Diagnostic usefulness of T-cell based assays for tuberculous meningitis in HIV-uninfected patients

Abstract

Early diagnosis and treatment of tuberculous meningitis (TBM) is essential for a positive outcome, but sensitive, specific, and rapid diagnostic tests for TBM are lacking. We evaluated the diagnostic utility of enzyme-linked immunosorbent spot (ELISPOT) assays in HIV-uninfected patients with suspected TBM. All HIV-uninfected patients with suspected TBM were prospectively enrolled at a tertiary care hospital in an intermediate TB-burden country, during a 6-year period. ELISPOT assays were performed on peripheral blood mononuclear cells (PBMC) and cerebrospinal fluid-mononuclear cells (CSF-MC). Of the 276 evaluable patients, 90 (33%) were classified as having TBM (30 definite cases, 19 probable, and 41 possible), and 186 (67%) as having non-TBM. When comparing definite TBM versus non-TBM, the sensitivity and specificity of the PBMC ELISPOT assay (≥6 spots; manufacturer's recommended cut-off) for diagnosing TBM were 96% (95% CI, 82-100) and 58% (95% CI, 50-66), respectively. The CSF-MC ELISPOT assay (≥38 spots; receiver operating characteristic [ROC]-derived cut-off) was a useful rule-in test with specificity of 95% (96% CI, 90-98). Its sensitivity was 68% (95% CI, 45-86), which was superior those of AFB smear microscopy (14%; P < 0.001) and CSF Mycobacterium tuberculosis PCR (41%; P = 0.07). Combining this assay with M. tuberculosis PCR, clinical score, and both together increased sensitivity to 86%, 91%, and 95%, respectively, while retaining about 95% specificity. The CSF-MC ELISPOT assay appears to be a rapid and accurate rule-in test for the diagnosis of TBM and a useful adjunct for diagnosing TBM in HIV-uninfected patients.