The urgent need for sensitive and specific diagnostic techniques for secondary bacterial meningitis remains a significant challenge in neurosurgical and intensive care units. Combining various clinical indicators, biochemical parameters, as well as biomarkers and metabolites found in cerebrospinal fluid (CSF), offers a promising approach for developing multivariate diagnostic models. In this research, 96 CSF samples from 53 patients suspected of secondary meningitis were examined. The first group included patients with residual effects of severe brain damage and chronic critical illness: 7 patients with secondary bacterial meningitis and 29 patients without it. The second group consisted of individuals following neurosurgical procedures: 10 patients with secondary bacterial meningitis and 7 patients without this complication. Combining the data, one group of 33 CSF samples represented patients with secondary bacterial meningitis, while another group of 63 CSF samples included those without the infection. These groups showed significant differences in cellular and biochemical composition, with higher levels of biomarkers such as interleukin-6, S100 protein, and some aromatic metabolites in the meningitis group. Univariate prognostic models based on three aromatic microbial metabolites of phenylalanine, tyrosine, and tryptophan achieved excellent discrimination with an AUC-ROC of more than 0.9. A comprehensive multivariate model incorporating all measured biomarkers and metabolites yielded an AUC-ROC of 0.94, with a sensitivity of 94% and specificity of 86%, representing the most accurate method currently available for diagnosing secondary bacterial meningitis in patients with both acute and chronic conditions.