Cookies on this website
We use cookies to ensure that we give you the best experience on our website. If you click 'Continue' we'll assume that you are happy to receive all cookies and you won't see this message again. Click 'Find out more' for information on how to change your cookie settings.

Tuberculous meningitis (TBM) remains difficult to diagnose. We prospectively evaluated a diagnostic algorithm for TBM in 205 HIV-negative patients with meningitis and a low CSF glucose. Patients were classified as having TBM or bacterial meningitis (BM) by two diagnostic methods: logistic regression method (LRM) and classification and regression tree (CART). We performed analyses of TBM versus BM and TBM versus non-TBM in all patients and in patients with microbiologically confirmed diagnoses. Diagnostic sensitivities for TBM were 99% (LRM) and 87% (CART). For BM, diagnostic sensitivities were 81.5% (LRM) and 86.5% (CART) in the primary analysis and 86.5% (LRM) and 74% (CART) in the secondary analysis. In microbiologically confirmed cases, similar rates were achieved. These figures are superior to microbiological confirmation rates in routine laboratories and support the use of this algorithm in high-prevalence TB settings with limited diagnostic facilities. Validation in an HIV-endemic setting is required.


Journal article


Am J Trop Med Hyg

Publication Date





555 - 559


Adult, Algorithms, Bacteria, Humans, Predictive Value of Tests, Reproducibility of Results, Sensitivity and Specificity, Tuberculosis, Meningeal