Diagnostic Accuracy and Potential Resource Savings of Pooled Sputum Testing with Xpert MTB/RIF Ultra for Tuberculosis among adults in Vietnam: A Cross-Sectional Study
- Publicada
- Servidor
- medRxiv
- DOI
- 10.64898/2026.03.31.26349825
Objectives
A pooled testing algorithm for tuberculosis (TB), in which sputum specimens from multiple individuals are tested in pools with individual testing of positive pools, can optimise diagnostic resources. This study evaluated the diagnostic accuracy and cartridge savings of pooled testing with the Xpert MTB/RIF Ultra assay (Xpert-Ultra) relative to individual Xpert-Ultra testing.
Methods
We conducted a cross-sectional study among 2,396 adults (≥15 years) with presumptive TB enrolled between July 2024 and February 2025, through facility-based case finding (FBCF) and community-based case finding (CBCF). Participants submitted two sputum specimens. The first underwent individual Xpert-Ultra testing; remnant specimens were combined into four-specimen pools and tested again with Xpert-Ultra. The second specimen was used to inoculate liquid culture (BACTEC MGIT). Data were used to simulate an up-front pooled testing strategy; sensitivity and specificity of this approach was estimated against culture, and cartridge use was compared with individual Xpert-Ultra testing.
Results
Of 2,396 participants, 395 (16.5%) had a positive Xpert-Ultra and/or culture, including 360/912 (39.5%) in FBCF and 35/1484 (2.4%) in CBCF. The pooled testing approach had sensitivity of 82.4% (95% confidence interval [CI], 77.9-86.3) and specificity of 98.5% (97.8-99.0) compared to culture, with lower sensitivity than individual Xpert-Ultra testing (86.5%, 82.4-89.9) but high specificity (98.1%, 97.4-98.7). Sensitivity of pooled testing was lower in CBCF (59.1%, 36.4–79.3) than in FBCF (84.0%, 79.5–87), whereas cartridge savings were greater in CBCF (69.1% vs 9.6%). The pooling strategy reduced Xpert-Ultra cartridge use by 46.5%, saving USD 14,447.
Conclusions
Pooled Xpert-Ultra testing among adults appears resource-efficient for TB screening in Vietnam. As sensitivity is lower compared to individual Xpert-Ultra testing, particularly for paucibacillary disease, these losses should be carefully weighed against gains in affordability and expand access to molecular testing. Careful, context-specific implementation is essential to maximise programmatic benefit while minimising missed persons with TB.