A Diagnostic Index To Distinguish Tuberculosis Meningitis From Other Infectious Meningitis In Adults

Objectives Diagnosis of tuberculosis meningitis still challenges clinicians. Microscopy or culture of Mycobacterium tuberculosis is either insensitive or time consuming, while nucleic acid amplification test is not available in resource limiting areas. We aim to derive diagnostic rules for early discriminating tuberculosis meningitis from other forms of meningitis basing on the clinical, laboratory and radiographic features.Methods 910 patients who satisfied the inclusion and exclusion criteria were recorded in this study retrospectively, including 349tuberculosis meningitis,481 viral meningitis,49bacterial meningitis and 31cryptococcal meningitis. We compared clinical, laboratory and radiographic features collected at admission in tuberculosis versus viral meningitis, tuberculosis versus bacterial meningitis, tuberculosis versus cryptococcal meningitis, tuberculosis versus non-tuberculosis meningitis. Features found significantly predicted in univariate analysis were included and modeled by multiple logistic regressions to derive the diagnostic rules. The diagnostic index were defined by rounded theβcoefficients.Results Four diagnostic rules were derived in this study for distinguishing tuberculosis meningitis from viral, bacterial, cryptococcal and non-tuberculosis meningitis. Features found to be independently predicted diagnosis of tuberculosis meningitis were:anorexia, weight loss, neck stiffness, low blood sodium, decreased cerebrospinal fluid glucose and chloride, elevation of cerebrospinal fluid protein, in combination with chest radiograph suggestive of lung tuberculosis, and basal meningeal enhancement, hydrocephalus, granuloma found in cerebral radiograph.Conclusions Diagnostic rules derived in this study may help in early diagnosis of tuberculosis meningitis by optimizing clinical, laboratory and radiographic features at admission. Although samples of bacteria and cryptococcal meningitis were relatively small and it was done retrospectively. We suggest that these rules be applied in patients with meningitis and further investigation should be done to validate these rules.