Objectives: Xpert MTB/RIF (Cepheid) assay has been introduced for the diagnosis of tuberculosis (TB) and RIF-resistance. The meta-analysis was used to establish the overall accuracy of Xpert MTB/RIF assay for diagnosing TB and RIF-resistance.
Methods: Based on comprehensive searches of the Pubmed and Embase, we identified outcome data from all articles estimating diagnostic accuracy with Xpert MTB/RIF assay. A summary estimation for sensitivity, specificity, diagnostic odds ratios (DOR) and the area under the summary ROC curve (AUC) was calculated by using the bivariate random-effects approach.
Results: The meta-analysis included 18 studies (10,224 suspected specimens). The summary estimate was 90.4% (95%CI 89.2%-91.4%) for sensitivity, 98.4% (95%CI 98.0%-98.7%) for specificity and 328.3/0.9822 for DOR/AUC in pulmonary tuberculosis (PTB). The sensitivity, specificity and DOR/AUC of detecting RIF-resistance were 94.1%, 97.0% and 177.8/0.9832, respectively. For extrapulmonary tuberculosis, the overall pooled sensitivity was 80.4% and specificity was 86.1%. The findings in subgroup analysis were as follows: the accuracy of Xpert MTB/RIF assay is higher in smear-positive specimens and the sensitivity of diagnosing PTB in adults was higher than that in children (90.8% versus 74.3%).
Conclusions: TB and RIF-resistance can be rapidly and effectively diagnosed with Xpert MTB/RIF assay.
Copyright © 2012 The British Infection Association. Published by Elsevier Ltd. All rights reserved.