The discovery of the polymerase chain reaction (PCR) as a means to generate detectable signals from as little as a single molecule of target deoxyribonucleic acid (DNA) [1] has led to the development of extremely powerful and rapid diagnostic tests for a number of infectious agents. PCR was therefore expected to improve significantly the sensitivity of tests for the diagnosis of tuberculosis (TB), compared with the standard stains for acid fast bacilli (AFB) on sputum smears. Since the first report by HANCE et al.[2] on the use of PCR amplification of the 65 kDa heat shock protein (HSP) gene for the diagnosis of Mycobacterium tuberculosis in clinical samples, a number of M. tuberculosis DNA sequences have been tested as amp-lification targets [3]. Over the past ten years, several PCR test formats have been employed and nested PCR tests have been described which were endowed with a detec-tion power 1,000-fold greater (lower detection limit, 0.1 M. tuberculosis colony-forming unit (cfu)·mL -1 ) than that of the standard PCR with 35 amplification cycles (detection limit, 100 cfu·mL -1 ) [4].Among DNA targets, the IS6110 insertion sequence, a M. tuberculosis complex-specific sequence which is repeated 10-16 times in the chromosome of M. tuberculosis [5], has been the most widely used. This test has the power of amplifying a fraction of a chromosome of M. tuberculosis and with this test it has been possible to accomplish something as spectacular as identifying mycobacteria buried in mummies thousands of years old [6]. However, the test is difficult to handle and its reproducibility across laboratories has been rather disappointing [7]. Even more importantly, all the amplification tests using either the PCR or other ribonucleic acid (RNA) and DNA amplification techniques that have been developed, bottled and marketed for the diagnosis of M. tuberculosis in routine laboratories do not seem sufficiently sensitive for the diagnosis of paucibacillary TB. Notwithstanding their extreme power in the laboratory setting and although the rapid amplification tests are more sensitive than AFB stains, home grown and commercial tests alike, both still fail to diagnose about 50% of microscopy-negative sputum samples [8] i.e., those samples containing less than the 5-10,000 M. tuberculosis organisms·mL -1 of sample that are required for a positive AFB smear.With this background, an American Thoracic Society (ATS) panel recently recommended that the rapid amplification tests should be used for the confirmation of rapid diagnosis of M. tuberculosis versus nontuberculous mycobacteria in individuals with an AFB-positive sputum smear. With a direct amplification test it can be rapidly determined (with >95% sensitivity) whether an individual has active, multibacillary TB (requiring isolation and contact screening), or a nontuberculous mycobacterial infection (not requiring any of the above). In contrast, the same ATS panel could not recommend the use of rapid amplification tests for the routine diagnosis of paucibacillary TB, due to thei...