Two commercial primer kits and detection systems, the Roche AMPLICOR MYCOBACTERIUM test and the Digene primer-probe kit with the SHARP Signal System, were compared to in-house PCR as well as standard culture techniques. For the 27 culture-positive specimens, the Roche AMPLICOR MYCOBACTERIUM test detected 20 specimens, the Digene system detected 19, and in-house PCR detected 21. Of the 86 culture-negative specimens, 13 were positive by the Roche AMPLICOR MYCOBACTERIUM test, 16 were positive by the Digene system, and 21 were positive by in-house PCR. When clinical situations were evaluated, 11 of 13 culture-negative Roche AMPLICOR MYCOBACTERIUM test-positive specimens, 10 of 16 culture-negative Digene system-positive specimens, and 13 of 21 culture-negative-in-house PCR-positive specimens were diagnosed as true-positive specimens. The sensitivities of Roche AMPLICOR MYCOBACTERIUM test, the Digene system, and in-house PCR were 73.81, 69.05, and 80.95%, and the specificities were 97.18, 91.55 and 88.73%, respectively. The positive predictive values were 93.94, 82.86, and 80.95%, and the negative predictive values were 86.25, 83.33, and 88.73%, respectively. For the commercial kits, the Roche AMPLICOR MYCOBACTERIUM test seems to be more sensitive and specific than the Digene system. However, the Roche AMPLICOR MYCOBACTERIUM test cannot be used on nonrespiratory specimens.