Tongue swab PCR misses many asymptomatic TB cases in community screening

Tuberculosis remains a major public health challenge, and researchers are exploring easier, non-invasive ways to find infections early. In a study led by Mark Hatherill, teams compared tongue swabs tested by high-volume qPCR for Mycobacterium tuberculosis in two different settings: community household screening and clinic-based triage. The community group (HHC Cohort) included adult household contacts of known TB patients, while the Clinic Cohort included symptomatic adults who came to clinics with presumptive TB. The work took place across eight South African sites. To define true TB cases, every participant submitted sputum for Xpert Ultra testing and for liquid culture; those with positive results were labeled TB Cases. Each TB Case was matched with several Controls without TB—about 1:3 in the household cohort and about 1:2 in the clinic cohort—so the investigators could compare how well the tongue swab qPCR identified real infections. In both cohorts tongue swabs were tested by high-volume qPCR, and in the Clinic Cohort an additional method, sequence-specific magnetic capture (SSMaC) with qPCR, was also applied. The study set out to see whether the non-invasive tongue swab approach performs differently when used to screen people who feel healthy versus those who already report symptoms.

The study used consistent laboratory definitions and tests across both groups: sputum Xpert Ultra and liquid culture were the reference standards, and tongue swabs were analyzed by high-volume qPCR for Mycobacterium tuberculosis. In the Clinic Cohort, tongue swabs were also processed using sequence-specific magnetic capture (SSMaC) combined with qPCR. Results reported a specificity of 91.9% for the tongue swab high-volume qPCR, and a statistical result reported as p = 0.0007. Sensitivity varied strongly with clinical context and measures of disease burden. Among household contacts with asymptomatic TB, high-volume qPCR on tongue swabs detected only 35.1% of cases. Sensitivity rose to 52.2% for TB cases that had an abnormal chest X-ray (CXR), and reached 100% for cases with High sputum Xpert Ultra grade. Overall, the study found that sensitivity of tongue swab high-volume qPCR for community-based, household screening of asymptomatic TB was low—approximately half that seen when used for facility-based triage of symptomatic patients—while performance improved when radiographic severity and sputum bacillary load were greater.

These findings matter for how tongue swab qPCR might be used in real-world TB programs. The low sensitivity for asymptomatic household contacts means that, as tested here, tongue swab high-volume qPCR would miss a majority of early or low-burden infections if deployed alone for community screening. On the other hand, the method performed much better when disease was more advanced or when sputum tests showed high bacillary load, suggesting tongue swabs could be useful for identifying the most contagious or radiographically obvious cases, particularly in clinics. The high specificity (91.9%) indicates positive tongue swab qPCR results are likely to reflect true infection, but the gap in sensitivity between community and clinic settings warns against relying on this single approach to rule out TB in people without symptoms. In short, tongue swabs and qPCR are promising as a non-invasive tool, but their current sensitivity limits their value for broad community screening of asymptomatic contacts without additional tests or strategies.