- Design
- Cochrane systematic review of diagnostic accuracy
- Population
- 3,291 people with presumptive pulmonary TB in five studies, nine countries
- Primary outcome
- Sensitivity and specificity against culture
- Effect
- Sensitivity 85.4% (95% CI 82.2–88.1); specificity 97.6% (96.9–98.1)
A Cochrane review commissioned by WHO assessed a new class of tuberculosis test: nucleic acid amplification run on small battery-operated devices that need no laboratory infrastructure and can be performed by health workers with basic training. Five studies with 3,291 people in nine moderate- to high-burden countries evaluated the one device meeting WHO's class criteria, using sputum.
Against culture, sensitivity was 85% and specificity 98%. In practical terms, among 1,000 symptomatic people of whom 100 have TB, the test would miss about 15 and falsely label about 22. Sensitivity fell to 79% in people with HIV and 67% in smear-negative, culture-positive disease. It was more sensitive than smear microscopy by 22 points but about 8 points less sensitive than Xpert MTB/RIF Ultra.
Most studies ran the test in laboratories, so performance at a primary health centre is not yet proven. For a family physician in a high-burden setting such as India, where smear microscopy is still used at many peripheral sites, the device could narrow the diagnostic gap. A negative result in a symptomatic patient with HIV or high clinical suspicion still needs a more sensitive test.
- Prefer a molecular test over smear microscopy for presumptive pulmonary TB wherever one is available.
- Do not rule out TB on a negative near point-of-care test in people with HIV or strong clinical suspicion.
- Send a sample for Xpert or culture when the near point-of-care result is negative but symptoms persist.
- Notify and start treatment promptly on a positive result according to national programme guidance.
Why it matters
Molecular TB testing could move from district laboratories to the clinic where patients first present.
Don't overread it
Most testing was done in laboratories, so accuracy at the periphery is not yet established.
The statistics, in plain English
Sensitivity of 85% means about one in seven people with culture-proven TB would test negative. Specificity of 98% means about two in a hundred without TB would test positive; in a low-prevalence group these false positives become a larger share of all positives.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for family medicine, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free