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

This cross-sectional study in Vietnam demonstrates that pooled Xpert MTB/RIF Ultra testing offers a resource-efficient strategy for tuberculosis screening, achieving a 46.5% reduction in cartridge use and significant cost savings while maintaining high specificity, though with a modest decrease in sensitivity compared to individual testing.

The Gist

Imagine you are trying to find a few hidden needles in a massive haystack. In the world of tuberculosis (TB) testing, the "haystack" is the population of people who might be sick, and the "needles" are the bacteria causing the disease.

Usually, doctors test every single person individually. They take a sample of their spit (sputum), put it in a special machine (the Xpert Ultra test), and get a result. This is like checking every single piece of hay one by one. It's very accurate, but it's also incredibly expensive and slow because the "test kits" (cartridges) cost money.

This study from Vietnam asked a clever question: What if we mixed the spit of four people together and tested that mixture first?

Here is the breakdown of what they found, using simple analogies:

1. The "Group Soup" Strategy

Instead of testing four people individually (using four test kits), the researchers mixed a little bit of spit from four people into one bowl and tested that bowl.

• If the bowl tests negative: Great! None of those four people have TB. You saved three test kits.
• If the bowl tests positive: Uh oh, at least one person in that group has TB. Now, you have to go back and test those four people individually to find out exactly who it is.

2. The Results: A Trade-Off

The study looked at nearly 2,400 people in two different settings:

• The Hospital (FBCF): People who were already sick enough to go to the doctor.
• The Community (CBCF): People who were screened in their neighborhoods, many of whom felt fine.

The Good News (The Savings):
By using this "Group Soup" method, they saved a massive amount of money.

• They reduced the number of test kits needed by nearly half (46.5%).
• In the community setting, where fewer people were actually sick, they saved almost 70% of the kits.
• The Analogy: Imagine you have a budget to buy 100 tickets to a concert. By pooling your friends together, you only need to buy 54 tickets to get the same information. That extra money can be used to buy tickets for more people who couldn't afford them before.

The Bad News (The Missed Cases):
The "Group Soup" method wasn't perfect. It missed some cases that the individual tests would have caught.

• The Sensitivity Drop: When the bacteria count in a person's spit was very low (like a "Trace" amount), mixing it with three other people's spit diluted it so much that the machine couldn't smell it anymore.
• The Analogy: If you put one drop of strong coffee into a cup of water, you can still taste it. But if you put that same drop into a bucket of water, the flavor disappears.
• In the community screening, the method missed about 40% of the cases it should have found. In the hospital, it was better, but still missed a few.

3. Why the Difference?

• In the Hospital: People were sicker, so their "coffee" (bacteria) was stronger. Even when diluted, the machine could still taste it.
• In the Community: People were healthier or in the early stages of illness. Their "coffee" was weak. When diluted, it vanished completely.

4. The Big Lesson

The researchers concluded that this strategy is a double-edged sword.

• If you use it to replace individual testing for everyone, you will save money but miss sick people. This is dangerous.
• If you use it to expand testing, it's a winner. The money saved by testing fewer people can be used to test more people who are currently being ignored because testing is too expensive.

The Final Verdict:
Think of pooled testing like a security checkpoint.

• Individual testing is a full-body scan for every single person. It finds everything but takes forever and costs a fortune.
• Pooled testing is a metal detector for groups. It's fast and cheap. If the group alarm goes off, you do a full scan on everyone in that group.

The Catch: The metal detector might miss a tiny, hidden knife if the group is too big or the knife is too small.

Recommendation:
Use the "Group Soup" method in low-risk areas (like community screenings) where you want to cast a wide net to find hidden cases, but be prepared to do follow-up tests for people who look suspicious (like those with abnormal X-rays) even if the group test was negative. Don't use it blindly; use it smartly to stretch your budget so you can help more people.

Technical Summary

1. Problem Statement

Tuberculosis (TB) remains a critical global health threat, with Vietnam being one of the highest burden countries. While the Xpert MTB/RIF Ultra (Xpert-Ultra) assay is the WHO-recommended standard for rapid molecular TB diagnosis, its widespread use in large-scale screening campaigns (particularly in community-based settings) is constrained by high costs, limited cartridge supply, and operational infrastructure requirements.
Pooled testing—combining sputum specimens from multiple individuals into a single test—has been proposed to improve efficiency and reduce costs. However, there is a lack of real-world evidence regarding the diagnostic accuracy of pooled Xpert-Ultra testing specifically in high-burden settings like Vietnam, and how this strategy performs across different screening contexts (facility-based vs. community-based).

2. Methodology

• Study Design: A cross-sectional, multi-site study conducted between July 2024 and February 2025.
• Setting & Population:
  • Location: Hanoi and Ho Chi Minh City (HCMC), Vietnam.
  • Participants: 2,396 adults (≥15 years) with presumptive TB.
  • Recruitment Strategies:
    • Facility-Based Case Finding (FBCF): 912 participants recruited from major hospitals (Vietnam National Lung Hospital and Pham Ngoc Thach Hospital).
    • Community-Based Case Finding (CBCF): 1,484 participants recruited through community screening events involving symptom screening and chest X-rays (CXR).
• Intervention & Protocol:
  • Each participant provided two sputum specimens.
  • Specimen 1: One portion was tested individually with Xpert-Ultra. The remaining material was pooled in groups of four specimens (0.5 mL each) and tested again with Xpert-Ultra.
  • Specimen 2: Used for liquid culture (MGIT 960), which served as the reference standard for diagnostic accuracy evaluation.
• Statistical Analysis:
  • Diagnostic accuracy (sensitivity and specificity) of a simulated "upfront pooled testing" strategy was calculated against culture.
  • In this simulation, only individuals with a positive pooled result would proceed to individual confirmatory testing.
  • Cartridge usage and cost savings were modeled based on local pricing (USD 12.98/cartridge).

3. Key Contributions

• Contextual Evidence: Provides the first large-scale, real-world data on four-specimen pooled Xpert-Ultra testing in a high-burden Asian setting, covering both facility and community screening.
• Comparative Analysis: Directly compares the performance and economic efficiency of pooled testing versus standard individual testing across different prevalence settings (high-prevalence FBCF vs. low-prevalence CBCF).
• Granular Sensitivity Data: Analyzes sensitivity losses specifically regarding bacillary burden (semi-quantitative grades) and setting types, offering nuanced guidance for policy implementation.

4. Key Results

• TB Prevalence:
  • Overall prevalence (based on expanded microbiological reference standard) was 16.5% (395/2,396).
  • FBCF: 39.5% (360/912).
  • CBCF: 2.4% (35/1,484).
• Diagnostic Accuracy (vs. Culture):
  • Pooled Testing: Sensitivity 82.4% (95% CI: 77.9–86.3); Specificity 98.5% (95% CI: 77.8–99.0).
  • Individual Testing: Sensitivity 86.5% (95% CI: 82.4–89.9); Specificity 98.1% (95% CI: 97.4–98.7).
  • Sensitivity Loss: Pooled testing resulted in a 4.1% absolute reduction in sensitivity compared to individual testing ($p < 0.001$).
  • Setting Disparity: Sensitivity loss was significantly more pronounced in CBCF (59.1% for pooled vs. 68.2% for individual) compared to FBCF (84.0% vs. 87.7%).
  • Bacillary Burden: Pooled testing maintained high detection rates for moderate-to-high bacillary loads (100% for Medium/High grades) but showed reduced sensitivity for "Trace" results and paucibacillary cases (culture-positive but Xpert-negative).
• Resource Savings:
  • Cartridge Reduction: Overall reduction of 46.5% (from 2,396 to 1,283 cartridges).
  • Cost Savings: Estimated saving of USD 14,447 (approx. USD 6.03 per participant).
  • Efficiency by Setting:
    • CBCF: Massive savings of 69.1% in cartridge use due to low prevalence.
    • FBCF: Minimal savings of 9.6%. In fact, in the higher-burden HCMC facility, pooling actually increased costs (+15.4%) due to a high rate of positive pools requiring re-testing.

5. Significance and Conclusion

• Strategic Implementation: Pooled testing is highly resource-efficient for low-prevalence, high-volume community screening (CBCF), where it can significantly expand testing capacity without proportionally increasing costs.
• Risk of Missed Diagnoses: The strategy comes at the cost of reduced sensitivity, particularly for paucibacillary disease (early-stage or low-bacterial-load TB). This risk is highest in community settings where early detection is crucial.
• Policy Recommendations:
  • Pooled testing should not be a "one-size-fits-all" replacement for individual testing.
  • It is best suited for low-prevalence contexts (e.g., CBCF) and should be integrated with risk stratification (e.g., using CXR or AI to triage high-risk individuals to individual testing).
  • To maximize public health benefit, the cost savings generated by pooling must be reinvested to expand overall screening coverage rather than simply replacing existing testing capacity.
  • Careful follow-up protocols (e.g., repeat testing for high-risk individuals with negative pooled results) are essential to mitigate the risk of missed diagnoses.

In summary, the study validates pooled Xpert-Ultra testing as a viable tool for scaling up TB screening in resource-limited settings, provided it is implemented with strict context-specific guidelines to balance economic efficiency against diagnostic sensitivity.