Improving the diagnosis of tuberculosis in hard to diagnose populations : clinical evaluation of GeneXpert MTB/RIF and alternative approaches in Ethiopia

Background: The GeneXpert MTB/RIF® assay (Xpert MTB/RIF) was endorsed as the initial diagnostic tool in people suspected of HIV-associated or drug resistant tuberculosis (TB). However, information regarding the performance of Xpert for diagnosing smear-negative TB in high burden settings remains limited. We evaluated the diagnostic accuracy of Xpert MTB/RIF and the impact of bleach-concentration on the performance of Xpert MTB/RIF using smear-negative sputum sample from HIV-negative patients.
Methods: One spot and one morning smear-negative sputum samples per patient were examined using Xpert MTB/RIF and culture at Mycobacteriology Research Center of Jimma University, Ethiopia. The sputum culture on both Löwenstein-Jensen (LJ) and/or mycobacteria growth indicator tube (MGIT) was the gold-standard.
Results: Of 185 smear-negative presumptive pulmonary TB cases, 19 (10.3%) had culture-proven TB. The sensitivity of Xpert MTB/RIF on spot and morning sputum was similar (63.2%). Testing two specimens per patient insignificantly increased the sensitivity of Xpert MTB/RIF. Bleach-concentration and pelleting improved sensitivity of Xpert MTB/RIF over unprocessed sputum in paired samples (73.8% vs 63.2%) without affecting the specificity (95%). Bleach-concentration and pelleting allowed an additional 7 cases of TB (missed on the first and second direct Xpert MTB/RIF) to be detected, 5 of which were from culture-negative cases.
Conclusions: Testing of single sputum sample by Xpert MTB/RIF can reach reasonable sensitivity and result would be available on the same day, avoiding loss of patients and treatment delay. The sensitivity of Xpert MTB/RIF was improved after bleach-concentration and pelleting, though its added value needs further study on a larger scale.
Key words: Smear-negative tuberculosis, Xpert MTB/RIF, diagnostic accuracy, bleach-concentration