Diagnostic accuracy of low and moderate complexity assays for pulmonary and extrapulmonary TB detection: systematic review and meta-analyses

BackgroundThe WHO estimates that globally in 2021, 10.6 million people fell ill with tuberculosis (TB). Nearly 4.2 million people were either not diagnosed or not reported during 2021, and this makes TB diagnosis the biggest challenge. With the emergence of molecular rapid tests for TB, there is some hope for improving case detection. We performed two systematic reviews on two technology classes - low and moderate complexity assays - for diagnosing pulmonary and extrapulmonary TB, respectively. MethodologyA comprehensive search of six databases for relevant citations was performed. Cross-sectional, case-control, cohort studies, and randomised controlled trials of any of the index tests were included. Data on Xpert and Ultra (low complexity assay) were extracted for extrapulmonary TB. Additionally, for moderate complexity assays for diagnosing pulmonary TB, five technologies were included. Study quality was assessed using Quadas instrument, and pooled test accuracy estimates were generated via bivariate random effects regression. Latent class analysis was used to estimate the accuracy of Xpert MTB/RIF and Ultra as culture is not a perfect reference standard for diagnosing extrapulmonary TB.ResultsLow complexity assays for EPTBFor this meta-analysis, we included 69 studies: 58 evaluated Xpert MTB/RIF, 2 evaluated Xpert Ultra, and 9 evaluated both tests. Overall, risk of bias was low for patient selection, index test, and flow and timing domains. For reference standard, 49% had low and 43% had unclear risk of bias. Applicability for the patient selection domain was unclear for most studies because clinical settings were uncertain. For CSF, Ultra pooled sensitivity and specificity (95% CrI) against culture were 89.4% (79.1 to 95.6) and 91.2% (83.2 to 95.7). Xpert pooled sensitivity and specificity against culture were 71.1% (62.8 to 79.1) and 96.9% (95.4 to 98.0). For pleural fluid, Ultra pooled sensitivity and specificity against culture were 75.0% (58.0 to 86.4) and 87.0% (63.1 to 97.9). Xpert pooled sensitivity and specificity against culture were 49.5% (39.8 to 59.9) and 98.9% (97.6 to 99.7) . For Lymph node aspirate, Ultra sensitivity and specificity (95% CI) against composite reference standard were 70% (51 to 85) and 100% (92 to 100). Xpert pooled sensitivity and specificity against composite reference standard were 81.6% (61.9 to 93.3) and 96.4% (91.3 to 98.6). For lymph node aspirates, Xpert MTB/RIF pooled specificity against culture was 86.2% (78.0 to 92.3. Using the latent class model, Xpert pooled specificity was 99.5% (99.1 to 99.7), similar to that observed with a composite reference standard.Moderate complexity assays for pulmonary TBA total of 21 studies were included. We could only meta-analyse data for 3 of the 5 assays identified, as data were limited for the remaining two. For TB detection, the included assays had a sensitivity of 91% or more and the specificity ranged from 97% to 100%. For rifampicin resistance detection, all the included assays had a sensitivity of more than 92%, with a specificity of 99% to 100%. Sensitivity for isoniazid resistance detection varied from 70% to 91%, with higher specificity of 99% to 100% across all index tests. Studies that included head-to-head comparisons of these assays with Xpert for detection of TB and rifampicin resistance suggested comparable diagnostic accuracy.ConclusionsUltra and Xpert may be helpful in diagnosing extrapulmonary tuberculosis. While sensitivity varies across different extrapulmonary specimens, for most specimen types, specificity is high, which is helpful in ruling in people for confirmed disease. Ultra and Xpert had similar sensitivity and specificity for rifampicin resistance. In people with symptoms of pulmonary TB, the moderate complexity assays demonstrate comparable diagnostic accuracy for detection of TB, rifampicin and isoniazid resistance to Xpert