Abstract
BACKGROUND: There are numerous immunotherapies that are effective in preventing relapses in neuromyelitis optica spectrum disorder (NMO-SD). With head-to-head clinical trials between immunotherapies lacking, Bayesian network meta-analysis can be used to compare treatment interventions. Previous network meta-analyses have compared monoclonal antibodies but either not included newer complement inhibitors or earlier immunotherapies such as rituximab or tocilizumab. OBJECTIVE: To compare immunosuppressive treatments used in relapse prevention in NMO-SD. METHODS: PubMed, EMBASE and Scopus were searched for randomised controlled trials until 20th September, 2024. Search terms strategy included neuromyelitis optica, antibody and relapse. Randomised controlled trials testing immunotherapies used in relapse reduction in NMO-SD were included. Of 550 studies screened, 8 clinical trials initially met inclusion criteria. The study was performed according to PRISMA guidelines by multiple observers. Bayesian fixed-effect network meta-analysis was conducted. The primary outcome was time to relapse. The secondary outcome was annualised relapse rate. Sensitivity analysis was undertaken in seropositive patients. Treatments were ranked using a probability measure called surface under the cumulative rank curve (SUCRA). RESULTS: Eight studies were included that contained a total 851 patients [716 (84%) seropositive]. There were six treatment interventions-ravulizumab, eculizumab, tocilizumab, rituximab, inebilizumab, satralizumab and the control arm (placebo/azathioprine). Ravulizumab was the ideal treatment (HR 0.00 (95%CrI 0.00-0.03), SUCRA 0.99) with a 98% probability of being the superior treatment in increasing time to relapse in NMO-SD. This was supported by secondary analysis of annualised relapse rate and the sensitivity analysis in seropositive patients. DISCUSSION: These findings suggest that ravulizumab had the highest probability of being the most superior treatment in decreasing relapse risk in NMO-SD.