Validation of diagnostic algorithm to differentiate between tuberculous meningitis and acute bacterial meningitis

Background: Discrimination between tuberculous and acute bacterial meningitis is difficult by clinical features alone and laboratory methods may only supplement the clinical suspicion. We aimed to validate the diagnostic criteria by Thwaites et al. [1] and construct our own diagnostic predictors based on the clinical and laboratory features. Methods: 380 patients of acute bacterial meningitis (ABM) and 210 patients of tuberculous meningitis (TBM) were enrolled retrospectively from June 2004 to June 2007 and prospectively from July 2007 to September 2008. HIV positive patients were excluded. Detailed history, clinical examination CSF analysis, haematological, biochemical investigations and imaging was performed in all patients. Results: Factors associated with the diagnosis of TBM in the present study included rural area of residence, longer duration of disease, presence of clear CSF, lower percentage of CSF neutrophils, presence of diplopia and hemiparesis. On validation, age did not appear as a significant factor in our population. The diagnostic algorithm from our study group had a sensitivity of 95.71% and specificity of 97.63%. Conclusions: The diagnostic criterion has a fair validation in our population when the age factor is excluded. The rule is useful in HIV negative patients with low CSF sugar and negative organism yield in the CSF. (C) 2012 Elsevier B.V. All rights reserved.