Xpert MTB/RIF assay for the diagnosis of extrapulmonary tuberculosis: a systematic review and meta-analysis

May combine two claims.

13 studies (839 samples; 10 with >10 samples, 159 culture positive) evaluated Xpert in CSF against culture.

“13 studies (839 samples; 10 with .10samples, 159 culture positive) evaluated Xpert in CSF against culture [14-16, 19, 20, 22-25, 27-29, 31].”

• cites [14] Al-Ateah, 2012, Evaluation of direct detection of Mycobacterium tuberculosis complex in respiratory and non-respiratory clinical specimens using the Cepheid Gene Xpert® system, Saudi Med J sourceNot checked — no full text of the source
• cites [15] sourceDisputed

  The Materials and Methods (Specimen processing; page unavailable) lists all specimen types studied, including 37 nonrespiratory specimens, but no CSF. This paper evaluated Xpert against culture in other specimen types, so it does not belong among the cited studies evaluating Xpert in CSF. The aggregate counts for the 13 studies cannot be assessed from this paper alone. (LLM, conf=0.88)

  Checked by a reasoning model — may be wrong.

  Experimental — accuracy not yet measured
• cites [16] sourceSupported
• cites [19] sourceNot checked — no full text of the source
• cites [20] sourceSupported
• cites [22] sourceNot checked — no full text of the source
• cites [23] sourceNot checked — no full text of the source
• cites [24] sourceSupported
• cites [25] sourceSupported
• cites [27] Safianowska, 2012, Zastosowanie wybranych komercyjnych testow molekularnych w mikrobiologicznej diagnostyce gruzlicy [Two selected commercially based nucleic acid amplification tests for the diagnosis of tuberculosis], Pneumonol Alergol PolNot checked — no full text of the source

May combine two claims.

The study by Vadwai et al. [29] had zero sensitivity (none out of 196 samples, three false negatives) and was considered an outlier.

“with the study by VADWAI et al. [29] (sensitivity none out of 196 samples, three false negatives) considered an outlier (fig.5a).”

• cites [29] sourceDisputed

  The Results report Xpert sensitivity of 81% (228/283) against the composite reference standard and 83% (125/150) against culture, contrary to the attributed zero sensitivity. The paper does not report the claimed 196-sample figure. Calling the study an outlier appears to be the citing authors’ assessment, not a finding attributed to Vadwai et al. (LLM, conf=0.88)

  Checked by a reasoning model — may be wrong.

  Experimental — accuracy not yet measured

May combine two claims.

Noninterpretable results for Xpert were reported in three studies, with only one study [28] having >2% (3.6%).

“Noninterpretable results for Xpert were reported in three studies, with only one study [28] having .2%(i.e.3.6%).”

• cites [28] sourceDisputed

  The Results section reports 17 indeterminate Xpert results, or 1.1% of samples—not the attributed 3.6% or a rate above 2%. This paper cannot establish how many other studies reported noninterpretable results. (LLM, conf=0.90)

  Checked by a reasoning model — may be wrong.

  Experimental — accuracy not yet measured

May combine two claims.

Existing tests for the diagnosis of extrapulmonary tuberculosis are limited in accuracy and time to diagnosis, and often require invasive procedures and special expertise.

“Existing tests for the diagnosis of EPTB are limited in accuracy and time to diagnosis, and often require invasive procedures and special expertise.”

• cites [4] sourcePartly supported

  The Background (pp. 2–3, 5) describes limited accuracy in extrapulmonary TB diagnosis, cultures that can take weeks, and diagnostic sampling by aspiration or biopsy. It does not establish that these tests often require special expertise, so that part of the claim is unsupported. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites [5] sourceNot checked — no full text of the source
• cites [6] sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Xpert MTB/RIF is a rapid, automated molecular test with high accuracy for pulmonary TB detection (sensitivity 89%, specificity 99%).

“The Xpert MTB/RIF assay (hereafter referred to as Xpert; Cepheid, Sunnyvale, CA, USA) is a rapid, automated molecular test with high accuracy for pulmonary TB detection (sensitivity 89%, specificity 99%) [7].”

• cites [7] sourcePartly supported

  The Background > Index test(s) describes Xpert as a rapid, automated molecular test, and Results > Findings characterizes its accuracy for TB detection as high. But the pooled estimates for TB detection were 88% sensitivity and 98% specificity, not the cited 89% and 99%; the 89% sensitivity reported in Abstract > Main results applies specifically to people without HIV infection. (LLM, conf=0.91)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

A protocol was developed before commencing the review, following standard guidelines.

“We developed a protocol before commencing the review, following standard guidelines [12,13].”

• cites [12] sourcePartly supported

  The Introduction describes this paper as methodological guidance for diagnostic test accuracy reviews, drawing on the Cochrane Handbook and updating earlier guidelines. It does not say that reviewers should develop a protocol before starting, and it cannot verify the citing authors’ statement that they did so. (LLM, conf=0.82)

  Experimental — accuracy not yet measured
• cites [13] Macaskill P Gatsonis C Deeks JJ . Cochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy Version. London, The Cochrane Collaboration, 2010Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

The reference standard could introduce bias due to misclassification of participants.

“We also considered the possibility that the reference standard could introduce bias due to misclassification of participants (fig.2) [34,35].”

• cites [34] sourceSupported
• cites [35] sourceSupported

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The majority of studies collected data prospectively (n=14; 78%) and only three studies used a case-control design.

“The majority of studies collected data prospectively (n514; 78%) (table 1) and only three studies used a case-control design [15,23,26].”

Probably not a substantial claim: may not pass S3 (load-bearing).

The percentages of HIV-positive patients included in remaining studies ranged from 1% to 87% of the study population.

“The percentages of HIV-positive patients included in remaining studies ranged from 1% to 87% of the study population (table 1).”

May combine two claims.

The review identified 18 studies involving 4461 samples.

“We identified 18 studies (fig. 1) [14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31] that included 4461 samples.”

Probably not a substantial claim: may not pass S3 (load-bearing).

One study only included HIV-positive patients.

“One study only included HIV-positive patients [30].”

The quote only partly states this claim.

Data were too limited to estimate specificity for smear-positive samples.

“Data were too limited to estimate specificity.”

Probably not a substantial claim: may not pass S3 (load-bearing).

11 (55%) studies reported a concentration step.

“11 (55%) studies reported a concentration step (table 2).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Six studies did not include any HIV-positive patients and for two studies HIV status was unknown.

“Six studies did not include any HIV-positive patients and for two studies, HIV status was unknown [20,32].”

Probably not a substantial claim: may not pass S3 (load-bearing).

One study reported noninterpretable Xpert results in pleural fluid: 5.4% (six out of 111).

“One study reported noninterpretable results for Xpert: 5.4% (six out of 111) [20].”

The quote only partly states this claim.

Xpert pooled specificity was consistently >98.7% against composite reference standard across different sample types.

“Xpert pooled specificity was consistently .98.7% against CRS across different sample types.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Eight (44%) studies were conducted in low/middle-income countries.

“Eight (44%) studies were conducted in low/ middle-income countries (table 1).”

May combine two claims.

Data for smear status of samples were not available for individual sample types, so samples included in the subgroups may be smear-positive, negative or unknown.

“Data for smear status of samples were not available for the individual sample types.Therefore, samples included in the subgroups may be smearpositive, negative or unknown.”

May combine two claims.

Studies using fresh samples showed higher sensitivity (86.4%, 95% CI 75.7-97.1%) than those using frozen samples (74.0%, 95% CI 56.5-91.5%) for lymph node TB, but precision was low.

“Studies that used fresh samples showed a higher sensitivity (86.4%, 95% CI 75.7-97.1%)than those that used frozen samples (74.0%, 95% CI 56.5-91.5%);however, the precision of these estimates was low as data were limited.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Most studies that included a mechanical homogenisation step (40%) also performed decontamination with N-acetyl-L-cysteine and sodium hydroxide solution.

“Most of the studies that included a mechanical homogenisation step (40%) also performed a decontamination procedure with N-acetyl-L-cysteine and sodium hydroxide solution.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Six studies used archived frozen samples, 11 used fresh samples, and one used both fresh and frozen samples.

“Six studies used archived frozen samples, 11 used fresh samples, and one used both fresh and frozen samples (table 2).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Five (25%) studies used a sample reagent/sample volume ratio of 3/1, while the remainder used 2/1.

“Five (25%) studies used a ratio of 3/1, while the remainder of studies used 2/1.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A univariate analysis for specificity in pleural fluid achieved a similar estimate (98.9%) with a slightly narrower confidence interval (97.9-99.8%) compared with the bivariate analysis.

“In comparison with the bivariate analysis, a univariate analysis for specificity achieved a similar estimate (98.9%) with a slightly narrower confidence interval (97.9-99.8%).”

Probably not a substantial claim: may not pass S3 (load-bearing).

The median number of samples per study was 137 (interquartile range 67-342).

“The median number of samples per study was 137 (interquartile range 67-342).”

May combine two claims.

In lymph node samples versus composite reference standard, Xpert pooled specificity was 99.1% (95% CI 94.5-99.9%), improved compared with the culture reference standard.

“As expected, the specificity was improved, at 99.1% (95% CI 94.5-99.9%), in comparison to the culture reference standard.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

10 studies included children, with percentages ranging from 2% to 34%.

“10 studies included children, with percentages ranging from 2% to 34%.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). The quote only partly states this claim.

In lymph node samples versus culture, Xpert pooled specificity was 93.6% (95% CI 87.9-96.8%).

“pooled specificity was 93.6% (95% CI 87.9-96.8%).”

Probably not a substantial claim: may not pass S3 (load-bearing).

All studies were performed either in tertiary care centres or reference laboratories.

“All studies were performed either in tertiary care centres or reference laboratories.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims. The quote only partly states this claim.

Only two studies reported noninterpretable Xpert results in lymph node samples: 1.4% and 10%.

“Only two studies reported any noninterpretable results for Xpert: 1.4% (five out of 353) and 10% (two out of 20) [15,30].”

Probably not a substantial claim: may not pass S3 (load-bearing).

Seven studies included only one sample type.

“Seven studies included only one sample type (e.g.pleural fluid only) [17,18,21,25,26,30,39] (table 1).”

May combine two claims.

Five studies (711 samples) assessed Xpert in CSF samples versus a composite reference standard (CRS) and found variable sensitivity (20-86%).

“Five studies (711 samples) that assessed Xpert in CSF samples versus a CRS found variable sensitivity (20-86%) (fig.5b) [24,25,[28][29]31].”

• cites [24,25,[28][29]31]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Prevalence of HIV and the condition of the specimen did not affect Xpert sensitivity or specificity in CSF.

“Prevalence of HIV and the condition of the specimen did not have an effect on Xpert sensitivity and specificity in CSF.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Study quality was assessed separately for culture and composite reference standard groups using QUADAS-2, a validated tool.

“We grouped studies according to the type of reference standard (culture versus CRS) and assessed study quality separately for the two groups using QUADAS-2, a validated tool for diagnostic studies [32].”

• cites [32] sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Xpert has been approved for TB detection in sputum by regulatory agencies, while Xpert for TB detection in nonrespiratory specimens is considered off-label use.

“While Xpert has been approved for TB detection in sputum by regulatory agencies, Xpert for TB detection in nonrespiratory specimens is considered ''off-label'' use [8][9][10].”

• cites [8][9][10]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Repeat examination of the cerebrospinal fluid or the use of NAATs may increase sensitivity for TB meningitis and is associated with high specificity (>98%).

“although repeat examination of the cerebrospinal fluid (CSF) or the use of NAATs may increase sensitivity and is associated with high specificity (.98%) [5,6].”

• cites [5] sourceNot checked — no full text of the source
• cites [6] sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Differences in study selection and data extraction between the two review authors were resolved by discussion with a third review author.

“Any differences in the study selection and data extraction process between the two review authors were resolved by discussion with a third review author (K.R. Steingart).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Two review authors independently assessed all articles for inclusion and extracted data using a standardised extraction form.

“Two review authors (C.M. Denkinger and S.G.Schumacher) independently assessed all articles for inclusion in the systematic review, and extracted data on study methodology, characteristics and test accuracy using a standardised extraction form (tables 1 and 2; extraction form available in the online supplementary material).”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The sponsors of the study had no role in study design, data collection, data analysis, data interpretation or writing of the report.

“The sponsors of the study had no role in study design, data collection, data analysis, data interpretation or writing of the report.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Two sensitivity analyses were performed by limiting inclusion to studies with consecutive patient selection or studies that did not use a case-control design.

“In addition, we performed two sensitivity analyses by limiting inclusion in the meta-analysis to: 1) studies in which patients were selected consecutively; or 2) studies that did not use a case-control design.”

Classified as probably not a substantial claim.

Differences in sample processing might affect estimates of the diagnostic accuracy of Xpert to varying degrees.

“We considered that differences in sample processing might affect estimates of the diagnostic accuracy of Xpert to varying degrees (table 2).”

May combine two claims.

Formal assessment of publication bias using funnel plots or regression tests was not carried out because such techniques are not considered valid for diagnostic accuracy reviews.

“We chose not to carry out formal assessment of publication bias using methods such as funnel plots or regression tests because such techniques are not considered to be valid for diagnostic accuracy reviews [13].”

• cites [13] Macaskill P Gatsonis C Deeks JJ . Cochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy Version. London, The Cochrane Collaboration, 2010Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Classified as probably not a substantial claim.

Subgroups by sample type were pre-specified: pleural fluid, lymph node samples, and cerebrospinal fluid.

“We expected heterogeneity in terms of the sample types.Therefore, we pre-specified subgroups by sample type: pleural fluid, lymph node samples (tissue and aspirate combined; reporting in studies did not allow separating tissue samples from aspirates) and CSF.”

Classified as probably not a substantial claim. May combine two claims.

The mechanical homogenisation step could be a source of variation in test accuracy because procedural differences might affect the quantity of sample particles and particles could clog cartridge valves leading to noninterpretable results.

“In particular, we were concerned that the mechanical homogenisation step could be a source of variation in test accuracy (n58, 40%) for two reasons: procedural differences might affect the quantity of sample particles in the sample volume and the particles could clog the cartridge valves leading to noninterpretable results.”

Classified as probably not a substantial claim. May combine two claims.

Based on this systematic review, the World Health Organization now recommends Xpert over conventional tests for diagnosis of TB in lymph nodes and other tissues, and as the preferred initial test for diagnosis of TB meningitis.

“Based on this systematic review, the World Health Organization now recommends Xpert over conventional tests for diagnosis of TB in lymph nodes and other tissues, and as the preferred initial test for diagnosis of TB meningitis.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Noninterpretable test results were excluded from the analyses for determination of sensitivity and specificity.

“We excluded noninterpretable test results (i.e.invalid or erroneous results) from the analyses for determination of sensitivity and specificity [37].”

• cites [37] Cepheid. Xpert MTB/RIF. http://tbevidence.org/documents/rescentre/sop/XpertMTB_Broch_R9_EU.pdf .Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Xpert MTB/RIF can improve the diagnosis of lymph node TB and TB meningitis.

“@ERSpublicationsXpert MTB/RIF can improve the diagnosis of lymph node TB and TB meningitis”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The data set involved comprehensive searching to identify studies and repeated correspondence with study authors to obtain additional data.

“This data set involved comprehensive searching to identify studies as well as repeated correspondence with study authors to obtain additional data on the studies.”

May combine two claims.

Xpert on pleural fluid has higher sensitivity than smear and provides a more rapid diagnosis than culture and histology.

“Where resources are available, Xpert on pleural fluid could still be considered in the work-up of pleural TB as it has higher sensitivity than smear and provides a more rapid diagnosis than culture and histology.”

Classified as probably not a substantial claim. May combine two claims.

The review used a standard protocol, strict inclusion criteria, standardised data extraction, independent reviewers, a bivariate random-effects model for meta-analysis, and pre-specified subgroups to account for heterogeneity.

“Strengths of our review include the use of a standard protocol, strict inclusion criteria, standardised data extraction, independent reviewers, a bivariate random-effects model for meta-analysis and pre-specified subgroups to account for heterogeneity.”

May combine two claims.

Sample processing was highly variable across and within studies, as there was no recommendation available on how to process nonrespiratory samples from the manufacturer or the WHO.

“Furthermore, sample processing was highly variable across and within studies, as there was no recommendation available on how to process nonrespiratory samples from the manufacturer or the WHO.”

Classified as probably not a substantial claim. May combine two claims.

WHO recommends Xpert over conventional tests for the diagnosis of TB in lymph nodes and other tissues, and as the preferred initial test for the diagnosis of TB meningitis.

“WHO now recommends Xpert over conventional tests for the diagnosis of TB in lymph nodes and other tissues, and as the preferred initial test for the diagnosis of TB meningitis.”

• cites [11] World Health Organization. Automated real-time nucleic acid amplification technology for rapid and simultaneous detection of tuberculosis and rifampicin resistance: Xpert MTB/RIF system for the diagnosis of pulmonary and extrapulmonary TB in adults and children. Geneva, WHO, 2013.Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

A concentration step for CSF increases Xpert sensitivity with unchanged specificity, probably by increasing the bacillary load in the cartridge input volume.

“Interestingly, we observed that a concentration step for CSF increases Xpert sensitivity with unchanged specificity, probably by increasing the bacillary load in the cartridge input volume.”

May combine two claims.

Patients in low-prevalence settings likely present earlier with more paucibacillary disease, which might result in decreased sensitivity of Xpert.

“It is likely that patients in these settings present earlier with more paucibacillary disease, which might result in a decreased sensitivity of Xpert [45].”

• cites [45] sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

The review may have missed some studies despite the comprehensive search.

“We acknowledge that we may have missed some studies despite the comprehensive search.”

Classified as probably not a substantial claim.

A detailed investigation of the best sample processing was not possible in this review due to limited data for the sample types and large variation in sample processing.

“Given the limited data for the sample types and the large variation in sample processing, a detailed investigation of the best sample processing was not possible in this review.”

May combine two claims.

The high sensitivity of Xpert in smear-positive samples (97.4%) and the low proportion of noninterpretable results (1.2%) support the use of Xpert in nonrespiratory samples in principle.

“The high sensitivity (97.4%, 95% CI 95.5-99.3%) of Xpert in smear-positive samples across sample types and the low proportion of noninterpretable results (1.2%) support the use of the test in nonrespiratory samples in principle.”

May combine two claims.

The poor sensitivity of Xpert in pleural fluid is probably due to the paucibacillary nature of the disease and the fact that pleural biopsy, not pleural fluid, is the sample of choice for the diagnosis of pleural TB.

“The poor sensitivity of Xpert in pleural fluid is probably due to the paucibacillary nature of the disease and the fact that not pleural fluid but rather pleural biopsy is the sample of choice for the diagnosis of pleural TB (as has been described for culture) [41].”

• cites [41] sourceNot checked — no full text of the source