Etiologic diagnosis of suspected tuberculous meningitis by multiplex PCR of cerebrospinal fluid

ABSTRACT Distinguishing tuberculous meningitis (TBM) from other causes of meningitis remains challenging in high-burden settings because clinical features overlap and existing assays often delay etiologic confirmation. We evaluated MeltArray CNS, a highly multiplexed PCR assay targeting 85 meningitis-associated pathogens, for etiologic diagnosis in patients with suspected meningitis. Analytical performance was assessed using pre-characterized targets. MeltArray was compared head-to-head with metagenomic next-generation sequencing (mNGS) in 79 cerebrospinal fluid (CSF) specimens, with discrepant results adjudicated by Sanger sequencing, and was then prospectively evaluated in 255 consecutive patients with suspected meningitis. Biomarker-based models using routine CSF and serum parameters were also explored. The assay achieved limits of detection of 5 copies/reaction for Mycobacterium tuberculosis (MTB) and 50 copies/reaction for other targets, with no false-positive results in analytical specificity testing. Within its targeted range, MeltArray yielded more confirmed detections than mNGS. In the prospective cohort, MeltArray showed 100.00% sensitivity for definite TBM (35/35; 95% confidence interval [CI], 90.11–100.00%) and 100.00% specificity for MTB detection among non-TBM meningitis cases (39/39; 95% CI, 89.32–100.00%). Among 49 MTB-positive TBM cases, 21 (42.86%) showed co-detections of additional pathogens. Test areas under the curve (AUCs) were 0.877 (95% CI, 0.760–0.994) for distinguishing infectious meningitis from noninfectious mimics and 0.776 (95% CI, 0.617–0.934) for distinguishing MTB-positive from MTB-negative cases. MeltArray enables rapid etiologic confirmation and may facilitate earlier TBM diagnosis in high-burden settings. Co-detections highlight microbiologic complexity with potential treatment implications, although viral findings should be interpreted cautiously. Biomarker models may aid triage and risk stratification but do not replace pathogen confirmation. IMPORTANCE Tuberculous meningitis (TBM) is a life-threatening infection that requires a rapid and accurate diagnosis to guide effective treatment. Conventional diagnostic methods are often slow or insufficiently sensitive, leading to delays in therapy and potential exposure to unnecessary medications. In this study, we evaluated a rapid multiplex molecular assay for patients with suspected tuberculous meningitis. Rapid detection of Mycobacterium tuberculosis together with alternative infectious causes of meningitis was achieved within approximately 2.5 h, supporting earlier etiologic clarification during initial clinical evaluation. Detection of additional pathogens in some patients further supported the value of broad-spectrum molecular testing in the differential diagnosis of central nervous system infections in high-burden settings. Routine laboratory biomarkers may assist clinical triage but do not replace rapid pathogen confirmation.