Diagnostic accuracy of low-complexity, manual nucleic acid amplification tests for the detection of pulmonary and extrapulmonary tuberculosis in adults and adolescents: a systematic review and meta-analysis∗

Background Low-complexity, manual nucleic acid amplification tests, such as loop-mediated isothermal amplification for tuberculosis (TB-LAMP), are among the molecular WHO-recommended rapid diagnostics and can provide results within a few hours, even in resource-limited settings. We aimed to synthesise evidence on the accuracy of these tests for the detection of pulmonary and extrapulmonary tuberculosis, to inform the 2024 update of the WHO consolidated guidelines on tuberculosis. Methods For this systematic review and meta-analysis, we searched the Cochrane Central Register of Controlled Trials, MEDLINE, Embase, the Science Citation Index and BIOSIS previews, WHO Global Index Medicus, and Scopus databases, for articles published from Jan 1, 1946, to Oct 2, 2023, using specific search terms such as "Tuberculosis", "mycobacterium tuberculosis", "pulmonary tuberculosis", "extrapulmonary tuberculosis", "Loopamp", "diagnostic test", "smear microscopy", and "TB-LAMP". We also examined the reference lists of the included articles to identify potentially eligible studies that were not found in the electronic searches. Additionally, we searched ClinicalTrials.gov and the WHO Clinical Trials Registry Platform for ongoing and unpublished studies. We also examined studies and data received through a WHO public call, made between Nov 30, 2023, and Feb 15, 2024, for eligibility. We included studies that evaluated design-locked, marketed technologies belonging to the class of low-complexity, manual nucleic acid amplification tests (ie, TB-LAMP) against microbiological or composite reference standards, in adults and adolescents (aged ≥10 years) with presumptive pulmonary or extrapulmonary tuberculosis. We excluded studies with case-control designs and those that used in-house methods, screening studies aimed at identifying individuals with active tuberculosis in community settings, and drug-resistance surveys. We extracted data using a standardised form and assessed risk of bias and applicability using the revised Quality Assessment of Diagnostic Accuracy Studies tool. We contacted study authors for further information and data as required. We conducted meta-analyses using bivariate random-effects models to estimate summary sensitivities and specificities for detecting pulmonary and extrapulmonary tuberculosis, and assessed the certainty of evidence using the GRADE approach. This study is registered with PROSPERO, CRD42023471548. Findings Our searches identified 2806 records from databases and seven records from other sources. Of these, we screened the full text of 151 articles and ultimately included 29 studies in our systematic review: 27 on pulmonary tuberculosis and three on extrapulmonary tuberculosis (one study evaluated both). The studies generally had low risk of bias and applicability concern. From 26 studies involving 18 297 participants, the summary sensitivity for the detection of pulmonary tuberculosis from respiratory specimens was 84·1% (95% CI 78·3-88·6) and the summary specificity was 96·1% (95% CI 94·2-97·4), both with high certainty of evidence. Three studies, involving 95 participants, assessed the accuracy of TB-LAMP for detecting lymph node tuberculosis using lymph node tissue from biopsy. The summary sensitivity was 94·3% (79·8-98·6) and the summary specificity was 90·0% (79·5-95·4), both with low certainty of evidence. Interpretation TB-LAMP has satisfactory performance for detecting pulmonary tuberculosis in adolescents and adults and is a potential alternative to molecular tests that require more advanced infrastructure. However, the inability to detect rifampicin resistance is an important limitation of TB-LAMP. Future research should focus on well powered studies to establish the diagnostic accuracy of TB-LAMP for extrapulmonary tuberculosis sites. Funding WHO.