ABSTRACT
Background
This study aimed to compare the real‐world efficacy and safety of rituximab and inebilizumab in patients with aquaporin‐4 immunoglobulin G (AQP4‐IgG) seropositive neuromyelitis optica spectrum disorder (NMOSD).
Methods
This retrospective study included patients treated with rituximab or inebilizumab at two tertiary hospitals in China between January 2015 and June 2025. Propensity score matching was conducted to reduce between‐group imbalance. The primary endpoint was time to first confirmed relapse. Secondary outcomes included changes in annualized relapse frequency, EDSS scores, and serum IgG and AQP4‐IgG levels. Safety profiles were also assessed.
Results
A total of 276 patients were analyzed (rituximab: 211; inebilizumab: 65), yielding 61 well‐balanced pairs after propensity score matching. In terms of efficacy, clinical outcomes and AQP4‐IgG dynamics were comparable between groups in pre‐ and post‐matching analyses, although inebilizumab exhibited a significantly greater reduction in serum IgG levels at 6 months. Safety profiles differed. Rituximab was associated with a higher incidence of overall adverse events, driven primarily by infusion‐related reactions, whereas other adverse events remained comparable.
Conclusions
In our real‐world study of 276 patients, rituximab and inebilizumab demonstrated comparable efficacy in the medium term but differed in their safety profiles, with a significantly higher incidence of infusion‐related reactions observed in the rituximab group.
Keywords: B‐cell depletion therapies, immunosuppression, inebilizumab, neuromyelitis optica spectrum disorder, rituximab
In our real‐world study of 276 patients, rituximab and inebilizumab demonstrated comparable efficacy in the medium term but differed in their safety profiles, with a significantly higher incidence of infusion‐related reactions observed in the rituximab group.
1. Introduction
Neuromyelitis optica spectrum disorder (NMOSD) is a rare yet highly relapsing autoimmune astrocytopathy of the central nervous system (CNS), mainly characterized by optic neuritis, longitudinally extensive transverse myelitis, and area postrema syndrome [1]. Unlike multiple sclerosis, NMOSD is notably prevalent in Asian populations [2]. Patients frequently suffer moderate‐to‐severe neurological sequelae following acute relapses, compromising quality of life and imposing substantial socioeconomic burdens [3, 4].
The discovery of aquaporin‐4 immunoglobulin G (AQP4‐IgG) underscores the pivotal role of B lymphocytes in NMOSD pathogenesis, likely involving mechanisms such as autoantibody production, proinflammatory cytokines secretion and efficient antigen presentation [5, 6]. In addition to interleukin‐6 receptor‐targeting therapies (e.g., tocilizumab and satralizumab) and complement inhibitors (e.g., eculizumab and ravulizumab) [7, 8, 9, 10], B‐cell depletion therapy is a widely used treatment option with extensive clinical experience, dating back to the empirical use of rituximab in NMOSD [11, 12]. Among them, two monoclonal antibodies (mAbs) are the most broadly utilized. Rituximab, a chimeric anti‐CD20 mAb, depletes pre‐B, mature, and memory B cells [13], and has demonstrated favorable efficacy in relapse prevention across diverse regimens [14, 15, 16, 17]. Inebilizumab, the first humanized anti‐CD19 mAb approved by the U.S. Food and Drug Administration (FDA) for seropositive NMOSD, targets a broader spectrum including early and late pro‐B cells, pre‐B cells, mature B cells, memory B cells, plasmablasts, and some plasma cells [18, 19, 20]. In March 2022, inebilizumab was approved by the National Medical Products Administration of China for the treatment of AQP4‐IgG seropositive NMOSD. Following its inclusion in the National Reimbursement Drug List in January 2023, it has been routinely prescribed in China.
Since AQP4‐IgG seropositive patients require lifelong immunosuppressive therapy, optimizing both affordability and treatment convenience is crucial—particularly in China, where healthcare resources are relatively constrained [21]. Although expanded insurance coverage has improved inebilizumab's accessibility, rituximab remains a pragmatic choice for patients with limited reimbursement capabilities due to its lower cost. Yet, this economic benefit contrasts with its more intensive monitoring requirements, whereas inebilizumab offers a streamlined fixed dosing regimen. Given these significant practical disparities, clarifying their comparative efficacy and safety is the fundamental prerequisite for balancing these trade‐offs. Only by establishing whether these agents deliver equivalent therapeutic outcomes can clinicians and patients make valid, value‐based decisions tailored to each patient's individual reality.
Currently, direct comparative evidence remains limited. In the absence of prospective head‐to‐head clinical trials, neurologists often rely on clinical experience rather than high‐level evidence. Furthermore, data derived from randomized controlled trials (RCTs) require validation in diverse, real‐world clinical settings. To bridge this knowledge gap, we conducted a retrospective comparative study evaluating the efficacy and safety of rituximab versus inebilizumab in AQP4‐IgG seropositive NMOSD patients, aiming to provide evidence‐based references for clinical decision‐making.
2. Methods
2.1. Study Population
This dual‐center retrospective cohort study enrolled NMOSD patients receiving B‐cell depletion therapies between January 2015 and June 2025 at Beijing Tiantan Hospital, Capital Medical University, and Tianjin Medical University General Hospital. Inclusion criteria: (1) diagnosis was based on the International Panel for Neuromyelitis Optica Diagnosis (IPND) 2015 revised diagnostic criteria [22]; (2) AQP4‐IgG seropositivity was confirmed using validated clinical assays; (3) receipt of ≥ 1 dose of either rituximab or inebilizumab as monotherapy throughout follow‐up; (4) available follow‐up information. Exclusion criteria: (1) prior use of monoclonal antibody therapies other than the index drug; (2) concomitant use of immunosuppressants other than oral corticosteroids during the treatment period. Data collected from medical records were supplemented by telephone interviews, outpatient visits, and re‐admissions. The study was approved by the ethics committees of both participating centers.
2.2. Clinical Data Collection
Baseline demographic characteristics, including age at NMOSD onset, age at mAb initiation, and gender, were collected. Comprehensive clinical data were recorded, covering initial core symptoms (optic neuritis, longitudinally extensive transverse myelitis, and area postrema syndrome), concomitant autoimmune disease, and pre‐treatment relapse activity. Prior immunotherapy history before mAb initiation was documented in detail, specifically including mycophenolate mofetil, cyclosporine A, azathioprine, intravenous immunoglobulin, and plasma exchange/immunoadsorption. Additionally, we documented the use of concomitant oral corticosteroids during mAb maintenance therapy. Longitudinal data collected included annualized relapse frequency (ARF) before and after treatment, EDSS scores, adverse events, and serological profiles (AQP4‐IgG and IgG levels). Regarding serological assessment, while seropositivity was confirmed via cell‐based assays (CBA) or enzyme‐linked immunosorbent assays (ELISA), quantitative titer analysis was strictly restricted to CBA measurements to ensure consistency. Strict time windows were applied: baseline values were defined as measurements obtained within 14 days prior to initiation, and follow‐up values at 6 months (±14 days) and 12 months (±14 days) post initiation.
2.3. Regimen of B Cell Depletion Therapy
Inebilizumab was administered as 300 mg by intravenous infusion on day 0 and day 14, followed by 300 mg every 6 months in maintenance phase. Rituximab followed Chinese guideline‐compliant regimens: the Beijing cohort received 500 mg rituximab infusions 14 days apart, followed by 500 mg every 6 months as subsequent treatment; the Tianjin cohort received 100 mg intravenous daily for 3 consecutive days as induction treatment, with re‐dosing triggered by CD19+ B cell levels exceeding 1%. To mitigate infusion‐related reactions (IRRs), standardized premedication protocols—including antipyretics, corticosteroids, and antihistamines—were consistently administered across both centers and treatment groups to all patients prior to each mAb infusion. Continuous real‐time vital signs monitoring during mAb infusion for all patients to ensure safety.
2.4. Clinical Outcome Assessment
The primary endpoint was time to first confirmed relapse. A relapse was defined as the onset of new or worsening neurological symptoms persisting for > 24 h in the absence of fever or infection, confirmed by new or enlarging T2‐weighted magnetic resonance imaging (MRI) lesions corresponding to clinical manifestations. However, disease exacerbations occurring within 2 months of mAb initiation were not considered confirmed relapses, as these may represent inflammatory reactions secondary to cytokine release of insufficient efficacy during the treatment latency period [23]. Time‐to‐event was defined as the interval from treatment initiation to the first confirmed relapse. For those who remained relapse‐free, data were censored at the last follow‐up. Secondary outcomes included changes in ARF, EDSS scores, serum IgG, and AQP4‐IgG titers. For statistical analysis, AQP4‐IgG titers were log‐transformed using lg (X + 1) to normalize distribution. Safety assessments classified AEs into infusion‐related reactions (IRRs) and other AEs. IRRs were defined as symptoms occurring during or within 2 h post‐infusion [24]. Other AEs included post‐discharge events such as infections and malignancies adjudicated by specialists as probably or definitely drug‐related.
2.5. Statistical Analyses
Statistical analyses utilized SPSS 25.0 software (International Business Machines Corporation, Chicago, IL, USA) and R (version 4.5.1; R Foundation for Statistical Computing, Vienna, Austria). Categorical data were presented as frequency and percentage and were analyzed using the chi‐squared test. Continuous data were described as median and range (minimum, maximum) and analyzed using the Mann–Whitney U test. Variables with a p‐value ≤ 0.1 in univariate were considered potential confounders. Secondary outcomes were analyzed using logistic (binary) or linear (continuous) regression. Relapse‐free survival was visualized via Kaplan–Meier curves and compared using the Log‐rank test, whereas multivariable Cox proportional hazards models estimated hazard ratios (HRs) and 95% confidence intervals (CIs).
To balance between‐group differences, propensity score matching (PSM) was performed using the R package “MatchIt”. Multivariable logistic regression was used to calculate propensity score. Patients were matched 1:1 using nearest‐neighbor matching with a caliper of 0.1 without replacement. Covariate balance was assessed via standardized mean differences (SMDs), with an absolute SMD < 0.2 considered indicative of acceptable balance [25]. Post‐matching survival was evaluated using Kaplan–Meier analysis (Log‐rank test) and univariate Cox proportional hazards regression. Other outcomes were compared between the matched groups using appropriate statistical tests.
Visualizations were generated using GraphPad Prism 10.0 (GraphPad Software, La Jolla, CA). A two‐sided p‐value of < 0.05 was considered as statistically significant.
3. Results
3.1. Demographic Characteristics Before PSM
A total of 211 patients in the rituximab group and 65 patients in the inebilizumab group were enrolled in this study. Demographic characteristics before PSM were presented in Table 1. Patients in the inebilizumab group were older at NMOSD onset (p = 0.016) and treatment initiation (p = 0.006). In terms of gender distribution, female predominance (rituximab vs. inebilizumab: 88.6% vs. 90.8%, p = 0.627) was observed in both groups with no apparent differences. While 59 rituximab‐treated patients and 14 inebilizumab‐treated patients had coexisting immune diseases, no significant inter‐group difference was observed (p = 0.305). Comparable proportions of patients in the rituximab (51.7%) and inebilizumab (52.3%) groups had a history of prior immunotherapy use (p = 0.927). In the total cohort, plasma exchange/immunoadsorption (32.6%) and mycophenolate mofetil (16.7%) were the most frequently utilized modalities. The proportion of patients receiving maintenance oral steroids was comparable between the two groups (rituximab vs. inebilizumab: 67.3% vs. 67.7%, p = 0.953). Compared to the inebilizumab group, patients in the rituximab group demonstrated both a higher number of pre‐treatment relapses (rituximab vs. inebilizumab: 2 vs. 1, p = 0.001) and more severe disability measured by EDSS scores (rituximab vs. inebilizumab: 3.5 vs. 3, p = 0.041) although with similar initial core symptoms (optic neuritis: p = 0.682; longitudinally extensive transverse myelitis: p = 0.231; area postrema syndrome: p = 0.145) and immunoglobulin G level (p = 0.136) inter groups. For ARF before mAb started, no significant statistical difference was found between two groups (p = 0.053).
TABLE 1.
Baseline demographic and clinical characteristics of the unmatched total population.
| Characteristics | Total (n = 276) | RTX (n = 211) | INE (n = 65) | p |
|---|---|---|---|---|
| Follow‐up duration, months, median (range) | 14 (6, 142) | 19 (6, 142) | 10 (6, 26) | NA |
| Age, years, median (range) | ||||
| Of onset | 41 (14, 82) | 39 (14, 72) | 46 (16, 82) | 0.016 |
| At mAb initiation | 45 (15, 82) | 44 (15, 74) | 52 (23, 82) | 0.006 |
| Gender (female), n (%) | 246 (89.1) | 187 (88.6) | 59 (90.8) | 0.627 |
| Concomitant autoimmune diseases, n (%) | 73 (26.4) | 59 (28.0) | 14 (21.5) | 0.305 |
| Immunotherapies before mAb initiation, n (%) | 143 (51.8) | 109 (51.7) | 34 (52.3) | 0.927 |
| Maintenance oral steroids after mAb onset, n (%) | 186 (67.4) | 142 (67.3) | 44 (67.7) | 0.953 |
| Relapses before mAb initiation, times, median (range) | 2 (1, 15) | 2 (1, 15) | 1 (1, 9) | 0.001 |
| Initial core symptoms, n (%) | ||||
| ON | 125 (45.3) | 97 (46.0) | 28 (43.1) | 0.682 |
| LETM | 152 (55.1) | 112 (53.1) | 40 (61.5) | 0.231 |
| APS | 46 (16.7) | 39 (18.5) | 7 (10.8) | 0.145 |
| EDSS score before mAb initiation, median (range) | 3 (0, 9) | 3.5 (0, 9) | 3 (0, 9) | 0.041 |
| ARF before mAb initiation, times per year, median (range) | 1.56 (0.08, 10) | 1.42 (0.08, 10) | 3.33 (0.12, 10) | 0.053 |
| Immunoglobulin G level before mAb initiation, g/L, median (range) | 10.6 (3, 49) | 10.7 (4, 49) | 9.85 (3, 26) | 0.136 |
Note: Bold values indicate statistical significance (p < 0.05).
Abbreviations: APS, area postrema syndrome; ARF, annualized relapse frequency; EDSS, Expanded Disability Status Scale; INE, inebilizumab; LETM, longitudinally extensive transverse myelitis; mAb, monoclonal antibody; NA, not applicable; ON, optic neuritis; RTX, rituximab.
3.2. Rituximab vs. Inebilizumab: Efficacy
In the total cohort, Kaplan–Meier survival analysis revealed no statistically significant difference in the time to confirmed relapse between the rituximab and the inebilizumab group (Log‐rank p = 0.334) (Figure 1). Consistent with the unadjusted analysis, multivariable Cox regression analysis—adjusted for potential confounders including age and EDSS score at initiation, pre‐treatment relapse activity, and prior immunotherapy history—showed no significant association between the treatment group and relapse risk (HR = 0.59, 95% CI [0.2, 1.7], adjusted p = 0.333) (Table 2). To address potential protocol heterogeneity regarding the two rituximab regimens, we also performed a center‐specific sensitivity analysis. The results were consistent with the primary analysis of the total cohort, confirming no statistically significant difference in relapse risk between the two therapies (Table S2). For ARF, while both groups showed reductions, the intergroup difference in ARF decline was not statistically significant (coefficient = 0.08, 95% CI [−0.15, 0.31], adjusted p = 0.483) after adjusting for factors mentioned above (Figure 2, Table 3). Similarly, regarding recovery, neither univariate nor multivariate linear regression analysis demonstrated statistically significant between‐group differences in EDSS score improvement at 6 months (adjusted p = 0.330) or 1 year (adjusted p = 0.573) post‐treatment (Figure 2, Table 3). Consistent with the observed clinical efficacy, the reduction in pathogenic antibody titers was comparable between the two arms. There were no statistically significant inter‐group differences in the AQP4‐IgG change at either 6 months or 1 year (adjusted p = 0.086 and 0.767 respectively) (Table 3).
FIGURE 1.
Kaplan–Meier analysis of time to first relapse in the unmatched total population. No significant difference was observed between the two groups (Log‐rank p = 0.334). The plot displays the follow‐up period up to 60 months to optimize the visualization of the concurrent data.
TABLE 2.
Univariate and multivariate Cox proportional hazards analyses of factors associated with the first confirmed relapse.
| Factors | Univariate p | Univariate HR (95% CI) | Multivariate p | Multivariate HR (95% CI) |
|---|---|---|---|---|
| Age at mAb initiation | 0.280 | 1.01 (0.99, 1.03) | 0.213 | 1.01 (0.99, 1.03) |
| Immunotherapies before mAb initiation | 0.849 | 1.06 (0.61, 1.83) | 0.830 | 0.94 (0.53, 1.67) |
| Number of relapses before mAb initiation | 0.033 | 1.10 (1.01, 1.21) | 0.026 | 1.11 (1.01,1.22) |
| EDSS score before mAb initiation | 0.397 | 1.06 (0.92, 1.22) | 0.547 | 1.05 (0.9, 1.22) |
| Type of mAb (INE vs. RTX) | 0.341 | 0.60 (0.21, 1.72) | 0.333 | 0.59 (0.2, 1.7) |
Abbreviations: CI, confidence interval; EDSS, Expanded Disability Status Scale; HR, hazard ratio; INE, inebilizumab; mAb, monoclonal antibody; RTX, rituximab.
FIGURE 2.
Comparison of clinical efficacy and safety outcomes between the rituximab and inebilizumab groups. (A) ARF after treatment initiation: The difference between groups was not statistically significant. (B) Decline in serum IgG levels at 6 months: The inebilizumab group showed a significantly greater reduction compared to the rituximab group (adjusted p = 0.001). (C) Improvement in EDSS scores at 6 and 12 months: No significant differences were observed between groups at either time point. (D) Frequency of AEs: Infusion‐related reactions were significantly more frequent in the rituximab group (adjusted p = 0.008), whereas other AEs were comparable (adjusted p = 0.370). Abbreviations: MAb, monoclonal antibody; ARF, annualized relapse frequency; IgG, immunoglobulin G; EDSS, Expanded Disability Status Scale; AEs, adverse events; IRRs, infusion‐related reactions.
TABLE 3.
Analysis of secondary efficacy and safety outcomes in the unmatched total population.
| Factors | Total (n = 276) | RTX (n = 211) | INE (n = 65) | Univariate p | Univariate estimate (95% CI) † | Multivariate p | Multivariate estimate (95% CI) † |
|---|---|---|---|---|---|---|---|
| EDSS score improvement, median (range) | |||||||
| 6‐month a | 0.5 (−2.5, 8) | 0.5 (−2.5, 8) | 1 (0, 4) | 0.876 | 0.04 (−0.44, 0.52) | 0.330 | 0.21 (−0.21, 0.62) |
| 1‐year b | 0.5 (−3.5, 8) | 0.5 (−3.5, 8) | 0.5 (0, 6) | 0.696 | 0.13 (−0.53, 0.79) | 0.573 | 0.16 (−0.40, 0.72) |
| AQP4‐IgG level change, median (range) | |||||||
| 6‐month c | 0.48 (−1, 2) | 0 (−1, 2) | 0.5 (−0.5, 2) | 0.118 | 0.41 (−0.11, 0.94) | 0.086 | 0.5 (−0.07, 1.07) |
| 1‐year d | 0.24 (−1.04, 2) | 0 (−1.04, 2) | 0.5 (−0.5, 0.96) | 0.645 | 0.25 (−0.92, 1.42) | 0.767 | 0.19 (−1.33, 1.72) |
| ARF decline, median (range) | 1.43 (−4.22, 10) | 1.36 (−4.22, 10) | 2.86 (−3, 10) | < 0.001 | 1.78 (0.86, 2.7) | 0.483 | 0.08 (−0.15, 0.31) |
| 6‐month IgG level decline, g/L, median (range) e | 1.4 (−8.1, 22.2) | 0.84 (−8.1, 22.2) | 2.05 (−1.16, 16.22) | 0.087 | 1.78 (−0.26, 3.82) | 0.001 | 1.96 (0.80, 3.13) |
| Adverse events, n (%) | |||||||
| Overall AEs | 55 (19.9) | 48 (22.7) | 7 (10.8) | 0.039 | 0.41 (0.18, 0.96) | 0.039 | 0.4 (0.17, 0.96) |
| Infusion‐related reactions | 42 (15.2) | 41 (19.4) | 1 (1.5) | 0.007 | 0.07 (0.01, 0.48) | 0.008 | 0.06 (0.01, 0.48) |
| Other AEs | 16 (5.8) | 10 (4.7) | 6 (9.2) | 0.183 | 2.04 (0.71, 5.86) | 0.370 | 1.67 (0.55, 5.09) |
Note: Bold values indicate statistical significance (p < 0.05).
Abbreviations: AEs, adverse events; AQP4‐IgG, aquaporin‐4 immunoglobulin G; ARF, annualized relapse frequency; CI, confidence interval; EDSS, Expanded Disability Status Scale; IgG, immunoglobulin G; INE, inebilizumab; OR, odds ratio; RTX, rituximab.
Estimates for continuous variables (EDSS, AQP4‐IgG, ARF, IgG) are presented as linear regression coefficients, representing the estimated difference (unadjusted or adjusted) between groups (INE vs. RTX). Estimates for binary variables (Adverse events) are presented as odds ratios (OR) (INE vs. RTX).
208 cases with available data.
162 cases with available data.
33 cases with available data.
11 cases with available data.
77 cases with available data.
3.3. Rituximab vs. Inebilizumab: Safety
Overall AEs were more frequent in the rituximab group (rituximab vs. inebilizumab = 48 vs. 7). This disparity remained statistically significant in multivariate binary logistic regression after adjusting for age and EDSS score at mAb initiation, and immunotherapies and relapse before mAb initiation (OR = 0.40, 95% CI [0.17, 0.96], adjusted p = 0.039). However, after stratifying AEs into IRRs and other AEs, we discovered the overall higher AEs rate in the rituximab group was primarily driven by IRRs (OR = 0.06, 95% CI [0.01, 0.48], adjusted p = 0.008). 41 IRRs occurred in the rituximab group (rash, n = 17; fever, n = 7; tight throat, n = 6; itch, n = 4; flush, n = 2; chest tightness, n = 2; gastrointestinal upset, n = 1; flu‐like symptoms, n = 1; dizziness, n = 1), 32 IRRs (78%) happened specifically during the first rituximab administration. In contrast, only one patient in the inebilizumab group experienced an IRR during the initial treatment, manifesting as mild nausea (Figure 2, Table 3).
Other AEs were comparable between two groups (OR = 1.67, 95% CI [0.55, 5.09], adjusted p = 0.370). For infections, urinary tract infections were more frequent in the rituximab group (rituximab: 5 vs. inebilizumab: 1). Pneumonia cases included Pneumocystis pneumonia (rituximab: 1 vs. inebilizumab: 2) and a single case of Escherichia coli pneumonia exclusively in the inebilizumab cohort. Herpes zoster occurred in one patient from each group. Regarding laboratory safety outcomes, at 6 months after mAb initiation, the inebilizumab group exhibited a greater median IgG reduction than the rituximab group (inebilizumab 2.05 g/L vs. rituximab 0.84 g/L). Multivariate adjustment confirmed a statistically significant difference (coefficient = 1.96, 95% CI [0.8, 3.13], adjusted p = 0.001) (Figure 2, Table 3).
Regarding malignancy, one patient in the rituximab group was diagnosed with breast cancer following the fifth cycle of rituximab, occurring approximately 20 months after treatment initiation. Upon diagnosis, the patient discontinued rituximab and underwent surgical resection, adjuvant radiotherapy, and systemic chemotherapy. The patient is currently maintained on oral prednisone (7.5 mg/day). Additionally, no confirmed cases of progressive multifocal leukoencephalopathy (PML) or deaths were identified in either group.
3.4. Propensity Score Matching
Multivariable logistic regression model including age at mAb initiation, gender, EDSS score, immunotherapies and times of relapses before mAb initiation, concomitant autoimmune diseases and maintenance oral steroids after mAb initiation was used to calculate propensity score. After PSM, 61 well‐balanced pairs from both groups were selected. The absolute SMDs for all covariates were reduced to < 0.2, indicating adequate balance (Table S3). The clinical characteristics and outcomes of the matched cohort are summarized in Table 4. In the matched cohort, Kaplan–Meier survival analysis showed no statistically significant difference in relapse‐free survival between the rituximab and the inebilizumab groups (Log‐rank p = 0.788) (Figure 3). Consistent with these findings, univariate Cox regression analysis confirmed that the risk of relapse was comparable between the two therapies (HR = 0.84, 95% CI [0.25, 2.91], p = 0.788). In line with pre‐matched findings, greater IgG reduction at 6 months in the inebilizumab group (p = 0.015) and higher incidence of overall AEs (p = 0.031), particularly infusion‐related reactions (p = 0.001) were seen in the rituximab group; all other indicators remained comparable between groups (all p > 0.05) (Table 4).
TABLE 4.
Clinical characteristics and efficacy and safety outcomes in the propensity score‐matched population.
| Characteristics | Total (n = 122) | RTX (n = 61) | INE (n = 61) | p |
|---|---|---|---|---|
| Age, years, median (range) | ||||
| Of onset | 45 (15, 82) | 45 (15, 72) | 45 (16, 82) | 0.922 |
| At mAb initiation | 50 (16, 82) | 48 (16, 74) | 51 (23, 82) | 0.605 |
| Gender (female), n (%) | 113 (92.6) | 58 (95.1) | 55 (90.2) | 0.488 |
| Concomitant autoimmune diseases, n (%) | 27 (22.1) | 13 (21.3) | 14 (23) | 0.827 |
| Immunotherapies before mAb initiation, n (%) | 57 (46.7) | 26 (42.6) | 31 (50.8) | 0.364 |
| Maintenance oral steroids after mAb onset, n (%) | 83 (68) | 42 (68.9) | 41 (67.2) | 0.846 |
| Relapses before mAb initiation, times, median (range) | 2 (1, 9) | 2 (1,7) | 2 (1,9) | 0.597 |
| Initial core symptoms, n (%) | ||||
| ON | 47 (38.5) | 21 (34.4) | 26 (42.6) | 0.352 |
| LETM | 79 (64.8) | 42 (68.9) | 37 (60.7) | 0.343 |
| APS | 21 (17.2) | 14 (23) | 7 (11.5) | 0.093 |
| EDSS score, median (range) | ||||
| before mAb initiation | 3 (0, 9) | 3 (0, 9) | 3 (0, 9) | 0.864 |
| 6‐month improvement a | 0.5 (0, 4) | 0 (0, 2.5) | 1 (0, 4) | 0.103 |
| 1‐year improvement b | 0.5 (−1, 6) | 0.5 (−1, 3.5) | 0.5 (0, 6) | 0.448 |
| ARF, times per year, median (range) | ||||
| before mAb treatment | 2.68 (0.11, 10) | 2.86 (0.11, 10) | 2.5 (0.12, 10) | 0.571 |
| ARF decline, median (range) | 1.9 (3.21, 10) | 1.78 (−3.21, 10) | 2.5 (−3, 10) | 0.399 |
| AQP4‐IgG level change, median (range) | ||||
| 6‐month c | 0.49 (−0.96, 1.52) | 0.48 (−0.96, 1.04) | 0.49 (−0.51, 1.52) | 0.591 |
| 1‐year d | 0 (−1, 0.96) | 0 (−1, 0.51) | 0.26 (−0.51, 0.96) | 0.486 |
| Immunoglobulin G level, g/L, median (range) | ||||
| before mAb initiation | 9.7 (3, 49) | 9.7 (5, 49) | 9.47 (3, 26) | 0.635 |
| 6‐month decline e | 1.5 (−4.2, 11.9) | 0.4 (−4.2, 3.74) | 2 (−1.16, 11.9) | 0.015 |
| Adverse events, n (%) | ||||
| Overall AEs | 21 (17.2) | 15 (24.6) | 6 (9.8) | 0.031 |
| Infusion‐related reactions | 14 (11.5) | 13 (21.3) | 1 (1.6) | 0.001 |
| Other AEs | 10 (8.2) | 5 (8.2) | 5 (8.2) | > 0.999 |
Note: Bold values indicate statistical significance (p < 0.05).
Abbreviations: AEs, adverse events; APS, area postrema syndrome; AQP4‐IgG, aquaporin‐4 immunoglobulin G; ARF, annualized relapse frequency; EDSS, Expanded Disability Status Scale; INE, inebilizumab; LETM, longitudinally extensive transverse myelitis; mAb, monoclonal antibody; ON, optic neuritis; RTX, rituximab.
90 cases with available data.
66 cases with available data.
17 cases with available data.
9 cases with available data.
41 cases with available data.
FIGURE 3.
Kaplan–Meier analysis of time to first relapse in the propensity score‐matched cohort. No significant difference was observed between the two groups (Log‐rank p = 0.788). The plot displays the follow‐up period up to 60 months to optimize the visualization of the concurrent data.
4. Discussion
This dual‐center study utilizing propensity score matching demonstrated that rituximab and inebilizumab share comparable efficacy in the medium term in preventing relapses and reducing disability in AQP4‐IgG seropositive NMOSD patients. While unadjusted analysis suggested a higher relapse‐free rate in the inebilizumab group (rituximab: 77.7% vs. inebilizumab: 93.8%), this was attributable to the shorter follow‐up duration resulting from its later clinical availability rather than superior potency. Importantly, this interpretation is confirmed by the consistent outcomes in both multivariate Cox regression and matched cohort analyses.
Currently, there is a lack of definitive evidence from direct comparisons between B cell–depleting therapies and of clear recommendations for switching between them. In contrast to our findings, recent studies by Lou et al. [26] and Feng et al. [27] suggested potential superiority of inebilizumab over rituximab in NMOSD. Currently, explanations for these divergent efficacy observations relative to our study cannot be established. Notably, Lou et al.'s study only enrolled 33 patients with short 6‐month follow‐up, while Feng et al.'s findings are currently restricted to conference data, lacking full methodological details. Consequently, more studies, preferably prospective, head‐to‐head trials are warranted for comparing the efficacy of both therapies in relapse prevention. In terms of switching strategies, current international consensus generally recommends switching to a different mechanism of action (e.g., complement inhibitors or interleukin‐6 receptor antagonists) for confirmed treatment failure, rather than cycling within the same therapeutic class [28]. Our finding of comparable overall efficacy between rituximab and inebilizumab aligns with this rationale. However, evidence from a post hoc analysis of the N‐MOmentum trial reveals a more complex landscape of therapeutic responsiveness. This study showed that inebilizumab was effective in preventing subsequent attacks in 17 participants with prior rituximab use, including seven patients who had experienced breakthrough attacks while on rituximab [29]. It suggested that some patients might further benefit from inebilizumab treatment, probably due to a broader spectrum of B‐cell depletion than that achieved with rituximab. These findings reflect that significant heterogeneity in individual treatment responsiveness exists in real‐world practice. Therefore, future research needs to prioritize identifying these specific patient subgroups and establishing evidence‐based criteria for choosing and switching between B‐cell depletion therapies through large‐scale prospective trials.
Given that, relative to rituximab, inebilizumab targets a broader lineage of B cells‐especially subpopulations of antibody‐generating plasma cells‐it was theoretically expected that patients in the inebilizumab group would exhibit lower anti‐AQP4‐IgG levels. However, we failed to observe this superior reduction in our cohort. It is probable that a significant portion of serum AQP4‐IgG is sustained by long‐lived plasma cells residing in tissue niches, which are effectively shielded from both therapies [30, 31]. Nevertheless, our results align with existing literature suggesting that effective B‐cell depletion does not always parallel reduced AQP4‐IgG levels, nor do these titers consistently predict clinical outcomes [32, 33]. Therefore, the therapeutic success of B‐cell depletion therapies likely extends beyond serum autoantibody reduction to broader immunomodulatory effects, including the suppression of antigen presentation, pro‐inflammatory cytokine release and B‐T cell interactions [34]. The precise mechanisms remain to be fully elucidated and warrant further investigation through both basic and clinical research.
Considering the equivalent efficacy of both agents in preventing relapses and minimizing disability, economic considerations and convenience merit particular attention in therapeutic decision‐making. Since NMOSD necessitates lifelong immunosuppression, financial toxicity is a substantial concern [35], particularly for patients facing disease‐related unemployment [36, 37]. Therefore, we suggest tailoring treatment to individual financial status and insurance coverage to minimize long‐term economic burden. Regarding convenience, inebilizumab offers fixed semi‐annual dosing, eliminating the need for frequent B‐cell repopulation surveillance required by rituximab.
In terms of safety, differences were primarily driven by IRRs. Consistent with prior studies [14, 38], a high incidence of IRRs remains the primary side effect of rituximab, with mechanisms yet to be fully elucidated. Unlike humanized inebilizumab, chimeric rituximab contains substantial murine antigen‐binding sites, which likely contribute to these reactions [39]. Another possible cause is cytokine release syndrome, which is triggered by the rapid lysis of targeted B cells [40]. Although most IRRs are mild and transient, effective management is crucial for optimizing patient adherence and confidence, particularly with rituximab. Thus, rigorous pre‐medication and close monitoring are essential. For patients unable to tolerate rituximab, inebilizumab may offer a favorable alternative.
The incidence of other adverse events was comparable between groups. Infections, primarily respiratory and urinary tract, were the most frequent adverse events, reflecting the expected consequences of immunosuppression and aligning with prior studies [19, 41]. Notably, inebilizumab induced a significantly greater reduction in serum IgG at 6 months, reflecting its broader depletion of CD19+ plasmablasts and plasma cells that maintain total serum antibody levels. While this did not translate into increased infection rates in our cohort during the study period, maintaining physiological IgG levels is critical for humoral immunity. Therefore, stringent monitoring of immunoglobulin profiles remains essential, particularly for patients on long‐term inebilizumab treatment [42]. Regarding malignancy, one case of breast cancer occurred in the rituximab group. Whereas large‐scale real‐world evidence indicates an overall safety profile of rituximab, which showed no increased invasive cancer risk in a nationwide cohort of over 4000 rituximab‐treated multiple sclerosis patients [43]. The 2023 French Multiple Sclerosis Society recommendations also supported the safety of rituximab regarding the cancer risk [44]. While derived from multiple sclerosis populations, these data provide a relevant safety reference for NMOSD, suggesting this isolated breast cancer case is likely coincidental. Nevertheless, long‐term pharmacovigilance remains essential to fully characterize the malignancy risk of inebilizumab given its relatively shorter clinical experience compared to rituximab [19].
Finally, it is necessary to acknowledge the limitations of this study. First, inherent limitations of the retrospective design include potential data incompleteness and underreporting of minor adverse events, although we attempted to mitigate this via telephone follow‐ups. Second, the exclusion of AQP4‐IgG seronegative cases due to inebilizumab's indication limits the generalizability of our findings. Consequently, future research addressing the unmet therapeutic needs of this specific cohort is warranted. Third, the later availability of inebilizumab (post‐January 2023) introduced potential calendar‐time bias due to evolving standards of care (e.g., diagnostics, supportive care, and relapse management). Although a sensitivity analysis restricted to the concurrent era yielded consistent results (Table S1), the potential influence of temporal changes in clinical practice cannot be completely ruled out. Lastly, the interpretation of our findings is limited by the relatively small sample size and shorter follow‐up duration of the inebilizumab group. Therefore, expanding the cohort and extending observation periods remain key priorities for our future research.
5. Conclusion
In conclusion, our real‐world study demonstrates comparable efficacy between rituximab and inebilizumab in the medium term for AQP4‐IgG seropositive NMOSD, while safety differences were primarily driven by a higher incidence of IRRs with rituximab. Given the complexities of real‐world management, clinical decision‐making should weigh factors such as tolerability, accessibility, and convenience to optimize patient‐tailored strategies.
Author Contributions
Fu‐Dong Shi and Wei Jiang: conceptualization and design; Ying Cui, Zihao Yu, Yutong Shi, Ai Guo, Chao Zhang, Dongmei Jia: data collection; Ying Cui and Hao Zhang: data analysis; Ying Cui drafted the manuscript; Fu‐Dong Shi, Wei Jiang, Hengri Cong, Yujing Li, and Kaibin Shi contributed to the revision of the manuscript. All the authors read and approved the final manuscript.
Funding
National Science Foundation of China (grant numbers 82,171,277, 82,320,108,007 and U25A2071).
Ethics Statement
This study was approved by the Ethics Committee of Beijing Tiantan Hospital. This was an observational, dual‐center, retrospective study, and the study did not damage patients' privacy and interest through the presented data. Thus, patients were not required to provide informed consent.
Conflicts of Interest
Fu‐Dong Shi received honoraria (lectures, advisory boards, consultations) from Alexion/AstraZeneca, Novartis, Hansoh Pharmaceutical, Sinomab, and Lundbeck. He received research grants from Zai Lab, AstraZeneca, Novartis, and Biogen. Fu‐Dong Shi was a co‐founder of New Terrain from August 2018 to December 2019; Akriva from November 2023 to September 2025. The other authors declare no conflicts of interest.
Supporting information
Table S1: Sensitivity analysis of relapse risk in patients initiating treatment after January 1st, 2023 (concurrent cohort).
Table S2: Sensitivity analysis of relapse risk stratified by study center.
Table S3: Confounders before and after propensity score matching.
Acknowledgements
We thank members of the Beijing‐Tianjin Center of Neuroinflammation (BTCN) for their valuable help.
Contributor Information
Fu‐Dong Shi, Email: fshi@tmu.edu.cn.
Wei Jiang, Email: jiangwei.med@gmail.com.
Data Availability Statement
The datasets generated and analyzed during the current study are not publicly available because patients' private information was included in our dataset but are available from the corresponding author upon reasonable request.
References
- 1. Wingerchuk D. M. and Lucchinetti C. F., “Neuromyelitis Optica Spectrum Disorder,” New England Journal of Medicine 387, no. 7 (2022): 631–639, 10.1056/NEJMra1904655. [DOI] [PubMed] [Google Scholar]
- 2. Papp V., Magyari M., Aktas O., et al., “Worldwide Incidence and Prevalence of Neuromyelitis Optica: A Systematic Review,” Neurology 96, no. 2 (2021): 59–77, 10.1212/wnl.0000000000011153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Kleiter I., Gahlen A., Borisow N., et al., “Neuromyelitis Optica: Evaluation of 871 Attacks and 1,153 Treatment Courses,” Annals of Neurology 79, no. 2 (2016): 206–216, 10.1002/ana.24554. [DOI] [PubMed] [Google Scholar]
- 4. Papp V., Wandall‐Holm M., Bacher Svendsen K., et al., “Socioeconomic Burden of AQP4‐Antibody Seropositive NMOSD: A Nationwide Registry‐Based Study,” Journal of Neurology, Neurosurgery & Psychiatry 96, no. 2 (2025): 184–187, 10.1136/jnnp-2024-333790. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Papadopoulos M. C. and Verkman A. S., “Aquaporin 4 and Neuromyelitis Optica,” Lancet Neurology 11, no. 6 (2012): 535–544, 10.1016/s1474-4422(12)70133-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Lee D. S. W., Rojas O. L., and Gommerman J. L., “B Cell Depletion Therapies in Autoimmune Disease: Advances and Mechanistic Insights,” Nature Reviews Drug Discovery 20, no. 3 (2021): 179–199, 10.1038/s41573-020-00092-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Yamamura T., Kleiter I., Fujihara K., et al., “Trial of Satralizumab in Neuromyelitis Optica Spectrum Disorder,” New England Journal of Medicine 381, no. 22 (2019): 2114–2124, 10.1056/NEJMoa1901747. [DOI] [PubMed] [Google Scholar]
- 8. Pittock S. J., Berthele A., Fujihara K., et al., “Eculizumab in Aquaporin‐4‐Positive Neuromyelitis Optica Spectrum Disorder,” New England Journal of Medicine 381, no. 7 (2019): 614–625, 10.1056/NEJMoa1900866. [DOI] [PubMed] [Google Scholar]
- 9. Pittock S. J., Barnett M., Bennett J. L., et al., “Ravulizumab in Aquaporin‐4‐Positive Neuromyelitis Optica Spectrum Disorder,” Annals of Neurology 93, no. 6 (2023): 1053–1068, 10.1002/ana.26626. [DOI] [PubMed] [Google Scholar]
- 10. Zhang C., Zhang M., Qiu W., et al., “Safety and Efficacy of Tocilizumab Versus Azathioprine in Highly Relapsing Neuromyelitis Optica Spectrum Disorder (TANGO): An Open‐Label, Multicentre, Randomised, Phase 2 Trial,” Lancet Neurology 19, no. 5 (2020): 391–401, 10.1016/s1474-4422(20)30070-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Holian A. H. and Weinshenker B. G., “Emerging Role of Targeted Monoclonal Antibodies in Neuromyelitis Optica Spectrum Disorders,” BioDrugs 39, no. 4 (2025): 573–589, 10.1007/s40259-025-00729-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Cree B. A., Lamb S., Morgan K., Chen A., Waubant E., and Genain C., “An Open Label Study of the Effects of Rituximab in Neuromyelitis Optica,” Neurology 64, no. 7 (2005): 1270–1272, 10.1212/01.Wnl.0000159399.81861.D5. [DOI] [PubMed] [Google Scholar]
- 13. Pittock S. J., Zekeridou A., and Weinshenker B. G., “Hope for Patients With Neuromyelitis Optica Spectrum Disorders ‐ From Mechanisms to Trials,” Nature Reviews. Neurology 17, no. 12 (2021): 759–773, 10.1038/s41582-021-00568-8. [DOI] [PubMed] [Google Scholar]
- 14. Tahara M., Oeda T., Okada K., et al., “Safety and Efficacy of Rituximab in Neuromyelitis Optica Spectrum Disorders (RIN‐1 Study): A Multicentre, Randomised, Double‐Blind, Placebo‐Controlled Trial,” Lancet Neurology 19, no. 4 (2020): 298–306, 10.1016/s1474-4422(20)30066-1. [DOI] [PubMed] [Google Scholar]
- 15. Iordanova R., Smith J., Gonzales E., Lee A., Li B., and Langer‐Gould A., “Effectiveness of Lower Rituximab Doses in Neuromyelitis Optica Spectrum Disorder: A Population‐Based Cohort Study (S50.006),” Neurology 100, no. 17_supplement_2 (2023): 2285, 10.1212/WNL.0000000000202445. [DOI] [Google Scholar]
- 16. Zhang C., Liu R., Gao B., et al., “Efficacy and Safety of Low‐ and Ultralow‐Dose Rituximab in Neuromyelitis Optica Spectrum Disorder,” Journal of Neuroimmunology 387 (2024): 578285, 10.1016/j.jneuroim.2024.578285. [DOI] [PubMed] [Google Scholar]
- 17. Yang C. S., Yang L., Li T., et al., “Responsiveness to Reduced Dosage of Rituximab in Chinese Patients With Neuromyelitis Optica,” Neurology 81, no. 8 (2013): 710–713, 10.1212/WNL.0b013e3182a1aac7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Cree B. A. C., Bennett J. L., Kim H. J., et al., “Inebilizumab for the Treatment of Neuromyelitis Optica Spectrum Disorder (N‐MOmentum): A Double‐Blind, Randomised Placebo‐Controlled Phase 2/3 Trial,” Lancet 394, no. 10206 (2019): 1352–1363, 10.1016/s0140-6736(19)31817-3. [DOI] [PubMed] [Google Scholar]
- 19. Cree B. A. C., Kim H. J., Weinshenker B. G., et al., “Safety and Efficacy of Inebilizumab for the Treatment of Neuromyelitis Optica Spectrum Disorder: End‐Of‐Study Results From the Open‐Label Period of the N‐MOmentum Trial,” Lancet Neurology 23, no. 6 (2024): 588–602, 10.1016/s1474-4422(24)00077-2. [DOI] [PubMed] [Google Scholar]
- 20. Forsthuber T. G., Cimbora D. M., Ratchford J. N., Katz E., and Stüve O., “B Cell‐Based Therapies in CNS Autoimmunity: Differentiating CD19 and CD20 as Therapeutic Targets,” Therapeutic Advances in Neurological Disorders 11 (2018): 1756286418761697, 10.1177/1756286418761697. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Yu Y., Zhong M., Quan C., and Ma C., “Treatment Access and Satisfaction on Disease‐Modifying Therapies of Neuromyelitis Optica Spectrum Disorder Patients in China: A Cross‐Sectional Survey,” Therapeutic Advances in Neurological Disorders 17 (2024): 17562864241239105, 10.1177/17562864241239105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Wingerchuk D. M., Banwell B., Bennett J. L., et al., “International Consensus Diagnostic Criteria for Neuromyelitis Optica Spectrum Disorders,” Neurology 85, no. 2 (2015): 177–189, 10.1212/wnl.0000000000001729. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Arnold D. M., Dentali F., Crowther M. A., et al., “Systematic Review: Efficacy and Safety of Rituximab for Adults With Idiopathic Thrombocytopenic Purpura,” Annals of Internal Medicine 146, no. 1 (2007): 25–33, 10.7326/0003-4819-146-1-200701020-00006. [DOI] [PubMed] [Google Scholar]
- 24. Chung C. H., “Managing Premedications and the Risk for Reactions to Infusional Monoclonal Antibody Therapy,” Oncologist 13, no. 6 (2008): 725–732, 10.1634/theoncologist.2008-0012. [DOI] [PubMed] [Google Scholar]
- 25. Stuart E. A., “Matching Methods for Causal Inference: A Review and a Look Forward,” Statistical Science 25, no. 1 (2010): 1–21, 10.1214/09-sts313. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Lou C. Y., Wang Y., Xing J. Y., et al., “Comparison of Inebilizumab or Rituximab in Addition to Glucocorticoid Therapy for Neuromyelitis Optica Spectrum Disorders,” International Journal of Ophthalmology 17, no. 6 (2024): 1073–1078, 10.18240/ijo.2024.06.12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. EBSCO , “ECTRIMS 2025 Abstracts Free Communication and Scientific Session,” Multiple Sclerosis 31, no. 3_suppl (2025): 3–135, 10.1177/13524585251358339.39784932 [DOI] [Google Scholar]
- 28. Paul F., Marignier R., Palace J., et al., “International Delphi Consensus on the Management of AQP4‐IgG+ NMOSD: Recommendations for Eculizumab, Inebilizumab, and Satralizumab,” Neurology Neuroimmunology & Neuroinflammation 10, no. 4 (2023): e200124, 10.1212/nxi.0000000000200124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Flanagan E. P., Levy M., Katz E., et al., “Inebilizumab for Treatment of Neuromyelitis Optica Spectrum Disorder in Patients With Prior Rituximab Use From the N‐MOmentum Study,” Multiple Sclerosis and Related Disorders 57 (2022): 103352, 10.1016/j.msard.2021.103352. [DOI] [PubMed] [Google Scholar]
- 30. Chen D., Gallagher S., Monson N. L., Herbst R., and Wang Y., “Inebilizumab, a B Cell‐Depleting Anti‐CD19 Antibody for the Treatment of Autoimmune Neurological Diseases: Insights From Preclinical Studies,” Journal of Clinical Medicine 5, no. 12 (2016): 107, 10.3390/jcm5120107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Hofmann K., Clauder A. K., and Manz R. A., “Targeting B Cells and Plasma Cells in Autoimmune Diseases,” Frontiers in Immunology 9 (2018): 835, 10.3389/fimmu.2018.00835. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Kim S. H., Jeong I. H., Hyun J. W., et al., “Treatment Outcomes With Rituximab in 100 Patients With Neuromyelitis Optica: Influence of FCGR3A Polymorphisms on the Therapeutic Response to Rituximab,” JAMA Neurology 72, no. 9 (2015): 989–995, 10.1001/jamaneurol.2015.1276. [DOI] [PubMed] [Google Scholar]
- 33. Bennett J. L., Pittock S. J., Paul F., et al., “B Cell and Aquaporin‐4 Antibody Relationships With Neuromyelitis Optica Spectrum Disorder Activity,” Annals of Clinical Translational Neurology 11, no. 10 (2024): 2792–2798, 10.1002/acn3.52171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Ochi H., Nakamura S., and Nakahara J., “B Cell Depletion as a Therapeutic Strategy for Neuromyelitis Optica Spectrum Disorder: Rationale, Evidence, and Challenges,” Frontiers in Immunology 16 (2025): 1635989, 10.3389/fimmu.2025.1635989. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Hill J. B., Mastick M. L., Chiong‐Rivero H., Grebenciucova E., and Mateen F. J., “Characterizing Financial Toxicity Among People Living With Neuromyelitis Optica Spectrum Disorder,” Multiple Sclerosis 31, no. 11 (2025): 1370–1380, 10.1177/13524585251365139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Levy M., Fujihara K., and Palace J., “New Therapies for Neuromyelitis Optica Spectrum Disorder,” Lancet Neurology 20, no. 1 (2021): 60–67, 10.1016/s1474-4422(20)30392-6. [DOI] [PubMed] [Google Scholar]
- 37. Han L., Hong P., Wan Y., et al., “Neuromyelitis Optica Spectrum Disorder in Western China Impacts Employment and Increases Financial Burden in Women,” Frontiers in Neurology 13 (2022): 973163, 10.3389/fneur.2022.973163. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Novi G., Bovis F., Sciolla C., et al., “Long‐Term Effectiveness and Safety of Rituximab in Neuromyelitis Optica Spectrum Disorder: A 5‐Year Observational Study,” CNS Drugs 39, no. 7 (2025): 693–705, 10.1007/s40263-025-01191-7. [DOI] [PubMed] [Google Scholar]
- 39. Johnson P. W. and Glennie M. J., “Rituximab: Mechanisms and Applications,” British Journal of Cancer 85, no. 11 (2001): 1619–1623, 10.1054/bjoc.2001.2127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Hauser S. L., Waubant E., Arnold D. L., et al., “B‐Cell Depletion With Rituximab in Relapsing‐Remitting Multiple Sclerosis,” New England Journal of Medicine 358, no. 7 (2008): 676–688, 10.1056/NEJMoa0706383. [DOI] [PubMed] [Google Scholar]
- 41. Wang H., Zhou J., Li Y., et al., “Adverse Events of Rituximab in Neuromyelitis Optica Spectrum Disorder: A Systematic Review and Meta‐Analysis,” Therapeutic Advances in Neurological Disorders 14 (2021): 17562864211056710, 10.1177/17562864211056710. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Rensel M., Zabeti A., Mealy M. A., et al., “Long‐Term Efficacy and Safety of Inebilizumab in Neuromyelitis Optica Spectrum Disorder: Analysis of Aquaporin‐4‐Immunoglobulin G‐Seropositive Participants Taking Inebilizumab for ⩾4 Years in the N‐MOmentum Trial,” Multiple Sclerosis 28, no. 6 (2022): 925–932, 10.1177/13524585211047223. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Alping P., Askling J., Burman J., et al., “Cancer Risk for Fingolimod, Natalizumab, and Rituximab in Multiple Sclerosis Patients,” Annals of Neurology 87, no. 5 (2020): 688–699, 10.1002/ana.25701. [DOI] [PubMed] [Google Scholar]
- 44. Collongues N., Durand‐Dubief F., Lebrun‐Frenay C., et al., “Cancer and Multiple Sclerosis: 2023 Recommendations From the French Multiple Sclerosis Society,” Multiple Sclerosis 30, no. 7 (2024): 899–924, 10.1177/13524585231223880. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1: Sensitivity analysis of relapse risk in patients initiating treatment after January 1st, 2023 (concurrent cohort).
Table S2: Sensitivity analysis of relapse risk stratified by study center.
Table S3: Confounders before and after propensity score matching.
Data Availability Statement
The datasets generated and analyzed during the current study are not publicly available because patients' private information was included in our dataset but are available from the corresponding author upon reasonable request.