Abstract
BACKGROUND: Natalizumab (NTZ) and ocrelizumab (OCR) can be used for the treatment of relapsing-remitting multiple sclerosis (RRMS). In patients treated with NTZ, screening for JC virus (JCV) is mandatory, and a positive serology usually requires a change in treatment after 2 years. In this study, JCV serology was used as a natural experiment to pseudo-randomize patients into NTZ continuation or OCR. METHODS: An observational analysis of patients who had received NTZ for at least 2 years and were either changed to OCR or maintained on NTZ, depending on JCV serology status, was performed. A stratification moment (STRm) was established when patients were pseudo-randomized to either arm (NTZ continuation if JCV negativity, or change to OCR if JCV positivity). Primary endpoints include time to first relapse and presence of relapses after STRm and OCR initiation. Secondary endpoints include clinical and radiological outcomes after 1 year. RESULTS: Of the 67 patients included, 40 continued on NTZ (60%) and 27 were changed to OCR (40%). Baseline characteristics were similar. Time to first relapse was not significantly different. Ten patients in the JCV + OCR arm presented a relapse after STRm (37%), four during the washout period, and 13 patients in the JCV-NTZ arm (32.5%, p = 0.701). No differences in secondary endpoints were detected in the first year after STRm. CONCLUSIONS: The JCV status can be used as a natural experiment to compare treatment arms with a low selection bias. In our study, switching to OCR versus NTZ continuation led to similar disease activity outcomes.