Abstract
INTRODUCTION: Autoimmune encephalitis/encephalopathy (AE) is a complex and heterogeneous disease, making it difficult to predict the prognosis. The neutrophil-to-lymphocyte ratio (NLR) has emerged as a potential prognostic tool, but its usefulness remains a matter of debate. This study aimed to explore prognostic factors in cases of clinically definite or probable AE, including those with autoantibody-negative, or unknown status. METHODS: Data on patients diagnosed with definite or probable AE, including those with autoantibody-negative, or unknown status, were retrospectively collected from the admission records of our department between January 2013 and December 2022. These patients were then categorized into either a good- or poor-response group, based on their short-term treatment response. Clinical characteristics, auxiliary examinations, and treatments were compared between the two groups. A multivariable logistic regression model was constructed to identify independent predictors of poor short-term treatment response by Akaike information criterion backward stepwise method. RESULTS: A total of 31 patients were included in the final analysis, with 18 of them included in the poor-response group. In the univariable analysis, the poor-response group had a higher proportion of patients with a modified Rankin Scale (mRS) high score upon admission, female, epileptic seizures, or NLRs of 3.93 or higher than the good-response group (all p < 0.10). Furthermore, the multivariable logistic regression analysis revealed that the mRS score upon admission [OR: 5.51, 95% confidence intervals (CI): 1.29-23.50, p = 0.02], epileptic seizures (OR: 10.01, 95% CI: 1.16-86.66, p = 0.04), and NLRs of 3.93 or higher (OR: 11.37, 95% CI: 1.12-114.68, p = 0.04) were significantly associated with poor short-term treatment response. CONCLUSION: The NLR may play a supplementary role in predicting the short-term treatment response in patients diagnosed with definite or probable AE, including those with autoantibody-negative, or unknown status.