Eculizumab in Myasthenia Gravis: A Multicenter Retrospective Real-World Study in China Authors: Yue Li, Huan Yang, Song Tan, Lijun Luo, Xuebing Cao, JinQuan Hu, Bitao Bu, Ting Chang, Zhijun Li Published: 2026 ABSTRACT AbstractIntroductionMyasthenia gravis (MG) is a chronic autoimmune neuromuscular disorder needing long-term immunosuppressive therapy. Eculizumab is a promising treatment, but real-world evidence on its efficacy and safety is limited.MethodsThis retrospective cohort study across six medical centers enrolled 60 patients with MG on eculizumab. Clinical data, including MG-Activities of Daily Living (MG-ADL) scores, quantitative MG scores (QMG), and oral corticosteroid dosage at baseline, weeks 1, 4, and 8, and final follow-up were obtained. Outcomes included score changes, post-intervention status (PIS), and safety profile.ResultsBy week 8, 30.95% of patients achieved minimal manifestation status (MMS) or better, increasing to 32.69% at final follow-up. Mean ADL scores significantly dropped from 8.18 ± 3.99 at baseline to 5.59 ± 3.72 at week 1, 3.78 ± 3.42 at week 4, 3.50 ± 2.92 at week 8, and 3.33 ± 3.56 during final follow-up (p < 0.0001). Clinically meaningful improvement (CMI) was observed in 56.9%, 70.0%, 85.0%, and 76.27% at the respective time points. Both thymoma-associated and non-thymoma subgroups showed significant ADL improvement (p ≤ 0.026) with mixed-effects modeling revealing significant group × time interaction (p = 0.01).The thymoma-associated MG (TAMG) group had a faster initial response, with subsequent outcomes comparable between the two groups. Among the 37 patients (61.67%) who switched from FcRn inhibitor therapy to eculizumab, clinical outcomes were non-inferior to those of direct eculizumab recipients and the magnitude of MG-ADL improvement was comparable to that of the overall treatment group. The mean daily corticosteroid dose decreased from 21.21 ± 13.68 mg at baseline to 13.32 ± 10.42 mg at final follow-up (p < 0.0001) Adverse events were rare, with no treatment-related serious events.ConclusionEculizumab provides rapid and sustained clinical improvement in patients with MG, with a favorable safety profile. Patients with TAMG show a faster initial response. For patients with an inadequate response to FcRn inhibitors, eculizumab remains an effective rescue strategy.Keywords: Myasthenia gravis, Eculizumab, Steroid-sparing, Clinical efficacy Key Summary Points Why carry out this study?Myasthenia gravis (MG) is a chronic autoimmune neuromuscular disorder requiring long-term immunosuppressive therapy. While eculizumab is a promising treatment, real-world evidence on its efficacy and safety remains limited.This study evaluated the effectiveness, safety, and subgroup-specific treatment responses of eculizumab in real-world patients with MG, including analysis of treatment patterns in refractory cohorts.What was learned from this study?Eculizumab demonstrated effectiveness in patients with MG, with clinical improvement achieved after 8 weeks or more of therapy and a sustained, relevant steroid-sparing effect. Notably, the TAMG group exhibited an accelerated initial therapeutic response compared to the non-TAMG group, with no differences observed in subsequent outcomes between the two cohorts.For patients with an inadequate response to FcRn inhibitor therapy, switching to eculizumab provides non-inferior clinical outcomes compared with direct C5 inhibitor therapy. Half of efgartigimod-resistant patients switched to eculizumab within 4 weeks, indicating its potential as an effective alternative for refractory cases.These findings provide valuable real-world evidence supporting eculizumab therapeutic positioning in MG management. Introduction Myasthenia gravis (MG) is a chronic autoimmune neuromuscular disorder caused by autoantibodies targeting components of the neuromuscular junction [1], most commonly anti-acetylcholine receptor antibodies (AChR-Abs) and anti-muscle-specific kinase antibodies (MuSK-Abs). The hallmark clinical manifestation of MG is fluctuating skeletal muscle weakness. Myasthenic crisis (MC) represents the most severe disease manifestation, leading to respiratory failure and imposing a substantial disease burden [2–4]. Conventional treatment strategies for MG include symptomatic therapy with cholinesterase inhibitors, corticosteroids, and nonsteroidal immunosuppressants. Rescue therapies, such as intravenous immunoglobulin (IVIG) and plasma exchange (PE), are frequently administered to patients experiencing MC, whereas thymectomy is recommended for patients with thymoma and for those younger than 65 years with non-thymomatous, AChR-Ab-positive generalized MG (gMG) [1]. Despite these therapeutic approaches, standard treatments achieve satisfactory and sustained clinical improvement in only 30–50% of patients [5]. Refractory MG is a chronic condition in which patients fail to achieve adequate relief of severe symptoms despite optimal use of prednisone and/or other immunosuppressants. Refractory MG affects approximately 10–20% of patients and is characterized by persistent disease activity, severe fatigue, significant functional disability, and recurrent MC requiring hospitalization [6, 7]. Moreover, the increased risk of treatment-related adverse events and complications may further exacerbate disease burden and impair quality of life. In recent years, novel biological therapies have emerged and driven a paradigm shift in MG management due to their high specificity and targeted mechanisms of action [5], particularly in patients with high disease activity, including patients with refractory MG. Earlier implementation of targeted therapies during the disease course may represent an important strategy [5]. These targeted agents include neonatal Fc receptor antagonists [8–11], complement inhibitors [12–14], B cell-depleting agents [15, 16], and B cell pathway modulators, such as telitacicept, a recombinant protein composed of the transmembrane activator and calcium-modulating cyclophilin ligand interactor (TACI) [17]. Effective management of MG requires an individualized, long-term therapeutic strategy that adapts to each patient’s evolving disease course and treatment response [5, 18]. Although glucocorticoids and conventional immunosuppressants remain widely used, a substantial proportion of patients experience recurrent exacerbations, refractory disease despite treatment escalation, or cumulative toxicity related to prolonged immunosuppression [12, 19, 20]. C5 complement inhibitors exert their therapeutic effects by blocking cleavage of complement component C5, thereby inhibiting activation of the downstream complement cascade and preventing formation of the membrane attack complex (MAC or C5b-9) [21]. To date, several C5 complement inhibitors, including eculizumab, ravulizumab, and zilucoplan [8, 13, 14, 22], have been approved for the treatment of AChR-Ab-positive generalized MG. Among these agents, eculizumab has demonstrated rapid symptom improvement and robust clinical efficacy in patients with refractory generalized MG, as shown in the REGAIN study. Furthermore, a recent real-world study reported that eculizumab was associated with a higher quantitative myasthenia gravis (QMG) response rate and greater steroid-sparing effect than efgartigimod in patients with AChR-Ab-positive MG [23]. Real-world studies of MG therapies often suffer from truncated follow-up durations. Eculizumab trials, in particular, have frequently prioritized short-term outcomes (e.g., 12–26 weeks) to demonstrate rapid symptom improvement. Consequently, large-scale investigations with extended follow-up remain limited, restricting insight into the durability of disease control, relapse patterns, the pace of concomitant immunosuppressant tapering, and cumulative adverse effects, particularly infection risk. This gap hinders a comprehensive understanding of the long-term role of eculizumab in sustaining remission. The present retrospective study, therefore, aimed to evaluate the effectiveness and safety of eculizumab in real-world clinical practice among patients with AChR-Ab-positive MG across multiple centers in China. Methods Patients and Study Design Our retrospective cohort study analyzed patients with MG treated with eculizumab at six Chinese hospitals—Tongji Hospital (Tongji Medical College, Huazhong University of Science and Technology), Tangdu Hospital of Air Force Medical University, Xiangya Hospital of Central South University, Sichuan Academy of Medical Sciences & Sichuan Provincial People’s Hospital, Wuhan No. 1 Hospital (Wuhan Hospital of Traditional Chinese and Western Medicine), and Taihe Hospital of Hubei University of Medicine, with treatment initiated before November 2024. The protocol was approved by all the necessary ethics committees and institutional review boards at each trial site. The institutional review board approval number from the leading institution for this study (Tongji Hospital) is TJ-JRB202411058. This retrospective study was granted a waiver of informed consent because it used de-identified data. All procedures rigorously adhered to the principles of the Declaration of Helsinki. The enrollment criteria were as follows: (1) Confirmation of a diagnosis of myasthenia gravis (MG) according to the 2020 diagnostic guidelines [24]; (2) receipt of at least four infusions of eculizumab to complete the induction phase (900 mg administered once weekly for the first four doses); and (3) availability of complete clinical follow-up data. Exclusion criteria included (1) receipt of fewer than four doses of eculizumab; (2) incomplete medical history data; and (3) a follow-up duration of less than 1 month after the first eculizumab infusion. Data Collection Demographic and clinical data were extracted from electronic medical records, encompassing key baseline characteristics such as age at disease onset, sex, disease duration (calculated from initial symptom onset to initiation of eculizumab), presence of thymoma, and autoantibody status (including AChR-positive and seronegative disease). Treatment history included prior use of immunosuppressive therapies (e.g., glucocorticoids, azathioprine), acetylcholinesterase inhibitors, and rescue therapies such as IVIG and PE. Clinical outcomes were assessed using the Myasthenia Gravis Activities of Daily Living (MG-ADL) score and the QMG score at baseline, weeks 1, 4, 8, and at the last follow-up before November 2024 after initiation of eculizumab. Safety data encompassed all treatment-emergent adverse events (TEAEs), including infections, infusion-related reactions, and laboratory abnormalities. Eculizumab Treatment Protocol Eculizumab was administered intravenously in accordance with standard clinical practice. The induction phase consisted of 900 mg administered weekly for four consecutive weeks. Maintenance therapy was subsequently continued at a dose of 1200 mg every 2 weeks, with dose adjustments made at the discretion of the treating physician on the basis of clinical response. Concomitant immunosuppressive therapies and pyridostigmine were continued on the basis of the treating physician’s clinical judgment, with dose modifications documented at weeks 1, 4, and 8, as well as at the final follow-up conducted before November 2024 after initiation of eculizumab. Follow-up and Outcome Measures Patients were followed up until December 2024 after initiation of eculizumab therapy. Clinical assessments were conducted at baseline (pre-infusion), week 1, week 4, week 8, and at the final follow-up visit. Evaluated outcomes included MG-ADL score, QMG score, change in MG-ADL score (with a reduction of ≥ 2 points defined as a CMI), change in QMG score (with a ≥ 3-point reduction defined as CMI), Myasthenia Gravis Foundation of America (MGFA) post-intervention status (PIS), and glucocorticoid dosage. In this retrospective study, MGFA-PIS was adopted for follow-up. It classifies patient status as complete stable remission, pharmacologic remission(PR), Minimal Manifestation Status (MMS), Improved, Unchanged, Worse, Exacerbation, and Death due to MG. Safety endpoints were also evaluated and included the incidence and severity of TEAEs, graded according to the Common Terminology Criteria for Adverse Events (CTCAE) version 5.0. Statistical Analysis Continuous variables were presented as mean ± standard deviation (SD) and were compared using the independent-samples t test or one-way analysis of variance (ANOVA), as appropriate. Longitudinal alterations in MG, ADL, QMG, and corticosteroid dosages were analyzed via linear mixed-effects models to assess temporal trends, group disparities, and group × time interactions. The model incorporated fixed effects for time, group, baseline score, and their interactions, along with a random intercept for each subject to account for within-patient correlations. Post hoc pairwise comparisons were adjusted by the Bonferroni method. All statistical analyses were performed using Prism version 10.1.2 and SPSS 27. A two-sided p value < 0.05 was considered statistically significant. Results Baseline Clinical Characteristics A total of 60 patients with MG who received at least four weekly infusions of eculizumab were enrolled in this retrospective cohort study before November 2024. These patients were distributed across six Chinese hospitals as follows: Tongji Hospital (n = 22), Tangdu Hospital (n = 21), Xiangya Hospital (n = 6), Sichuan Provincial People’s Hospital (n = 5), Wuhan No.1 Hospital (n = 4), and Taihe Hospital (n = 2). The cohort comprised 35 men (58.33%) and 25 women (41.67%). The mean age at diagnosis was 42.25±17.71 years (range 1–74 years), and the mean age at Eculizumab initiation was 47.47±15.75 years (range 15–84 years). The median disease duration was 5.61 ± 6.52 years (Table 1). All patients were monitored for more than 4 weeks. Table 1Baseline clinical characteristics of patients with MG (N = 60)Clinical variablesNumberPercentAge at Eculizumab initiation (years)47.47 ± 15.75Mean age at diagnosis (years)42.25 ± 17.71Disease duration (years)5.61 ± 6.52Sex Male3558.33% Female2541.67%MGFA classification at enrollment I35.00% II2338.33% III1220.00% IV2033.33% V23.33%Antibody status AChR4880.00% AChR + titin1118.33% Seronegative11.67%Ocular MG35.00%Early-onset MG (< 50 years)3761.67%Late-onset MG (> 50 years)2338.33%Thymoma2440.00%WHO histology B13 B26 B34 AB2 B1 + B21 B2 + B33 Missing5Previous treatment Glucocorticoid5693.33% Tacrolimus4066.67% Mycophenolate mofetil1728.33% Azathioprine58.33% Rituximab46.67% Cyclosporine11.67% IVIG1016.67% PE2948.33%Treatment at baseline Glucocorticoid5490.00% Tacrolimus3355.00% Mycophenolate mofetil1220.00% Azathioprine23.33%Mean corticosteroid dosage at baseline, mg21.21 ± 13.68ADL score at baseline, median ± SD8.18 ± 3.99QMG score at baseline, median ± SD14.48 ± 6.59 Baseline clinical characteristics of patients with MG (N = 60) Regarding clinical phenotypes, 3 patients (5.00%) had ocular MG (OMG). For age at onset, 37 patients (61.67%) had early-onset MG (EOMG; before 50 years) and 23 (38.33%) had late-onset MG (LOMG; at or after 50 years). Thymoma was present in 24 patients (40.00%). Serologically, 48 patients (80.00%) were AChR-Ab positive, 11 (18.33%) were dual positive for AChR and titin antibodies, and 1 (1.67%) was seronegative. Among the 60 patients, 52 (86.67%) had disease worsening, and 20 (33.33%) developed MC. Before eculizumab initiation, all patients had received corticosteroids or at least one non-corticosteroid immunosuppressive agent. Glucocorticoids were the main therapy, with 90% (54/60) taking oral prednisone at baseline and a mean daily dose of 21.21 ± 13.68 mg. In addition, 54 patients (90%) had received non-steroidal immunosuppressants, with tacrolimus most frequently prescribed, followed by mycophenolate mofetil, rituximab, and azathioprine. More than 60% of patients switched to eculizumab after failure of efgartigimod treatment due to disease exacerbation or refractory symptoms. The mean duration of eculizumab treatment was 14.82 ± 12.65 weeks (range 4–49 weeks). A total of 42 patients completed the initial 8-week treatment course. As of December 2024, 23 patients remained on continuous eculizumab therapy, with a mean treatment duration of 26.02 ± 13.13 weeks (range 11–49 weeks). Efficacy and Safety Across the cohort, the distribution of clinical status based on MGFA PIS demonstrated dynamic changes over time (Fig. 1). At week 1, the majority of patients (71.67%) were classified as “improved,” 1.67% achieved “PR,” and 25.00% remained “unchanged.” By week 4, the proportion of patients attaining “MMS or better” increased to 21.67%, while 61.67% were categorized as “improved.” Concurrently, 13.33% remained “unchanged,” and 3.33% were classified as “worse.” Two patients received IVIG as rescue therapy within the first 8 weeks of treatment. At week 8, 30.95% of patients had achieved “MMS or better,” 42.86% were “improved,” 16.67% remained “unchanged,” 7.14% were “worse,” and 2.38% experienced exacerbation. At the final follow-up, 32.69% of patients maintained “MMS or better,” 40.38% were “improved,” 11.54% remained “unchanged,” 11.54% were classified as worse, 1.92% experienced exacerbation, and 1.92% succumbed to MG-related complications. Longitudinal distribution of Myasthenia Gravis Foundation of America (MGFA) post-intervention status (PIS). a Percentage of patients classified as pharmacologic remission(PR), minimal manifestation status(MMS), improved, unchanged, worse, or deceased at baseline, weeks 1, 4, 8, and the last follow-up. b Percentage of patients achieving MMS + status at baseline, week 1, 4, 8, and the last follow-up For the entire cohort, the MG-ADL score decreased significantly from 8.18 ± 3.99 at baseline to 5.59 ± 3.72 at week 1, further declining to 3.78 ± 3.42 at week 4, 3.50 ± 2.92 at week 8, and 3.33 ± 3.56 at the final follow-up (p < 0.0001) (Fig. 2). CMI (≥ 2-point reduction in MG-ADL) was achieved by 56.9%, 70.0%, 85.0%, and 76.27% of patients at the respective time points. Correspondingly, QMG scores decreased from 14.48 ± 6.59 at baseline to 11.09 ± 6.01 at week 1, 9.24 ± 5.63 at week 4, 9.00 ± 4.96 at week 8, and 8.18 ± 6.24 at the final follow-up (p < 0.0001). Corresponding CMI rates (≥ 3-point reduction) in QMG were 50.91%, 65.52%, 76.92%, and 75.44%, respectively. Myasthenia Gravis Activities of Daily Living(ADL) and Quantitative Myasthenia Gravis (QMG) score changes following initiation of eculizumab. a Changes in ADL score. The dark blue bars represent the mean ADL score for the entire cohort at baseline and at weeks 1, 4, 8, and the last follow-up; the light blue bars represent the maintenance therapy (MT) subgroup. b Changes in QMG score. Symbols indicate the mean QMG score for the overall cohort (dark blue) and the MT subgroup (light blue) at the corresponding time points. c Proportion of patients achieving clinically meaningful improvement (CMI) in MG-ADL score (≥ 2-point reduction). Dark blue bars = overall cohort (n = 60); light blue bars = MT subgroup (n = 23). CMI rates for the overall cohort: 56.9% (W1), 70.0% (W4), 85.0% (W8), 76.3% (final follow-up). CMI rates for the MT subgroup: 71.4% (W1), 91.3% (W4), 100% (W8), 91.3% (final follow-up). d Corresponding proportions achieving CMI in QMG score (≥ 3-point reduction). Dark blue bars = overall cohort (n = 60); light blue bars = MT subgroup (n = 23). CMI rates for the overall cohort: 50.9% (W1), 65.5% (W4), 76.9% (W8), 75.4% (final follow-up). CMI rates for the MT subgroup: 61.9% (W1), 69.6% (W4), 85.7% (W8), 87.0% (final follow-up) Among the 23 patients who continued eculizumab therapy, the MG-ADL score declined from 9.30 ± 4.11 at baseline to 6.19 ± 4.37 at week 1, followed by further reductions to 3.61 ± 3.50, 3.42 ± 3.23, and 3.13 ± 3.51 at subsequent follow-up visits (weeks 4, 8, and the final follow-up) (p < 0.0001). Corresponding CMI rates reached 71.43%, 91.30%, 100%, and 91.30% at weeks 1, 4, 8, and the final follow-up, respectively. Similarly, QMG scores declined from 17.11 ± 6.54 to 12.33 ± 6.45, 10.05 ± 5.79, 9.56 ± 4.95, and 8.10 ± 5.96 (p < 0.0001), with CMI rates of 61.90%, 69.57%, 85.71%, and 86.96% at corresponding time points. Mixed-effects modeling demonstrated a significant group × time interaction (F4,205 = 5.19, p = 0.01), indicating differential trajectories of ADL improvement between the thymoma-associated MG (TAMG) and non-TAMG groups (Table 2). Baseline MG-ADL scores were 1.99 points higher in the TAMG cohort (9.38 ± 0.72 vs. 7.39 ± 0.59). During the first week, patients with TAMG exhibited a pronounced 3.56-point reduction (95% CI 1.53–5.58), nearly double the 1.77-point reduction observed in the non-TAMG group (95% CI 0.12–3.42). This divergence persisted over time: by week 8, the mean reduction was 7.18 points in the TAMG group versus 3.48 points in the non-TAMG group, and at the last follow-up, reductions were 6.88 and 3.59 points, respectively (all within-group p < 0.001). Simple-effects analyses confirmed significant improvement in both subgroups (TAMG: F4,203 = 35.42; non-TAMG: F4,205 = 12.23; both p < 0.001). Between-group differences at individual post-baseline visits ranged from − 1.71 to 0.20 points and did not reach statistical significance (p ≥ 0.11). Both thymoma-associated and non-thymomatous MG groups demonstrated CMIs in ADL scores, with a more rapid early response observed in the TAMG cohort. Table 2ADL scores of the TAMG and non-TAMG cohorts across five time points (Bonferroni-adjusted)VisitTAMG (n = 24)Non-TAMG (n = 36)Between-group difference in reduction from baseline (TAMG − non-TAMG)Mean ADL ± SDReduction from baseline (95% CI)Mean ADL ± SDReduction from baseline (95% CI)Δ (95% CI)pBaseline9.38 ± 0.727.39 ± 0.59W15.80 ± 0.733.56 (1.53–5.58)†5.62 ± 0.601.77 (0.12 to 3.42)*1.79 (− 0.54 to 4.12)0.83W43.00 ± 0.726.38 (4.37–8.38)4.31 ± 0.643.08 (1.45 to 4.72)†3.30 (0.87–5.73)0.16W82.33 ± 0.897.18 (4.98–9.37)†3.93 ± 0.803.48 (1.56 to 5.41)†3.70 (1.14–6.26)0.11WL2.50 ± 0.726.88 (4.87–8.88)†3.78 ± 0.713.59 (1.84 to 5.34)†3.29 (0.87–5.71)0.18CI confidence interval, WL week of last follow-up†p < 0.001 versus baseline (Bonferroni-corrected pairwise comparisons at each visit)*p = 0.026 versus baseline ADL scores of the TAMG and non-TAMG cohorts across five time points (Bonferroni-adjusted) CI confidence interval, WL week of last follow-up †p < 0.001 versus baseline (Bonferroni-corrected pairwise comparisons at each visit) *p = 0.026 versus baseline Consistent with the trend observed for MG-ADL scores, mixed-effects modelling demonstrated a significant group × time interaction for QMG scores (F4,195.7 = 6.50, p < 0.001) (Table 3). The TAMG cohort exhibited a pronounced decline, with mean QMG scores decreasing from 16.3 ± 1.2 at baseline to 7.5 ± 1.2 at the last follow-up (mean reduction, 8.8 points; p < 0.001 for all post-baseline time points). In contrast, the non-TAMG group showed a reduction from 13.2 ± 1.0 to 8.7 ± 1.1 (maximum reduction, 4.5 points; p ≤ 0.01 from week 4 onward). Consequently, between-group differences at individual follow-up visits ranged from −2.41 to −0.30 points, with 95% CIs consistently crossing zero (p ≥ 0.05). Table 3Longitudinal change in QMG scores from the baselineVisitTAMG (n = 24)Non-TAMG (n = 36)Between-group difference (TAMG − non-TAMG)Mean QMG ± SDReduction from baseline (95% CI)Mean QMG ± SDReduction from baseline (95% CI)Δ (95% CI)pBaseline16.29 ± 1.23–13.20 ± 1.02––W111.19 ± 1.255.11 (2.62–7.59)†11.49 ± 1.031.71 (− 0.34 to 3.76)−0.30 (−3.52 to 2.92)0.85W48.50 ± 1.237.79 (5.37–10.21)†9.85 ± 1.033.35 (1.32–5.38)†− 1.35 (− 3.69 to 0.99)0.40W86.94 ± 1.299.36 (6.70–12.01)†9.35 ± 1.123.85 (1.46–6.25)†− 2.41 (− 4.96 to 0.14)0.16WL7.54 ± 1.238.75 (6.33–11.17)†8.72 ± 1.074.48 (2.28–6.67)†− 1.18 (− 3.63 to 1.27)0.47CI confidence interval, WL week of last follow-up†p < 0.001 versus baseline (Bonferroni-corrected pairwise comparisons at each visit) Longitudinal change in QMG scores from the baseline CI confidence interval, WL week of last follow-up †p < 0.001 versus baseline (Bonferroni-corrected pairwise comparisons at each visit) Over the 8-week treatment period, patients with TAMG—who began with higher baseline scores (ADL 9.38 ± 4.42, QMG 16.82 ± 7.04)—demonstrated more rapid improvements compared to patients with non-TAMG. ADL scores in patients with TAMG decreased to 2.72 ± 2.65 at week 8 and 2.50 ± 3.43 at week last follow-up (WL) (both p < 0.0001), whereas patients with non-TAMG showed a slower decline to 4.14 ± 3.10 (p = 0.0002) and 3.79 ± 3.23 (p < 0.0001). Similarly, QMG scores in patients with TAMG fell to 8.53 ± 4.61 at week 8 and 7.78 ± 6.34 at WL (both p < 0.0001), while patients with non-TAMG maintained higher scores at 9.71 ± 5.27 (p = 0.0013) and 8.81 ± 5.74 (p < 0.0001). Patients with TAMG consistently demonstrated higher CMI rates than patients with non-TAMG at early time points. ADL-CMI rates were 73.91% versus 25.71% at week 1, 91.67% versus 47.22% at week 4, and 100% versus 60.87% at week 8. Corresponding QMG-CMI rates were 77.27% versus 45.45%, 91.67% versus 61.76%, and 100% versus 71.43%, respectively. By the last follow-up, this difference had narrowed (91.67% vs. 74.07%) (Fig. 3). Thus, although both groups ultimately achieved substantial clinical improvement, the TAMG cohort demonstrated a more rapid and pronounced early response. Myasthenia Gravis Activities of Daily Living (ADL) and Quantitative Myasthenia Gravis (QMG) score changes after initiation of eculizumab in the thymoma-associated MG (TAMG) and non-TAMG groups. a Changes in ADL score. The dark blue bars represent the mean ADL score for the entire cohort at baseline, weeks 1, 4, 8, and the final follow-up in the TAMG group; light blue bars represent the non-TAMG group. b Changes in QMG score. Symbols indicate the mean QMG score for the TAMG group (dark blue) and the non-TAMG group (light blue) at the corresponding time points. c Dark blue bars represent the percentage of patients in the TAMG group meeting the clinically meaningful improvement (CMI) criteria (≥ 2-point reduction in MG-ADL) at post-baseline time points (week 1, week 4, week 8, and last follow-up). Light blue bars represent the non-TAMG group at the corresponding time points. d Dark blue bars represent the percentage of patients in the TAMG group achieving CMI in QMG score (≥ 3-point reduction) at post-baseline time points (week 1, week 4, week 8, and last follow-up). Light blue bars represent the non-TAMG group at the corresponding time points In our cohort, eculizumab treatment was associated with a sustained reduction in corticosteroid dosage throughout the observation period. Among the 60 treated patients, the mean daily prednisone dosage decreased significantly from 21.21 ± 13.68 mg at baseline to 14.23 ± 9.85 mg at week 8 (p = 0.0008) and further to 13.32 ± 10.42 mg at the final follow-up (p < 0.0001). Glucocorticoids remain the first-line therapy for MG, and at enrollment, 90% of participants were steroid-dependent. Introduction of eculizumab was accompanied by a rapid and sustained reduction in steroid burden. Among the 23 patients receiving long-term eculizumab therapy, prednisone doses declined from 19.02 ± 13.37 mg at baseline to 14.35 ± 10.80 mg at week 8 (p = 0.0617) and to 11.25 ± 9.81 mg at the final assessment (p = 0.0016). Similarly, in the subgroup with TAMG, mean daily glucocorticoid doses decreased from 21.67 ± 12.48 mg at baseline to 13.89 ± 9.93 mg at week 8 (p = 0.0303) and to 12.28 ± 9.68 mg at the final follow-up (p = 0.0009). Compared with baseline, when 28.33% (17/60) of patients were receiving ≤ 10 mg/day of prednisone, a significantly higher proportion had achieved this threshold at the last follow-up (56.0%, 28/50) (Fig. 4). Conversely, the proportion receiving ≥ 20 mg/day declined from 43.33% (26/60) to 22.0% (11/50). A similar pattern was observed among the 23 patients maintained on long-term eculizumab therapy, in whom 63.64% (14/22) were receiving ≤ 10 mg/day at the final follow-up compared with 34.78% (8/23) at baseline (p = 0.039). The group × time interaction was not statistically significant (F2,98.7 = 0.32, p = 0.73), indicating comparable steroid-tapering trajectories between TAMG and non-TAMG groups. Nevertheless, both groups exhibited a significant overall reduction in corticosteroid exposure over time (F2,98.7 = 15.9, p < 0.001). Pairwise comparisons confirmed that mean daily dose reduced from baseline to the final assessment in each group (TAMG: 21.7 → 12.7 mg, p < 0.001; non-TAMG: 20.9 → 14.2 mg, p = 0.001), whereas between-group differences calculated at any single visit were negligible (all p ≥ 0.65). Linear mixed-effects modelling further demonstrated that during the first treatment year, patients with TAMG achieved steroid tapering comparable to that of non-thymomatous patients, indicating equivalent treatment responsiveness. Proportions of subjects with different daily corticosteroid doses from enrollment to the last follow-up. a Percentage distribution of daily prednisone doses (0–10 mg, 10–20 mg, 20–30 mg, 30–40 mg, > 40 mg) across the entire cohort of 60 patients with myasthenia gravis (MG). b Percentage distribution of daily prednisone doses among the 23 patients who maintained continuous eculizumab therapy throughout the observation period. c Proportions of patients with thymoma-associated MG, stratified by corticosteroid dose category, throughout the study period. d Proportions of patients with non-thymoma-associated MG, stratified by corticosteroid dose category, throughout the study period Notably, no patient required a daily prednisone dose exceeding 40 mg at either week 8 or the final assessment. Two patients achieved complete corticosteroid withdrawal by week 8, and six had discontinued steroids entirely by the final follow-up, indicating a sustained downward shift in corticosteroid exposure across the observation period. Within the cohort, the 37 patients who switched from efgartigimod to eculizumab were demographically and clinically indistinguishable from the entire group of 60 eculizumab-treated patients (Table 4). The switch subgroup and the overall group showed nearly identical distributions of EOMG (62.16% vs. 61.67%) and LOMG (37.83% vs. 38.33%), as well as comparable rates of TAMG (48.65% vs. 40%) and non-TAMG (51.35% vs. 60%) diseases. MGFA severity at enrollment was balanced: the majority of patients in both populations presented with class II–IV, with only isolated class I or V cases recorded. Baseline AChR-antibody titers, ADL, and QMG scores were likewise similar, indicating that the patients who switched treatment at later stages entered efgartigimod treatment with the same clinical burden as the broader eculizumab population. In both groups, the reductions in MG-ADL scores were highly synchronous across all time points. From baseline to week 4, the MG-ADL score decreased from 8.49 ± 3.86 to 3.76 ± 3.47 (a reduction of 4.73) in the efgartigimod-to-eculizumab switch group, and from 8.18 ± 3.99 to 3.78 ± 3.42 (a reduction of 4.40) in the eculizumab group. At week 8, the scores further decreased to 3.56 ± 2.96 (reduction of 4.93 from baseline) in the switch group and to 3.50 ± 2.92 (reduction of 4.68) in the eculizumab group. At the last follow-up, the scores were 3.38 ± 3.37 (reduction of 5.11) and 3.33 ± 3.56 (reduction of 4.85), respectively. Regarding corticosteroid reduction, baseline doses were comparable between the two groups (20.72 mg vs. 21.21 mg). At week 8, doses decreased to 14.45 mg and 14.23 mg, respectively, and further declined to 12.35 mg and 13.32 mg at the last follow-up, demonstrating nearly identical tapering trajectories. Overall, these data demonstrate that the treatment effect in the efgartigimod-to-eculizumab switch group is non-inferior to that in the eculizumab therapy group, with comparable corticosteroid-sparing effects observed in both groups. Of note, although some patients did not achieve rapid or adequate clinical improvement during initial FcRn inhibitor therapy, subsequent switching to eculizumab still resulted in effective symptom control and corticosteroid reduction. Among the 37 patients transitioning from efgartigimod to eculizumab, nearly half (48.65%) did so within the first 4 weeks of efgartigimod initiation, reflecting early inadequate response or loss of efficacy (Table 5). An additional 21.62% switched between week 4 and 8, whereas only 8.11% transitioned at week 8–12, 11% at week 12–24, and 14% beyond week 24. Thus, the majority of treatment transitions (70.27%) occurred within the first 2 months, underscoring the importance of early therapeutic reassessment in patients with an insufficient response. Table 4Demographic characteristics of patients who switched from efgartigimod to eculizumab versus the overall cohortSwitch groupEculizumab groupEOMG62.16% (23/37)61.67% (37/60)LOMG37.83% (14/37)38.33% (23/60)TAMG48.65% (18/37)40% (24/60)Non-TAMG51.35% (19/37)60% (36/60)MGFA classification at enrollment I05.00% (3/60) II35.14% (13/37)38.33% (23/60) III24.32% (9/37)20.00% (12/60) IV37.84% (14/37)33.33% (20/60) V2.70% (1/37)3.33% (2/60)Antibody status AChR antibody titer (nmol/L)40.59 ± 50.8230.79 ± 47.61ADL score, median ± SD Baseline8.49 ± 3.868.18 ± 3.99 W43.76 ± 3.473.78 ± 3.42 W83.56 ± 2.963.50 ± 2.92 WL3.38 ± 3.373.33 ± 3.56QMG score, median ± SD Baseline16.09 ± 5.9714.48 ± 6.59 W49.33 ± 5.649.24 ± 5.63 W89.21 ± 4.909.00 ± 4.96 WL8.38 ± 6.198.18 ± 6.24Mean corticosteroid dosage, mg Baseline20.72 ± 13.2621.21 ± 13.68 W814.45 ± 9.8714.23 ± 9.85 WL12.35 ± 9.9813.32 ± 10.42 Demographic characteristics of patients who switched from efgartigimod to eculizumab versus the overall cohort Table 5Number of patients who switched from efgartigimod to eculizumab at different time intervals after efgartigimod initiationTime interval (weeks)NumberPercent0–41848.65%4–8821.62%8–1238.11%12–24410.81%24–36513.51% Number of patients who switched from efgartigimod to eculizumab at different time intervals after efgartigimod initiation In this retrospective cohort of patients with myasthenia gravis treated with eculizumab, the overall incidence of treatment-emergent adverse events (TEAEs) was 8.3% (5/60). All AEs were classified according to CTCAE version 5.0. Adverse events considered attributable to eculizumab were observed in five patients. One patient developed a low-grade fever (38.3 °C) accompanied by a mild headache 24 h after the initial infusion. The next scheduled dose was temporarily withheld, and no recurrence was observed upon resuming therapy. One patient developed a low-grade fever (38.3 °C) accompanied by a mild headache 24 h after the initial infusion. The next scheduled dose was temporarily withheld, and no recurrence was observed upon resuming therapy. One patient experienced an upper respiratory tract infection that resolved promptly with empirical treatment and without complications. One patient experienced an upper respiratory tract infection that resolved promptly with empirical treatment and without complications. A single episode of pulmonary embolism occurred during the treatment period; the patient received anticoagulation therapy and experienced no further thrombotic events. A single episode of pulmonary embolism occurred during the treatment period; the patient received anticoagulation therapy and experienced no further thrombotic events. One patient developed recurrent urinary tract infections requiring repeated antibiotic courses. One patient developed recurrent urinary tract infections requiring repeated antibiotic courses. Another patient experienced recurrent post-infusion fevers accompanied by positive blood cultures; sepsis was promptly diagnosed, leading to the permanent discontinuation of eculizumab. Four AEs (low-grade fever/headache [Grade 1], upper respiratory infection [grade 2], pulmonary embolism [grade 3], and recurrent urinary tract infections [grade 2]) were deemed unrelated to eculizumab. One AE (sepsis grade 4]) was considered treatment-related, yielding a treatment-related AE incidence of 1.7% (1/60), which led to permanent discontinuation. No grade 5 events occurred. Another patient experienced recurrent post-infusion fevers accompanied by positive blood cultures; sepsis was promptly diagnosed, leading to the permanent discontinuation of eculizumab. Four AEs (low-grade fever/headache [Grade 1], upper respiratory infection [grade 2], pulmonary embolism [grade 3], and recurrent urinary tract infections [grade 2]) were deemed unrelated to eculizumab. One AE (sepsis grade 4]) was considered treatment-related, yielding a treatment-related AE incidence of 1.7% (1/60), which led to permanent discontinuation. No grade 5 events occurred. One 63-year-old man diagnosed with anti-AChR antibody-positive gMG, with a disease duration of 7 years, had previously received corticosteroids, tacrolimus, rituximab, and IVIG. Although initiation of efgartigimod resulted in transient clinical improvement, symptoms subsequently worsened, prompting initiation of eculizumab after 6 months. After 1 month of complement inhibition, no meaningful clinical improvement was observed, necessitating rescue treatment with intravenous immunoglobulin. Three months later, the patient died as a consequence of myasthenic complications. This fatal outcome was not considered related to the adverse effects of eculizumab. All patients received meningococcal vaccination prior to initiation of eculizumab therapy. Despite the recognized risk of meningococcal infection associated with terminal complement inhibition, no cases of invasive meningococcal disease were observed in this cohort. During the final telephone follow-up period, one patient was reported to have died from a MG-related pulmonary infection approximately 2 months after the last eculizumab infusion. This event was assessed as unrelated to eculizumab treatment. Discussion The introduction of targeted biologic therapies—most notably eculizumab—has therefore represented a major therapeutic advance in the management of MG [25]. In this context, the present multicenter retrospective study evaluated the clinical efficacy, glucocorticoid-sparing potential, and safety profile of eculizumab in patients with MG. Before initiation of eculizumab, all enrolled patients (n = 60) had been treated with corticosteroids or at least one conventional immunosuppressive agent. A total of 10 patients were refractory despite sequential exposure to two or more non-steroidal immunosuppressants, whereas 37 patients transitioned to eculizumab following inadequate disease control with efgartigimod. These findings underscore the substantial therapeutic burden and unmet clinical need within this cohort. Although the PIS framework is widely used to assess disease control in MG, attainment of MMS or better remains challenging in routine clinical practice [26]. Following 8 weeks of eculizumab therapy, 30.95% of patients had achieved MMS or better, and 42.86% demonstrated clinical improvement; by the final follow-up, these proportions were 32.69% and 40.38%, respectively. This near-overlap indicates that clinical benefit plateaus rapidly and is maintained throughout the extended observation period [22]. The MG-ADL score declined from a baseline median of 8.18 ± 3.99 to 3.50 ± 2.92 at week 8 (p < 0.0001). Similarly, QMG scores decreased from 14.48 ± 6.59 at baseline to 9.00 ± 4.96 at week 8. Our retrospective eculizumab cohort demonstrated robust efficacy. In this real-world cohort, 85% of patients achieved CMI in MG-ADL, and 76.92% achieved CMI in QMG by week 8, exceeding the corresponding response rates of 35–45% (MG-ADL) and 20–30% (QMG) reported in real-world studies of ravulizumab and efgartigimod over comparable 8- to 10-week treatment duration [27]. Real-world datasets are inherently more heterogeneous; they encompass OMG, patients with mild symptoms and low baseline ADL scores, and therapeutically complex TAMG [28]. This diversity in clinical manifestations inevitably amplifies the variance observed in any population-level analysis. Among the 23 patients who continued long-term eculizumab therapy, the median follow-up duration was 25.74 ± 13 weeks (range 10–49). At the final assessment, two patients had completely discontinued corticosteroids, four were maintained on glucocorticoid monotherapy, two were maintained exclusively on tacrolimus, and 14 remained on combination therapy consisting of prednisolone plus one steroid-sparing agent. Overall, 63.64% of patients achieved a daily prednisone dose of ≤ 10 mg. Across the cohort, the median daily prednisolone dose declined significantly from 19 mg at baseline to 11 mg at the last follow-up—representing a robust 42% reduction (8 mg absolute decrease). Notably, this extent of glucocorticoid tapering proved very similar to that observed in another retrospective cohort receiving efgartigimod for MG, with the mean dosage decreasing notably from 22 to 15 mg by week 12, marking a significant 32% reduction—equivalent to an absolute decrease of 7 mg [29]. Collectively, these findings demonstrate that eculizumab not only arrests disease activity but also demonstrates a sustained, clinically relevant steroid-sparing effect throughout the first treatment year. TAMG remains a particularly challenging disease subtype, characterized by therapeutic refractoriness and prolonged dependence on glucocorticoids, often necessitating exploration of alternative treatment paradigms [30]. In a prior prospective study involving 24 patients with TAMG, the mean prednisone-equivalent dose declined from 41 to 23 mg over a 12-week period, accompanied by a reduction in the ADL score from 11.7 to 4.0 [27]. In the present cohort, a similar trajectory was observed: by week 8, the mean daily steroid dose had decreased from 21.67 mg at baseline to 13.89 mg. Thereafter, steroid doses declined only modestly, with no statistically significant change beyond week 8 (p = 0.7194), indicating that clinical improvement plateaued after the second month. At the final assessment, approximately 50% of patients were maintained on a daily prednisone dose of ≤ 10 mg. Two patients with persistent disease activity were transitioned to rituximab therapy. A total of 12 patients were maintained on complement inhibition alone. One patient experienced disease exacerbation, whereas 75% (9/12) achieved MMS or better status, and 16.67% (2/12) were classified as improved. Corticosteroids were discontinued entirely in two individuals, and all patients were maintained on daily prednisone doses below 20 mg. Concomitant therapies included low-dose corticosteroids in two patients, monotherapy with an additional immunosuppressant in two patients, and combination regimens involving corticosteroids and another immunosuppressive agent in seven patients. Consistent with prior reports, disease control in TAMG remained challenging, often necessitating combination therapy with corticosteroids, immunosuppressants, and eculizumab [27, 31, 32]. Although clinical data remain limited, accumulating evidence, including the data presented in this study, shows that eculizumab delivers significant clinical improvement and enables corticosteroid reduction in patients with TAMG. This cohort exemplifies a population with highly refractory MG. All patients had previously received conventional immunosuppressive therapies, yet adequate disease control had not been achieved, and 90% remained dependent on glucocorticoids (mean prednisone dose 21 mg/day). Two-thirds of the patients had received tacrolimus, one-third mycophenolate, and four patients had additionally been treated with rituximab, yet stable disease control remained elusive. In the year preceding eculizumab initiation, 48.33% of patients required plasma exchange and 16.67% required intravenous immunoglobulin, underscoring the cohort’s severity and treatment resistance. Notably, 61.67% of patients had already demonstrated an inadequate response to efgartigimod, the most recently introduced targeted therapy, before eculizumab was initiated. Previous studies have also investigated treatment switching strategies. In the Huashan TAMG study, approximately 40% of patients with TAMG switched from efgartigimod to eculizumab because of poor initial response or high corticosteroid burden, achieving significant clinical improvement and corticosteroid reduction, with non-inferior efficacy compared to the non-switch group [31]. Similarly, the larger SPOTLIGHT registry study found that over one-third of patients with gMG who switched from eculizumab to ravulizumab—a long-acting complement C5 inhibitor—maintained or even slightly improved their clinical outcomes. Specifically, the proportion of patients achieving minimal symptom expression (MSE) increased from 27.0% before the switch to 35.1% after the switch, and 64% of patients preferred the every-8-week dosing regimen of the longer-acting agent [33]. Consistent with these findings, our cohort demonstrated that baseline characteristics of the switch group were generally comparable to those of the eculizumab monotherapy group, and that both groups achieved comparable or even superior symptom control and corticosteroid reduction after switching. Furthermore, 70.27% of switches occurred within the first 2 months, underscoring the importance of close clinical monitoring during the initial 8 weeks of FcRn inhibitor therapy. Together, these results confirm that eculizumab serves as an effective rescue therapy for patients who show an inadequate response to novel targeted agents. Limited evidence exists concerning the efficacy of immunosuppressive therapy in altering the natural progression of OMG [34–36]. In the present cohort, two patients with OMG failed to respond to eculizumab treatment with no discernible benefit and discontinued therapy by week 8. In contrast, a third patient with anti-AChR-positive OMG achieved MMS within 4 weeks of initiating eculizumab in combination with tacrolimus; tacrolimus monotherapy subsequently maintained PIS-defined MMS throughout extended follow-up. A recent study by Shi et al. specifically confirmed significant and rapid improvement with eculizumab in refractory OMG and ocular-predominant MG, further supporting its clinical benefit in this population [37]. Although only one patient in our cohort showed an encouraging treatment response, this finding suggests that eculizumab may still represent a future therapeutic option for selected patients with OMG. Nevertheless, owing to the small number of OMG cases, no definitive conclusions regarding treatment responsiveness can be drawn and larger-scale studies are warranted to further evaluate its efficacy. One death occurred in a patient with refractory gMG who had previously received multiple immunosuppressive agents, including corticosteroids, conventional immunosuppressants, several biologics, and repeated courses of IVIG without disease control. The fatal outcome was attributed to disease-related complications and was not considered related to the treatment. Among the 60 enrolled patients, five adverse events were reported, and only one patient discontinued therapy because of recurrent sepsis. Overall, eculizumab demonstrated an acceptable safety profile in this cohort. This study has inherent limitations. Its retrospective design and relatively small sample size restricted statistical power and limited the ability to perform robust subgroup analyses. In addition, reliance on routine clinical documentation resulted in incomplete longitudinal data for some patients, particularly regarding precise timing of disease fluctuations and the stepwise tapering—or discontinuation—of glucocorticoids and conventional immunosuppressants. These limitations constrained the ability to construct definitive treatment algorithms for individualized therapy. Conclusions This retrospective study evaluated the efficacy and safety of adding eculizumab to fundamental therapy in patients with MG. Findings underscore significant improvements in clinical status, functional outcomes, and glucocorticoid burden, alongside a favorable tolerability profile. These results further support the role of complement inhibition as a viable strategy for managing refractory autoimmune neuromuscular disorders. Acknowledgements We thank the participants of the study. We also thank the patients and their families for their valuable contribution and support. Author Contributions Zhijun Li, Ting Chang and Bitao Bu designed the study and methods. Formal analysis was done by Yue Li and Zhijun Li. Clinical data were by provided Bitao Bu, Zhijun Li, Ting Chang, Huan Yang, Song Tan, Lijun Luo, Xuebing Cao and JinQuan Hu. Yue Li and Zhijun Li wrote the original draft. Bitao Bu, Zhijun Li, Ting Chang, Huan Yang, Song Tan, Lijun Luo, Xuebing Cao, and JinQuan Hu reviewed and edited the manuscript. Figures were created by Yue Li and Zhijun Li,. Supervision by Ting Chang and Huan Yang Guarantor of the study: Bitao Bu. All authors read and approved the final manuscript. Funding This work was supported by the Development and Transformation Fund of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology (Grant No. 25-2ZJZ08002-02).The Rapid Service Fee was covered by the corresponding author's research fund (Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology; Grant No. 25-2ZJZ08002-02). Data Availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Conflict of Interest Yue Li, Huan Yang, Song Tan, Lijun Luo, Xuebing Cao, JinQuan Hu, Bitao Bu, Ting Chang, and Zhijun Li declare no competing interests. Ethical Approval The protocol was approved by all the necessary ethics committees and institutional review boards at each trial site. The institutional review board approval number from the leading institution for this study (Tongji Hospital) is TJ-JRB202411058. This retrospective study was granted a waiver of informed consent because it used de-identified data. All procedures rigorously adhered to the principles of the Declaration of Helsinki. Contributor Information Ting Chang, Email: changting1981@163.com. Zhijun Li, Email: lizhijun@tjh.tjmu.edu.cn. References Associated Data Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.