IntroductionKnowledge about the effect of disease modifying treatment (DMT) in late-onset multiple sclerosis (LOMS, onset ≥50 years-old) is scarce. This study aims to evaluate the association between DMT use and multiple sclerosis (MS) evolution in a LOMS cohort. MethodsThis multicentre, retrospective and observational study included LOMS patients with ≥2 years of follow-up. Data on demographics, clinical/paraclinical (baseline and follow-up), DMT and adverse events were collected. Primary outcomes were irreversible EDSS 4.0 and 6.0 achievement and first year ARR. Univariate and multivariate regression models were conducted, with treated and/or relapsing phenotypes (RMS) subgroups analyses. ResultsWe included 232 patients (53.4% with RRMS phenotype; 84.9% submitted to DMT; median follow-up time of 141.5 (IQR 92.7-193.1) months). Treatment versus non-treatment did not affect EDSS milestones in multivariate analysis (adjusted to phenotype, baseline EDSS, age, and ARR), but initially receiving monoclonal antibodies (MAbs) was associated with lower odds of EDSS 4.0 (OR 0.13). In treated patients, starting with high efficacy DMT (HE-DMT) was related to a lower chance of EDSS 4.0 (OR 0.05) and 6.0 (OR 0.26) compared with being exclusively treated with moderate efficacy DMT (ME-DMT), with similar results when analysing the subgroup of RMS treated patients. In multivariate models, initial treatment with MAbs (vs. non-treatment) and with HE-DMT (vs. ME-DMT) were related to a lower first year ARR; when considering only RMS patients, every DMT class analysed reduced first year ARR vs. non-treatment. During DMT, we documented a rate of 0.6% serious infections, 0.07% opportunistic infections and 0.7% neoplasm diagnosis per patient year. ConclusionDMT type and therapeutic strategy influenced LOMS disability accumulation and relapses in our cohort. Our findings support the importance of investment in LOMS treatment optimization.