Abstract Rituximab, an anti-cluster of differentiation 20 (anti-CD20) monoclonal antibody, has been widely used off-label for multiple sclerosis (MS) treatment, demonstrating high effectiveness and a favorable safety profile. However, real-world data from Brazil, particularly within the public healthcare system, remain scarce. To assess the real-world effectiveness and safety of rituximab for MS treatment in a Brazilian tertiary center. We conducted a single-center, observational study at Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo (HCFMUSP), including MS patients treated with rituximab between January 2016 and October 2024. The primary outcome was the change in annualized relapse rate (ARR) following rituximab initiation. The secondary outcomes included disability progression, radiological activity, and safety. The statistical analyses employed non-parametric tests and time-to-event analysis. Among the 45 patients (25 with relapsing-remitting MS RRMS, 18 with secondary progressive MS, and 2 with primary progressive MS), rituximab reduced the ARR by 89.8% (from 0.49–0.05; p < 0.001), with 90.9% remaining relapse-free. In RRMS patients, the ARR significantly decreased, from 0.56 to 0.17, after rituximab initiation (p = 0.019), with 88% remaining relapse-free (95%CI: 68.8–97.5). Radiological activity was observed in 6.8% of the sample. Patients with breakthrough disease activity (n = 4; 8.9%) presented more frequent prior natalizumab use than those without it (75.0% versus 43.9%; p = 0.039), as well as higher pretreatment relapse rates (1.8 versus 0.7; p = 0.033). Infusion-related reactions affected 12 (26.7%) patients, and they were mostly mild, while infections were the most common non-infusion adverse event (20; 57.8%). In total, 4 patients discontinued rituximab, mainly due to access limitations. Rituximab demonstrated high effectiveness and acceptable safety for MS treatment in a Brazilian public healthcare setting. Its favorable risk-benefit profile supports broader adoption in resource-limited contexts.
Simon-Nogueira et al. (2026) studied this question.