Performance of conventional histopathology and GeneXpert MTB/RIF in the diagnosis of spinal tuberculosis from bone specimens: A prospective clinical study.

Objective: Xpert MTB/RIF is recommended to detect pulmonary tuberculosis; however, there is insufficient data on its utility for bone samples. This study aimed to assess the accuracy of Xpert MTB/RIF compared with conventional histopathology in diagnosing spinal tuberculosis (STB) based on bone specimens in high burden settings.
Materials and Methods: Totally, 128 suspected STB participants were enrolled into this study. The bone specimens were obtained through puncture or operation for histological and Xpert MTB/RIF analyses, so as to compare their accuracy in diagnosing STB by the composite reference standard (CRS).
Results: Finally, 106 subjects with suspected STB were recruited into the analysis, including 27 confirmed and 33 clinically diagnosed STB patients. Relative to histopathology, Xpert MTB/RIF achieved a 86.7% sensitivity, and 12 out of 30 STB patients were positive, while the negative results in them were obtained upon histopathology. Based on CRS, Xpert MTB/RIF yielded a 63.3% sensitivity, which significantly elevated relative to that obtained upon histopathological test (50.0%, p < 0.001). In addition, the pooled sensitivity obtained using the above 2 approaches was as high as 95.0%, which was higher than that of any of the 2 approaches alone. The pooled specificity was 97.8%. Moreover, the area under the curve (AUC) value was 0.75 for Xpert MTB/RIF and 0.81 for histopathology, with no statistical significance. The two methods showed moderate concordance in the diagnosis of STB.
Conclusions: The Xpert MTB/RIF test achieves superior specificity and fair sensitivity, which can not be recommended to replace the conventional examinations for the diagnosis of STB. The combined application of these 2 approaches can improve the pooled diagnostic sensitivity and accuracy for STB.
(Copyright © 2020 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.)