The performance of the DR. MTBC PCR-based assay and the BD ProbeTec ET Mycobacterium tuberculosis Complex Direct Detection (DTB) assay for the direct detection of Mycobacterium tuberculosis was evaluated using 1,066 consecutive clinical respiratory samples collected from 494 patients who did not have old cases of pulmonary tuberculosis and were not receiving antituberculosis treatment at National Taiwan University Hospital from January to February 2005. The results of both assays were compared to the "gold standard" of combined culture results and clinical diagnosis. The overall sensitivity and specificity of the DR. MTBC Screen assay were 56.6% and 98.9%, respectively, and of the DTB assay were 63.2% and 98.4%, respectively. The positive and negative predictive values for the DR. MTBC Screen assay were 84.5% and 95.4%, respectively, and for the DTB assay were 81.7% and 96.0%, respectively. The DR. MTBC Screen assay produced 11 false-positive results for 11 patients, including three samples yielding non-M. tuberculosis mycobacteria (one each for M. abscessus, a mixture of M. abscessus and M. chelonae, and unidentified non-tuberculosis mycobacteria). The DTB assay produced 15 false-positive results for 13 patients, including five samples from four patients yielding non-tuberculosis mycobacteria (two for M. abscessus, one for a mixture of M. abscessus and M. chelonae, and two for unidentified non-tuberculosis mycobacteria). This study demonstrated that the DR. MTBC Screen assay has a similar diagnostic value but fewer false-positive results than the DTB assay for respiratory specimens.