Near point-of-care molecular TB test: 85% sensitive, 98% specific, below Xpert Ultra

Design

Cochrane systematic review and meta-analysis of diagnostic accuracy

Population

3,291 people with presumptive pulmonary TB in 9 moderate- to high-burden countries (5 studies)

Primary outcome

Sensitivity and specificity against culture

Effect

Sensitivity 85.4% (82.2–88.1), specificity 97.6% (96.9–98.1); vs Xpert Ultra sensitivity −7.5 points

This Cochrane review, supported by WHO, assessed near point-of-care nucleic acid amplification tests for pulmonary TB: battery-powered instruments that need no special infrastructure and can be run by staff with basic training. Five studies with 3,291 participants from nine moderate- to high-burden countries met the criteria. All assessed the same device, the only one meeting WHO's definition of the class, using sputum.

Against culture, sensitivity was 85.4% and specificity 97.6% (moderate certainty). In a group of 1,000 symptomatic people of whom 100 have TB, that means 15 missed and 22 wrongly diagnosed. Sensitivity was 79.2% in people with HIV and 66.7% in smear-negative, culture-positive TB. Head to head with Xpert MTB/RIF Ultra, it was 7.5 percentage points less sensitive with similar specificity; against smear microscopy, it was 22 points more sensitive.

The case for this class of test is reach: a molecular result at a clinic that cannot run Xpert. Where Xpert or Truenat is already accessible, as across much of India's programme, it is not an upgrade. Most studies ran the test in laboratories rather than peripheral clinics, data in children were almost absent, and it does not report rifampicin resistance in the evidence summarised here.

• Where Xpert Ultra or an equivalent is accessible, keep using it; it was more sensitive.
• Consider a near point-of-care molecular test where the alternative is smear microscopy; it detected far more TB.
• Expect lower sensitivity in HIV and smear-negative disease, and follow a negative result with culture when suspicion is high.
• Do not use these results to guide testing in children; paediatric data were almost absent.
• Arrange separate drug-susceptibility testing; this review reported detection, not rifampicin resistance.

Why it matters

It places a new device class between smear and Xpert, which is where peripheral clinics actually sit.

Don't overread it

Most testing was done in laboratories, so real-world peripheral accuracy may be lower.

The statistics, in plain English

Sensitivity of 85% means 15 of every 100 people with TB test negative; specificity of 98% means about 2 of every 100 without TB test positive. The 'per 1,000' figures assume 10% of those tested have TB; in a setting where fewer do, false positives make up a larger share of positive results.

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.