Clinical and laboratory profiles of acute central nervous system infections in a Chinese pediatric cohort: a five-year retrospective analysis.

BACKGROUND: Acute central nervous system (CNS) infections in children cause severe neurological morbidity, but early diagnosis and risk assessment remain challenging. This study aimed to characterize the clinical and laboratory features, pathogen spectrum, and risk factors for severe disease in a five-year pediatric cohort. METHODS: We retrospectively analyzed children diagnosed with acute CNS infections at the Shandong Regional Children’s Medical Center from January 2020 to December 2024. Clinical, laboratory, and outcome data were extracted from electronic medical records. RESULTS: Among 422 enrolled children, 132 (31.3%) had clinically diagnosed bacterial and 290 (68.7%) nonbacterial CNS infections. Bacterial infections were associated with significantly younger age (median 0.22 vs. 4.02 years), longer hospitalization (23.5 vs. 13 days), and higher rates of systemic complications (all P < 0.001). To distinguish bacterial from nonbacterial CNS infections, ROC analysis demonstrated that CSF protein had excellent discriminative ability (AUC = 0.917), followed by serum albumin (AUC = 0.869) and CSF white blood cells (AUC = 0.837). CRP (AUC = 0.801) and PCT (AUC = 0.682) were also evaluated. A total of 175 children (41.5%) required pediatric intensive care unit (PICU) admission. In addition to impaired consciousness, multivariate analysis identified respiratory failure (aOR = 39.76), somnolence (aOR = 15.03), vomiting (aOR = 3.93), seizure (aOR = 2.53), pneumonia (aOR = 2.51), age (aOR = 1.09), and length of hospital stay (aOR = 1.02) as independent predictors of PICU admission (all P < 0.05). Pathogens were identified in 94 cases (54 bacterial and 40 viral). Among bacterial pathogens, Streptococcus pneumoniae (n=13) and Escherichia coli (n=10) predominated, with 68.5% (37/54) occurring in infants ≤1 year. mNGS detected pathogens missed by conventional culture and identified co-infections. CONCLUSIONS: In addition to CSF analysis, CRP can help distinguish bacterial from nonbacterial CNS infections; serum albumin and total protein (~50% sensitivity) can also help but should not be used alone. Impaired consciousness, respiratory failure, somnolence, vomiting, seizure, and pneumonia were key risk factors for PICU admission.The relatively high rate of bacterial infections in infants warrants particular clinical attention. These findings can guide clinical practice and improve outcomes in this vulnerable population.