Abstract
BACKGROUND: Blood stream tuberculosis (TB), caused by Mycobacterium tuberculosis (MTB) is common among HIV-positive patients, turning rapidly fatal unless detected and treated promptly. Blood culture is currently the standard test for the detection of MTB in whole blood but results take weeks; patients deteriorate markedly and often die before a diagnosis of blood stream TB is made. Rapid molecular tests on whole blood, with potential for same day diagnosis of blood stream TB usually show low sensitivity due to the problem of insufficient MTB DNA template when extraction is performed directly on low blood volumes. This study assessed the influence of blood volume on the sensitivity of a HyBeacon PCR assay-the FluoroType MTB (Hain Lifescience, Nehren, Germany) on direct detection of MTB in whole blood. METHODS: Prospective recruitment of HIV-positive patients with clinical suspicion of blood stream TB but not on anti-TB or HIV drug treatment was done. Venous blood samples were collected and DNA extracted using the MolYsis (Molzym, Bremen, Germany) methods; for study A, from duplicate 1 ml (42 patients) and for study B (31 patients) from 9 ml EDTA blood samples. The FluoroType MTB PCR assay targeting an IS6110 sequence was performed and results compared with blood culture. RESULTS: The diagnostic sensitivity and specificity of the FluoroType MTB PCR in study A was 33% and 97%, respectively. Corresponding values in study B were 71% and 96%, respectively. In both studies, one case each of blood culture-negative blood stream TB was detected with the FluoroType MTB PCR assay. The median time to positivity of blood culture was 20.1 (range 12-32) for study A and 19.9 days (range 15-30) for study B. CONCLUSION: Larger blood volumes (9 ml) improved and gave acceptable sensitivity of direct PCR diagnosis of blood stream TB.