小児インフルエンザ関連脳炎の予後因子特定とモデル構築：後ろ向き症例対照研究

To identify risk factors for poor outcomes in IAE and to guide early diagnosis and intervention.

We retrospectively collected clinical, laboratory, imaging, treatment, and outcome data of children with IAE. Univariate and survival analyses identified prognostic factors and supported the development of a prediction model.

Among 42 children with IAE, 16 (38.1 %) had poor outcomes, including 6 (14.3 %) in-hospital deaths. Poor outcomes were associated with deep coma (75 %), structural brain damage in MRI (86.7 %), and EEG abnormalities (100 %). Univariate analysis revealed that CSF IL-10, CSF pro, blood ALT, AST, CK-MB, and LDL-H were significant discriminators between the good outcome and poor outcome groups (all P < 0.001). Subsequent survival analysis showed AST had the best predictive value, while CK-MB was identified as the only independent risk factor of mortality. A predictive model using optimal cutoffs for AST (>132 IU/L) and CK-MB (>136 IU/L) was developed to stratify patients into good outcome, severe sequelae, and death groups, achieving prediction accuracies of 92.3 %, 100 %, and 66.7 %, respectively.

Our findings indicate that elevated levels of CSF IL-10, CSF pro, and blood AST, ALT, CK-MB, and LDL-H at admission are significantly associated with poor outcomes in pediatric IAE patients. Specifically, AST ≥132 IU/L and CK-MB ≥136 IU/L may serve as early warning indicators of poor outcomes, highlighting the need for timely recognition and intervention in high-risk patients.