Efficacy and safety of different dosages of rituximab for myasthenia gravis: a single-arm meta-analysis

Abstract

Background

Rituximab (RTX) is one of the treatment options for refractory myasthenia gravis (MG), yet the optimal dosing schedule remains undetermined. Our study aims to explore this issue and offer a valuable reference for clinical dosing.

Methods

This is a single-arm meta-analysis. Studies in adults with myasthenia gravis published before 31 December 2023 were searched in PubMed, Web of Science, and other databases. Two primary effectiveness outcomes were analyzed: (1) Proportion of patients achieving minimal manifestation status (MMS) or better, (2) Change in Quantitative MG Score (QMGs) after RTX treatment. Safety outcomes included the incidence and description of serious adverse events (SAEs) and adverse events (AEs). Forest plots were generated to provide an overview and detailed combined effects. Publication bias was evaluated using funnel plots and the Egger test. Conventional dose refers to an RTX regimen similar to that used for the treatment of B-cell lymphoma: 375 mg/m² per week for 4 weeks or 1000 mg for Weeks 1 and 3. Dosing regimens below the conventional dose in a treatment cycle are defined as low dose.

Results

A total of 1037 MG patients received RTX treatment. Overall, 59.0% (95% CI: 48.2–69.8%, n = 599) of patients achieved MMS or better, with a mean decrease in QMGs of 6.81 (95% CI, -9.27 to -4.35, n = 222). The low-dose group showed a higher proportion of patients achieving MMS or better (76.6% vs 51.6%) and a more significant decrease in QMGs from baseline (-9.04 vs -3.62) compared to the conventional dose group (P < 0.01). Differences in the incidence of SAEs and AEs between the two groups were not significant (P > 0.05). Univariate meta-regression analyses showed that the dose administered was significantly associated with the proportion of MMS or better and the change in QMGs, whereas the proportion of Musk patients was not significantly associated with any of the outcomes. Stepwise logistic regression analyses showed that non-refractory MG, mild disease severity (MGFA classification), and low-dose were significant predictors for achieving an MMS or better prognosis, whereas for achieving improvement or better, only low dose was an independent predictor.

Conclusion

RTX can improve clinical symptoms, reduce QMGs in MG patients and the use of oral glucocorticoids and other immunosuppressants. The efficacy of low-dose RTX in treating MG patients is more effective than conventional-dose RTX and demonstrates a better safety profile. Mild disease severity, non-refractory MG, low dose, and MuSK-MG over AChR-MG predict better efficacy. Large randomized controlled trials are necessary to evaluate the efficacy and safety of RTX in MG patients and its various subtypes.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s40199-025-00557-y.

Introduction

Myasthenia gravis (MG) is an autoimmune disease caused by antibodies against the acetylcholine receptor (AChR), muscle-specific kinase (MuSK), or other AChR-related proteins in the muscle membrane. Localized or general muscle weakness is the predominant symptom [1]. Approximately 80% of MG patients have detectable antibodies against AChR, while a small percentage show antibodies against MuSK or lipoprotein-receptor-related protein 4 (LRP4) [2]. The management of MG focuses on restoring muscle strength and improving patients' well-being by controlling disease activity, monitoring treatment-related adverse events, and implementing individualized supportive measures. Current treatments are either symptomatic or cause nonspecific immunosuppression. Anticholinesterases are the first-line medications for MG patients, and those with symptoms that exert functional impairment or reduce quality of life despite optimal acetylcholinesterase inhibitor treatment should receive immunotherapy based on steroids and other immunosuppressants [1].

Rituximab (RTX) is a chimeric monoclonal antibody that targets CD-20, leading to the depletion of B lymphocytes through cell-mediated or complement-dependent toxicity [3]. B cell depletion using RTX is a treatment option for refractory MG and has shown significant and long-lasting benefits, particularly in patients with MUSK myasthenia gravis (MuSK-MG). It has been proposed as an early treatment option after the failure of first-line therapies [4–6]. In some countries, it is used as a second-line therapy for patients with moderate to severe MG [4]. However, there is currently no consensus on the optimal dosing schedule for RTX in MG.

Previous uncontrolled reports have demonstrated the effectiveness of rituximab in all subgroups of MG, albeit with varying response rates [6–8]. However, these studies had different dose regimens and enrollment criteria. Most of them followed a treatment protocol similar to that used in B-cell lymphoma, consisting of either 375 mg/m² per week for 4 weeks or 1000 mg at weeks 1 and 3.

The efficacy and safety of different RTX dosing regimens for the treatment of MG are inconclusive. In patients with refractory AChR myasthenia gravis (AChR-MG), no significant difference was found in the proportion of improvement in clinical status or experiencing adverse reactions between the low and conventional dose RTX groups, although adverse reactions were potentially lower in the low-dose group [9]. In patients with refractory MG, a single-arm meta-analysis conducted in 2021 indicated that efficacy appeared to be independent of RTX dose, while another meta-analysis published in 2014 suggested significantly better effectiveness in the high-dose group [10, 11]. In summary, no relevant studies have been found that compare various RTX dosing regimens across all patients with MG. Moreover, discrepancies exist in the outcomes of previous studies conducted on patients with refractory MG. Consequently, in a single-arm meta-analysis, we broadened the scope of our study population to encompass all MG patients. Additionally, we carried out subgroup analyses to assess the efficacy and safety of different RTX doses in the treatment of MG. The aim of these efforts is to offer valuable insights for the determination of the optimal RTX dose.

Methods

Study selection and data collection

This meta-analysis was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) [12]. The review protocol was registered on PROSPERO (registration number: CRD42023430601). Two reviewers (Jianchun Li and Di Chen) independently conducted a comprehensive search in multiple databases, including PubMed, Embase, Cochrane Library, China National Knowledge Infrastructure (CNKI), China Science and Technology Journal Database (VIP Data), Wanfang Data, Web of Science, Scopus, and ClinicalTrails.gov for studies investigated the effectiveness and safety of RTX in treating MG in adults. The search was limited to studies published in Chinese or English up to December 31, 2023. Human studies and clinical trials were included. Detailed search strategies for each database can be found in the Supplementary Material 1.

Inclusion and exclusion criteria

Participants

The study included adult patients diagnosed with MG.

Interventions

The study included investigations that examined the effectiveness and safety of regular RTX treatment and follow-up for MG patients.

Outcome measures

The primary effectiveness outcomes included: (1) the proportion of patients achieving minimal manifestation status (MMS) or better, as defined by international consensus, which means the absence of symptoms or functional limitations from MG patients but with weakness in some muscles upon examination [5]; (2) the change in Quantitative MG score (QMGs) after RTX therapy from baseline. Secondary outcomes included: (1) the proportion of patients achieving improvement or better; (2) the proportion of patients who discontinued oral immunosuppressants and glucocorticoids (GC); (3) the change in GC doses after RTX therapy from baseline. Safety outcomes encompassed the incidence and descriptions of serious adverse events and adverse events.

Study design

The study considered human randomized controlled trials (RCTs), observational studies, and single-arm studies published in Chinese or English. Small case series with fewer than ten patients and studies that did not report the desired outcomes were excluded. The full text of relevant studies was assessed for inclusion in the meta-analysis by two reviewers (Jianchun Li and Di Chen).

Data extraction

Information such as author names, publication year, country, sample size, patient characteristics, RTX regimens, and outcome measures was extracted for each study. Patient characteristics, including the presence of MG-related autoantibodies, proportion of females, age at initiation of RTX therapy, disease duration, and follow-up duration, were retrieved when available. For studies that reported individual data, age, sex, MFGA classification, antibody type, whether MMS or better was achieved, and whether improvement or better was achieved were extracted for individuals. Two reviewers independently extracted the data using a predetermined form, and a third reviewer (Pengfei Jin) verified the data. Any discrepancies were resolved through discussion.

Statistical analysis

The meta-analysis was conducted using R software 4.3.3. The meta function package was used to obtain summary means and 95% confidence intervals (CIs) for changes in QMGs and GC doses. Other outcomes were summarized as percentages or incidence with 95% CIs. Forest plots were generated to provide an overview and detailed combined effects. Heterogeneity between studies was assessed using the χ² test and I² statistics, and Meta regression was used to explore sources of heterogeneity. When I² is less than 50 %, a fixed effects model is selected, otherwise, a random effects model is implemented. Publication bias was evaluated using funnel plots and the Egger test. Sensitivity analyses were performed to assess the impact of each study on the pooled results. In conjunction with the results of Meta regression, subgroup analyses were conducted if sufficient subgroup data were available to evaluate outcome measures among different MG subtypes. Univariate logistic analyses and stepwise regression logistic analyses were used to explore predictive factors affecting response to RTX treatment. P < 0.05 was defined as Statistical significance.

Results

Study characteristics

A flow chart of the study selection process is presented in Fig. 1. After removing duplicates and screening titles and abstracts, 166 relevant studies were identified. Following a full-text review, 42 studies were included in the meta-analysis. Among the excluded studies, 17 were excluded because updated or original trial data were retrieved (BeatMG).

Fig. 1.

Flow chart of study selection

Of the 42 studies included, 4 were randomized controlled trials, and the rest were observational studies. A total of 1037 MG patients treated with RTX were enrolled, including 367 men and 670 women. Among these patients, 838 had clear MG antibodies, with 577 being AChR-IgG positive, 234 MuSK-IgG positive, and 93 defined as "Double Seronegative" (DN). The mean or median age at MG diagnosis ranged from 22.6 to 54 years for all patients, and the age at initiation of RTX therapy ranged from 30.6 to 59.3 years.

The dosing regimen for RTX treatment varied among the studies. A treatment regimen similar to that used for B-cell lymphoma of either 375 mg/m² per week for 4 weeks or 1000 mg at weeks 1 and 3 was defined as a conventional dose. Among the patients, 687 (66.2%) received the conventional induction dose of RTX, 285 (27.5%) received a low induction dose of RTX, and 65 (6.3%) received an unknown dose. The reperfusion regimen typically depended on the clinical presentation, circulating CD19+ B-cell repopulation, or clinical relapse (Table 1).

Table 1.

Summary on the demographic and clinical characteristics of the patients included in the meta-analysis

References	Country	Sample size	AchR	Musk	Female/Total	Age at onset, mean ± SD	Follow-up in months after RTX, mean ± SD	Age at initiation of RTX, mean ± SD	Disease duration in years before RTX, mean ± SD	RTX regimen
Fei Fei, 2020 [13]	China	28	NA	NA	13/28	NA	1	41.23 ± 10.22	6.9 ± 1.5	500 mg of RTX per week for 4 weeks
Anning Li, 2022 [14]	China	13	7	6	10/13	36.38 ± 13.07	12	47.38 ± 10.47	NA	375mg/m2 of RTX per week for 4 weeks, then 375mg/m2 of RTX per month for 2 months. Temporary RTX treatment when muscle weakness symptoms return
Jiexi Wen, 2020 [15]	China	21	NA	NA	11/21	NA	1	55.3 ± 17.4	8.5 ± 6.9	500 mg of RTX per week for 4 weeks
Sijia Zhao, 2018 [16]	China	17	10	0	13/17	42.5 ± 12.1	6	44.4 ± 12.20	38.8 ± 25.0(month)	100 mg of RTX per week for 3 weeks
Brauner, Susanna, 2020 [17]	Sweden	72	60	NA	31/72	54 ± 19	40 ± 19	60 ± 18	NA	A single infusion of 500 mg of RTX per 6months, with some patients receiving only a single infusion of 100 mg. The infusion interval was sometimes extended or discontinued in patients who demonstrated a favorable clinical response
Vadim Afanasiev, 2017 [18]	France	28	21	3	15/28	39.1 ± 14.6	27.2 ± 16.6	50.6 ± 12.0	NA	The treatment with RTX consisted of an induction treatment using either the “rheumatologic disease-like” protocol, i.e. 1000 mg on day 1 (D1) and D15, or the “lymphoma-like” protocol, i.e. 375 mg/m2 on D1, D7, D15 and D21 The induction treatment was followed by a maintenance treatment, 1000 mg or 375 mg/m2 infusion, with a 6 months periodicity. MG symptoms being the major decision criterion
Fredrik Piehl, 2022 [19]	Sweden	25	23	NA	7/25	NA	6	67.4 ± 13.4	NA	Intravenous infusion of 500 mg of RTX disolved in sodiumchloride to 2 mg RTX/mL
Tess Litchman, 2020 [20]	America	33	17	16	24/33	35.9 ± 15.6	1861.5  ± 953.4, d	NA	NA	375mg/m2 of RTX per week for 4 weeks. The interval between cycles is 6 months. 2–4 cycles
Kimberly R, 2017 [8]	America	16	16	NA	10/16	NA	> 12m	40.5 ± 16.79	NA	375mg/m2 of RTX per week for 4 weeks. The interval between cycles is 6 months. 2–4 cycles
Christopher Nelke, 2022 [21]	German	57	NA	NA	35/57	40.8 ± 18.2	NA	46.5 ± 17.10	NA	A dose of 1000 mg was given 14 days apart. The treatment interval for maintenance therapy with 1000 mg was 6–9 months, depending on clinical response
Sisi Jing, 2019 [22]	China	33	31	1	25/33	NA	6M	39.79 ± 14.26	NA	A total dose of 600 mg: 100 mg on the first day and 500 mg on the second day
Mehmet Fatih GÖL, 2022 [23]	Turkey	19	10	6	9/19	NA	NA	NA	NA	1000 mg RTX on days 1 and 15, and 1000 mg RTX per 6 months depending on clinical status
A. Dos Santos, 2020 [24]	France	29	20	5	17/29	41.3 ± 22.1	20.06 (0.17–68.93) Mean (range)	49.6 ± 16.3	8.8 (0.41–33) Mean (range)	Protocol A: two infusions of 1g separated by 2 weeks, followed by 1 g infusion every 6 months; Protocol B: two infusions of 1g separated by 2 weeks and an infusion at 6 months. New infusion was realized only in case of relapse, defined by reappearance of myasthenic symptoms limiting daily activity; Protocol C: 375 mg/m2 every week for 1 month and new infusion only if patient relapsed; and Protocol D: infusion of 1 g every 2 months for 1 year and then 1 g every 6 months
Océane Landon-Cardinal, 2018 [25]	France	11	NA	NA	8/11	NA	> 18M	44 (24–61) Median (range)	13 (3–32) Median (range)	1 g of RTX at day 0, day 14, and 6-month follow-up (M6)
Chao Zhang, 2020 [26]	China	49	43	4	27/49	NA	3.1 (2.2‐3.9) Median (IQR)	59.3 (44.3‐69.5) Median (IQR)	NA	NA
Elena Cortés-Vicente, 2018 [27]	Spain	25	NA	25	24/25	NA	NA	NA	NA	①(4 + 2) 375 mg/m2 every week for four consecutive weeks and then monthly for the next 2 months; ②(1 + 1) two 1 g doses separated by 2 weeks; ③375 mg/m2 every week for four consecutive weeks. RTX re-infusions were administered only if patients relapsed. A relapse was defined as the reappearance of myasthenic symptoms that limited daily activity
Ying Du, 2022 [28]	China	13	13	NA	7/13	NA	25 ± 11.12	58.92 ± 10.01	5.86(month)	Individualized low-dose RTX monotherapy protocol consisted of an induction with intravenous administration of 100 mg RTX once a week no more than three cycles, followed by reinfusions (100 mg once) every 3 months, sequentially according to whether achieving “MMS or better status” (primary endpoint) and peripheral CD19 + B-cell repopulation ≥ 1% of total lymphocytes in each visit
Díaz-Manera, J, 2012 [29]	Spain	17	11	6	15/17	33.82 ± 15.97	31.12 ± 20.05	44.35 ± 15.58	NA	RTX was administered at the standard dose of 375 mg/m2 every week for 4 consecutive weeks and then monthly for the next 2 months. Repeat RTX infusions were administered only when myasthenic symptoms reappeared and interfered with daily life activities
Fiona Chan, 2019 [30]	Australia	38	28	6	26/38	42.74 ± 17.10	54.90 ± 30.26	NA	NA	Most (80%) patients received 375 mg/m2 of RTX per week for 4 weeks, with lower doses being infrequently reported
Castiglione, Juan I., 2022 [31]	Argentina	16	8	8	13/16	45.5 ± 16.2	39.1 ± 25.1	NA	NA	① 3 weekly cycles dose = 375 mg/m2 body surface area (BSA), ② 2 cycles (dose = 375 mg/m2 BSA) separated by a 2-week interval, ③ 2 cycles (500 mg/m2 BSA) also separated by 2 weeks The number of RTX maintenance treatment cycles was not dictated by B-cell counts but rather based on clinical improvement
Jun Lu, 2019 [32]	China	12	12	NA	10/12	NA	NA	30.6 ± 29.6	NA	600 mg RTX intravenously every 6 months with three infusions at 0, 6, and 12 months. Once RTX was initiated, all 12 patients ceased other immunosuppressive agents, except for prednisolone
Huahua Zhong, 2020 [33]	China	12	12	NA	10/12	NA	NA	30.75 ± 13.51	54.83 ± 40.83(month)	After baseline evaluation, a single 600 mg RTX infusion was administered to each patient. Once RTX was initiated, all 12 patients ceased other immunosuppressive agents, except for prednisolone
Huining Li, 2021 [34]	China	19	19	NA	12/19	47.08 ± 15.54	51.55 ± 11.32	47.1	NA	The initial doses of RTX were administered to the patients according to the baseline CD19 + B cell counts. The percentage of CD19 + B cells was monitored at 12-week intervals. The indicated dose reinfusion regimen was based on CD19 + B cells repopulation, which was defined as the circulating CD19 + B cell proportion > 0.5% of PBMCs CD19 + B cells % in PBMCs: Dosage of rituximab 0–0.5: NA 0.5–1.5: 100mg,i.v.,QD 1.5–5.0: 100 mg,i.v.,QD, 2 consecutive days 5.0–15.0: 100 mg,i.v.,QD, 3 consecutive days  > 15.0: 100 mg, i.v.,QD,4 consecutive days
Richard J. Nowak, 2022 [35]	America	25	25	NA	11/25	46.6 ± 18.7	12	53.2 ± 17.5	NA	The treatment group received a 2-cycle RTX regimen separated by 6 months. Each cycle was defined as 1 weekly infusion (375 mg/m2) for 4 consecutive weeks. Cycle 1 was administered in weeks 0–3 and cycle 2 was given in weeks 24–27
Kyomin Choi, 2019 [36]	Korea	17	9	6	11/17	NA	24.47 ± 11.29	50.53 ± 15.56	10.18 ± 7.36	375 mg/m2 twice with a 2-week interval followed by additional single infusions (375 mg/m2 once) as indicated.The retreatment with low-dose RTX was based on either circulating B-cell repopulation or clinical relapse
Richard J. Nowak, 2011 [37]	America	14	8	6	11/14	NA	> 12m	43.43 ± 15.95	NA	RTX was given at a standard dose of 375 mg/m2. Each cycle is defined as one infusion per week for four consecutive weeks. Interval between cycles was set at 6 months
Christopher T. Doughty, 2021 [38]	America	40	28	9	18/40	50.0 ± 19.9	45.1 ± 40.7	55.5 ± 18.1	2.7 IQR, 1.5–7.7	1000 mg × 2 (> 80%) OR 375mg/m2 × 4
Bentolhoda Ziaadini, 2022 [39]	Iran	59	28	24	41/59	29.20 ± 4.69	12	40.31 ± 13.53	6.83 ± 5.37	RTX treatment included the use of either 1000 mg on day 1 and day 15. The induction treatment was followed by maintenance treatment, including 500 or 1000 mg infusion with 6 months or more periodicity
Hehir,Michael K, 2017 [40]	America	24	NA	24	21/24	31.3 ± 14.9	45 (6–116) Median (range)	NA	NA	The initial dose was 375 mg/m2 per week for four sessions. During the observation period, 15 patients received more than one course of treatment
Sumanth Shivaram, 2022 [41]	India	13	10	2	7/13	NA	> 3M	44.84 ± 15.73	NA	500 mg × 4 doses OR 1gm x 2doses
Marta Caballero-Ávila, 2022 [42]	Spain	30	18	12	27/30	37.1 ± 19.2	85.5 ± 48.0	NA	NA	All patients received the standard dose of RTX (375 mg/m2) administered weekly for four consecutive weeks and then monthly for the next two months
Nicolas Collongues, 2012 [43]	France	20	12	4	11/20	NA	NA	NA	NA	RTX was given according to 2 different protocols: ①375 mg/m2 weekly for 4 consecutive weeks (induction stage), and subsequently 375 mg/m2 every 3 months;② 2 infusions of 1 g each, 2 weeks apart (induction stage), and subsequently 1 g, as required if symptoms worsened
Grayson Beecher, 2018 [44]	Canada	22	10	9	11/22	NA	28.8 ± 19.0	49.4 ± 13.4	51.4 ± 53.7(month)	Patients received 1 of 2 RTX induction regimens. In regimen 1, infusions of 375 mg/m2 were given once weekly for 4 weeks and then once every 4 weeks for 2 additional infusions. In regimen 2, infusions of 750 mg/m2 (up to a maximum of 1 g per dose) were given twice, with 2 weeks between infusions. When relapse occurred, patients received a maintenance regimen of 2 doses of 750 mg/m2 (up to a maximum of 1 g per dose), with 2 weeks between infusions
A.A. Rashid, 2020 [45]	Iraq	24	NA	NA	15/24	NA	> 6M	33.3 ± 10.1	6.3 ± 3.6	A standard dose of 1g. Every cycle is estimated as one infusion for two weeks. The space between cycles was adjusted as 6 months
Yufan Zhou, 2021 [46]	China	12	0	12	11/12	NA	6	40.0 ± 13.9	21.42 ± 22.10(month)	Low dose RTX (600 mg over 2 consecutive days, 100 mg on day 1 and 500 mg on day 2) was administrated
Jeannine M. Heckmann, 2022 [47]	South Africa	17	10	5	16/17	NA	18 (12–27) Median (IQR)	36.38 ± 15.17	11.18 ± 9.04	Single infusion to the patient's body surface area (375 mg/m2)
Ricardo H. Roda, 2019 [48]	America	27	10	13	22/27	NA	NA	NA	NA	Either 375 mg/m2 weekly for four consecutive weeks or rarely 1000 mg at weeks 1 and 3
Stefan Blum, 2011 [49]	Australia	14	11	3	9/14	NA	14.36 ± 11.31	51.14 ± 18.42	10.64 ± 12.66	In most patients 1 g of RTX, in two divided doses, was given
Paul Maddison, 2010 [50]	England	10	7	3	10/10	22.60 ± 16.45	NA	32.70 ± 12.21	NA	Rituximab treatment was given at a standard dose of 375 mg/m2 or equivalent in all patients. The commonest dosing schedule was weekly infusions on four consecutive weeks (eight patients), continued as once-monthly infusions in three. Four patients (three from one institution) only received one or two initial infusions of RTX
Dustin Anderson, 2016 [7]	Canada	14	5	6	8/14	NA	22.6 ± 2.4	50.9 ± 3.7	47.1 ± 15.0 (month)	RTX was either administered at a dose of 375 mg/m2 every week for four consecutive weeks then monthly for 2 months or at dose of 750 mg/m2 every 2 weeks for 1 month
Jingru Ren, 2023 [51]	China	22	9	4	18/22	40.77 ± 16.98	NA	48.45 ± 16.29	5.5 (3.75–12.00)	All patients were treated with intravenous RTX at initial dosage of 100 mg, 200 mg, or 500 mg, according to the risk of infection and disease severity. CD20 + B cells were evaluated at baseline and every 2–3 months during the treatment interval. When CD20 + B cells percentage exceeded 0, they were tested every month. Once the peripheral blood CD20 + B-cell percentage exceeded 1%, patients were treated with the next dose of RTX. The dosage of follow-up RTX varied between 100 and 500 mg
Zinovia-Maria Kefalopoulou, 2024 [52]	Greece	30	16	6	20/30	40.5 ± 13.9	NA	NA	11.7 ± 7.8	Anti-AChR + or DSN patients had induction cycle with 4 weekly infusions of 375 mg/m2, and a default repeat cycle with 2 weekly infusions of 375 mg/m2 after 6–8 months. Anti-MuSK + patients had induction with 375 mg/m2 infusions every week for 4 weeks, without a default repeat cycle

^† AChR Acetylcholine receptor, IQR Interquartile range, Musk Muscle-specific kinase, NA Not Available, RTX Rituximab, SD Standard Deviation, RTX Rituximab

Efficacy

Proportion of patients achieving MMS or better

23 studies reported the proportion of patients achieving MMS or better. The overall proportion was 59.0% (95% CI: 48.2–69.8%, n=599), as shown in Fig. 2. Significant heterogeneity was observed between studies (I²=95.28%). Sensitivity analysis showed stable results, and the Egger test indicated significant publication bias (P = 0.0125).

Fig. 2.

Forest plot showing the mean effect size and 95% confidence interval (CI) for the proportion of patients achieving minimal manifestation status (MMS) or better

Univariate meta-regression analyses showed that the dose of RTX was significantly associated with the proportion of MMS or better. When considering the dosing regimen, a significantly higher proportion of patients in the low-dose group (76.6%; 95% CI: 61.8–91.4%) achieved MMS or better compared to the conventional dose group (51.6%; 95% CI: 39.0–64.3%) (P=0.01) (Fig. 2). However, this difference was not observed when analyzing AChR-MG (P=0.13) and MuSK-MG (P=0.46) subgroups separately (Fig. S3 and Fig. S4). In patients with refractory MG, the proportion of patients achieving MMS or better was similar in the conventional dose group (51.4%; 95% CI: 20.8–81.4%) versus the low dose group (50.5%; 95% CI: 13.9–86.8%), (Fig. S5).

Subgroup analysis based on MG subtype showed that 60.7% (95% CI: 39.6–80.2%; I² = 88.54%, n=238) of AChR-MG patients and 79.6% (95% CI: 58.2–95.6%; I² = 75.41%, n=113) of MuSK-MG patients achieved MMS or better. The difference between the two groups was not statistically significant (P=0.24), (Fig. S6).

Change in QMGs

10 studies assessed changes in QMGs. The meta-analysis showed a mean reduction in QMGs of 6.81 (95% CI, −9.27 to −4.35, n = 233). However, there was high heterogeneity between the studies (I²=96.5%). No significant publication bias was detected (P = 0.4883). Univariate meta-regression analyses showed that the dose of RTX was significantly associated with changes in QMGs. When comparing the low-dose group (−9.04; 95% CI: −11.41 to −6.68) to the conventional dose group (−3.62; 95% CI: −6.57 to −0.68), the decrease in QMG score was more significant in the low-dose group (P < 0.01), as shown in Fig. 3. In patients with refractory MG, although the difference was not significant, the low-dose group (−7.17; 95% CI: −12.4 to −1.95) demonstrated a trend towards greater values of decreased QMGs compared with the conventional-dose group (−3.34; 95%: −7.49 to 0.82), P=0.26, (Fig S9).

Fig. 3.

Forest plot showing the mean difference and 95% confidence interval (CI) for the reduction of Quantitative Myasthenia Gravis Score (QMGs)

Improvement or better

22 studies reported that 83.1% (95% CI: 75.1–89.9%, n=562) of patients achieved improvement or better. Significant heterogeneity was observed between studies (I²=77.8%). Sensitivity analysis showed consistent results, and the Egger test indicated no significant publication bias (P=0.2835). The proportion of patients achieving improvement or better was higher in the low-dose group (87.7%; 95% CI: 78.9%−94.7%) compared to the conventional dose group (81.2%; 95% CI, 70.8–90.0%), but the difference between was not statistically significant (P=0.30) (Fig. 4).

Fig. 4.

Forest plot showing the mean difference and 95% confidence interval (CI) the proportion of patients achieving improvement or better

In patients with refractory myasthenia gravis, the difference between groups showed the same trends: the proportion of patients achieving improvement or better was higher in the low dose group (91.6%; 95% CI: 81.7–98.3%) than in the conventional dose group (89.0%; 95% CI: 69.8–99.8%), but the differences were not significant (P=0.77), (Fig S12).

Proportion of immunosuppressive drugs discontinued

11 studies reported the proportion of patients who discontinued both oral glucocorticoids (GC) and immunosuppressive drugs. The meta-analysis showed that 34.4% (95% CI: 16.3%−55.2%) of patients discontinued both oral GC and immunosuppressants, (Fig. S13). Pooled results from 21 studies showed that 45.1% (95% CI: 34.0%−56.4%, n=378) of patients discontinued oral GC (Fig. S14), and pooled results from 13 studies showed that 62.2% (95% CI: 42.2%−80.5%) of patients discontinued oral immunosuppressants (excluding GC), (Fig. S15). There were no significant differences between the conventional and low-dose groups (P > 0.05).

GC doses

16 studies assessed changes in GC dose. The mean decrease in GC dose was 20.6 mg (95% CI, −25.0 to −16.1, n=337). High heterogeneity was observed between studies (I²=86.88%). No significant publication bias was detected (P=0.54). The decrease in GC dose was more significant in the conventional dose group (−22.4; 95% CI: −27.7 to −17.2) compared to the low-dose group (−13.9; 95% CI, −19.1 to −8.7) (P<0.01) when compared to baseline values (Fig. 5).

Fig. 5.

Forest plot showing the mean difference and 95% confidence interval (CI) the reduction in Glucocorticoids (GC) dose

Safety

The incidence of serious adverse events (SAEs) was 4.0% (95% CI: 1.1%−8.0%, n=542) based on data from 19 studies . The incidence of SAEs was lower in the low-dose group (2.8%, 95% CI: 0.2–7.2%) than in the conventional dose group (5.1%; 95% CI: 0.3–13.6%), but the difference was not statistically significant (P=0.85), (Fig. S16). Pooled results from 17 studies showed an incidence of adverse events (AEs) of 30.0% (95% CI, 19.2%−42.1%, n=434). The incidence of AEs was lower in the low-dose group (25.3%; 95% CI: 7.5–49.1%) than in the conventional dose group (35.1%; 95% CI: 23.0–48.2%), but the difference was not significant (P=0.58), (Fig. S17). The most common adverse events included infections (24.2%), infusion reactions (18.0%), allergic reactions (5.4%), cytopenia (6.9%), and death (3.6%).

Meta regression

Table 2 provides a summary of the meta-regression, which was conducted to assess the effect of the included covariates on outcomes such as the proportion of patients achieving MMS or better. The univariate meta-regression analyses revealed a significant association between the proportion of females and the change in GC dose, the proportion of SAEs, the proportion of AEs, and the proportion of oral immunosuppressants other than GC that were discontinued. Furthermore, the administered dose was significantly associated with the proportion of MMS or better and the change in QMGs. However, the proportion of Musk patients was not significantly associated with any of the outcomes.

Table 2.

Summary of results from univariate meta-regression

	Dosage-Low				Musk				Female
Outcomes(Proportion)	Estimate	LCI	UCI	P-value	Estimate	LCI	UCI	P-value	Estimate	LCI	UCI	P-value
MMS or Better	0.24	0.03	0.46	0.03*	−0.12	−0.60	0.36	0.62	−0.11	−0.78	0.57	0.75
QMGs	−5.42	−9.15	−1.68	0.00*	−2.60	−11.23	6.03	0.56	−5.92	−19.22	7.39	0.38
Improvement or Better	0.10	−0.12	0.32	0.37	0.47	−0.06	1.01	0.08	0.33	−0.26	0.92	0.27
Immunosuppressive Drugs Discontinued	−0.33	−0.86	0.21	0.23	−0.09	−0.90	0.73	0.84	0.56	−1.07	2.18	0.50
Reduction in GC dose	8.12	−2.01	18.25	0.12	8.20	−8.03	24.43	0.32	35.55	5.85	65.24	0.02*
SAEs	−0.04	−0.20	0.12	0.64	−0.23	−0.47	0.01	0.06	−0.51	−0.86	−0.16	0.00*
AEs	−0.10	−0.38	0.17	0.46	−0.24	−0.73	0.25	0.34	−0.74	−1.40	−0.08	0.03*
Discontinued GC	0.09	−0.15	0.33	0.46	−0.16	−0.53	0.20	0.37	−0.25	−0.99	0.49	0.51
Discontinued oral immunosuppressants (excluding GC)	−0.17	−0.63	0.29	0.47	0.28	−0.37	0.93	0.40	1.89	0.75	3.04	0.00*

^† AE Adverse Events, GC Glucocorticoids, LCI Lower Confidence Interval, MMS Minimal Manifestation Status, QMGs Quantitative Myasthenia Gravis Score, SAEs Serious Adverse Events, UCI Upper Confidence Interval, * P < 0.05

Prognostic factors of MG

Univariate analyses

The univariate analyses demonstrated that non-refractory MG, age greater than 50.75 years, lower MGFA grading, and low-dose dosing regimens were significantly associated with patients achieving MMS or better prognosis. Specifically, patients with refractory MG were only 0.28 times more likely to achieve an MMS or better prognosis than non-refractory MG patients (OR = 0.28, 95% CI = 0.16–0.49). Patients aged 50.75 years or above were observed to be 2.14 times more likely to achieve MMS or better prognosis in comparison to patients aged 50.75 years or below. Patients with moderate severity (MGFA class III) exhibited a 0.26-fold increased likelihood of achieving an MMS or better prognosis in comparison to patients with mild severity (MGFA classes I and II). Conversely, patients with severe severity (MGFA classes IV and V) demonstrated a 0.15-fold increased likelihood. Patients who received low-dose therapy were 3.47 times more likely to achieve an MMS or better prognosis than those who were treated with conventional doses. No significant association was observed between gender and MG antibody type and the achievement of an MMS or better prognosis.

With regard to the achievement of an improved or better prognosis, only the dose of RTX was found to be significantly associated with the outcome. Patients treated with low-dose therapy were 2.79 times more likely to achieve this prognosis than patients treated with conventional doses. The distinction between non-refractory MG and refractory MG was more pronounced, yet did not attain statistical significance (P=0.06), as illustrated in Table 3.

Table 3.

Univariate analysis of factors predicting response to rituximab treatment in myasthenia gravis patients

Factor	OR(95%CI)	P Value	Factor	OR(95%CI)	P Value
	MMS or Better			Improvement or Better
Refractory VS. Non-Refractory	0.28(0.16–0.49)	< 0.01*	Refractory VS. Non-Refractory	0.51(0.24–1.03)	0.06
Female VS. Male	0.68(0.38–1.20)	0.18	Female VS. Male	1.10(0.50–2.28)	0.80
Age at Rituximab Treatment, > VS. < 50.75y	2.14(1.25–3.71)	< 0.01*	Age at Rituximab Treatment, > VS. < 35.25y	1.76(0.86–3.58)	0.12
MG Grade, MGFA			MG Grade, MGFA
Moderate VS. Mild	0.26(0.12–0.54)	< 0.01*	Moderate VS. Mild	0.56(0.23–1.40)	0.43
Severe VS. Mild	0.15(0.07–0.30)	< 0.01*	Severe VS. Mild	0.68(0.28–1.65)	0.67
Severe VS. Moderate	0.58(0.31–1.11)	0.22	Severe VS. Moderate	1.22(0.55–2.70)	0.88
Antibody			Antibody
MUSK VS. AChR	1.90(0.93–3.88)	0.18	MUSK VS. AChR	1.97(0.66–5.94)	0.45
DN VS. AChR	0.51(0.15–1.76)	0.54	DN VS. AChR	0.66(0.17–2.57)	0.82
MUSK VS. DN	3.71(0.95–14.54)	0.14	MUSK VS. DN	3.00(0.57–15.87)	0.40
Low VS. Conventional	3.47(2.01–6.10)	< 0.01*	Low VS. Conventional	2.79(1.30–6.52)	0.01*

^† AChR Acetylcholine Receptor, CI Confidence Interval, DN Double Seronegative, MMS Minimal Manifestation Status, MG Myasthenia Gravis, MGFA Myasthenia Gravis Foundation of America, MUSK Muscle-specific kinase, OR Odd Ratio, * P < 0.05

Multivariate analysis

A stepwise logistic regression analysis was employed to identify the optimal combination of factors for predicting improvement. As with the findings of the univariate analyses, the results demonstrated that non-refractory MG, mild disease severity (classified according to the MGFA classification) and a low-dose medication regimen were significant predictors of achieving MMS or better. Conversely, for improvement or better, the medication dose alone was identified as an independent predictor. Moreover, age demonstrated a considerable impact on both prognostic indicators, although this difference was not statistically significant. It is noteworthy that, in the multivariate analyses, patients with MuSK antibody type were significantly more likely to achieve MMS or better by 4.98-fold compared to AChR antibody type. This differs from the results of the univariate analyses. Further details can be found in Table 4.

Table 4.

Multivariate stepwise logistic regression analysis of factors predicting response to rituximab in myasthenia gravis patients

Factor	OR(95%CI)	P Value	Factor	OR(95%CI)	P Value
	MMS or Better			Improvement or Better
Refractory VS. Non-Refractory	0.40(0.21–0.75)	< 0.01*	Refractory VS. Non-Refractory
Female VS. Male			Female VS. Male
Age at Rituximab Treatment, > VS. < 50.75y	1.72(0.91–3.30)	0.10	Age at Rituximab Treatment, > VS. < 35.25y	1.95(0.93–4.04)	0.07
MG Grade, MGFA			MG Grade,MGFA
Moderate VS. Mild	0.34(0.15–0.77)	0.01*	Moderate VS. Mild
Severe VS. Mild	0.14(0.06–0.32)	< 0.01*	Severe VS. Mild
Severe VS. Moderate			Severe VS. Moderate
Antibody			Antibody
MUSK VS. AChR	4.98(2.14–12.24)	< 0.01*	MUSK VS. AChR
DN VS. AChR	0.44(0.10–1.70)	0.25	DN VS. AChR
MUSK VS. DN			MUSK VS. DN
Low VS. Conventional	2.94(1.51–5.81)	< 0.01*	Low VS. Conventional	2.97(1.37–7.02)	0.01*

^† AChR Acetylcholine Receptor, CI Confidence Interval, DN Double Seronegative, MMS Minimal Manifestation Status, MG Myasthenia Gravis, MGFA Myasthenia Gravis Foundation of America, MUSK Muscle-specific kinase, OR Odd Ratio, * P < 0.05

Discussion

Consistent with previous research [9, 11, 53–55], our findings indicate that RTX is a safe and effective treatment option for patients with MG, even in refractory MG patients. Notably, the low-dose regimen demonstrated a trend towards better efficacy than the conventional dose regimen, and in certain aspects, such as the proportion of patients achieving MMS or better, the difference was statistically significant. Also, the low-dose group had a lower incidence of adverse events and a better safety profile. In addition, non-refractory MG, MUSK-MG over AChR-MG and milder disease severity were identified as predictors of better efficacy. In conclusion, RTX has been demonstrated to be an effective treatment for MG, with a reduction in QMGs and the dosage of immunosuppressive agents, including glucocorticoids. Furthermore, the lower dose has been shown to be more advantageous than the conventional dosing regimen, in addition to lowering GC dosage.

The improvement in symptoms observed in patients was found to be closely correlated with the administered dose of RTX, with the presence or absence of refractory MG also exerting a significant influence. The conventional dosage regimen proved more efficacious than the low-dose regimen in patients with refractory MG, as evidenced by the achievement of MMS or better. The achievement rates for conventional versus low-dose administration were 54.1% versus 46.3% (refractory AchR-MG), 67% versus 48% (refractory MG), and 51.4% versus 50.5% (refractory MG in this study), respectively [9, 11]. In contrast, when the patient range was extended to include all MG patients, a significantly higher proportion of patients treated with low-dose RTX achieved MMS or higher (76.6% versus 51.6%). When improvement or betterment was used as an indicator, the proportion of patients achieving improvement or betterment in the conventional dose group was comparable to that of the low dose group in patients with refractory MG (89.0% versus 91.6%) and 76.8% versus 77.1% (refractory AchR-MG) [9]. Whereas, the patient range was extended to encompass all MG patients, a higher proportion of patients in the low-dose group exhibited improvement or better than those in the conventional-dose group (87.7% versus 81.2%). The aforementioned results indicate that the efficacy of RTX treatment may vary between refractory and non-refractory MG patients. It may be the case that conventional doses of RTX are more effective in patients with refractory MG, whereas this trend may be reversed or lost in non-refractory patients. In all patients with myasthenia gravis, the efficacy of the low dose was superior to that of the conventional dose [9].

RTX treatment can reduce the need for other immunosuppressive drugs [9, 17, 53]. Compared to conventional immunosuppressive therapy, RTX leads to a shorter median time to remission for both new and refractory cases, as well as the proportion of patients who relapse or require further immunosuppression within the first 24 months [56]. As for the comparative role of different administered doses in reducing the need for immunosuppression, a previous study reported a higher rate of discontinuation or dose reduction to ≤10 mg in the conventional dose group compared to the low-dose group (75.9% versus 64.2%) [9]. However, our study demonstrated a higher rate of GC discontinuation in the low-dose group compared to the conventional dose group (52.6% versus 42.6%). The differences in trends regarding GC discontinuation may be attributed to variations in the populations included in different studies as well as the fact that data from patients reduced to ≤10 mg were not included in this study. As for the reduction in GC dose, the decrease was more significant in the conventional dose group than in the low-dose group (−24.05 versus −13.88, P=0.01). While the greater reduction in GC dose in the conventional dose group may be associated with a higher initial GC dose compared to the low-dose group (mean: 31.32 mg versus 26.22 mg).

Furthermore, when considering MMS or better and improvement or better as indicators, both parameters favored the low-dose group over the conventional dose group in MG patients. However, for GC dose reduction, the conventional dose group exhibited superior results compared to the low-dose group. Consider that proportion of refractory MG patients included in the analysis of GC dose changes was significantly higher compared to the analysis of MMS or better and improvement or better (P<0.001). This suggests again, to some extent, that the effects of different RTX doses may differ between refractory and non-refractory MG patients.

Previous studies have shown that RTX reduces QMG score in patients with refractory MG, but comparisons of changes in QMG score in all MG patients and between different dose groups are lacking. In previous studies QMG score were reduced from baseline in refractory patients by 1.55 (SMD; 95% CI: −2.22 to −0.88) and 4.16 points (MD; 95% CI: −5.32 to −2.99) [10, 11]. In our study, QMG score were reduced by 4.85 points from baseline in patients with refractory MG and 6.81 points in patients with MG. In addition, low-dose RTX reduced QMG score significantly better compared to the conventional dose group (P<0.01). Due to limited data, we were unable to explore whether this trend differed between refractory and non-refractory MG patients, The observation that the reduction in QMG score among patients with refractory MG was less pronounced than that observed in the overall population provides some indication that the decline in QMG score was more significant among patients with non-refractory MG.

Therapeutic decisions and outcomes in MG are significantly influenced by the clinical phenotype and the type of circulating antibodies present [57]. Rituximab has been shown to decrease MuSK antibodies and coincided with clinical remission, but its effect on AchR antibodies is not significant [17, 56]. Studies have shown that a higher proportion of RTX-treated MuSK-MG patients achieve MMS or better compared to AchR-MG patients [11, 14, 17].The response rates were also higher in MuSK-MG patients than in AchR-MG patients (88.8% versus 80.4%, P > 0.05) [17].The 2020 International Consensus Guidelines for Myasthenia Gravis recommend considering RTX as an early treatment option for MuSK-Ab+ MG patients who have a poor response to initial immunotherapy. For patients with refractory AChR-Ab+ MG, RTX is also an option if they fail to respond or are unable to tolerate other immunosuppressive drugs [58]. Our study similarly demonstrated that among all MG patients, a higher proportion of those with MuSK MG achieved MMS or better than those with AChR MG (79.6% versus 60.7%, P=0.24). Furthermore, meta-regression demonstrated a robust correlation between the proportion of MuSK patients and the proportion of individuals who exhibited improvement or better (P=0.08). Furthermore, multivariate stepwise logistic regression analyses based on individual data revealed that MuSK-MG patients were 4.98 times more likely to achieve MMS or better than AChR-MG patients.

The efficacy of RTX in the treatment of MG patients is contingent upon a number of factors. The study by Tandan et al. demonstrated that the administration of RTX to patients with MUSK-MG at an earlier age and with less severe disease was associated with more favourable outcomes [6]. Our findings were consistent with those of Tandan et al., indicating that milder disease severity (MGFA grading) and MUSK over AChR were predictors of achieving MMS or better. Furthermore, our findings indicated that non-refractory MG and a low-dose drug regimen were also predictive of achieving MMS or better. With regard to the age at which RTX treatment was initiated, although no significant difference was observed, our study indicated a trend towards enhanced efficacy at an earlier age.

RTX demonstrated a favorable safety profile in MG patients, with a lower rate of adverse event discontinuations compared to conventional immunosuppressive therapy (3% versus 46%, P < 0.001) [17]. The incidence of adverse events associated with RTX for MG varied between studies, ranging from 4% to 19.6%, but most symptoms were mild [10, 11, 59]. Consistent with previous research [10] , the incidence of serious adverse events or adverse events was lower in the low-dose group than in the conventional dose group, but the difference was not statistically significant. Overall, RTX appears to have a favorable safety profile for the treatment of MG, and low-dose RTX may be a safer option compared to conventional doses.

In conclusion, a low-dose regimen is recommended for MG patients who are receiving RTX, particularly those with non-refractory MG. The low-dose regimen has been demonstrated to be more efficacious and to result in a lower incidence of adverse events in patients when compared to the conventional dose regimen. Furthermore, it is evident that the low-dose regimen is a more cost-effective option than the conventional dose regimen.

However, there are some limitations to consider in our study. Firstly, as a single-arm meta-analysis, we were unable to compare RTX with other existing therapeutic agents for MG. Secondly, most of the included studies were observational studies, which may introduce confounding factors and increase the risk of bias and heterogeneity between studies. For instance, Egger's test demonstrated a notable publication bias between studies reporting MMS or better in all MG patients. This bias could be attributed to the observation that the incidence of MMS or better was 100% in the articles by Kimberly R [8], Ying Du [28], and Huining Li [34], which had a considerable impact on the publication bias. Following the exclusion of the aforementioned three articles, the evidence indicated that the publication bias was no longer statistically significant (P=0.8525). The results remained consistent with those observed prior to exclusion, indicating that the proportion of MG patients achieving MMS or better was significantly higher in the low-dose group than in the conventional dose group (65.1% vs. 48.2%, P=0.03)..

Additionally, the sample size, especially for low-dose patients, was small, resulting in insufficient data to fully support the results of subgroup analysis. Therefore, large randomized controlled trials are necessary to further clarify the differences in efficacy and safety between different RTX doses in MG patients.

Conclusion

In conclusion, our study provides evidence that RTX is effective in improving clinical symptoms and reducing QMGs in MG patients. In addition, RTX treatment reduces the use of oral glucocorticoids and other immunosuppressive drugs. Low-dose RTX is more effective than conventional-dose RTX in treating MG patients and exhibits a better safety profile. Mild disease severity, non-refractory MG, low dose, and MUSK-MG over AChR-MG predict better efficacy. Large randomized controlled trials are necessary to further evaluate the efficacy and safety of RTX in MG patients and its various subtypes.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AChR

Acetylcholine Receptor

AEs

Adverse Events

CI

Confidence Intervals

DN

Double Seronegative

GC

Glucocorticoids

MG

Myasthenia Gravis

MGFA

Myasthenia Gravis Foundation of America

Musk

Muscle-specific kinase

MMS

Minimal Manifestation Status

QMGs

Quantitative Myasthenia Gravis Score

RTX

Rituximab

SAEs

Serious Adverse Events

Authors’ contributions

Jianchun Li: Conceptualization(equal), Data curation(lead), Formal analysis(lead), Investigation(lead), Methodology(equal), Software(lead), Visualization(lead), Writing – original draft(lead). Di Chen: Conceptualization(lead), Data curation(equal), Funding acquisition(lead), Investigation(equal), Methodology(equal), Resources(lead), Validation(lead), Writing – original draft(equal). Fei Zhao: Methodology(lead), Software(equal), Supervision(supporting), Writing – review & editing(equal). Weihang Cao: Data curation(supporting), Investigation(supporting), Validation(equal), Writing – review & editing(supporting). Pengfei Jin: Conceptualization(equal), Data curation(supporting), Project administration(lead), Resources(equal), Supervision(lead), Writing – review & editing(lead). Jianchun Li and Di Chen contributed equally to this work.

Funding

This work was supported by the National High Level Hospital Clinical Research Funding (BJ-2021–230) and the Beijing Natural Science Foundation (L242150). The funders had no role in the study design, data interpretation, or manuscript preparation, etc.

Data availability

The datasets supporting the conclusions of this article are available in the Baidu Netdisk repository, unique persistent hyperlink to datasets in https://pan.baidu.com/s/1tI81hdnkCiLS8u1exLd_Lw. The extracted code is 3d2y.

Code availability

The code used to analyse the data for this study can be found in Supplementary Material 2.

Declarations

Ethics approval and consent to participate

The data used in this study are all from the published literature. These data are publicly available and do not contain any sensitive personal data. Therefore, we believe that ethical approval and consent to participate are not applicable to this study.

Consent for publication

Not applicable.

Competing interests

The authors have no competing interests to declare that are relevant to the content of this article.