Clinical Question

Does adding long-term mycophenolate mofetil (MMF) maintenance therapy to first-line treatment reduce relapse rates and improve outcomes in patients with newly diagnosed anti-NMDA receptor encephalitis compared to first-line treatment alone?

Inclusion Criteria

• New acute onset of ≥1 of 8 major manifestations within <3 months: psychosis, memory deficit, speech disturbance, seizures, movement disorder, disturbance of consciousness, autonomic dysfunction, central hypoventilation
• Positive CSF test for NMDAR antibodies (cell-based assay)
• Age ≥14 years
• Received ≥1 first-line treatment within 2 weeks of hospitalization before randomization
• Modified Rankin Scale (mRS) score ≥2 for at least 24 hours before assessment

Exclusion Criteria

• Pregnancy or breastfeeding
• Reproductive potential patients not agreeing to effective contraception through 24 months post-treatment
• Clinically significant infection (bacterial, fungal, parasitic, TB, HSV, VZV, CMV, EBV, HIV, hepatitis, syphilis)
• Encephalopathy secondary to sepsis or systemic inflammatory response syndrome
• Pre-existing epilepsy, cerebral trauma, or other nervous system diseases
• Positive tests for other neuronal surface or intracellular antigens
• Untreated teratoma before randomization
• History or current presence of cancer or malignancy
• Surgical procedure within 4 weeks prior to baseline (excluding teratoma resection)
• Insufficient heart, liver, kidney, or bone marrow function (WBC <3.0×10^9/L, ANC <2.0×10^9/L, ALC <0.5×10^9/L, platelets <100×10^9/L, AST/ALT >1.5× ULN)
• Previous treatment with anti-CD20, anti-CD19, anti-IL6R, T-cell depletion, complement inhibitors, Fc receptor antagonists, anti-BLyS, cyclophosphamide, azathioprine, or bortezomib
• Lack of key clinical data
• Contraindications to MMF
• Inability to give informed consent
• Participation in another interventional trial within 3 months

Baseline Characteristics

Arms

Outcomes

Criticisms

• Open-label design may introduce observer bias and placebo effects, though blinded endpoint adjudication mitigates this
• Study conducted only in China - generalizability to other populations uncertain
• Did not reach planned sample size (100 vs target 118) due to interim analysis showing significant effect and low NMDARE prevalence
• Strict exclusion criteria (no patients <14 years, severe infections, untreated teratomas) limit real-world representativeness
• Unexpected early benefit at 4 weeks is atypical for MMF mechanism (usually slow-acting) - may reflect non-immunological factors
• No significant difference in long-term functional outcomes (mRS ≤2) despite relapse prevention benefit
• MOCA may lack sensitivity/specificity for detecting cognitive deficits in this population
• MMF benefit may be limited to high-risk/relapsing patients; subgroup analysis showed no benefit in non-relapsing patients
• Cost-effectiveness of 24-month maintenance therapy not evaluated
• No comparison with other maintenance agents (rituximab, azathioprine, cyclophosphamide)
• Control group had higher proportion of patients receiving both IVIg and IVMP (85.7% vs 70.5%)
• Class II evidence only due to open-label design

Funding

National Key R&D Program of China (2022YFC2503800), National Science Foundation of China (82471388), Clinical Research Incubation Program West China Hospital, Sichuan University (2022HXFH022), Postdoctoral Fellowship Program of CPSF (GZC20241130), China Postdoctoral Science Foundation (2024M752236), Postdoctor Research Fund of West China Hospital (2024HXBH053), Natural Science Foundation of Sichuan Province (2025ZNSFSC1635), Sichuan Science and Technology Program (2024YFHZ0056). Funders had no role in design, conduct, analysis, or decision to publish.