Limits of commercial molecular tests for diagnosis of pulmonary tuberculosis.

Several studies report high specificity, but variable sensitivity, of Amplified Mycobacterium tuberculosis Direct Test (AMTDT, Gen-Probe) based on ribosomal ribonucleic acid (rRNA) amplification and Amplicor Mycobacterium tuberculosis test (Amplicor, Roche) based on deoxyribonucleic acid (DNA) amplification for diagnosis of pulmonary tuberculosis. We have retrospectively evaluated these assays on selected acid-fast bacilli (AFB)-positive and -negative smear specimens and compared the results obtained from nucleic acid amplification with those of AFB staining of semi-quantitative cultures as determined by radiometric Bactec and Löwenstein-Jensen cultures. In comparison to cultures, Amplicor and AMTDT assays exhibited identical overall sensitivities of 80%, while the staining had a lower sensitivity of 62%. The sensitivities of Amplicor and AMTDT were 98% and 100%, respectively, for the AFB-positive specimens, and 50 and 46%, respectively for the AFB-negative specimens in comparison to cultures. The sensitivities of both assays appeared similar, and were directly related to the number of bacilli in the specimens studied. The low sensitivity (50%) for smear-negative specimens showed that current amplification assays may be unsuitable to replace cultures for diagnosis of tuberculosis. The decision to perform these molecular techniques should result from close co-operation between clinicians and microbiologists, taking into account the sensitivity results reported here, as well as the expense of the assays.