Background: The Xpert® MTB/RIF assay (Xpert; Cepheid, Sunnyvale, CA, USA) is a cartridge-based nucleic acid amplification assay for rapid tuberculosis diagnosis and assessment of antibiotic sensitivity. Although previous evidence supports the use of Xpert for diagnosing extrapulmonary tuberculosis (EPTB) in adults, information regarding the accuracy of Xpert for EPTB only in children is lacking. This meta-analysis aimed to assess the accuracy of Xpert for detecting EPTB in children.
Methods: We searched the MEDLINE, EMBASE, and the Cochrane Infectious Diseases Group Specialized Register from January 1, 2010, to July 16, 2019, for studies on diagnostic performance wherein Xpert was analyzed against cultures or composite reference standards for <18-year-old children with EPTB.
Results: Exclusively on the basis of pediatric studies, 7 studies including 560 samples were selected. The pooled sensitivity and specificity of Xpert for all samples were 67% (95% CI 0.51-0.79) and 94% (95% CI 0.87-0.98), respectively. The area under the summary receiver operating characteristic (sROC) curve was 0.89. For lymph node tissues or aspirates, the pooled sensitivity and specificity of Xpert were 75% (95% CI 0.54-0.89) and 90% (95% CI 0.77-0.96), respectively; for cerebrospinal fluid (CSF), they were 40% (95% CI 0.23-0.60) and 96% (95% CI 0.85-0.99), respectively.
Conclusion: Overall, Xpert displayed high specificity but modest specificity across various samples in the diagnosis of pediatric EPTB in comparison with the composite reference standard. Xpert sensitivity varied with the sampling site, and was especially lower for CSF samples. Positive Xpert results may be considered to indicate a presumptive case of pediatric EPTB, whereas negative test results indicate that the possibility of pediatric EPTB should not be excluded.