Evaluation of Alternative Diagnostic Methods for the Diagnosis of Smear Negative Pulmonary Tuberculosis: A cross-sectional comparative study at Jimma University specialized hospital, Jimma, Southwest Ethiopia

Abstract

Background: Smear-negative pulmonary tuberculosis (SNPTB) is an increasing problem worldwide. Its diagnosis remains a challenge especially in developing countries like Ethiopia, where majority of SNPTB has been diagnosed only on the basis of clinical and chest radiographic findings which lead to high rate of misdiagnosis. Even though culture is the gold standard, results may take 2-8 weeks, and delay in initiation of treatment may allow further transmission of disease. Little is known about the effect of bleach concentration for the detection of acid fast bacilli by fluorescent microscopy and GeneXpert MTB/RIF assy. Thus, this study is intended primarily to determine effect of simple bleach concentration method for the diagnosis of smearnegative tuberculosis by fluorescent microscopy and GeneXpert MTB/RIF. Objective: To evaluate the diagnostic yield of different methods (GeneXpert and fluorescent microscopy) for the detection of M. tuberculosis in smear negative sputum samples Methods: A cross-sectional comparative study was conducted at Jimma University Specialized Hospital from February 14 to August 1, 2014. A total of two hundred and two patients suspected of having SNPTB were enrolled. Two sputum specimens (spot and morning specimen) were collected from each suspect. The spot sputum was analysed for culture (L-J media& MGIT 960) and direct GeneXpert test. The morning sputum was processed for direct and indirect GeneXpert test (bleach treated sediment) as well as fluorescent microscopy (FM). Half of the morning sputum was treated with 5% bleach and centrifuged at 3000g for 15 minutes. The concentrated sediment was used for Xpert and FM. Data was analysed by using SPSS version 20.0. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the assays with 95%CI was calculated by using culture as reference method. Result: Complete data were available for 185 samples. The detection rate of direct GeneXpert was 5.4 % (10/185) on spot and 6.5% (12/185) on morning sputum. Similarly indirect Gene Xpert detected M. tuberculosis in 11.4% (21/185) and LED-FM in 7.6% (14/185) of the suspected cases. L-J media supported the growth ofII mycobacteria in 4.9% (9/185) and that of MGIT 960 in 8.6% (16/185) of the cases. Based on the gold standard method (combination of L-J and/or MGIT 960 media), 16(8.6%) cases were confirmed as smear negative tuberculosis. Direct GeneXpert had sensitivity of 50.0%, specificity of 97.6%, PPV of 66.7%, and NPV of 95.4%. While on bleach treated sputum, GeneXpert had sensitivity of 56.2%, specificity of 92.9%, PPV of 42.9% and NPV of 95.7%. LED-FM on bleach treated specimen had 25% sensitivity, 94.1% specificity, 28.6% PPV and 92.9% NPV. The sensitivity of direct GeneXpert on both spot and morning sputum was equal (50.0%) and GeneXpert on bleach treated sputum had 4.9% incremental yield than the direct one. All GeneXpert positive cases were sensitive for rifampicin. Conclusions and recommendation: GeneXpert on bleach treated sputum had highest detection yield. A significant proportion of suspected cases missed on conventional Ziehl-Neelsen method were detected on LED-FM. Bleach concentration method for GeneXpert had relatively higher sensitivity than the direct one and can potentially improve the diagnosis of SNPTB suspects. The use of only one sputum sample could be sufficient for TB diagnosis with Gene Xpert MTB/RIF