Evaluation of the diagnostic performance of the Xpert® MTB/RIF assay in pulmonary and extrapulmonary samples

Introduction This study aimed to evaluate the performance of the Xpert MTB/RIF assay in the identification of M. tuberculosis in pulmonary and extrapulmonary clinical samples by taking the results of the BACTEC MGIT 960TB culture system as a reference. Materials and methods A total of 11,341 specimens sent to Sivas Cumhuriyet University Application and Research Hospital Tuberculosis Laboratory for microbiological examination with suspicion of tuberculosis infection between January 2013 and December 2019 were examined, and 6847 clinical specimens that underwent culture (BACTEC MGIT 960TB), Xpert MTB/RIF and AFB (Acid-fast bacilli) testing were selected and included in our study. Of the samples included in the study, 5096 samples were pulmonary, and 1751 were extrapulmonary samples. Result In our study, sensitivity, specificity, PPV and NPV values of Xpert MTB/ RIF and AFB were calculated by taking TB culture test as reference test. The sensitivity of the Xpert MTB/RIF assay was calculated as 96.1%, specificity as 99.7%, positive predictive value (PPV) as 88.2%, and negative predictive value (NPV) as 99.9%. These values for pulmonary samples were determined as 98.3%, 99.7%, 89.9%, and 99.9%, respectively. For extrapulmonary samples, the sensitivity of the assay was found as 89.4%, specificity as 99.5%, PPV as 82.9%, and NPV as 99.7%. The sensitivity and PPV values for AFBpositive samples were found to be 99.0% and 97.1%, respectively. For AFB negative samples, the sensitivity, specificity, PPV, and NPV values were determined as 90.5%, 99.7%, 73.8%, and 99.9%, respectively. Conclusions A large number of clinical samples were studied with the Xpert MTB/RIF test in our study. It can be a guide in determining the performance of the test under the conditions of our country. Especially in the diagnosis of extrapulmonary TB, the effectiveness of the Xpert MTB/RIF assay has not been certainly proven in countries having a moderate prevalence of TB, such as Turkey. In most of the published studies, only a small part of the samples is extrapulmonary samples. So, our study provides valuable results in terms of evaluating a large number of extrapulmonary samples.