[Evaluation of the Xpert MTB/RIF assay for the diagnosis of pulmonary and extrapulmonary tuberculosis in an intermediate-prevalence setting].

Early and accurate detection of tuberculosis (TB) is a global priority for TB control. In order to obtain results in a short period of time, nucleic acid amplification tests are increasingly used worldwide for the rapid diagnosis of tuberculosis. The Xpert MTB/RIF® (Cepheid, USA) is a commercially available, real-time PCR-based assay, which can detect both TB and resistance to rifampicin directly in clinical samples. The aim of this study was to evaluate the performance of Xpert MTB/RIF assay for M.tuberculosis detection in pulmonary and extrapulmonary clinical samples in routine laboratory practice in Turkey, an intermediate-prevalence setting. A total of 2639 clinical specimens, 1611 of which were pulmonary and 1028 were extrapulmonary, were included in the study. The results of Xpert MTB/RIF assay were evaluated by comparing the results with those obtained by culture [BACTEC MGIT 960 (Becton Dickinson, USA) and Löwenstein Jensen medium]. Overall sensitivity, specificity, positive and negative predictive values of Xpert MTB/RIF assay were determined as 73.9%, 98.6%, 79.6% and 98.1%, respectively. These values were calculated as 80.8%, 98.8%, 84.9% and 98.4% for pulmonary specimens, and 58.2%, 98.4%, 66.7% and 97.7% for extrapulmonary specimens. The sensitivity and specificity were 100% and 58.1%, respectively, for acid-fast bacilli (AFB) smear-positive specimens, 39.7% and 99.1%, respectively for smear-negative specimens. The sensitivity and specificity were 100% and 76.2% for smear-positive pulmonary specimens; 100% and 20% for smear-positive extrapulmonary specimens; 47.8% and 99.1% for smear-negative pulmonary specimens; and 28.2% and 99.2% for smear-negative extrapulmonary specimens, respectively. The sensitivity and specificity of microscopic examination were found to be 56.7% and 98.7% for all specimens; 63.2% and 98.6% for pulmonary specimens; and 41.8% and 99% for extrapulmonary specimens, respectively. Rifampicin resistance was detected by Xpert MTB/RIF assay in only two specimens, however, rifampicin resistance was failed to be detected by BACTEC MGIT 960 TB method in one of these samples. Xpert MTB/RIF assay appeared to be a reliable method for the diagnosis of TB for AFB smear-positive samples, but less sensitive for smear-negative samples, particularly for extrapulmonary samples which include low numbers of bacilli. However, we concluded that the MTB/RIF is a useful assay for rapid diagnosis of tuberculosis, considering that the results can be given in the same day of sample collection and the assay is superior in sensitivity than microscopic examination.