Diagnosis of cryptococcal and tuberculous meningitis in a resource-limited African setting

OBJECTIVES Cryptococcal meningitis (CM) and tuberculous meningitis (TBM) are common in HIV-infected adults in Africa and difficult to diagnose. Inaccurate diagnosis results in adverse outcomes. We describe patterns of meningitis in a Malawian hospital, focusing on features which differentiate CM and TBM with the aim to derive an algorithm using only clinical and basic laboratory data available in this resource-poor setting. METHODS Consecutive patients admitted with meningitis were prospectively recruited, clinical features were recorded and cerebrospinal fluid (CSF) was examined. RESULTS A total of 573 patients were recruited, and 263 (46%) had CSF consistent with meningitis. One hundred and twelve (43%) had CM and 46 (18%) had TBM. CM was associated with high CSF opening pressure and low CSF leukocyte count. Fever, neck stiffness and reduced conscious level were associated with TBM. A diagnostic index was constructed demonstrating sensitivity 83% and specificity 79% for the differentiation of CM and TBM. An algorithm was derived with 92% sensitivity for the diagnosis of CM, but only 58% specificity. CONCLUSIONS Although we demonstrate features associated with CM and TBM, a sufficiently sensitive and specific diagnostic algorithm could not be derived, suggesting that the diagnosis of CM and TBM in resource-limited settings still requires better access to laboratory tools.