Objective: To improve the sensitivity and specificity of polymerase chain reaction (PCR) for detecting M. tuberculosis in uncultural clinical samples.
Methods: PCR amplification products were identified with agarose gel electrophoresis (PCR-electrophoresis), then Southern blotted onto a membrane and hybridized with a digoxigenin-labeled M. tuberculosis DNA probe (PCR-probe).
Results: The sensitivity of the purefied DNA for PCR-electrophoresis was 1 picogram, that for PCR-probe was increased to 100 femtogram. In 24 species tested, only M. tuberculosis complex and M. xenopi DNA produced 245-bp amplification bands, but the band from M. bovis did not hybridize with 188-bp digoxigenin-labeled probe. 225 clinical samples were examined by smear, PCR-electrophoresis and PCR-probe, 51 nontuberculous clinical samples were all negative, the positive rates of 173 tuberculous samples were 16.2%, 37.6% and 50.3%, respectively. The smear and culture of 1 sputum sample from a patient infected by nontuberculous mycobacteria were all positive, but its PCR-electrophoresis and PCR-probe were all negative. Identification of species showed that it was a quickly growing mycobacterium. Otherwise, the figures of electrophoresis were classified as four types, and influential factors were also discussed.
Conclusions: The results showed that PCR was a valuable detective tool for early diagnosis and differential diagnosis of tuberculosis, and PCR-probe could increase positive rate of the detection and avoid misjudgements of results.