Pooling sputum cuts TB test costs but misses some cases

Tuberculosis (TB) remains a major global health concern, and molecular tests like the Xpert MTB/RIF Ultra assay (Xpert-Ultra) are important for rapid diagnosis. However, cartridge costs and laboratory capacity limit how widely these tests can be used. To explore a more efficient approach, Han Thi Nguyen and colleagues evaluated a pooled testing algorithm in Vietnam. In pooled testing, sputum specimens from several people are combined and tested together; if a pool is positive, the members of that pool are then tested individually. The research team enrolled 2,396 adults (aged 15 and older) with symptoms suggesting TB between July 2024 and February 2025. Participants were found through facility-based case finding (FBCF) and community-based case finding (CBCF). Each person provided two sputum specimens: the first was tested individually with Xpert-Ultra and the leftover material was combined into four-specimen pools for repeat Xpert-Ultra testing; the second specimen was used for liquid culture using BACTEC MGIT. The study was designed to compare pooled testing with individual testing, looking at diagnostic accuracy and potential savings in Xpert-Ultra cartridges.

The study used the collected data to simulate an up-front pooled testing strategy and compared results against culture. Of 2,396 participants, 395 (16.5%) had a positive Xpert-Ultra and/or culture result. Detection varied sharply by setting: 360 of 912 people (39.5%) in facility-based case finding (FBCF) were positive, while only 35 of 1,484 (2.4%) in community-based case finding (CBCF) were positive. When compared to culture, the pooled testing approach had a sensitivity of 82.4% (95% confidence interval [CI] 77.9–86.3) and specificity of 98.5% (97.8–99.0). Individual Xpert-Ultra testing had slightly higher sensitivity, 86.5% (82.4–89.9), and similar specificity, 98.1% (97.4–98.7). Pooled testing performed worse in the low-prevalence CBCF group (sensitivity 59.1%, 36.4–79.3) than in FBCF (84.0%, 79.5–87). Cartridge savings were also setting-dependent: CBCF saw 69.1% savings versus 9.6% in FBCF. Overall, the pooling strategy reduced Xpert-Ultra cartridge use by 46.5%, saving USD 14,447.

These results show that pooled Xpert-Ultra testing can be a resource-efficient way to expand molecular TB screening in Vietnam, especially where many tests are needed and a low proportion of people actually have TB. The trade-off is lower sensitivity compared with individual Xpert-Ultra testing, meaning some cases, particularly those with low amounts of bacteria (paucibacillary disease), may be missed. That loss in sensitivity was most pronounced in community-based screening where TB prevalence was low, while cartridge savings were greatest in that same setting. The authors emphasize that decisions to use pooled testing should balance affordability and wider access to molecular testing against the risk of missed diagnoses. Careful, context-specific implementation — choosing where pooling makes sense and monitoring for missed cases — will be essential to maximize programmatic benefit while minimizing the number of people with TB who are not diagnosed.