Alexander J. Yan

The WHO has recommended tongue swabs (TS) as alternative samples for microbiological diagnosis of tuberculosis. We evaluated the effects of oral hygiene and food/drink intake on TS performance in South Africa. Food/drink intake prior to sampling marginally decreased Mycobacterium tuberculosis DNA signal strength, but neither behaviour decreased diagnostic sensitivity.

Tongue swabs hold promise as a non-invasive sample for diagnosing tuberculosis (TB). However, their utility as replacements for sputum has been limited by their varied diagnostic performance in PCR assays compared to sputum. The use of silica-based DNA extraction methods may limit sensitivity due to incomplete lysis of Mycobacterium tuberculosis (MTB) cells and co-extraction of non-target nucleic acid, which may inhibit PCR. Specificity may also be compromised because these methods are labor-intensive and prone to cross-contamination. To address these limitations, we developed a sample preparation method that combines sonication for MTB lysis and a sequence-specific MTB DNA capture method using hybridization probes immobilized on magnetic beads. In spiked tongue swabs, our hybridization capture method demonstrated a 100-fold increase in MTB DNA yield over silica-based Qiagen DNA extraction and ethanol precipitation. In a study conducted on clinical samples from South Africa, our protocol had 74% (70/94) sensitivity and 98% (41/42) specificity for detecting active pulmonary TB with sputum Xpert MTB/RIF Ultra as the reference standard. While hybridization capture did not show improved sensitivity over Qiagen DNA extraction and ethanol precipitation, it demonstrated better specificity than previously reported methods and was easier to perform. With integration into point-of-care platforms, these strategies have the potential to help enable rapid non-sputum-based TB diagnosis across key underserved patient populations.

Tongue swab (TS) sampling combined with qPCR to detect Mycobacterium tuberculosis (MTB) DNA is a promising alternative to sputum testing for tuberculosis (TB) diagnosis. In prior studies, the sensitivity of tongue swabbing has usually been lower than sputum. In this study, we evaluated two strategies to improve sensitivity. In one, centrifugation was used to concentrate tongue dorsum bacteria from 2-mL suspensions eluted from high-capacity foam swab samples. The pellets were resuspended as 500-µL suspensions, and then mechanically lysed prior to dual-target qPCR to detect MTB insertion elements IS 6110 and IS 1081 . Fractionation experiments demonstrated that most of the MTB DNA signal in clinical swab samples (99.22% ± 1.46%) was present in the sedimentable fraction. When applied to archived foam swabs collected from 124 South Africans with presumptive TB, this strategy exhibited 83% sensitivity (71/86) and 100% specificity (38/38) relative to sputum MRS (microbiological reference standard; sputum culture and/or Xpert® Ultra). The second strategy used sequence- specific magnetic capture (SSMaC) to concentrate DNA released from MTB cells. This protocol was evaluated on archived Copan FLOQSwabs® flocked swab samples collected from 128 South African participants with presumptive TB. Material eluted into 500 µL buffer was mechanically lysed. The suspensions were digested by proteinase K, hybridized to biotinylated dual-target oligonucleotide probes, and then concentrated ∼20-fold using magnetic separation. Upon dual-target qPCR testing of concentrates, this strategy exhibited 90% sensitivity (83/92) and 97% specificity (35/36) relative to sputum MRS. These results point the way toward automatable, high-sensitivity methods for detecting MTB DNA in TS. Importance: Improved testing for tuberculosis (TB) is needed. Using a more accessible sample type than sputum may enable the detection of more cases, but it is critical that alternative samples be tested appropriately. Here, we describe two new, highly accurate methods for testing tongue swabs for TB DNA.