Abstract
Background: Using blood to diagnose tuberculosis (TB) infection has been advocated for years, but the results were unsatisfactory. In this study, we modified the nucleic acid extraction and amplification methods to see if diagnostic sensitivity could improve.
Materials and Methods: We prospectively enrolled 74 patients suspected of having TB infection. The blood was collected from each patient. Modified Triton X-100 differentiated lysis method was used to purify mycobacterial nucleic acid from the cytoplasm of mononuclear blood cells. Real-time polymerase chain reaction (PCR) of mpt64 gene fragment were carried out on both mononuclear cells and plasma samples of each patient.
Results: The sensitivity, specificity, positive and negative predictive values of peripheral blood
real-time PCR for diagnosing TB infection was 30.8, 88.8, 95.2 and 15.1%, respectively. The diagnostic sensitivities of pulmonary, extra-pulmonary and disseminated TB were 26.9, 45.4 and 50%, respectively. Of the 20 TB patients whose mycobacterial DNA were detected in blood, 11 (55%) had detectable DNA only in their plasma, 7 (35%) only in mononuclear cells, and 2 (10%) in both plasma and cells. When the PCR results of both plasma and white cell were put together, the sensitivity (30.8%) was more than that of either one alone (13.8% for white cell only, 20% for plasma only).
Conclusion: Peripheral blood mpt64 PCR amplification can diagnose some TB patients, especially those with extra-pulmonary and disseminated TB. Although the method is not sensitive enough to diagnose pulmonary TB, it may be a potentially useful tool in diagnosing extra-pulmonary and disseminated TB. Mycobacterial DNA could be detected in both plasma and white cells. Combining the PCR results of both plasma and white cells could increase the diagnostic sensitivity for TB.
Key words: blood buffy coat, clinical laboratory techniques, real-time polymerase chain reaction, tuberculosis