Near point-of-care TB testing: 85% sensitive, needs no lab

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)

Near point-of-care nucleic acid amplification tests are battery-powered molecular devices designed for health centres without laboratory infrastructure. This Cochrane review, supported by the WHO, found five studies with 3,291 people in nine high-burden countries. Only one device, the MiniDock system, met the class definition.

Against culture, sensitivity was 85.4% (95% CI 82.2–88.1) and specificity 97.6%. In 1,000 symptomatic people of whom 100 have tuberculosis, that means about 15 cases missed and 22 false positives. Sensitivity was 79% in people with HIV and 67% in smear-negative disease. Head-to-head, it was less sensitive than Xpert MTB/RIF Ultra (85.6% vs 93.1%) but far more sensitive than smear microscopy (86.7% vs 64.7%), with similar specificity to Ultra.

The case for this kind of test is reach: it could replace smear microscopy at the periphery, where it would find about a third more cases. It should not replace Xpert where Xpert is available. Most studies ran the test in laboratories, data in children were almost absent, and its regulatory status and availability in India are unknown.

• Where only smear microscopy is available, a molecular test like this would detect substantially more TB.
• Where Xpert MTB/RIF Ultra is available, keep using it; it is more sensitive.
• Be cautious with negative results in people with HIV or smear-negative presentations.
• Data in children are almost absent; do not rely on this test to exclude TB in a child.
• Check national programme guidance before adopting any new TB diagnostic.

Why it matters

Molecular TB diagnosis could reach clinics that today rely on a microscope.

Don't overread it

Most testing was done in laboratories, so accuracy in peripheral clinics is not yet proven.

The statistics, in plain English

Sensitivity is the share of true TB cases the test finds; specificity is the share without TB it correctly clears. The difference from Ultra, about 7.5 percentage points, has an interval of 3.5 to 11.6, so Ultra is genuinely more sensitive. Moderate certainty means the estimates are likely close, but real-world peripheral performance could differ.

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.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.