Near point-of-care nucleic acid amplification tests (NPOC-NAATs) are a newer class of diagnostic assays intended for decentralized use in peripheral health settings. These instrument-based, battery-operated tests require minimal infrastructure and can be run by health workers with basic technical skills. The review assessed whether NPOC-NAATs provide accurate detection of pulmonary tuberculosis when applied to respiratory specimens.
The primary objective was to determine the diagnostic accuracy of NPOC-NAATs for pulmonary TB using respiratory samples from adults, adolescents, and children with presumptive tuberculosis. Secondary objectives included comparisons of NPOC-NAATs with smear microscopy and with low complexity automated NAATs (LC-aNAATs), and exploration of potential sources of heterogeneity such as smear status, semi-quantitative index test categories, HIV status, age groups, and history of TB. The review also aimed to report the frequency of inconclusive or invalid index test results.
The authors searched multiple databases (including CENTRAL, MEDLINE, and Embase) and two trial registers between 21 and 25 July 2025, and issued a WHO Public Call for ongoing and unpublished studies between 29 May and 15 July 2025. In consultation with WHO, studies published since 2024 were specifically sought to capture recently developed tests. Eligibility required studies reporting diagnostic accuracy of NPOC-NAATs on respiratory specimens in participants with presumptive pulmonary TB. Reference standards were culture and a composite standard that included culture or the clinical decision to treat.
Five studies met inclusion criteria, enrolling a total of 3,291 participants from nine countries with moderate to high TB burden. All included studies evaluated the same design-locked, marketed NPOC-NAAT: the MTB NAT Card processed on battery-powered Pluslife Thermolyse and MiniDock devices. The only respiratory specimen reported across studies was sputum. The review notes that testing in most included studies occurred in laboratory settings rather than peripheral clinics.
Reference standards used in the included studies were culture and a composite reference standard that incorporated culture or clinical decision to treat. The review assessed risk of bias using QUADAS-2 and QUADAS-C tools. Most studies were judged to have a low risk of bias across QUADAS-2 domains; however, concerns about applicability led to downgrading the certainty of evidence.
Using a bivariate meta-analytic model, the pooled performance of NPOC-NAAT (MiniDock) compared with culture was:
The evidence was rated as moderate certainty, with one level of downgrading for applicability because most testing was performed in laboratory rather than peripheral settings. The authors provided a worked example: in a hypothetical cohort of 1,000 symptomatic people with 100 true TB cases, NPOC-NAATs would miss 15 TB cases and falsely identify 22 people without TB as positive.
In people living with HIV (five studies, 506 participants), pooled results against culture were:
For participants who were smear-negative but culture-positive (three studies, 2,241 participants), performance was:
The review could not perform several other preplanned subgroup analyses—such as performance in people with prior TB or in children—because too few studies addressed those questions.
Five studies provided head-to-head data comparing NPOC-NAATs with a low complexity automated NAAT, Xpert MTB/RIF Ultra. Pooled comparison showed lower sensitivity for the NPOC-NAAT relative to Xpert MTB/RIF Ultra (85.6% vs 93.1%; difference −7.54 percentage points, 95% CI −11.6 to −3.47) and similar specificity (97.6% vs 97.3%; difference 0.34 percentage points, 95% CI −1.16 to 1.84).
Overall certainty of the main diagnostic estimates was rated moderate. The primary reason for downgrading was applicability: most included studies executed NPOC-NAAT testing in laboratory environments rather than the intended peripheral clinical settings where battery-operated devices aim to expand access. The review also highlights gaps: only sputum specimens were evaluated and data were insufficient for some important subpopulations (children, people with prior TB) and for real-world peripheral implementation outcomes.
The pooled estimates suggest that the evaluated NPOC-NAAT (MTB NAT Card on MiniDock) has high specificity and good sensitivity for detecting pulmonary TB on sputum, with somewhat reduced sensitivity in people living with HIV and in smear-negative disease. However, because most evidence comes from laboratory-based evaluations, the applicability to peripheral, point-of-care use is uncertain. The review identifies the need for diagnostic accuracy studies conducted in intended peripheral settings, evaluation in children and people with prior TB, and studies reporting inconclusive/invalid result rates and operational performance in real-world workflows.
(Details not reported in the source: the abstract does not provide granular data on the frequency of invalid/indeterminate index test results, specific country-level data, or detailed operational outcomes in peripheral clinics.)