Diagnostic accuracy of the Xpert MTB/RIF assay for bone and joint tuberculosis using tissue specimens

Objectives To evaluate the diagnostic accuracy of the Xpert MTB/RIF assay for bone and joint tuberculosis (BJTB) using tissue specimens, and to compare the diagnostic accuracy of different types of tissue specimens. Methods This study involved 242 patients admitted with suspected BJTB between May 2018 and March 2020. The Xpert MTB/RIF assay was performed on surgically excised tissue. Diagnostic accuracy of the Xpert MTB/RIF assay was evaluated by culture, histopathology and a composite reference standard (CRS). Results One hundred and seventy-five patients were excluded (91 based on the exclusion criteria, and 84 as pus specimens were used instead of tissue specimens). Of the 67 patients enrolled, 14 were confirmed as BJTB, 20 as probable BJTB, 11 as possible BJTB, and 22 as non-BJTB. Using culture as the reference standard, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (PLR), negative likelihood ratio (NLR) and area under the curve (AUC) of the Xpert MTB/RIF assay were 92.9% (88.2-97.6%), 62.3% (56.1-68.5%), 39.4% (33.2-45.6%), 97.1% (92.3-100.0%), 2.464 (1.077-3.851), 0.114 (0.025-0.203) and 0.776 (0.654-0.897), respectively. When histopathology was used as the reference standard, the Xpert MTB/RIF assay had sensitivity, specificity, PPV, NPV, PLR, NLR and AUC of 79.3% (73.5-85.1%), 73.7% (67.8-79.6%), 69.7% (63.8-75.6%), 82.4% (76.5-88.3%), 3.015 (1.184-4.846), 0.281 (0.141-0.421) and 0.765 (0.646-0.884), respectively. Sensitivity, specificity, PPV, NPV, PLR, NLR and AUC obtained when using CRS as the reference were 73.3% (67.9-78.7%), 100.0% (100.0-100.0%), 100.0% (100.0-100.0%), 64.7% (58.5-70.9%), +∞, 0.267 (0.129-0.405) and 0.867 (0.781-0.952), respectively. Tissue samples were classified, and the positive rate of the Xpert MTB/RIF assay for BJTB using granulation tissue specimens was found to be significantly higher than that for caseous necrotic tissue, sequestrum and other necrotic connective tissues (P Conclusion The Xpert MTB/RIF assay showed high sensitivity and specificity for the diagnosis of BJTB from tissue specimens.