Combined Double-Filtration Plasmapheresis and Corticosteroids are Associated with Improved Outcomes in Acute Myelin Oligodendrocyte Glycoprotein Antibody-Associated Disease: An Ambispective Multicenter Cohort Study

Abstract

Background

Double-filtration plasmapheresis (DFPP) and intravenous methylprednisolone (IVMP) are established treatments for acute myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD) attacks, but comparative efficacy data remain limited. This study aimed to compare the efficacy between IVMP monotherapy and DFPP+IVMP therapy in acute MOGAD and identify factors associated with treatment response.

Methods

Patients with acute MOGAD attacks receiving DFPP+IVMP or IVMP alone were enrolled in the ambispective, multicenter cohort study. Primary outcome was change in Expanded Disability Status Scale (EDSS) score. Secondary outcome included absolute EDSS and modified Rankin Scale (mRS) score, the change in mRS, improvement rates of EDSS and mRS, immunological parameters, and adverse events.

Results

36 MOGAD patients were enrolled (DFPP+IVMP: 15; IVMP: 21). At 3-month follow-up (3M-FU), the DFPP+IVMP group showed greater EDSS reduction [2.00 (1.00, 2.00) vs 1.00 (0.75,1.75), p=0.012] and mRS reduction [2.00 (1.00, 3.00) vs 1.00 (1.00,1.50), p=0.009] compared with IVMP alone. In the IVMP group, higher baseline EDSS score correlated with greater EDSS improvement (ρ=0.468, p=0.032). In the DFPP+IVMP group, a higher baseline EDSS score and higher serum IgG were associated with greater EDSS improvement (baseline EDSS: ρ=0.738, p=0.002; serum IgG: ρ=0.685, p=0.010). Meanwhile, longer delay to IVMP and DFPP initiation were associated with reduced benefit (delay from onset to IVMP: ρ=−0.705, p=0.003; delay from onset to DFPP: ρ=−0.660, p=0.007). In high-serum IgG patients, DFPP+IVMP showed larger EDSS and mRS improvement at 3M-FU compared with IVMP [ΔEDSS: 3.50 (1.75, 5.00) vs 1.00 (1.00, 1.00), p=0.003; ΔmRS: 3.50 (2.00, 4.00) vs 1.00 (1.00, 1.00), p=0.003], while no significant difference was observed in low-IgG patients. No severe adverse events led to treatment discontinuation.

Conclusion

DFPP+IVMP was associated with better functional recovery at 3M-FU compared to IVMP group for acute MOGAD, particularly in patients with high baseline serum IgG, and early intervention was important.

Graphical Abstract

Introduction

Myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD) is a distinct inflammatory demyelinating disorder of the central nervous system, recognized as separate from multiple sclerosis and aquaporin-4 (AQP4) antibody seropositive neuromyelitis optica spectrum disorder (NMOSD) based on clinical phenotype, imaging, laboratory and pathological findings.1,2 The prevalence of MOGAD is approximately 1.3–2.5/100,000, and the annual incidence is approximately 3.4–4.8 per million.3 The disease affects both adults and children, with no clear sex or racial predominance.2,4 Clinically, MOGAD is characterized by optic neuritis, transverse myelitis, acute disseminated encephalomyelitis, cerebral cortical encephalitis, or brainstem and cerebellar syndromes.5 Pathogenic MOG-autoantibodies are believed to mediate demyelination, leading to either a monophasic or relapsing disease course.6

First-line treatments for acute MOGAD attacks include high-dose intravenous methylprednisolone (IVMP), therapeutic plasma exchange (TPE), and intravenous immunoglobulin.7 Among them, IVMP is most widely used owing to the good anti-inflammatory effects. Most MOGAD attacks respond well to corticosteroids, but a subset of MOGAD patients is steroid-refractory and have persistent residual deficits. TPE is recommended for its ability to remove pathogenic antibodies and other pro-inflammatory factors from the circulation. In NMOSD, an inflammatory demyelinating disease mediated by AQP4 autoantibodies, TPE has shown significant improvement in clinical outcomes, and is strongly recommended for the treatment of NMOSD attacks.8,9 By contrast, evidence on TPE for MOGAD remains relatively limited. Several retrospective studies have reported that TPE yielded favorable clinical improvement and emphasized the importance of early initiation of TPE.10–12

Double-filtration plasmapheresis (DFPP) is a semi-selective membrane-based apheresis technique that removes high-molecular-weight immunoglobulins while conserving beneficial plasma components. This approach eliminates the need for exogenous plasma replacement, thereby reducing the risk of transfusion-related infections and enhancing tolerability.13 Several studies have demonstrated the efficacy and safety in the treatment of NMOSD, autoimmune encephalitis, myasthenia gravis,14–16 but its efficacy and safety in MOGAD still lack robust evidence. Additionally, limited studies have directly compared the efficacy of DFPP combined with IVMP versus IVMP alone in MOGAD and factors associated with DFPP still require further investigation.

In this ambispective, multicenter cohort study, we compared functional outcomes between IVMP monotherapy and a combined DFPP+IVMP therapy in patients with acute MOGAD attacks. We also explored the factors associated with treatment response in IVMP and DFPP+IVMP groups. Moreover, immunological changes and DFPP-related adverse events were summarized. Our findings aim to clarify whether DFPP combination therapy enhances recovery better and to provide insights into application of DFPP in MOGAD.

Methods

Ethics

The retrospective cohort study protocol received approval from the Ethics Committee of Renji Hospital, Shanghai Jiao Tong University School of Medicine in January 2026 (Retrospective cohort period: January 2015 to December 2025. Number: LY2025-435-B). The prospective cohort study received approval from the Ethics Committee of Renji Hospital, Shanghai Jiao Tong University School of Medicine in November 2022 (Prospective cohort period: November 2022 to December 2025. Number: LY2022-051-B). The study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants.

Study Design

This was an ambispective, multi-center, observational cohort study conducted between January 2015 and December 2025. This ambispective study combined prospectively collected data from a cohort of patients with inflammatory demyelinating diseases (initiated in November 2022) and retrospectively identified cases from electronic medical records (from 2015 to October 2022). Diagnosis of MOGAD was retrospectively adjudicated by two neurologists according to the 2023 international diagnostic criteria,5 based on clinical presentation, MRI findings, and positive serum MOG-IgG testing using cell-based assay. This study aimed to evaluate the efficacy of combination use of DFPP and IVMP compared with IVMP alone in the treatment of patients with MOGAD acute attacks.

Study Participants

Patients were enrolled from Renji Hospital, School of Medicine, Shanghai Jiao Tong University; Punan Branch of Renji Hospital, School of Medicine, Shanghai Jiao Tong University; and Anhui No. 2 Provincial People’s Hospital between January 2015 and December 2025. Inclusion criteria were: (1) diagnosis meeting the 2023 Criteria;5 (2) with acute attack (including patients with first onset and relapse; relapse was defined as a new clinical episode or worsening of pre-existing symptoms occurring at least 30 days after the onset date of the last acute attack); (3) use of IVMP alone or DFPP+IVMP during the acute phase; (4) with complete clinical data. Exclusion criteria were: (1) receipt of other treatments during the acute phase (such as intravenous immunoglobulin); (2) with heart, liver, lung, and renal failure, severe infections and malignant tumors. For patients who experienced multiple attacks during the study period, only the first attack was included to avoid statistical dependency from repeated measurements. Serum MOG-IgG testing was performed using a commercial live cell-based assay using full-length human MOG before treatment initiation.

Treatment

Treatment Protocol Selection

In this observational cohort, the treatment strategy for an acute attack was determined by the treating physician based on the attack’s severity, clinical phenotype, and the patient’s prior treatment response. Patients with severe attacks (eg, high baseline mRS/EDSS, severe visual loss, limb weakness, or impaired consciousness) were more likely to receive the combination therapy (DFPP+IVMP). Some patients refused plasma exchange due to its invasiveness and received corticosteroid therapy alone. All treatment decisions followed a pre-specified institutional protocol.

Double-Filtration Plasmapheresis Procedure

Patients in the DFPP+IVMP group underwent DFPP according to a standardized protocol. The DFPP initiated 2–5 days after the first IVMP dose. The treatment employed advanced apheresis platforms, including the Plasauto Σ (Asahi Kasei Medical) and KM-9000 (Kawasumi) systems, in conjunction with specialized Plasmacure™/Plasmafo™ plasma separators and Evafux™/Cascadefo™ EC component filters, enabling selective depletion of high-molecular-weight pathogenic substances. The plasma exchange volume was set at 1.0–1.5 times the individual’s estimated plasma volume. Anticoagulation was achieved using unfractionated heparin or nafamostat mesilate. Each session lasted approximately 2–3 hours and was repeated every other day. The total number of sessions (typically 3–5) was determined by the clinical response, tolerability, and adverse events. Safety laboratory tests were performed within 24 hours after each session.

Corticosteroid Therapy

All patients received an initial IVMP dose of 500–1000 mg daily for 3–5 days, followed by halving the dose every 3 days until reaching 60 mg once daily for oral administration. After switching to oral prednisone, the dosage was tapered by 5 mg every two weeks until complete discontinuation, or was maintained at a daily dose of 5–10 mg. Concomitant prophylactic medications (eg, proton pump inhibitors, calcium/vitamin D supplementation) were administered to mitigate steroid-related adverse effects.

Detection of Immunological Parameters

Immunological parameters were detected at baseline prior to treatment initiation for all enrolled patients. For participants receiving DFPP+IVMP, immune profiles were re-examined after completion of the final DFPP treatment to capture serial dynamic changes induced by apheresis. In contrast, nearly all patients receiving IVMP monotherapy only underwent a single baseline immune measurement; repeat immunological testing at discharge was not routinely performed and was unavailable for the majority of this group. Scatter turbidimetry was used to measure the levels of immunoglobulin and complement components. Flow cytometry was used to measure the levels of serum cytokines and lymphocyte subsets. All tests were conducted in clinical laboratory in hospital on the day of sample collection according to standardized process.

Outcome Measure Definition

Clinical assessments, including the Expanded Disability Status Scale (EDSS) and modified Rankin Scale (mRS), were performed at admission, discharge, and 3-month follow-up (3M-FU). All EDSS and mRS ratings were completed by neurologists trained in unified rating criteria, blinded to patients’ assigned treatment regimen to reduce assessment bias.

Treatment response was evaluated by the change in EDSS and mRS scores from baseline. The primary outcome was the change in EDSS score (ΔEDSS) at discharge and 3M-FU. The primary endpoint was defined as ΔEDSS at 3M-FU. The 3M-FU was selected as the primary assessment time point because neurological status is generally stable by this stage, allowing assessment of residual disability and comparison with previous apheresis studies.12,17 Secondary outcomes included: (1) absolute EDSS scores and EDSS improvement rates at discharge and 3M-FU; (2) absolute mRS scores, change in mRS (ΔmRS), and mRS improvement rates at discharge and 3M-FU; (3) immunological parameters (immunoglobulins, complement, lymphocyte subsets, cytokines); and (4) adverse events, graded according to the Common Terminology Criteria for Adverse Events (CTCAE) v5.0. Improvement rate was defined asΔEDSS or mRS ÷ EDSS or mRS at admission.

Statistical Analysis

Statistical analysis was conducted using SPSS 26.0. A sample size of 36 patients was estimated to provide 80% power to detect a 1-point difference in EDSS change (α=0.05, two-tailed), based on prior pilot data. All analysis were performed on the available cases only. Continuous variables were assessed for normality using the Shapiro–Wilk test. For Normally distributed variables, data were reported as mean ± standard deviation (SD) and Student’s t-test was used. For non-normally distributed variables, data were reported as median and interquartile range (IQR) and Mann–Whitney U-test was used. Categorical variables were expressed as frequencies and percentages and compared using the Fisher’s exact test. In the analysis of factors associated with therapeutic efficacy, Spearman correlation and point biserial correlation analysis were conducted for continuous variables and categorical variables respectively. Correlation and subgroup analysis were exploratory. No adjustment for multiple comparisons was applied; therefore, the reported p values are nominal and should be interpreted cautiously. For other prespecified comparisons, two-tailed p values < 0.05 were considered statistically significant; p values from correlation and subgroup analyses were interpreted as nominal and exploratory.

Results

Patient Characteristics

During the study period, 50 patients were screened in the MOGAD cohort. 36 patients met the inclusion criteria and were enrolled in the research. Of these, 21patients were treated with IVMP alone, and 15 patients were treated with DFPP and IVMP (Figure 1). Most baseline demographics and clinical characteristics were balanced between the two groups (Table 1), with no significant differences in sex, proportion of first attacks, prior use of disease-modifying therapy, or coexisting autoimmune diseases, but age differed significantly between two groups (50.24±18.96 vs 38.33±13.20, p=0.044). Clinical manifestations—including optic neuritis [8 (38.1%) vs 9 (60.0%)], myelitis [4 (19.0%) vs 7 (46.1%)], and encephalitis [13 (61.9%) vs 9 (60.0%)]—did not differ significantly between the IVMP and DFPP+IVMP groups, respectively. The delay from symptom onset to IVMP initiation, initial IVMP dose, and serum MOG-IgG titers were also comparable between groups. Baseline EDSS [2.00 (IQR 2.00–3.25) vs 2.00 (IQR 2.00–4.00), p=0.326] and mRS [2.00 (IQR 2.00–2.50) vs 2.00 (IQR 2.00–3.00), p=0.111] scores were not significantly different.

Figure 1.

Cohort flowchart of patient enrollment, treatment allocation, and follow-up. Patients with multiple attacks during the study period were included only once, based on their first eligible attack.

Abbreviations: IVMP, Intravenous methylprednisolone; DFPP, Double-filtration plasmapheresis.

Table 1.

Baseline Characteristics of the Included MOGAD Patients

	IVMP	IVMP+DFPP	p
Patients	21	15
Male, n (%)	8 (38.1)	8 (53.3)	0.500
Age, mean (SD), years	50.24 (18.96)	38.33 (13.20)	0.044
First attack, n (%)	17 (81)	10 (66.7)	0.443
Use of DMT, n (%)	18 (85.7)	13 (86.7)	1.000
Coexisting autoimmune diseases, n (%)	1 (4.8)	0 (0.0)	1.000
Clinical manifestation, n (%)
Optic neuritis involvement	8 (38.1)	9 (60.0)	0.311
Myelitis involvement	4 (19.0)	7 (46.7)	0.141
Encephalitis involvement	13 (61.9)	9 (60.0)	1.000
Delay from onset to IVMP (IQR), days	21.00 (6.00, 35.00)	15.00 (10.00, 30.00)	0.700
Initial dose of IVMP (IQR), mg	500 (500, 500)	500 (500, 1000)	0.186
Delay from onset to PE (IQR), days	/	17.00 (13.00, 34.00)	/
Serum MOG antibody tilter	1:32 (1:10, 1:100)	1:100 (1:10, 1:320)	0.492
EDSS at admission, median (IQR)	2.00 (2.00, 3.25)	2.00 (2.00, 4.00)	0.326
mRS at admission, median (IQR)	2.00 (2.00, 2.50)	2.00 (2.00, 3.00)	0.111

Abbreviations: IVMP, intravenous methylprednisolone; DFPP, Double-filtration plasmapheresis; DMT, disease-modifying therapy; MOG, Myelin oligodendrocyte glycoprotein; mRS, modified Rankin Scale; EDSS, Expanded Disability Status Scale; IQR, interquartile range; SD, standard deviation.

Efficacy Comparison Between IVMP and DFPP+IVMP

EDSS and mRS scores over time are shown in Figure 2A, B and Supplementary Table 1. No significant between-group differences were observed in EDSS at discharge [2.00 (IQR0.50–2.00) vs 1.00 (IQR 1.00–2.50), p=0.856] or at 3M-FU [1.00 (IQR 0.00–2.00) vs 1.00 (IQR 0.00–2.00), p=0.421; Figure 2A]. Similarly, mRS at discharge or at 3M-FU did not differ significantly [discharge: 2.00 (IQR 1.00–2.00) vs 2.00 (IQR 1.00–2.00), p=0.611; 3M-FU: 1.00 (IQR 1.00–1.00) vs 1.00 (IQR 0.00–1.00), p=0.133; Figure 2B]. However, ΔEDSS and ΔmRS was significantly greater in the DFPP+IVMP group at 3M-FU [ΔEDSS: 2.00 (IQR 1.00–2.00) vs 1.00 (IQR 0.75–1.75), p=0.012; ΔmRS: 2.00 (IQR 1.00–3.00) vs 1.00 (IQR 1.00–1.50), p=0.009; Figure 2C and D], though not at discharge [ΔEDSS: 1.00 (IQR 0.00–2.00) vs 1.00 (IQR 0.00–1.00), p=0.199; ΔmRS: 1.00 (IQR 0.00–1.00) vs 1.00 (IQR 0.00–1.00), p=0.346]. The EDSS improvement rate was marginally significantly higher in the DFPP+IVMP group at 3M-FU [0.67 (IQR 0.50–1.00) vs 0.50 (IQR 0.27–1.00), p=0.051; Figure 2E], but not at discharge. The mRS improvement rate was significantly higher in the DFPP+IVMP group at 3M-FU [0.67 (IQR 0.50–1.00) vs 0.50 (IQR 0.50–0.58), p=0.027; Figure 2F].

Figure 2.

Comparison of EDSS and mRS changes between IVMP and DFPP+IVMP groups. (A and B) The EDSS and mRS at admission, discharge and 3M-FU in IVMP and DFPP+IVMP group. (C and D) The ΔEDSS and ΔmRS at discharge and 3M-FU compared to the baseline in two groups. (E and F) The improvement rate of EDSS and mRS at discharge and 3M-FU compared to the baseline in two groups. Values are presented as median ± IQR. Improvement rate of mRS or EDSS at discharge or 3M-FU = ΔmRS or EDSS at discharge or 3M-FU÷mRS or EDSS at admission.

Abbreviations: IVMP, intravenous methylprednisolone; DFPP, Double-filtration plasmapheresis; 3M-FU, 3-month follow-up; EDSS, Expanded Disability Status Scale; IQR, interquartile range; mRS, modified Rankin Scale; ΔEDSS or mRS, change of EDSS or mRS from admission; EDSS or mRS improvement rate, ΔEDSS or mRS÷ EDSS or mRS at admission; ns, not significant.

Considering that patients in DFPP+IVMP group were younger, analysis of covariance was used to reduce confounding factors. After adjusting for age using analysis of covariance, the between-group difference in ΔEDSS at 3M-FU was marginally significant (p=0.079), and ΔmRS at 3M-FU remained significant (p=0.042).

Factors Associated with the Therapeutic Efficacy

In the IVMP group, baseline EDSS scores were positively correlated with ΔEDSS at 3M-FU (ρ=0.468, p=0.032; Table 2, results with p<0.1 were marked in bold). Additionally, there was a trend toward a negative correlation between IVMP initiation time with ΔEDSS (ρ=−0.391, p=0.079; Table 2) and a positive correlation between CSF leucocyte count with ΔEDSS (ρ=0.425, p=0.089). In the DFPP+IVMP group, a lower baseline EDSS score and a longer delay to IVMP initiation as well as DFPP initiation were associated with less EDSS improvement (EDSS at admission: ρ=0.738, p=0.002; delay from onset to IVMP: ρ=−0.705, p=0.003; delay from onset to DFPP: ρ=−0.660, p=0.007; Table 3, results with p<0.1were marked in bold). Additionally, analysis showed a positive correlation between serum IgG with ΔEDSS at 3M-FU (ρ=0.685, p=0.010; Table 3). Collectively, these findings indicate that earlier treatment initiation was associated with greater neurological recovery in both groups (Figure 3A–C).

Table 2.

Factors Associated with EDSS Decrease at 3M-FU in IVMP Group

	Spearman Correlation Coefficient (ρ)	P value
Age	−0.300	0.186
Sex	0.203	0.369
Disease course	−0.053	0.819
Use of DMT	−0.251	0.273
Optic neuritis involvement	0.258	0.259
Myelitis involvement	0.287	0.207
Encephalitis involvement	−0.353	0.117
EDSS at admission,	0.468	0.032
Delay from onset to IVMP	−0.391	0.079
CSF leucocyte count	0.425	0.089
CSF protein	−0.305	0.233
CSF IgG	−0.405	0.106
Serum MOG antibody tilter	0.061	0.798
Serum IgG	−0.046	0.853
Lymphocyte count	0.191	0.462
B lymphocyte count	0.135	0.605

Abbreviations: DMT, disease-modifying therapy; mRS, modified Rankin Scale; IVMP, intravenous methylprednisolone; CSF, cerebrospinal fluid; MOG, Myelin oligodendrocyte glycoprotein.

Table 3.

Factors Associated with EDSS Decrease at 3M-FU in DFPP+IVMP Group

	Spearman Correlation Coefficient (ρ)	P value
Age	−0.25	0.369
Sex	−0.114	0.686
Disease course	−0.069	0.807
Use of DMT	−0.167	0.551
Optic neuritis involvement	−0.315	0.253
Myelitis involvement	0.13	0.644
Encephalitis involvement	0.133	0.637
EDSS at admission,	0.738	0.002
Delay from onset to IVMP	−0.705	0.003
Delay from onset to DFPP	−0.660	0.007
DFPP times	0.270	0.330
CSF leucocyte count	0.398	0.141
CSF protein	−0.181	0.520
CSF IgG	0.076	0.823
Serum MOG antibody tilter	−0.041	0.884
Serum IgG	0.685	0.010
Lymphocyte count	−0.416	0.231
B lymphocyte count	−0.261	0.389

Abbreviations: DMT, disease-modifying therapy; mRS, modified Rankin Scale; IVMP, intravenous methylprednisolone; DFPP, Double-filtration plasmapheresis; CSF, cerebrospinal fluid; MOG, Myelin oligodendrocyte glycoprotein.

Figure 3.

Correlation analysis between treatment initiation time and efficacy in IVMP group and DFPP+IVMP group. (A) The correlation analysis between delay from onset to IVMP and ΔEDSS in IVMP group. (B and C) The correlation analysis between delay from onset to IVMP or DFPP and ΔEDSS in IVMP+DFPP group.

Abbreviations: IVMP, intravenous methylprednisolone; DFPP, Double-filtration plasmapheresis; EDSS, Expanded Disability Status Scale; ΔEDSS, change of EDSS from admission.

Given the correlation between serum IgG and outcomes in the DFPP+IVMP group (Figure 4A), we further compared efficacy between treatments stratified by serum IgG level. Patients were divided into high-level IgG and low-level IgG groups based on the median IgG level (10.65 g/L) of the entire cohort. In the high-IgG subgroup (>10.65 g/L), DFPP+IVMP led to significantly greater EDSS improvement as well as mRS improvement than IVMP alone [ΔEDSS: 3.50 (IQR 1.75–5.00) vs 1.00 (IQR 1.00–1.00), p=0.003; ΔmRS: 3.50 (IQR 2.00–4.00) vs 1.00 (IQR 1.00–1.00), p=0.003; Figure 4B and C]. In the low-IgG subgroup (≤10.65 g/L), no significant difference was observed [ΔEDSS: 1.00 (IQR 1.00–2.00) vs 1.00 (IQR 0.63–3.13), p=0.821; ΔmRS:1.50 (IQR 1.00–2.00) vs 1.50 (IQR 1.00–2.00), p=0.814]. Baseline characteristics of the IgG-based subgroups are shown in Supplementary Table 1and Table 3. In the high-IgG subgroup, patients in the IVMP group were older and had lower baseline EDSS as well as mRS than DFPP+IVMP group. After adjusting for age and EDSS at admission using analysis of covariance respectively, the between-group difference in ΔEDSS remained significant (p=0.003 and 0.043 respectively). After adjusting for age and mRS at admission respectively, the between-group difference in ΔmRS also remained significant (p=0.003 and 0.001 respectively).

Figure 4.

Relationship of serum IgG and therapeutic efficacy. (A) The correlation analysis between IgG and ΔEDSS at 3M-FU in IVMP+DFPP group. (B and C) comparison of ΔEDSS and ΔmRS at 3M-FU between IVMP and DFPP+IVMP group in patients with low level (≤ 10.65 g/L) or high level (> 10.65 g/L) of serum IgG. Values are presented as median ± IQR.

Abbreviations: IVMP, intravenous methylprednisolone; DFPP, Double-filtration plasmapheresis; 3M-FU, 3-month follow-up; EDSS, Expanded Disability Status Scale; mRS, modified Rankin Scale; ΔEDSS or mRS, change of EDSS or mRS from admission; IQR, interquartile range; ns, not significant.

Immunological Changes

DFPP+IVMP treatment led to significant reductions in immunoglobulin levels (IgG, IgM, IgA; Figure 5A) and complement components (C4, CH50, C1Q; Figure 5B–D). Changes in serum cytokines and lymphocyte subsets were also measured. Results showed a decreased proportion of NK cells, with no significant difference in serum cytokines. (Figure 5E and F).

Figure 5.

Immunological changes after DFPP+IVMP. (A-F) The comparison of immunoglobulin, C4, CH50, C1Q, serum cytokines and lymphocyte subsets before and after DFPP+IVMP. Values are presented as mean ± SD.

Abbreviations: IVMP, intravenous methylprednisolone; DFPP, Double-filtration plasmapheresis; SD, standard deviation; ns, not significant.

Adverse Event Associated with DFPP

DFPP was generally well tolerated. A total of 17 adverse events were recorded during 15 DFPP treatment courses (Table 4). The most common event was decreased fibrinogen (73.33% of treatments). Other events included elevated alanine aminotransferase (26.67%) and decreased platelet count (6.67%). No thrombosis, severe bleeding, shock, or serious infections occurred, and no patient discontinued DFPP due to adverse events.

Table 4.

Severity of Adverse Events Associated with DFPP

Grading	Adverse Events	Number	Percentage of Total Treatments
Grade I	Surgical and medical procedures - Other, specify (bleeding at the catheter site)	1	6.67%
	Alanine aminotransferase increased	2	13.33%
	Platelet count decreased	1	6.67%
Grade II	Alanine aminotransferase increased	2	13.33%
	Fibrinogen decreased	3	20.00%
Grade III	Fibrinogen decreased	7	46.67%
Grade IV	Fibrinogen decreased	1	6.67%

Discussion

In this ambispective multicenter cohort study, we demonstrated that combining DFPP with IVMP for acute MOGAD attacks showed significantly greater functional improvement at 3 months compared with IVMP monotherapy, particularly in patients with high baseline serum IgG levels. The combination therapy exhibited a favorable safety profile, consistent with previous reports on apheresis techniques.

Although corticosteroids exert potent anti-inflammatory effects, their capacity to remove pre-existing pathogenic autoantibodies is limited. In contrast, TPE directly and rapidly removes circulating pathogenic factors. Previous retrospective studies have reported neurological improvement following TPE in MOGAD,10,11 but comparative evidence has been scarce. Our study found that the benefit of DFPP+IVMP became significant only at the 3-month follow-up, suggesting that the therapeutic advantage of combination therapy may manifest over a longer recovery period. This finding aligns with a prospective NMOSD cohort study in which IVMP+TPE led to better and sustained neurological improvement over 6 months compared to IVMP alone.8 The DFPP+IVMP group had a significantly younger age compared with IVMP group, potentially constituting a source of confounding in the comparison of treatment efficacy. After adjusting for age by analysis of covariance, the intergroup difference in 3-month ΔEDSS reached only borderline statistical significance, while ΔmRS remained significantly improved in the combination group. This borderline finding may reflect limited statistical power stemming from the modest total sample size of 36 patients rather than missing follow-up data, and this result requires validation in larger prospective cohorts.

Correlation analysis highlighted the critical role of early treatment initiation in both groups. In the IVMP alone group, a longer time to treatment initiation was associated with less EDSS reduction, though this did not reach statistical significance. In the DFPP+IVMP group, both time to DFPP initiation and time to IVMP initiation were negatively correlated with the EDSS improvement. A large retrospective MOGAD study identified early apheresis and prior immunosuppression as predictors of complete attack recovery.10 Another retrospective analysis also highlighted younger age and early apheresis as favorable factors,11 echoing findings in NMOSD and autoimmune encephalitis.9,15 Collectively, the present results and previous evidence support early treatment initiation as a key modifiable factor for optimizing functional outcomes in MOGAD attacks.

In the DFPP+IVMP group, higher serum IgG level was associated with greater EDSS improvement. No such correlation was found in the IVMP group. Our further subgroup analysis suggested that patients with high IgG derived substantial benefit from DFPP+IVMP compared with IVMP, whereas those with low IgG did not. The association between serum IgG with treatment efficacy may reflect the mechanism of DFPP, which directly removes circulating pathogenic antibodies. Higher IgG levels may indicate greater humoral autoimmunity, making the disease process more responsive to antibody depletion, although this is just a hypothesis and a direct causal link cannot be established from our data. These results suggest that serum IgG may help identify patients more likely to benefit from add-on DFPP. However, the high-IgG subgroup had imbalanced age and baseline disability, and although we adjusted for age, baseline mRS and EDSS, residual confounding may persist. In addition, the IgG cutoff in this study was based on the cohort median and is not a clinically validated threshold. Given the small sample size, larger prospective cohorts are needed to confirm the observational finding and explore the optimal cutoff value.

Immunological profiling demonstrated that DFPP+IVMP significantly reduced immunoglobulin and complement levels, consistent with its putative mechanism of action.18,19 Beyond antibody removal, we observed a decreased proportion of NK cells. The shift suggested that DFPP may promote a more regulatory environment, potentially contributing to sustained recovery.

The safety profile of DFPP in our cohort was consistent with previous reports.8,16,20,21 The most frequent laboratory abnormality was decreased fibrinogen, underscoring the importance of post-procedural monitoring and supplementation. Notably, no severe complications occurred, and no patient required treatment discontinuation. The avoidance of exogenous plasma with DFPP may reduce transfusion-related risks, making it a valuable and well-tolerated alternative to conventional TPE.

This ambispective multicenter cohort bears several limitations: Firstly, the non-randomized treatment assignment and the small sample size may have introduced selection bias, such as patients with more severe attacks prefer to DFPP+IVMP. Despite adjustment for age and baseline disability using analysis of covariance, residual confounding cannot be excluded. This observational study also cannot suggest a causal relationship. Additionally, the small sample of 36 patients further weakened subgroup analysis and generalizability of the study. Secondly, visual outcomes, such as standardized best-corrected visual acuity and optical coherence tomography assessment were not systematically obtained for optic neuritis patients, resulting in incomplete visual function evaluation; efficacy was solely measured via EDSS and mRS without quality-of-life, cognitive or neuropsychological metrics. Thirdly, follow-up only lasted 3 months with no long-term relapse or persistent disability data. Fourthly, immune sampling was unbalanced, as DFPP+IVMP patients had paired pre- and post-treatment biomarkers while most IVMP-only patients lacked post-treatment immune tests, hindering separate analysis of each therapy’s immunomodulatory effects. Finally, the association between IgG and DFPP treatment efficacy was hypothesis-generating and cannot establish a casual link. The IgG cutoff of 10.65 g/L was based on internal cohort median without external validation, carrying overfitting risks and limiting generalizability.

Conclusion

In this small observational cohort, DFPP plus IVMP was associated with greater improvement in some 3-month functional outcomes than IVMP alone, although the adjusted difference in the primary endpoint did not reach conventional statistical significance. The IgG subgroup findings are exploratory and require external validation before clinical application. DFPP was generally well tolerated, with no serious adverse events leading to treatment discontinuation. Larger prospective studies are needed to confirm these findings and determine whether baseline serum IgG can identify patients more likely to benefit from DFPP.

Funding Statement

This study is supported by National Natural Science Foundation of China (No. 82301527, KW); The Young Medical Talents Training Program of Shanghai Pudong New Area Health Commission (No. PWRq2025-23, KW); The Municipal Commission of Health and Family Planning Foundation of Shanghai Pudong New Area (No. PW2022E-01, YG); New Quality Clinical Specialties of High-end Medical Disciplinary Construction in Pudong New Area (No. 2024-PWXZ-16, YG). The funders were not involved in study conception, data acquisition and interpretation, manuscript preparation, or the decision to submit the article for publication.

Abbreviations

MOGAD, Myelin oligodendrocyte glycoprotein antibody-associated disease; AQP4, aquaporin-4; NMOSD, neuromyelitis optica spectrum disorder; IVMP, intravenous methylprednisolone; TPE, therapeutic plasma exchange; DFPP, Double-filtration plasmapheresis; 3M-FU, 3-month follow-up; EDSS, Expanded Disability Status Scale; mRS, modified Rankin Scale; IQR, interquartile range; SD, standard deviation.

Data Sharing Statement

Due to ethical restrictions, the raw data cannot be made publicly available. However, de-identified data may be obtained from the corresponding author (Yangtai Guan) upon reasonable request.

Ethics Approval and Consent to Participate

The study protocol received approval from the Ethics Committee of Renji Hospital, Shanghai Jiao Tong University School of Medicine (LY2025-435-B, LY2022-051-B). The study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.