Abstract
INTRODUCTION: The aim was to study the accuracy of Xpert(®) (Cepheid Inc., Sunnyvale, CA, USA) Mycobacterium tuberculosis/rifampicin (MTB/RIF) assay as compared to a composite gold standard (urine culture, imaging, and biopsy) and to asses its utility as the initial test compared to smear microscopy to diagnose urinary tuberculosis. METHODS: This prospective study included adult patients suspected to have urinary tuberculosis from March 2014 to December 2017. Three urine samples were collected from each patient and were subjected to Xpert MTB/RIF assay, acid-fast bacillus (AFB) smear microscopy, and liquid media (BACTEC Mycobacteria Growth Indicator Tube [MGIT] 960) culture. Imaging and tissue biopsies were performed as clinically indicated. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated using the bootstrap method for 95% confidence intervals for the Xpert assay. RESULTS: Xpert MTB/RIF assay was found to be superior to the currently best available light-emitting diode fluorescent smear microscopy as the initial test for urinary tuberculosis (sensitivity of 69.09% vs. 32.72%). The Xpert MTB/RIF polymerase chain reaction test was found to have a moderate sensitivity (69.09%) and high specificity (100%) as compared to the composite reference standard. The sensitivity of liquid AFB culture MGIT 960 as compared to the reference standard was 90.32%. CONCLUSIONS: Xpert MTB/RIF assay on an early morning first void urine specimen can replace smear microscopy as the initial diagnostic test for urinary tuberculosis.