The reliability of the Gen-Probe enhanced Amplified Mycobacterium Tuberculosis Direct Test (MTD) for identification of Mycobacterium tuberculosis complex (MTBC) in BACTEC 12B broth cultures of respiratory specimens was evaluated by testing aliquots from 268 bottles with a growth index of 50. MTD results were compared to those obtained by usual laboratory protocol, whereby MTBC was identified by DNA probe (Gen-Probe, Inc.) testing sediment from broth samples or colonies on a solid medium. For the first 134 cultures, from which 68 mycobacterial isolates (including 27 MTBC isolates) were recovered, both fresh and frozen aliquots were tested. MTD results for the frozen aliquots agreed with the identification by usual protocol in all cases, whereas there was one false-negative MTD result with fresh aliquots. For the remaining 134 cultures, only frozen aliquots were tested. Of the total 268 broth cultures (from 210 patients) evaluated, 137 (51.1%) grew mycobacteria, including 60 MTBC isolates. All 60 isolates were MTD positive, as was one additional culture that grew Mycobacterium gordonae. The latter culture was from a patient who was diagnosed with tuberculosis a few months earlier and was on therapy; therefore, the MTD result was considered a true positive. Sensitivity, specificity, and positive and negative predictive values of MTD were 100%. The mean times from specimen receipt to identification of MTBC were 15 (±1) days (range, 4 to 27 days) for BACTEC plus MTD and 19 (±1) days (range, 6 to 36 days) for the usual protocol (P < 0.001). These data indicate that the MTD is a rapid, reliable method for identification of MTBC in fresh or frozen aliquots of broth from positive BACTEC 12B cultures of respiratory specimens.