Introduction: The Xpert MTB/RIF test detects DNA from Mycobacterium tuberculosis complex and susceptibility to rifampin. It has been evaluated repeatedly under “ideal” conditions including centrifugation of sputum and bronchoalveolar lavage, Ziehl Neelsen (ZN) and auramine/rhodamine staining, as well as with solid and liquid automated culture methods. Results from such evaluations cannot be extrapolated to low-income countries that do not routinely use all these processes.Objective: To assess the performance of the Xpert MTB/RIF test in respiratory samples under “real” conditions of work in a low-income country and its correlation with phenotypic susceptibility testing.Materials and methods: We conducted a cross-sectional study to assess the performance of the Xpert MTB/RIF test in ≥12 year-old patients with suspected pulmonary tuberculosis. In routine sample processing at the Hospital we do not use sputum centrifugation, staining with auramine/rhodamine or automated liquid culture.Results: We screened 152 patients of whom 108 were eligible for the study and 103 were included in the analysis; 34% of the samples were positive. The overall test sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were 91%, 92%, 83% and 96%, respectively. In ZN-negative samples the sensitivity, specificity, PPV and NPV were 87%, 91%, 68% and 97%, respectively. The results of sensitivity and resistance to rifampin were concordant with susceptibility testing using the multiple proportions method (kappa=1, p<0.0001).Conclusions: The Xpert MTB/RIF test overall performance was similar to the one achieved under ideal conditions. Its performance in ZN-negative samples was better under “real” conditions of work in a low-income country.