Clinical significance of a polymerase chain reaction assay for the detection of Mycobacterium tuberculosis.

Various investigators have reported rapid detection of Mycobacterium tuberculosis (MTB) in clinical samples by polymerase chain reaction (PCR). To improve the specificity and efficiency of PCR, the authors adopted a variety of conditions, then analyzed sputum specimens from 217 patients clinically suspected of having MTB. Sputum samples were sonicated to obtain MTB DNA. The DNA was subjected to PCR using primer sets from the region of 650-900 in MTB. The PCR product was detected by direct gel electrophoresis and Southern blot hybridization using digoxigenin-labeled MTB-specific probe. The results of smears, cultures, and PCR were concordant in 93% (202) of the 217 specimens and discordant in 7% (15). Fifteen of the discordant specimens, all from patients who had received antituberculosis medications for days to months, were PCR positive. Of these, 11 were culture negative and 3 were smear negative. Only one specimen was false negative on PCR. Our results indicate that PCR is the method of choice when clinical suspicion is high, but smears or cultures are negative. When smears are positive, PCR permits rapid distinction between MTB and other mycobacterial infections. Because PCR can detect nonviable MTB DNA, positive PCR should be interpreted in conjunction with clinical information.