Abstract
Diagnostic test accuracy of the loop-mediated isothermal amplification (LAMP) assay for culture proven tuberculosis is unclear. We searched electronic databases for both cohort and case-control studies that provided data to calculate sensitivity and specificity. The index test was any LAMP assay including both commercialized kits and in-house assays. Culture-proven M. tuberculosis was considered a positive reference test. We included 26 studies on 9330 sputum samples and one study on 315 extra-pulmonary specimens. For sputum samples, 26 studies yielded the summary estimates of sensitivity of 89.6% (95% CI 85.6-92.6%), specificity of 94.0% (95% CI 91.0-96.1%), and a diagnostic odds ratio of 145 (95% CI 93-226). Nine studies focusing on Loopamp MTBC yielded the summary estimates of sensitivity of 80.9% (95% CI 76.0-85.1%) and specificity of 96.5% (95% CI 94.7-97.7%). Loopamp MTBC had higher sensitivity and lower specificity for smear-positive sputa compared to smear-negative sputa. In-house assays showed higher sensitivity and lower specificity compared to Loopamp MTBC. LAMP promises to be a useful test for the diagnosis of TB, however there is still need to improve the assay to make it simpler, cheaper and more efficient to make it competitive against other PCR methods already available.