Objective: This study was conducted to investigate the applicability of Thwaites' diagnostic criteria in order to differentiate tuberculous from bacterial meningitis. Material and Methods: The study was done retrospectively in HIV-negative patients (n = 126) with meningitis. Of the 126 patients, 103 had bacterial meningitis and 23 patients had tuberculous meningitis. The diagnosis was confirmed microbiologically in 59 and 12 patients with bacterial and tuberculous meningitis, respectively. The prediction of tuberculous meningitis was determined by Thwaites' diagnostic scoring using parameters such as age, history of illness, white blood cell count, total cerebrospinal fluid (CSF) white cell count and the percent of neutrophils in CSF. The diagnostic value of the model was assessed by calculating the area under the receiver operating characteristic (ROC) curves. Results: The sensitivity and the specificity of Thwaites' diagnostic scoring were 95.6% (22/23) and 70.8% (73/103), respectively. In microbiologically proven cases the sensitivity and the specificity were 91.7% (11/12), 79.7% (57/59), respectively. The area under the ROC curve value for the diagnostic scoring was 0.92. Conclusions: Thwaites' diagnostic score was found to be helpful in differential diagnosis of tuberculous meningitis, however, the usefulness of diagnostic scoring should be validated in large series especially in patients with viral meningitis and low CSF glucose levels. Copyright (C) 2005 S. Karger AG, Basel.