Background Anti-leucine-rich glioma-inactivated 1 (LGI1) encephalitis is clinically heterogeneous, especially at presentation, and though it is sometimes found in association with tumor, this is by no means the rule. Methods Clinical data for 10 people with anti-LGI1 encephalitis and 3 people with anti-N-Methyl-D-aspartate receptor (NMDAR) encephalitis with teratoma were collected. Microscopic pathological examination and immunohistochemical (IHC) assay of the LGI1 antibody were performed on teratoma tissue obtained by laparoscopic oophorocystectomy. Results In our teratoma associated anti-LGI1 encephalitis case, teratoma pathology was characterized by mostly thyroid tissue and IHC assay confirmed partial or focal positive nuclear staining of LGI1 in some tumor cells. The case was similar to the non-teratoma (NT) group in many ways: age at onset; percent presenting with rapidly progressive dementia (RPD) and psychiatric symptoms; hyponatremia; normal cerebrospinal fluid (CSF) results except for positive LGI1 antibody; bilateral hippocampal hyperintensity on magnetic resonance imaging (MRI); diffuse slow waves on electroencephalogram (EEG); good response to immunotherapy and mild residual cognitive deficit. Her chronic anxiety and status epilepticus (SE) were the biggest differences compared with NT group. Interestingly, the case presented many differences compared with anti-NMDAR encephalitis with teratoma: older onset age, prominent anxiety, SE, hyponatremia, normal CSF cell count, hippocampal hyperintensity on MRI and slowly recovered and residual short-term memory impairment. Conclusion In our series, anti-LGI1 encephalitis included common clinical features: RPD, faciobrachial dystonic seizures, behavioral disorders, hyponatremia, T2-MRI hyperintensity of hippocampus and residual cognitive deficit, but a larger accumulation of cases is needed to improve our knowledge base.

Aim: To determine the risk factors of valproic acid (VPA)-induced tremor, with particular attention on characterizing cerebellar atrophy and identifying tremor-susceptible gene mutations. Methods: Epileptic patients taking VPA were divided into two groups, a tremor and a non-tremor group, based on self-reported or clinically assessed tremors. A mutation of rs9652490 in the leucine-rich repeat and immunoglobulin domain-containing Nogo-receptor-interacting protein 1 (LINGO-1) gene was determined by Sanger sequencing. Cerebellar atrophy was assessed and various cerebellar dimensions were measured on magnetic resonance imaging (MRI) scans. Results: Among 200 subjects enrolled, 181 were included for analysis (mean age 33.28±11.78 years old, male:female=2.77:1). In the tremor group, the percentage of females (p=0.036), positive tremor family history (p=0.001), and incidence of polytherapy (p=0.034), treatment duration (>12 months [p=0.013] or >24 months [p=0.008]), and daily dosage (>1,000 mg/d; p=0.003) of VPA, were significantly higher than in the non-tremor group. Treatment with VPA magnesium (p=0.030), alone or in combination with carbamazepine (p=0.040), reduced the incidence of tremor. Furthermore, 176 gene sequencing results ruled out any significant difference between the two groups in the mutation of rs9652490 within LINGO-1 (p=0.443); 86 subjects’ MRI scans indicated no significant differences in the ratio of cerebellar atrophy or the cerebellar-dimension values (p>0.05). However, mutation of rs9652490 within LINGO-1 was correlated with increased cerebellar atrophy (p=0.001), reduced cerebellar-hemisphere thickness (p=0.025), and right-cerebellar-hemisphere longitudinal diameter (p=0.047). Conclusion: Our cohort indicated risk and protective factors of VPA-induced tremor. Although mutation of rs9652490 within LINGO-1 correlated with cerebellar atrophy, neither was correlated with VPA-induced tremors.