RESUMO
PURPOSE: Our objective was to investigate whether diagnosis of tuberculous meningitis (TBM) with microbiological and molecular analysis was affected by prior empirical non anti tuberculosis antibiotics or by duration of symptoms before lumbar puncture. MATERIALS AND METHODS: We retrospectively evaluated medical records of patients with TBM confirmed by positive culture, nucleic acid amplification techniques (NAATs) or Ehrlich-Ziehl-Neelsen staining (EZNs) from the cerebrospinal fluid (CSF) or by characteristic results of biochemical analysis of CSF combined with a typical clinical manifestation. RESULTS: 68 adult patients were analyzed. The isolation rates for NAATs, Lowenstein-Jensen (LJ) culture, BACTEC and EZNs were 70.6%, 69.1%, 67.6% and 26.5%, respectively. Biochemical analysis of CSF samples revealed: pleocytosis (median 224 [range 78-380]cells/mm3) with lymphocyte predominance (76 [45-90]%), elevated levels of protein (2.43 [1.50-3.84]g/l) and lactic acid (5.0 [3.9-7.2]mmol/l). Forty (65%) patients received no anti-tuberculosis antibiotic treatment before the diagnostic lumbar puncture. The were no significant differences in the microbiological and biochemical CSF analyses, between the patients who received and those who did not receive non anti-TB empirical antibiotic treatment. The median duration of symptoms before the diagnostic lumbar puncture was 24 (range 11-61) days. No significant differences in microbiological and biochemical analysis of CSF where found when comparing patients with duration of symptoms lasting above and less than the median time. CONCLUSIONS: Neither prior non anti-TB antibiotic therapy, nor the duration of symptoms before diagnostic lumbar puncture have any effect on confirmation of TBM by microbiological and biochemical CSF analysis.