Optimal processing of tongue swab samples for Mycobacterium tuberculosis detection by the Xpert MTB/RIF Ultra assay

Tongue swabs represent a potential alternative to sputum as a sample type for detecting pulmonary tuberculosis (TB) with molecular diagnostic tests. The methods used to process tongue swabs for testing in the World Health Organization-recommended Xpert MTB/RIF Ultra (Xpert Ultra) assay vary greatly. Here, we aimed to identify the optimal tongue swab processing for Xpert Ultra testing. We compared four methods for treating dry tongue swabs with the Xpert Sample Reagent (SR) mixed with various concentrations of Tris-EDTA-Tween to treatment with SR alone or to a commonly used heat inactivation protocol. In each condition, swabs obtained from volunteers without TB were placed into test buffer spiked with known amounts of Mycobacterium tuberculosis (Mtb) H37Rv-mc²6230. Swabs processed with 1:1 diluted SR buffer had the lowest Mtb limit of detection (LOD) at 22.7 CFU/700 µL (95% confidence interval (CI) 14.2-31.2), followed by 2:1 diluted SR buffer at 30.3 CFU/700 µL (95% CI 19.9-40.7), neat SR at 30.9 CFU/700 µL (95% CI 21.5-40.3), and SR prefilled in the Xpert Ultra at 57.1 CFU/700 µL (95% CI 42.4-71.7). Swabs processed using the heat-based protocol had the highest LOD (77.6 CFU/700 µL; 95% CI 51.2-104.0). Similar findings were observed for the LOD of RIF susceptibility. The 2:1 diluted SR buffer condition produced similar LODs when swabs were tested in the presence of sputum matrix or phosphate buffer saline. Further studies are needed to assess the performance of this processing protocol in a clinical setting.IMPORTANCEXpert MTB/RIF Ultra (Xpert Ultra) is approved by the World Health Organization for the diagnosis of tuberculosis (TB). This test is typically performed using sputum specimens obtained from people with presumptive TB. In order to inactivate Mycobacterium tuberculosis (Mtb) and aid liquefaction, sputum must be mixed with Xpert SR prior to transfer into the Xpert Ultra. However, some people under evaluation for TB are unable to produce sputum. Alternative sample types for TB diagnosis would, therefore, be of value. Oral swabs, including tongue swabs, have shown promise, but there are technical challenges associated with sample processing. In this study, several new tongue swab processing conditions were evaluated by utilizing SR, either neat or diluted in buffer. The ability of Xpert Ultra to detect TB was improved under these conditions compared with the previously published heat-processing method; processing steps were simplified; and technical challenges were overcome.