This study aims to examine the differential efficacy of first-line versus second-line immunotherapies in antibody-confirmed AE.
Autoimmune encephalitis (AE) is a neurologic syndrome characterized by inflammation and dysfunction of the central nervous system which is precipitated by autoantibodies against neuronal antigens. Immunotherapy is its first line treatment yet the comparative value of escalation strategies beyond first-line remains uncertain. We examined the differential efficacy of first-line versus second-line immunotherapies in antibody-confirmed AE.
A systematic review was conducted in adherence with PRISMA standards, employing the PICO framework. Electronic databases (PubMed, Embase, Cochrane CENTRAL, Web of Science, ClinicalTrials.gov) were requeried for studies evaluating first-line therapies (corticosteroids, intravenous immunoglobulin [IVIG], plasmapheresis) against second-line agents (rituximab, cyclophosphamide, tocilizumab) in adults (≥18 years) diagnosed with AE (including NMDAR, LGI1, CASPR2, and GAD65 subtypes). The primary outcome was functional recovery, assessed by the modified Rankin Scale (mRS); secondary outcomes included seizure control, relapse incidence, and mortality. A random-effects meta-analysis was employed.
Eleven studies met inclusion criteria. Second-line therapy yielded a greater mean mRS improvement (reduction: 2.67; 95% CI: 2.04–3.3; p<0.001). Rituximab was associated with the highest functional recovery (80% achieved mRS ≤2 vs. 53% with first-line agents; p=0.002) and a lower relapse rate (HR=0.29; 95% CI: 0.09–0.85; p=0.024). IVIG significantly improved seizure control (75% vs. 22%; OR=10.5; p=0.044). Subgroup analysis demonstrated antibody dependent variability: LGI1-positive individuals responded more favorably to second-line therapy (OR=6.70; p=0.024), whereas NMDAR-associated cases showed minimal difference (OR=1.03; p=0.944). Mortality across all groups remained below 1%. Heterogeneity was substantial (I²=65%; p<0.01) likely due to protocol variation and antibody heterogeneity.