The laboratory diagnosis of tuberculosis (TB) on extrapulmonary specimens is particularly challenging. A number of commercial nucleic acid amplification tests able to detect and identify Mycobacterium tuberculosis (MTB) complex directly from respiratory secretions have been developed, but their use on extrapulmonary samples still calls for validation. The BDProbeTec ET Mycobacterium tuberculosis Complex Direct Detection Assay (DTB) was applied to 918 consecutive extrapulmonary specimens (collected from 863 patients), including 84 gastric aspirates, 145 urine, 136 sterile body fluids, 83 cerebrospinal (CSF) fluids, 237 fine-needle aspirates, 175 pus, 56 biopsies, and two stool specimens. The results were compared with those of acid-fast staining and culture (solid plus liquid media), setting the combination of culture and clinical diagnosis as the gold standard. Ninety-two specimens yielded culture positive for MTB and 24 (smear- and culture-negative) were from patients with TB clinical diagnosis. Of these, 96 were DTB-positive, including all of those from culture-negative TB cases. From 26 specimens, nontuberculous mycobacteria were grown. Two of these specimens were positive by the DTB assay. Finally, of the 776 samples that were smear- and culture-negative for acid-fast bacilli (AFB), collected from patients for whom the diagnosis of TB was excluded, six were DTB-positive. The overall sensitivity, specificity, and positive and negative predictive values (PPV and NPV, respectively) of extrapulmonary samples were 82.7, 99.0, 92.3, and 97.8%, respectively. Although, at present, amplification assays cannot replace culture techniques, DTB proved to be rapid and specific for the detection of MTB in extrapulmonary samples.