancy and lectures.