Efficacy and Safety of Rituximab Therapy in Neuromyelitis Optica Spectrum Disorders: A Systematic Review and Meta-analysis

Key Points

Question  Is rituximab an efficacious and safe therapy for patients with neuromyelitis optica spectrum disorders (NMOSDs)?

Findings  In this systematic review and meta-analysis, rituximab therapy significantly reduced the annualized relapse rate ratio and neurological disability in patients with NMOSDs. Adverse effects were recorded in 26% of patients.

Meaning  Rituximab therapy may reduce the frequency of NMOSD relapses and neurological disability of NMOSDs; however, the safety profile suggests caution in prescribing rituximab as a first-line therapy.

Importance  Neuromyelitis optica spectrum disorders (NMOSDs) are autoimmune astrocytopathies characterized by predominant involvement of the optic nerves and spinal cord. In most patients, an IgG autoantibody binding to astrocytic aquaporin 4, the principal water channel of the central nervous system, is detected. Rituximab, a chimeric monoclonal antibody specific for the CD20 B-lymphocyte surface antigen, has been increasingly adopted as a first-line off-label treatment for patients with NMOSDs.

Objective  To perform a systematic review and a meta-analysis of the efficacy and safety of rituximab use in NMOSDs, considering the potential predictive factors related to patient response to rituximab in this disease.

Evidence Review  English-language studies published between January 1, 2000, and July 31, 2015, were searched in the MEDLINE, Central Register of Controlled Trials (CENTRAL), and clinicaltrials.gov databases. Patient characteristics, outcome measures, treatment regimens, and recorded adverse effects were extracted.

Findings  Forty-six studies were included in the systematic review. Twenty-five studies that included 2 or more patients with NMOSDs treated with rituximab were included in the meta-analysis. Differences in the annualized relapse rate ratio and Expanded Disability Status Scale score before and after rituximab therapy were the main efficacy measures. Safety outcomes included the proportion of deaths, withdrawals because of toxic effects, and adverse effects.

Results  Among 46 studies involving 438 patients (381 female and 56 male [sex was not specified in 1 patient]; mean age at the outset of treatment, 32 years [age range, 2-77 years]), rituximab therapy resulted in a mean (SE) 0.79 (0.15) (95% CI, −1.08 to −0.49) reduction in the mean annualized relapse rate ratio and a mean (SE) 0.64 (0.27) (95% CI, −1.18 to −0.10) reduction in the mean Expanded Disability Status Scale score. A significant correlation was observed between disease duration and the Expanded Disability Status Scale score. Adverse effects were recorded in 114 of 438 (26%) patients treated with rituximab. Specifically, 45 patients (10.3%) experienced infusion-related adverse effects, 40 patients (9.1%) had an infection, 20 patients (4.6%) developed persistent leukopenia, 2 patients (0.5%) were diagnosed as having posterior reversible encephalopathy, and 7 patients (1.6%) died.

Conclusions and Relevance  This systematic review and meta-analysis provides evidence that rituximab therapy reduces the frequency of NMOSD relapses and neurological disability in patients with NMOSDs. However, the safety profile suggests caution in prescribing rituximab as a first-line therapy.

Neuromyelitis optica spectrum disorders (NMOSDs) are autoimmune astrocytopathies characterized by optic neuritis and transverse myelitis and, in most patients, by IgG autoantibodies binding to astrocytic aquaporin 4 (AQP4), the predominant water channel of the central nervous system.^1-4 It has been demonstrated that AQP4-IgG has pathogenic potential: the antibody binds to the extracellular domain of AQP4, activates complement that leads to complement-mediated destruction of astrocytes, induces internalization of the water channel, and mediates antibody-dependent cell cytotoxicity.^5,6 The detection of AQP4-IgG predicts relapses of myelitis and optic neuritis, with cumulative neurological disability, and justifies prompt initiation of immunosuppressive drugs.^7-10 Current treatment options are corticosteroids and immunosuppressive drugs, including (but not exclusively) azathioprine, mycophenolate mofetil, and methotrexate.⁴ These therapies may be effective because they can prevent relapses in most patients. However, this outcome often requires prolonged and even lifelong immunosuppression. Moreover, some patients have refractory disease and continue to experience frequent relapses or require high dosages of corticosteroids or other immunosuppressive drugs, with deleterious adverse effects.

Rituximab is a mouse and human chimeric IgG1 monoclonal antibody that binds to CD20 B-lymphocyte surface antigen, which is involved in B-cell activation, differentiation, and growth. Studies have shown the efficacy of rituximab in treating autoimmune diseases,^11,12 and the drug has been increasingly administered to patients with refractory or severe NMOSDs.^13,14 To date, experience with the use of rituximab in NMOSDs is still based on single-cohort studies. Moreover, recent studies^15,16 have shown that rituximab treatment may increase NMOSD relapse frequency in some patients, especially soon after the outset of treatment. To our knowledge, data are lacking on the efficacy of rituximab in AQP4-IgG–seropositive and –seronegative patients with NMOSDs and the influence of disease duration and severity on clinical response. In the present study, we performed a systematic review and a meta-analysis to evaluate the efficacy and safety of rituximab use for the treatment of NMOSDs.

Two of us (V.D. and R.L.) independently searched the MEDLINE, Central Register of Controlled Trials (CENTRAL), and clinicaltrials.gov databases (published between January 1, 2000, and July 31, 2015) using the terms neuromyelitis optica and rituximab or Devic disease and rituximab. A flowchart of the search strategy is shown in Figure 1. The search was limited to English-language studies of humans. Because no randomized clinical trial was identified, only uncontrolled observational studies were included. The studies were read in their entirety to assess the appropriateness for their inclusion in the meta-analysis.

Case reports and studies that included fewer than 2 patients were excluded from the meta-analysis. Information extracted included study design, participant characteristics, treatment regimens, and outcome measures. For each study, the following patient characteristics were retrieved, when available: mean age, proportion of women, follow-up duration, mean disease duration, annualized relapse rate (ARR) ratio, Expanded Disability Status Scale (EDSS) score before and after rituximab therapy, AQP4-IgG serostatus, rituximab regimen, mean number of rituximab reinfusions, and adverse effects, as well as the proportion of patients who at the time of rituximab treatment received immunomodulatory drugs, corticosteroids plus another immunosuppressant, plasma exchange, or intravenous immunoglobulin (IVIG).

Data were abstracted by one of us (V.D.) using a standardized data extraction form and were checked by another of us (R.L.). Any discrepant data were rereviewed, and disagreement was resolved by discussion and consensus.

In this systematic review and meta-analysis, 2 primary efficacy outcome measures were assessed, namely, differences in the ARR ratio and the mean EDSS score before and after rituximab therapy. Safety outcomes included the proportion of deaths, withdrawals because of toxic effects, and adverse effects.

The efficacy outcome measures were pooled using the method of inverse variance, with random effects on the logit-transformed proportions. The combined estimates are reported with 95% CIs. The I² test was used to assess the presence of between-study heterogeneity. Representative forest plots showing the ratios of the individual studies and the combined effect were generated to provide an overview of the results.

A meta-regression with random effects was performed to assess the influence of covariables on the ARR ratio and EDSS score. These included the mean disease duration, AQP4-IgG serostatus (frequency of AQP4-IgG–seropositive patients in each study), mean number of rituximab reinfusions, proportion of patients receiving immunomodulatory drugs (glatiramer acetate or interferon beta), immunosuppressive drugs, plasma exchange or IVIG, and the different rituximab regimens (375 mg/m² weekly for 4 weeks or 1 g weekly for 2 weeks). The meta-analysis was performed using a software program (Comprehensive Meta-Analysis; Biostat).

Forty-six studies were identified and included in the systematic review.^13,14,16-61 The data from the 3 studies by Kim et al^30,44,58 were pooled because the authors described the same patient population at different time points. Single-case reports and articles without sufficient clinical data were excluded from the quantitative synthesis. Twenty-five studies^{13,16,17,19,22,29,31,33,34,36-38,42,43,46-48,53-55,57-61} were included in the meta-analysis.

The combined data sets of all studies included a total of 438 patients (381 female and 56 male, with sex not specified in 1 patient) treated with rituximab. The main characteristics of the patients included are summarized in the Table. The mean age of all patients at the outset of treatment was 32 years (age range, 2-77 years). The AQP4-IgG serostatus was reported for 387 patients, of whom 320 (82.7%) were AQP4-IgG seropositive. The mean disease duration at first infusion of rituximab was 50 months (range, 1.5-276 months), and the mean follow-up after rituximab therapy was 27.5 months (range, 3-272 months). In 57 patients (13%), rituximab was used as a first-line therapy, while 124 patients (28.3%) were treated with immunomodulatory drugs before rituximab, 143 patients (32.6%) were receiving immunosuppressive drugs at the time of first infusion of rituximab, and 58 patients (13.2%) had plasma exchange or IVIG before rituximab therapy. The rituximab regimen was available for 313 patients and varied among studies: 139 patients (44.4%) received 375 mg/m² weekly for 4 weeks, 156 patients (49.8%) were treated with 1 g every 2 weeks for 2 times, and 9 patients (2.9%) received 500 mg weekly for 2 weeks, while different therapeutic regimens were used in another 9 patients (2.9%).

A forest plot of the standardized mean difference in the ARR ratio before and after rituximab therapy is shown in Figure 2. The mean (SE) reduction in the mean ARR ratio after rituximab therapy was 0.79 (0.15) (95% CI, −1.09 to −0.50).

Moderate heterogeneity was detected (I² = 53%). To investigate the reasons for such heterogeneity and to evaluate the effect of the different covariates on the ARR ratio reduction, a meta-regression was performed. No significant correlation was detected between the outcome (ARR ratio change) and the following variables: mean number of rituximab reinfusions (P = .96; 95% CI, −0.40 to 0.42), immunomodulatory drugs before rituximab (P = .23; 95% CI, −0.40 to 0.42), IVIG (P = .42; 95% CI, −9.08 to 3.80), plasma exchange (P = .69; 95% CI, −4.96 to 7.52), different rituximab regimens (375 mg/m² weekly for 4 weeks [P = .30; 95% CI, −3.87 to 1.20] and 1 g every 2 weeks for 2 times [P = .68; 95% CI, −4.97 to 7.52]), disease duration (P = .71; 95% CI, −0.01 to 0.02), and AQP4-IgG serostatus (P = .40; 95% CI, −7.97 to 3.15).

The EDSS score was reported in 18 studies^{13,16,17,19,29,31,33,34,36-38,42-44,46,47,57,60} included in the meta-analysis (Figure 3). Rituximab treatment resulted in a mean (SE) reduction in the mean EDSS score by 0.64 (0.27) (95% CI, −1.18 to −0.10). Substantial between-study heterogeneity was detected by the I² test (I² = 62%). To explore the reasons for this heterogeneity, a meta-regression was performed. A significant correlation was observed between disease duration and the EDSS score (P = .04; 95% CI, −0.02 to 0.10). No significant correlation was detected between the standardized mean difference of the EDSS score and the following variables: mean number of rituximab reinfusions (P = .67; 95% CI, −1.24 to 1.90), therapy with immunomodulatory drugs before rituximab (P = .59; 95% CI, −5.60 to 21.60), IVIG (P = .73; 95% CI, −18.50 to 12.59), plasma exchange (P = .76; 95% CI, −13.40 to 19.40), different rituximab regimens (375 mg/m² weekly for 4 weeks [P = .64; 95% CI, −7.48 to 4.07] and 1 g every 2 weeks for 2 times [P = .56; 95% CI, −8.10 to 4.07]), and AQP4-IgG serostatus (P = .27; 95% CI, −29.90 to 16.03).

Adverse effects were recorded in 114 of 438 (26%) patients treated with rituximab. Specifically, 45 patients (10.3%) experienced infusion-related adverse effects, 40 patients (9.1%) had an infection, 20 patients (4.6%) developed persistent leukopenia, 2 patients (0.5%) were diagnosed as having posterior reversible encephalopathy, and 7 patients (1.6%) died. None of the patients developed progressive multifocal leukoencephalopathy.

In patients with NMOSDs, disability is attack related because each disease relapse causes an accumulation of disability.⁶² Within 5 years of the disease onset, half of the individuals diagnosed as having neuromyelitis optica require the use of a wheelchair or become functionally blind.⁴ For this reason, the main goal of NMOSD therapy is to prevent disease relapses. This systematic review and meta-analysis provides sufficient data to support the efficacy of rituximab therapy in reducing relapse rates and disability in patients with NMOSDs.

Rituximab was originally approved for the treatment of B-cell lymphoma in adults, but it has been increasingly used in autoimmune diseases in which B cells are considered to have a prominent role, such as in systemic lupus erythematosus, rheumatoid arthritis, immune thrombocytopenic purpura, and myasthenia gravis.^11,12 In NMOSDs, the pathogenic role of AQP4-IgG, as demonstrated in in vitro and in vivo studies,^4,5 justifies therapies targeting antibody-producing B cells. Rituximab primarily acts by depleting plasma cell precursors because the expression of CD20—the rituximab target antigen—is restricted to the late pre–B-cell stage, and it is maintained until the B lymphocytes differentiate into antibody-producing plasma cells, when the expression is usually downregulated. On average, B-cell depletion after rituximab therapy lasts for 12 months in the peripheral blood before the generation of a new B-lymphocyte population. It has been demonstrated that rituximab therapy does not alter the frequencies of autoreactive and polyreactive B cells; therefore, it does not reset the defective early B-cell tolerance checkpoints.⁶³ This finding may explain the occurrence of NMOSD relapses after rituximab therapy in some patients and may justify rituximab reinfusions during follow-up to avoid the reexpansion of autoreactive B cells and reduce the risk of NMOSD relapses. The meta-regression analysis performed in this study showed that the number of rituximab reinfusions does not affect the ARR ratios and EDSS scores in patients with NMOSDs. However, how to monitor the biological effects of rituximab therapy to decide when treatment should be repeated is a matter of debate. Kim et al⁴⁴ proposed a bimonthly assessment of CD19/CD27-positive memory B-cell frequency, which needs validation in further studies. Rituximab has been increasingly used as a first-line therapy in NMOSDs.

A significant correlation was observed herein between disease duration of patients with NMOSDs and the EDSS score after rituximab therapy, suggesting that early treatment may reduce disability. However, treatment adverse effects were observed in 26% (114 of 438) of patients. These adverse effects were minor in most cases, and 9.1% (40 of 438) of patients had an infection. While receiving rituximab treatment, 7 patients died. However, patient death may reflect the natural history of NMOSDs because the disease has been associated with a mortality rate of up to 12%.⁶⁴ These data suggest that, until the results of controlled trials become available, the risk-benefit ratio of rituximab treatment should be carefully assessed in individual patients with NMOSDs.

A limitation of this systematic review and meta-analysis is the inclusion of observational studies with high heterogeneity. However, when we performed a meta-regression, only disease duration showed a significant correlation with the efficacy measures, suggesting that the observed heterogeneity may be mainly due to the variability of sample sizes (ie, the number of patients enrolled) in the studies included in the meta-analysis.

In summary, this systematic review and meta-analysis provides evidence that rituximab therapy reduces the frequency of disease relapses and neurological disability in patients with NMOSDs. It also suggests caution in prescribing rituximab as a first-line therapy until randomized trials determine the safety of the drug in this patient population.

Corresponding Author: Raffaele Iorio, MD, PhD, Institute of Neurology, Department of Neuroscience, Fondazione “A. Gemelli,” Catholic University, Largo Gemelli, 8, 00168 Rome, Italy (raffaele.iorio@policlinicogemelli.it).

Accepted for Publication: April 13, 2016.

Published Online: September 26, 2016. doi:10.1001/jamaneurol.2016.1637

Author Contributions: Dr Iorio had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Damato, Iorio.

Acquisition, analysis, or interpretation of data: Damato, Iorio.

Drafting of the manuscript: Iorio.

Critical revision of the manuscript for important intellectual content: Evoli.

Obtained funding: Evoli.

Administrative, technical, or material support: Evoli.

Study supervision: Iorio.

Conflict of Interest Disclosures: Dr Evoli reported serving on a scientific advisory board for UCB Biosciences GmbH. No other disclosures were reported.

Funding/Support: This work was supported by institutional funds from Catholic University, Rome, Italy.

Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.