Performance Characteristics of the Cepheid Xpert MTB/RIF Test in a Tuberculosis Prevalence Survey
Susan E. Dormana b✉, Violet N Chihotac, James J. Lewisc d, Maunank Shahb, David Clarkc, Alison D. Grantd, Gavin Churchyardc d e, Katherine Fieldingd
- aJohns Hopkins University School of Medicine, Baltimore, Maryland, United States of America. dsusan1@jhmi.edu
- bJohns Hopkins University School of Medicine, Baltimore, Maryland, United States of America
- cAurum Institute for Health Research, Johannesburg, South Africa
Papers open inside SciRah from the original publisher or repository; SciRah never hosts them.
Abstract
BACKGROUND: Xpert MTB/RIF ("Xpert") is a molecular test for detection of Mycobacterium tuberculosis (MTB) in sputum. Performance characteristics have been established for its use during passive tuberculosis (TB) case detection in symptomatic TB suspects, but Xpert performance has not been assessed in other settings. Objectives were to determine Xpert performance and costs in the context of a TB prevalence survey. METHODOLOGY/PRINCIPAL FINDINGS: This was a diagnostic sub-study of a TB prevalence survey conducted in gold mining companies in South Africa. Sputa (one per participant) were tested using smear microscopy, liquid culture (reference comparator), and Xpert. Costs were collected using an ingredients approach and analyzed using a public health program perspective. 6893 participants provided a sputum specimen. 187/6893 (2.7%) were positive for MTB in culture, 144/6893 (2.1%) were positive for MTB by Xpert, and 91/6893 (1.3%) were positive for acid fast bacilli by microsocopy. Sensitivity, specificity, positive predictive value, and negative predictive value for detection of MTB by Xpert were 62.6% (95% confidence interval [CI] 55.2, 69.5), 99.6% (99.4, 99.7), 81.3% (73.9, 87.3), and 98.9 (98.6, 98.8); agreement between Xpert and culture was 98.5% (98.2, 98.8). Sensitivity of microscopy was 17.6% (12.5, 23.9). When individuals with a history of TB treatment were excluded from the analysis, Xpert specificity was 99.8 (99.7, 99.9) and PPV was 90.6 (83.3, 95.4) for detection of MTB. For the testing scenario of 7000 specimens with 2.7% of specimens culture positive for MTB, costs were $165,690 for Xpert and $115,360 for the package of microscopy plus culture. CONCLUSION: In the context of a TB prevalence survey, the Xpert diagnostic yield was substantially higher than that of microscopy yet lower than that of liquid culture. Xpert may be useful as a sole test for TB case detection in prevalence surveys, particularly in settings lacking capacity for liquid culture.
TL;DRXpert may be useful as a sole test for TB case detection in prevalence surveys, particularly in settings lacking capacity for liquid culture. (Semantic Scholar)