Comparative study of the performance of GeneXpert MTB/RIF in bronchoalveolar lavage compared to bronchial lavage in patients with clinical suspicion of tuberculosis

ABSTRACT

Abstract Tuberculosis (TB) is one of the ten leading causes of death worldwide, and the main cause from a single infectious agent. Early detection of the Mycobacterium tuberculosis complex (MTC) and of mutations conferring resistance to the main drugs used in antituberculous treatment contributes to reducing the transmission of the infection, and consequently the spread of resistant TB. The GeneXpert MTB/ RIF test identifies the MTC and simultaneously detects mutations most frequently associated with rifampicin resistance, through real-time PCR testing. The purpose of this study was to compare the performance of the GeneXpert MTB/RIF method in bronchoalveolar lavage (BAL) with bronchial lavage (LB) in immunocompetent patients with clinical suspicion of pulmonary TB without any previous microbiological documentation. Materials and Methods: We prospectively enrolled patients with radiologic pulmonary infiltrates compatible with active or residual TB without previous treatment, with negative direct bacilloscopy or nonproductive cough, for the assessment of active disease. We identified the most affected segment through computed axial tomography and bronchoscopy with BAL in said segment, followed by BL of the affected lobe. A BAL recovery > 40% was considered significant. The samples obtained were processed for bacilloscopy, culture and GeneXpert MTB/RIF. We analyzed sensitivity (S), specificity (SP), positive predictive value (PPV) and negative predictive value (NPV), taking the solid culture medium as reference for the diagnosis of MTC. Results: We included 20 patients; 3 were excluded because they didn't have a representative BAL sample. 17 patients were evaluated (11 women, 65%), age 37.2 ± 16.3. The MTC was identified through conventional methods in 10 patients: 10 with positive culture in BL and 9 in BAL. In comparison with the conventional methods, 6 out of 17 samples obtained through BAL had a positive result for GeneXpert MTB/ RIF: S = 60.0% (CI 31%-83%), SP = 100% (CI 65%-100%), PPV = 100% (CI 61%-100%) and NPV = 64% (CI 35.4%-84.8%). With BL, 9 out of 17 had a positive result for Xpert MTB/RIF: S = 90.0% (CI 60%-98%), SP = 100% (CI 65%-100%), PPV = 100% (CI 70%-100%) and NPV = 88% (CI 53%-98%). All the cases identified with GeneXpert MTB/RIF were true positives in relation to conventional cultures. Conclusion: Considering the solid culture as reference method, the BL was more sensitive than the BAL for the diagnosis of tubercu lous infection through the GeneXpert MTB/RIF method in patients with suspected TB without previous microbiological documentation.