LGI1 antibody-associated encephalitis without evidence of inflammation in CSF and brain MRI

Objective This study aimed to explore the frequency and distinct characteristics of adult patients with LGI1 antibody-associated encephalitis in the absence of inflammatory abnormalities in both routine CSF analysis and brain MRI. Methods We conducted a retrospective study of adult patients with antibodies targeting LGI1 and then screened patients with no evidence of inflammation in brain MRI and normal results in routine CSF analysis, including white blood cell count, protein concentration, IgG, and oligoclonal bands. Results Among 80 patients with LGI1 antibody-associated encephalitis in our center, 31 (38.8%) fulfilled the screening criteria. For these patients, the onset age was 57.0 +/- 14.7 years, and 19 (61.3%) were female. Viral prodrome occurred in 5 patients (16.1%). Faciobrachial dystonic seizures (FBDS) were the most predominant symptom (38.7%), followed by seizure onset (22.6%) and memory deficits (19.4%). The sensitivity of antibody detection in serum was higher than CSF (96.8% vs. 48.4%, p < 0.001). Most patients (30/31, 96.8%) benefited from the first-line immunotherapy, and 23 patients (74.2%) achieved complete recovery, yet 3 patients (9.7%) had clinical relapses in 2-year follow-up after discharge. The patients had a higher prevalence of females (61.9% vs. 26.7%, p = 0.003) and were more frequently associated with FBDS during the disease course (38.7% vs. 10.2%, p = 0.004). However, there was no difference in treatment outcomes and recurrence ratio between the two groups (p = 0.144 and p = 0.515). Moreover, we divided all 80 patients into four groups according to antibody titer levels in serum and CSF at the time of diagnosis, respectively. WBC and protein concentrations in CSF showed no difference among the four groups. Conclusions The absence of evidence of inflammation in routine CSF analysis and brain MRI did not rule out anti-LGI1 associated encephalitis. FBDS and the subacute onset of cognitive dysfunction should push forward with autoantibody testing for patients even without inflammatory abnormalities. The routine inflammatory indicators in CSF seemed to be unrelated to antibody titer levels.