Double Filtration Plasmapheresis Shows Superior Short-Term Efficacy to Intravenous Methylprednisolone in Acute Autoimmune Encephalitis: A Prospective Observational Study

Introduction

Autoimmune encephalitis (AE) represents a group of immune-mediated disorders characterized by autoantibodies targeting neuronal surface or intracellular antigens, leading to diverse neuropsychiatric manifestations such as cognitive impairment, seizures, and movement disorders. With an estimated prevalence of 13.7/100000, AE poses significant diagnostic and therapeutic challenges due to its heterogeneous clinical presentation and severity, often necessitating intensive care in up to 70% of cases.¹ Early intervention is critical, as prompt immunotherapy within four weeks of symptom onset has been associated with improved outcomes. Current first-line treatments include intravenous methylprednisolone (IVMP), intravenous immunoglobulin (IVIG), and therapeutic plasma exchange (TPE), which collectively aim to mitigate immune-mediated neuronal dysfunction.^2,3 However, the comparative efficacy of these modalities, particularly novel TPE-derived techniques like double filtration plasmapheresis (DFPP), remains inadequately explored.

DFPP, a semi-selective apheresis technology, offers distinct advantages over conventional TPE by selectively removing pathogenic macromolecules (eg, autoantibodies, immune complexes) while retaining beneficial plasma components like albumin. Unlike TPE, DFPP eliminates the need for exogenous plasma replacement, thereby circumventing supply limitations and reducing transfusion-related risks.⁴ Preliminary studies in other autoimmune neurological conditions, such as myasthenia gravis and multiple sclerosis, suggest DFPP’s potential to enhance clinical outcomes when combined with immunosuppressants.^5–8 For instance, DFPP combined with oral methylprednisolone demonstrated superior efficacy to oral methylprednisolone alone in myasthenia gravis.⁹ Improved symptoms of steroid-refractory multiple sclerosis were noted after DFPP.¹⁰ Despite these promising findings, robust evidence comparing DFPP directly with IVMP in acute-phase AE is scarce, highlighting a knowledge gap in clinical guidance for optimal first-line therapy selection.

This study addresses this gap by prospectively evaluating the efficacy and safety of DFPP versus IVMP in antibody-positive patients with acute-phase AE. While prior research has predominantly examined TPE as an adjunct to IVMP or IVIG,^11,12 confounding the assessment of its standalone utility, our study design isolates the comparative effectiveness of DFPP monotherapy. Given the American Society for Apheresis’s endorsement of TPE for anti-NMDAR encephalitis,¹³ yet limited data on DFPP’s role, our findings aim to clarify whether DFPP’s mechanistic advantages translate into superior clinical outcomes. We hypothesized that DFPP, by rapidly removing pathogenic autoantibodies, would lead to superior short-term clinical improvement compared to IVMP in acute-phase AE. By analyzing functional improvement and symptom resolution, this study seeks to inform evidence-based therapeutic decisions and highlight potential avenues for optimizing AE management.

Methods

Ethics

This study was approved by the Ethics Committee of Renji Hospital, Shanghai Jiao Tong University School of Medicine (LY2024-031-B). All data were collected prospectively for the purpose of this study. Written informed consent was waived for the use of serum, cerebrospinal fluid (CSF), and clinical data, as this was a prospective observational study. The study was not registered as a clinical trial as it did not involve random assignment to interventions. All treatment decisions were made independently by clinicians based on disease severity, antibody profile, and patient-specific factors, in accordance with standard care guidelines. Patient data were anonymized to ensure confidentiality. This study was conducted in accordance with the principles of the Declaration of Helsinki.

Study Design

A prospective observational cohort study was conducted at Renji Hospital from July 2018 to May 2024. The study aimed to compare the efficacy and safety of DFPP and IVMP as first-line immunotherapies for AE. The primary outcome was improvement in the mRS, and secondary outcomes included changes in the CASE and treatment-related adverse events.

Study Subjects

Inclusion Criteria: 1) Diagnosis of AE based on established consensus criteria; 2) Serum and/or CSF positivity for AE-related antibodies; and 3) Treatment with either IVMP or DFPP as initial monotherapy therapy during the acute phase of hospitalization. Exclusion Criteria: 1) Incomplete or uncertain clinical or treatment data; and 2) Receipt of other first-line immunosuppressive therapies (eg, IVIG) or second-line immunotherapy.

Definition of Acute Phase of AE and Treatment Window

The acute phase of AE was defined as the period within the first 4 weeks after symptom onset, characterized by active and progressive neuropsychiatric deficits requiring hospitalization and initiation of first-line immunotherapy. This definition aligns with established consensus guidelines emphasizing the critical importance of early intervention to mitigate neuronal injury and improve long-term outcomes.

The treatment window refers to the timeframe within which first-line immunotherapy (either DFPP or IVMP) was initiated. In this study, all treatments were started within the acute phase (ie, ≤4 weeks from symptom onset). This ensures that the comparative efficacy of the two interventions is evaluated during the period when immunotherapeutic agents are believed to have the greatest impact on disease activity and subsequent recovery.

Immunomodulatory agents such as mycophenolate mofetil (MMF) or azathioprine (AZA) were initiated only after completion of the acute-phase treatment and outcome assessment, and thus did not confound the short-term efficacy evaluation.

Treatment Allocation

Treatment allocation was based on clinical judgment and multidisciplinary team discussion, considering patient-specific factors such as antibody type, disease severity, comorbidities, and patient/family preferences. DFPP was typically considered for patients with moderate to severe disease or those with prominent antibody-mediated features, based on multidisciplinary discussion. Neither random assignment nor blinding was employed, in accordance with the observational design.

Treatment Protocols

DFPP Procedure: Patients underwent 3–5 sessions of DFPP, administered every other day. Plasmapheresis was performed using Plasauto Σ (Asahi Kasei Medical) or KM-9000 (Kawasumi) devices, with Plasmacure™ or Plasmaflo™ as plasma separators and Evaflux™ or Cascadeflo™ EC as component separators. The treated plasma volume was calculated using Nadler’s formula and hematocrit levels, targeting 1.3–1.5 times the total plasma volume. Anticoagulants included unfractionated heparin, low-molecular-weight heparin, or nafamostat mesilate.

IVMP Protocol: Patients received an initial IVMP dose of 480/500 mg/day, tapered by 50% every 3–5 days, followed by oral prednisone acetate (60 mg) with further adjustments based on clinical response.

Outcome Measures

Primary Outcome (Functional Assessment)

Proportion of patients with improved modified Rankin Scale (mRS) scores (ΔmRS = pre-treatment mRS-post-treatment mRS).

Secondary Outcomes (Symptomatic Assessment)

Proportion of patients with improved Clinical Assessment Scale for Autoimmune Encephalitis (CASE) scores (ΔCASE = pre-treatment CASE - post-treatment CASE), and incidence of treatment-related adverse events. Improvement was defined as ΔmRS> 0.

All outcome assessments, including the mRS, CASE and relevant laboratory parameters, were conducted at admission (pre-treatment) and upon completion of the acute treatment phase (post-treatment, typically 7–10 days after treatment initiation). Although assessors were not blinded to treatment allocation, mRS and CASE scores were assessed using standardized criteria to minimize bias.

All unexpected medical events occurring during the treatment period and within 7 days post-treatment were recorded as adverse events. These events were monitored throughout hospitalization and graded according to the Common Terminology Criteria for Adverse Events (CTCAE), version 5.0. Infections were defined as clinical events necessitating antimicrobial therapy, corroborated by supporting laboratory or imaging findings. A serious adverse event was defined as any event that was fatal, life-threatening, resulted in prolonged hospitalization, or caused significant disability.

Statistical Analysis

Owing to the relative rarity of autoimmune encephalitis, the sample size (N = 38) was determined by the number of eligible cases available during the study period. A post-hoc power analysis confirmed that the sample provided adequate power for the primary comparisons. All statistical analyses were performed using SPSS version 22.0 (IBM Corporation, Armonk, NY). The normality of continuous variables was assessed using the Kolmogorov–Smirnov test. Normally distributed data were expressed as mean ± standard deviation (SD) and compared using Student’s t-test; non-normally distributed data were reported as median with interquartile range (IQR) and compared using the Mann–Whitney U-test. Categorical variables were summarized as frequencies and percentages, and group differences were examined using the chi-square test or Fisher’s exact test, as appropriate. For paired comparisons within groups (eg, pre- vs post-treatment), paired t-tests or Wilcoxon signed-rank tests were applied. A complete-case analysis strategy was adopted, whereby patients with missing primary outcome data (mRS at discharge) were excluded.

Multivariable logistic regression was employed to adjust for potential confounding variables, including age, sex, baseline mRS score, antibody type, and treatment group. Variables were included in the model based on clinical relevance and univariate associations (p < 0.1). To evaluate whether baseline disease severity influenced treatment response, an interaction term between treatment group and baseline mRS was incorporated into the regression model. A two-tailed p-value < 0.05 was considered statistically significant. Specific statistical tests used for each outcome are reported in the corresponding Results section. As the subgroup analyses were exploratory in nature, no correction for multiple comparisons was applied, and these findings should be interpreted with caution.

Results

Study Population and Baseline Characteristics

From July 2018 to May 2024, 41 consecutive patients with antibody-confirmed acute-phase AE were prospectively enrolled and allocated to first-line immunotherapy (DFPP: n=24; IVMP: n=17). Three participants discontinued treatment due to intolerance (DFPP: n=2; IVMP: n=1), yielding a final analytic cohort of 38 patients (DFPP: n=22; IVMP: n=16) who completed the therapeutic protocols and outcome assessments. A flowchart of the study is shown in Figure 1.

Figure 1 A flowchart of the study.

The median age at onset was 46.00 years (IQR: 27.00–64.25), with 55.3% (21/38) being female. Two patients had coexisting autoimmune diseases (one had systemic lupus erythematosus and the other had autoimmune thyroiditis), and three had tumors. Baseline characteristics were comparable between groups, except for antibody profiles and hospitalization duration. Baseline characteristics and detailed antibody profiles of study participants were showed in Table 1. The IVMP group exhibited a balanced distribution of cell-surface (50%) and intracellular antigens (50%), while the DFPP group had a higher prevalence of cell-surface antibodies (68.2%, p=0.024), including anti-NMDAR encephalitis (40.91% vs 6.25%, p=0.025). Total hospitalization duration was longer in the DFPP group (18.23±0.96 vs 14.81±1.18 days, p=0.029), likely due to procedural requirements.

Table 1 Baseline Characteristics and Detailed Antibody Profiles of Study Participants

Treatment Efficacy

Both therapies demonstrated clinical efficacy in the acute phase, with 52.6% (20/38) achieving mRS improvement and 60.5% (23/38) showing CASE score reduction. However, DFPP outperformed IVMP significantly. 1) Functional improvement (change in mRS) (Figure 2A): 68.2% (15/22) of DFPP patients exhibited mRS reduction versus 31.3% (5/16) with IVMP (p=0.047). The median ΔmRS was 1.00 (IQR: 0.00–1.00) for DFPP versus 0.00 (IQR: 0.00–1.00) for IVMP (p=0.060). 2) Symptomatic improvement (change in CASE) (Figure 2B): A trend favored DFPP [72.7% (17/22) vs 43.8% (7/16) p=0.099], with a median ΔCASE of 1.00 (IQR: 0.00–2.00) versus 0.00 (IQR: 0.00–1.00) for IVMP (p=0.125).

Figure 2 Clinical improvement rates. DFPP showed higher mRS improvement (68.2% vs 31.3%, p=0.047) (A) and CASE improvement (72.7% vs 43.8%, p=0.099) (B) than IVMP in the acute phase. Gray bars denote non-responders.

In our study, univariate analysis was performed using a significance threshold of p < 0.1 to identify potential confounding factors, accounting for the limited sample size. This analysis indicated that baseline mRS (p = 0.056) and treatment type (p = 0.028) were potential determinants of the outcome. Gender and age were included as regular confounding factors in the multivariate analysis. Multivariate logistic regression (Tables 2 and 3) identified DFPP as the sole independent predictor of mRS improvement (OR: 5.234, 95% CI: 1.179–23.235, p=0.030), adjusting for sex, age, baseline mRS, and treatment type.

Table 2 Univariate Analysis of Treatment Efficacy

Table 3 Multivariate Regression Analysis for Effects of DFPP on Acute Autoimmune Encephalitis

Subgroup Analysis

Subgroup analyses were performed to evaluate the consistency of treatment effects across different patient populations. Among patients with anti-NMDAR encephalitis—a subgroup of particular interest due to its predominance in the DFPP group—a higher rate of functional improvement (mRS reduction) was observed compared to the IVMP group, consistent with the overall cohort; however, formal statistical comparison was limited by the small sample size. Similarly, analyses based on other antibody subtypes (eg, LGI1, GABAR) suggested a trend toward better outcomes with DFPP, though these did not reach statistical significance. Sensitivity analysis including only patients receiving their first immunotherapy episode (n=32) also showed results aligned with the primary analysis, with DFPP remaining associated with greater improvement in mRS scores (68.2% vs 31.3%, p = 0.047).

Safety Outcomes

No adverse events occurred in the IVMP group. In the DFPP group, three events were reported (13.6% [3/22]), including two cases of deep venous thrombosis (9.09% [2/22]) and one heparin-induced hepatic impairment (4.55% [1/22]), all resolved with anticoagulation or heparin withdrawal. No cases resulted in bleeding, and no treatment discontinuations or serious adverse events occurred.

Discussion

This prospective observational study provides compelling evidence that DFPP may offer superior clinical efficacy compared to IVMP as a first-line immunotherapy for antibody-positive acute-phase AE. Our findings demonstrate that DFPP was associated with significantly higher rates of functional improvement, and a strong trend toward greater symptomatic relief. These results contribute to the growing body of evidence supporting the use of DFPP in AE and highlight the need for further research to optimize treatment strategies.

The observed superiority of DFPP in rapidly removing pathogenic autoantibodies and inflammatory mediators. Unlike conventional TPE, DFPP selectively filters large molecular weight substances, including immunoglobulins and immune complexes, while preserving beneficial plasma components such as albumin.⁴ This targeted approach may explain the higher efficacy of DFPP in our study, particularly in patients with cell-surface antibody-mediated AE, who constituted the majority of the DFPP group. Previous studies have similarly reported favorable outcomes with TPE in anti-NMDAR encephalitis.¹⁴ However, many of these studies evaluated TPE in combination with steroids or IVIG,^15–17 complicating the interpretation of its standalone efficacy. In addition, passively acquired autoantibodies may be detected after infusion of IVIG, which may mislead patients.¹⁸ Our work addresses this gap by directly comparing DFPP monotherapy with IVMP, thereby minimizing confounding effects from adjunctive treatments.

The differential response rates between DFPP and IVMP may also reflect the distinct pathophysiological mechanisms underlying AE. Cell-surface antibody-mediated AE, which predominated in the DFPP group, is thought to involve reversible neuronal dysfunction driven by humoral immunity, making it particularly amenable to antibody removal via apheresis. In contrast, intracellular antibody-associated AE, more common in the IVMP group, often involves T-cell-mediated cytotoxicity, which may respond better to broad immunosuppression with steroids.¹⁹ This hypothesis is supported by the significant intergroup differences in antibody profiles and the higher prevalence of anti-NMDAR encephalitis in the DFPP group. Nevertheless, the multivariate analysis adjusted for these baseline differences, reinforcing the robustness of our findings.

Despite its efficacy in the acute phase, DFPP was associated with a longer total hospitalization duration and a higher incidence of adverse events, including deep venous thrombosis and heparin-induced hepatic impairment. These complications, though manageable, underscore the invasive nature of DFPP and the need for careful patient selection and monitoring.²⁰ The extended hospitalization in the DFPP group likely reflects the logistical demands of apheresis, such as catheter placement and equipment availability, rather than inferior clinical outcomes.²¹ Future studies should explore strategies to streamline DFPP administration, such as early catheter placement or standardized protocols, to reduce hospitalization times and improve cost-effectiveness.

The discrepancy between mRS and CASE score improvements warrants further discussion. While both scales demonstrated a treatment advantage for DFPP, the effect was statistically significant only for mRS. This may reflect the broader symptom dimensions captured by the CASE score, which includes psychiatric, cognitive, and motor domains, compared to the mRS’s focus on global disability.²² Some patients may exhibit symptomatic improvements (CASE) without achieving a corresponding reduction in disability (mRS), particularly in milder cases. Alternatively, the smaller sample size may have limited the statistical power to detect significant differences in CASE scores. Larger studies with longer follow-up periods are needed to clarify this relationship and assess the durability of treatment effects.²³

Our study has several limitations. First, the non-randomized, observational design introduces potential selection bias, as treatment allocation was based on clinical judgment rather than randomization. The predominance of cell-surface antibodies in the DFPP group suggests that clinicians may have favored apheresis for patients perceived to have a higher likelihood of response, potentially inflating the observed treatment effect. Second, the small sample size precluded subgroup analyses by specific antibody types or disease severity, limiting the generalizability of our findings. Third, the prolonged startup time from symptom onset to treatment—median 66.5 days—was influenced by referral delays, diagnostic confirmation, and multidisciplinary decision-making. This delay may have affected treatment outcomes and should be considered when interpreting the results. Fourth, the short-term follow-up period did not allow for assessment of long-term outcomes or relapse rates, which are critical for evaluating the sustained benefits of DFPP. Finally, the single-center design may limit the external validity of our results, particularly in settings with differing apheresis protocols or patient populations.

Despite these limitations, our study adds to the emerging literature supporting the use of DFPP in AE. The findings are consistent with prior reports of apheresis efficacy in refractory AE and extend these observations to first-line treatment. Future research should prioritize multicenter, randomized controlled trials to validate our results and address the limitations of the current study. Key areas for investigation include (1) standardized protocols for DFPP administration, (2) long-term outcomes and relapse rates, (3) cost-effectiveness analyses comparing DFPP with other immunotherapies, and (4) biomarker-driven approaches to predict treatment response. Additionally, studies exploring combination therapies, such as DFPP followed by IVMP or IVIG, may further optimize outcomes for severe or refractory cases.

In conclusion, our study suggests that DFPP is a promising first-line therapy for acute-phase AE, offering superior short-term efficacy compared to IVMP. These findings underscore the importance of early and targeted immunotherapy in AE and highlight the need for further research to refine treatment algorithms and improve patient outcomes.