Diagnostic accuracy of a uniform tuberculous meningitis research case definition in children

BACKGROUND: Bacteriological confirmation of tuberculous meningitis (TBM) is problematic and rarely guides clinical management. A uniform TBM case-definition has been proposed for research purposes, but its diagnostic accuracy in a clinical setting has not been evaluated. METHODS: We prospectively enrolled patients with meningitis confirmed by cerebrospinal fluid analysis aged 3 months to 13 years at Tygerberg Children’s Hospital, Cape Town, South Africa. Participants were investigated for TBM, bacterial and viral meningitis. Criteria that differentiated TBM from other causes were explored and the accuracy of a probable TBM score assessed by comparing bacteriologically-confirmed cases to “non-TBM” controls. RESULTS: Of 139 meningitis suspects, 79 were diagnosed with TBM (35 bacteriologically-confirmed), 10 with bacterial meningitis and 50 with viral meningitis. Among bacteriologically-confirmed TBM, 15 were M. tuberculosis culture positive, 20 were culture negative but positive on GenoType MTBDRplus® and Xpert MTB/RIF®, 18 were positive on only a single commercial nucleic acid amplification test. A probable TBM score provided good diagnostic accuracy; sensitivity 74% (95% confidence interval (57-88%) and specificity 97% (95% confidence interval 86-99%) compared to bacteriologically-confirmed TBM. CONCLUSION: A probable TBM score demonstrated excellent specificity compared to bacteriological confirmation, justifying consideration as a “rule-in” test. However, accurate diagnosis of early TBM remains a challenge. Childhood tuberculous meningitis algorithm 55 C H A PT ER 3 .2 INTRODUCTION In 2013, there was an estimated 9.0 million new tuberculosis (TB) cases world-wide. Central nervous system involvement, mostly tuberculous meningitis (TBM), is considered to account for approximately 1% of all cases. Although TB is predominantly a pulmonary disease, extrapulmonary involvement is particularly common in young children and immune-compromised individuals. In TB endemic areas with access to routine Haemophilus influenza type-B and pneumococcal vaccination, such as Cape Town (South Africa), TBM has surpassed other causes of bacterial meningitis. In clinical practice, diagnosis is based on a combination of clinical, laboratory, and neuroimaging findings. The early clinical presentation of TBM is often non-specific. With delayed treatment initiation TBM outcome is often poor, emphasizing the importance of early and accurate diagnosis. TBM is a paucibacillary disease and the sensitivity of cerebrospinal fluid (CSF) microscopy is low. The sensitivity of CSF culture is increased compared to microscopy, but remains low and the result rarely influences clinical management due to delays of up to 6 weeks. Nucleic-acid amplification tests (NAATs) offer the prospect of rapid and specific diagnosis, but sensitivity remains suboptimal. A recent meta-analysis assessing the accuracy of commercial NAATs for the diagnosis of TBM showed a sensitivity of 64% and specificity of 98%, compared to CSF culture. In clinical practice, TBM is diagnosed based on a combination of clinical, laboratory, and neuroimaging findings. Previous studies identified key characteristics suggestive of TBM. Initial symptoms and signs are non-specific but persistent, and include poor weight gain or weight loss, fever, and lethargy. More specific signs have a sub-acute onset and include meningeal irritation, reduced consciousness, focal neurological deficits, raised intracranial pressure, brainstem dysfunction, and cranial nerve palsies. Typical CSF features include a moderately increased white cell count with lymphocyte predominance, increased protein, and decreased glucose. Apart from hyponatremia, no peripheral blood test result has consistently been associated with TBM. On neuroimaging, hydrocephalus, basal ganglia infarctions, and basal meningeal enhancement are common features. Clinical prediction rules that differentiate TBM from other forms of meningitis in adults include five parameters: young age (<36 yrs), sub-acute presentation (symptom duration >5 days), normal peripheral white blood cell count, moderately raised CSF white cell count (<500 cells/ml), and lymphocyte predominance (CSF neutrophil proportion <50%). Good diagnostic accuracy was obtained using both a composite clinical