Amino acid dysregulation in cerebrospinal fluid of stroke patients serving as diagnostic biomarkers in post-neurosurgical bacterial meningitis.

Post-neurosurgical bacterial meningitis (PNBM) is a severe complication following neurosurgical operations. However, clinical diagnosis of PNBM is difficult because of the complicated pathological conditions. This study aims to investigate alterations in amino acid metabolism in hemorrhagic stroke patients with PNBM through targeted metabolomics analysis. Cerebrospinal fluid (CSF) samples were collected from 66 hemorrhagic stroke patients who underwent neurosurgical operation in our department. Baseline data were retrospectively analyzed for two patient groups: the post-neurosurgical bacterial meningitis group (PNBM, n = 40) and the non-post-neurosurgical bacterial meningitis (non-infected control group, n = 26), classified based on established diagnostic criteria for intracranial infection. A targeted analysis of 36 amino acids and their derivatives in CSF was performed using liquid chromatography-mass spectrometry (LC-MS). Candidate biomarkers were identified through Student's t-test, fold change (FC) analysis, Variable Importance in Projection (VIP), and logistic Least Absolute Shrinkage and Selection Operator (LASSO) regression. The diagnostic performance was evaluated using Receiver Operating Characteristic (ROC) curve analysis. Twenty-one amino acids and their derivatives were found to be significantly downregulated in the PNBM group. Logistic LASSO regression identified Glycine (p