DOI: 10.1177/17562864261477197 ISSN: 1756-2864

Relapse risk and functional outcomes in anti-LGI1 encephalitis: A multicenter retrospective analysis

Chun-Hong Shen, Bo Jin, Ye Du, Wen-Jing Shen, Rui Li, Yin-Xi Zhang, Qi-Lun Lai, Hao Song, Shan Wang, Mei-Ping Ding, Shuang Wang, Yi Guo, Meng-Ting Cai

Background

Anti-leucine-rich glioma-inactivated protein 1 (LGI1) encephalitis presents relapse risk and potential functional impairment, with no established maintenance treatment guidelines.

Objectives

The study aimed to uncover prognostic signatures, and identify predictors of relapse and poor outcome using real-world evidence.

Design

This multicenter retrospective study included anti-LGI1 encephalitis patients from five tertiary hospitals in Eastern China between January 2015 and January 2024, with ≥ 24-month follow-up.

Methods

Relapse was defined as the presence of new or worsening clinical features after at least 3 months of stability or improvement. Poor outcome was defined as a modified Rankin Scale score > 2 at last follow-up.

Results

A total of 104 patients were included (median age 56.5 years, 60.6% male). Following the acute phase, the majority (66.3%) received prolonged corticosteroid treatment (≥6 months) without steroid-sparing maintenance immunotherapy (SSMI), whereas 16.3% received SSMI. Twelve patients (11.5%) experienced relapse. Hyponatremia at diagnosis was more common in relapsing patients ( p = 0.017), and was significantly associated with an increased risk of relapse in Cox regression (hazard ratio = 4.67, 95% confidence interval [CI] = 1.41–15.53, p = 0.012). Poor outcome was observed in 17 patients (16.3%), with male sex (odds ratio [OR] = 5.12, 95% CI = 1.27–34.48, p = 0.041) and hyponatremia (OR = 4.69, 95% CI = 1.54–15.47, p = 0.008) as potential risk factors.

Conclusion

This study suggests that hyponatremia is associated with an increased risk of relapse and poor outcome in anti-LGI1 encephalitis. These findings may facilitate individualized risk stratification, but further validation in larger prospective cohorts is needed.

More from our Archive