Tongue Swabs with Xpert Ultra Offer Alternative TB Test

Tuberculosis diagnosis often relies on sputum, which some people cannot produce or provide easily. Tongue swabs have emerged as a simpler specimen that can be collected without coughing, and an earlier study using a consensus protocol found tongue swabs tested with Xpert MTB/RIF Ultra (Xpert Ultra, Cepheid, USA) did better than sputum smear microscopy but still produced a notable share of non-actionable results (6.1%). To address that problem, researchers led by corresponding author Adithya Cattamanchi evaluated a revised protocol for processing tongue swabs in four high TB burden countries. The study enrolled people aged 12 years and older who were suspected of having TB at outpatient clinics in the Philippines, South Africa, Nigeria, and Zambia. Between March and November 2024 the team tested whether a changed preparation step could reduce invalid or error results and whether tongue swab testing could reliably detect TB compared with culture-based methods and with sputum-based tests. The work aimed to find a practical alternative diagnostic approach for settings and patients where sputum collection is difficult.

The multi-country diagnostic accuracy study included 1,168 participants (median age 37, interquartile range 28–48 years), of whom 46.7% were female, 21.8% were living with HIV, and 18.5% had culture-confirmed TB. Tongue swabs were processed using Sample Reagent (SR, Cepheid, USA) diluted 2:1 with phosphate buffer or phosphate-buffered saline and then tested with Xpert MTB/RIF Ultra (Xpert Ultra, Cepheid, USA). Diagnostic performance was measured against a culture-based microbiological reference standard and compared to sputum-based tests. The revised protocol yielded an overall non-actionable result rate of 5.6%, improving on the prior 6.1% figure; non-actionable results were under 4% in all countries except South Africa, where the rate was 15.4%. Tongue swab testing had a sensitivity of 66.0% (95% CI 59.0–72.5) and a specificity of 99.6% (95% CI 98.9–99.9). These results show a moderate ability to detect TB cases and a very low false-positive rate when using the revised processing approach with Xpert Ultra.

The findings suggest that tongue swabs processed with the revised Sample Reagent dilution and run on Xpert MTB/RIF Ultra can serve as a practical alternative diagnostic option when sputum is unavailable. High specificity (99.6%) means positive tongue swab results are very likely to reflect true TB disease, while the moderate sensitivity (66.0%) indicates that some cases would still be missed compared with culture. The overall low rate of non-actionable results in most countries supports the protocol’s operational reliability, but the substantially higher error rate observed in South Africa (15.4%) highlights that local factors can affect performance and that sites should monitor error rates during implementation. By testing the approach in the Philippines, South Africa, Nigeria, and Zambia, the study provides multi-country evidence that tongue swab Xpert Ultra testing using the revised protocol can expand diagnostic options for people who cannot produce sputum, though it is not a full replacement for sputum-based testing or culture.