Study objectives: To investigate the clinical usefulness of amplification (COBAS AMPLICOR; Roche Diagnostics Systems; Branchburg, NJ) on bronchoscopic aspirate specimens in the diagnosis of pulmonary tuberculosis, with particular regard to the possibility of false-positive results in subsequent specimens due to residual Mycobacterium tuberculosis DNA.
Design and setting: A prospective clinical study at a tertiary referral medical center.
Participants and methods: Four hundred fiberoptic bronchoscopic procedures were performed, using seven bronchoscopes on 335 consecutive patients, for therapeutic or diagnostic purposes. Serial bronchial aspirates were collected and tested for M tuberculosis, using COBAS AMPLICOR (CA). Bronchoscopes were cleaned and disinfected automatically, between patient use, by the same endoscope washer. The name of each bronchoscope and the sequence of its use were recorded, together with the sequence of washing. The CA results were compared with the bacteriologic and histologic results for M tuberculosis infection. When there was a suspicion of contamination, outward polymerase chain reaction analysis was performed.
Results: Of 392 specimens (332 subjects), excluding the 8 specimens (4 subjects) in which bacteriologic and histologic analyses were omitted, a smear-positive result for acid-fast bacilli (AFB), culture-positive or biopsy-positive results, and CA-positive results were obtained in 16, 49, and 32 specimens, respectively. In AFB smear-positive subjects, the sensitivity, specificity, positive predictive values (PPVs), and negative predictive values (NPVs) were 92%, 67%, 92%, and 67%, respectively. In AFB smear-negative subjects, the sensitivity, specificity, PPV, and NPV values were 38%, 99%, 74%, and 94%, respectively. The CA test was more sensitive than the AFB smears for the diagnosis of pulmonary tuberculosis (53% vs 27%, respectively; p < 0.05). False-positive CA results were seen in only six specimens. Three of these six subjects received a diagnosis of pulmonary tuberculosis on clinical and radiologic grounds, and none of the six results seemed to be associated with bronchoscopic cross-contamination.
Conclusions: Adequately cleaned and disinfected bronchoscopes did not cause false-positive amplification test results for M tuberculosis on bronchial aspirates by cross-contamination. Furthermore, sensitivity was greater with the CA tests. Therefore, CA tests on bronchial aspirates seem to be useful in the diagnosis of pulmonary tuberculosis.