The purpose of the study is to assess the efficacy of immunoadsorption therapy (IA) on improving the neurological status of severe pediatric anti-NMDAR encephalitis patients.

Anti-NMDA-Receptor (NMDAR) encephalitis, the most frequent autoimmune encephalitis after Acute Demyelinating encephalomyelitis (ADEM), affects children with predominant movement disorders, decline of consciousness, psychiatric symptoms, language dysfunction, seizures, dysautonomic symptoms. The cerebrospinal fluid (CSF) is most often abnormal with lymphocytic pleocytosis, CSF-specific oligoclonal bands with intrathecal synthesis of anti-NMDAR antibodies. Antibody titres in CSF and serum seem correlated with clinical outcome. Early start of immunotherapy has been reported to improve clinical outcome and associated with less relapses. In a recent large series (211 children/577), 77% of the patients were admitted to Intensive Care Unit (ICU) at the beginning. Within the group of children, first-line immunotherapy (95%) consisted of corticosteroids (89%), and/or intravenous immunoglobulins (IgIV) (83%), and/or plasma exchange (28%) with failure in 46%. The second-line immunotherapy consisting in rituximab (24%) and/or cyclophosphamide (16%) was proposed in 32%, and tended to be associated with good outcome (OR=3.35, CI: 0.86-12.98, p=0.081 for 53 children; statistical significance was achieved for the entire population including adults (OR: 2.69, CI: 1.24-5.80, p = 0.012) and less relapses.

In investigators' experience, the clinical benefit of rituximab is delayed over one month, while children go on worsening (50% admitted in ICU) thus claiming for faster removal of the antibodies. Plasma exchange is proposed in most of the series as alternative or combined treatment in the acute stage (first-line immunotherapy); recently, another plasmatherapy, immunoadsorption therapy (IA), has been reported as an efficient therapeutic approach in 11/13 patients. In this retrospective study, patients received a median of 6 IA sessions within a median period of 8 days with relevant clinical improvement. However these encouraging results and investigators' experience in few children need further prospective and standardized evaluation.

In IANMDAR study, each patient will receive 10 IA sessions during 28 days maximum. Rituximab will be given each week for 4 weeks (one injection by week +/- 3 days):

• at least 1 day before each IA session
• the 4 injections should be done before V2 (Day 28 after the inclusion)

To assess the efficacy of IA-therapy at short term, the neurological status of patients will be evaluated before and after the 10 IA sessions using the Pediatric Cerebral Performance Category Scale (PCPCS) and the modified Rankin Scale (mRS).

To assess the efficacy of IA-therapy at long term, patients will have a standardized follow-up during two years including neuropsychological evaluation at 1 year and at 2 years (see below for further details).

Inclusion Criteria:

• Age: 0-18 years inclusive
• Autoimmune encephalitis with positive anti-NMDAR antibodies in CSF (definite anti-NMDAR encephalitis according to Graus's criteria (Graus et al., 2016).
• PCPCS and mRS at 4 or over at the inclusion after first line therapy (steroids and/or IgIV) when Rituximab therapy is warranted
• Parents or legal guardians signed the Informed consent form
• Social insurance affiliation

Exclusion Criteria:

• Autoimmune encephalitis without NMDAR antibodies

• PCPCS and mRS scores under 4 after first-line therapy

• Contraindication to perform central vascular access

• Pregnancy, breastfeeding or absence of effective contraception (including abstinence) in a pubertal patient.

• Contraindication to perform IA therapy :

  • Clinical conditions that prohibit transitory volume changes
  • Indications that prohibit anticoagulation using Heparin and/or ACD-A solutions
  • History of hypercoagulability
  • Generalized viral, bacterial and/or mycotic infections
  • Severe immune deficiencies (e.g. AIDS)
  • Suspected allergies against sheep antibodies or agarose