Diagnostic accuracy of nucleic acid amplification based assays for tuberculous meningitis: A meta-analysis

Comparative diagnostic utility of metagenomic next-generation sequencing, GeneXpert, modified Ziehl-Neelsen staining, and culture using cerebrospinal fluid for tuberculous meningitis: A multi-center, retrospective study in China

Background

Early diagnosis of tuberculosis meningitis (TBM) remains a great challenge during clinical practice. The diagnostic efficacies of cerebrospinal fluid (CSF)‐based mycobacterial growth indicator tube (MGIT) culture, modified Ziehl–Neelsen (ZN) staining, Xpert MTB/RIF, and metagenomic next‐generation sequencing (mNGS) for TBM remained elusive.

Methods

A total of 216 adult patients with suspicious TBM were retrospectively enrolled in this multi‐cohort study. The diagnostic performances for MGIT, modified ZN staining, Xpert MTB/RIF, and mNGS using CSF samples were evaluated.

Results

Uniform clinical case definition classified 88 (40.7%) out of 216 patients as the definite TBM, 5 (2.3%) patients as probable TBM cases, and 24 (11.1%) patients as possible TBM cases. The sensitivities of MGIT, modified ZN staining, Xpert MTB/RIF, and mNGS for TBM diagnosis against consensus uniform case definition for definite TBM were 25.0%, 76.1%, 73.9%, and 84.1%, respectively. Negative predictive values (NPVs) were 66.0%, 85.9%, 84.8%, and 90.1%, respectively. The sensitivities of MGIT, modified ZN staining, Xpert MTB/RIF, and mNGS for TBM diagnosis against consensus uniform case definition for definite, probable, and possible TBM were 18.8%, 57.3%, 55.5%, and 63.2%, respectively. Negative predictive values (NPVs) were 51.0%, 66.4%, 65.6%, and 69.7%, respectively. mNGS combined with modified ZN stain and Xpert could cover TBM cases against a composite microbiological reference standard, yielding 100% specificity and 100% NPV.

Conclusion

Metagenomic next‐generation sequencing detected TBM with higher sensitivity than Xpert, ZN staining and MGIT culture, but mNGS cannot be used as a rule‐out test. mNGS combined with Xpert or modified ZN staining could enhance the sensitivity of diagnostic tests for TBM.

Xpert MTB/RIF Ultra and Xpert MTB/RIF assays for extrapulmonary tuberculosis and rifampicin resistance in adults

BACKGROUND

Xpert MTB/RIF Ultra (Xpert Ultra) and Xpert MTB/RIF are World Health Organization (WHO)-recommended rapid nucleic acid amplification tests (NAATs) widely used for simultaneous detection of Mycobacterium tuberculosis complex and rifampicin resistance in sputum. To extend our previous review on extrapulmonary tuberculosis (Kohli 2018), we performed this update to inform updated WHO policy (WHO Consolidated Guidelines (Module 3) 2020).

OBJECTIVES

To estimate diagnostic accuracy of Xpert Ultra and Xpert MTB/RIF for extrapulmonary tuberculosis and rifampicin resistance in adults with presumptive extrapulmonary tuberculosis.

SEARCH METHODS

Cochrane Infectious Diseases Group Specialized Register, MEDLINE, Embase, Science Citation Index, Web of Science, Latin American Caribbean Health Sciences Literature, Scopus, ClinicalTrials.gov, the WHO International Clinical Trials Registry Platform, the International Standard Randomized Controlled Trial Number Registry, and ProQuest, 2 August 2019 and 28 January 2020 (Xpert Ultra studies), without language restriction.

SELECTION CRITERIA

Cross-sectional and cohort studies using non-respiratory specimens. Forms of extrapulmonary tuberculosis: tuberculous meningitis and pleural, lymph node, bone or joint, genitourinary, peritoneal, pericardial, disseminated tuberculosis. Reference standards were culture and a study-defined composite reference standard (tuberculosis detection); phenotypic drug susceptibility testing and line probe assays (rifampicin resistance detection).

DATA COLLECTION AND ANALYSIS

Two review authors independently extracted data and assessed risk of bias and applicability using QUADAS-2. For tuberculosis detection, we performed separate analyses by specimen type and reference standard using the bivariate model to estimate pooled sensitivity and specificity with 95% credible intervals (CrIs). We applied a latent class meta-analysis model to three forms of extrapulmonary tuberculosis. We assessed certainty of evidence using GRADE.

MAIN RESULTS

69 studies: 67 evaluated Xpert MTB/RIF and 11 evaluated Xpert Ultra, of which nine evaluated both tests. Most studies were conducted in China, India, South Africa, and Uganda. Overall, risk of bias was low for patient selection, index test, and flow and timing domains, and low (49%) or unclear (43%) for the reference standard domain. Applicability for the patient selection domain was unclear for most studies because we were unsure of the clinical settings. Cerebrospinal fluid Xpert Ultra (6 studies) Xpert Ultra pooled sensitivity and specificity (95% CrI) against culture were 89.4% (79.1 to 95.6) (89 participants; low-certainty evidence) and 91.2% (83.2 to 95.7) (386 participants; moderate-certainty evidence). Of 1000 people where 100 have tuberculous meningitis, 168 would be Xpert Ultra-positive: of these, 79 (47%) would not have tuberculosis (false-positives) and 832 would be Xpert Ultra-negative: of these, 11 (1%) would have tuberculosis (false-negatives). Xpert MTB/RIF (30 studies) Xpert MTB/RIF pooled sensitivity and specificity against culture were 71.1% (62.8 to 79.1) (571 participants; moderate-certainty evidence) and 96.9% (95.4 to 98.0) (2824 participants; high-certainty evidence). Of 1000 people where 100 have tuberculous meningitis, 99 would be Xpert MTB/RIF-positive: of these, 28 (28%) would not have tuberculosis; and 901 would be Xpert MTB/RIF-negative: of these, 29 (3%) would have tuberculosis. Pleural fluid Xpert Ultra (4 studies) Xpert Ultra pooled sensitivity and specificity against culture were 75.0% (58.0 to 86.4) (158 participants; very low-certainty evidence) and 87.0% (63.1 to 97.9) (240 participants; very low-certainty evidence). Of 1000 people where 100 have pleural tuberculosis, 192 would be Xpert Ultra-positive: of these, 117 (61%) would not have tuberculosis; and 808 would be Xpert Ultra-negative: of these, 25 (3%) would have tuberculosis. Xpert MTB/RIF (25 studies) Xpert MTB/RIF pooled sensitivity and specificity against culture were 49.5% (39.8 to 59.9) (644 participants; low-certainty evidence) and 98.9% (97.6 to 99.7) (2421 participants; high-certainty evidence). Of 1000 people where 100 have pleural tuberculosis, 60 would be Xpert MTB/RIF-positive: of these, 10 (17%) would not have tuberculosis; and 940 would be Xpert MTB/RIF-negative: of these, 50 (5%) would have tuberculosis. Lymph node aspirate Xpert Ultra (1 study) Xpert Ultra sensitivity and specificity (95% confidence interval) against composite reference standard were 70% (51 to 85) (30 participants; very low-certainty evidence) and 100% (92 to 100) (43 participants; low-certainty evidence). Of 1000 people where 100 have lymph node tuberculosis, 70 would be Xpert Ultra-positive and 0 (0%) would not have tuberculosis; 930 would be Xpert Ultra-negative and 30 (3%) would have tuberculosis. Xpert MTB/RIF (4 studies) Xpert MTB/RIF pooled sensitivity and specificity against composite reference standard were 81.6% (61.9 to 93.3) (377 participants; low-certainty evidence) and 96.4% (91.3 to 98.6) (302 participants; low-certainty evidence). Of 1000 people where 100 have lymph node tuberculosis, 118 would be Xpert MTB/RIF-positive and 37 (31%) would not have tuberculosis; 882 would be Xpert MTB/RIF-negative and 19 (2%) would have tuberculosis. In lymph node aspirate, Xpert MTB/RIF pooled specificity against culture was 86.2% (78.0 to 92.3), lower than that against a composite reference standard. Using the latent class model, Xpert MTB/RIF pooled specificity was 99.5% (99.1 to 99.7), similar to that observed with a composite reference standard. Rifampicin resistance Xpert Ultra (4 studies) Xpert Ultra pooled sensitivity and specificity were 100.0% (95.1 to 100.0), (24 participants; low-certainty evidence) and 100.0% (99.0 to 100.0) (105 participants; moderate-certainty evidence). Of 1000 people where 100 have rifampicin resistance, 100 would be Xpert Ultra-positive (resistant): of these, zero (0%) would not have rifampicin resistance; and 900 would be Xpert Ultra-negative (susceptible): of these, zero (0%) would have rifampicin resistance. Xpert MTB/RIF (19 studies) Xpert MTB/RIF pooled sensitivity and specificity were 96.5% (91.9 to 98.8) (148 participants; high-certainty evidence) and 99.1% (98.0 to 99.7) (822 participants; high-certainty evidence). Of 1000 people where 100 have rifampicin resistance, 105 would be Xpert MTB/RIF-positive (resistant): of these, 8 (8%) would not have rifampicin resistance; and 895 would be Xpert MTB/RIF-negative (susceptible): of these, 3 (0.3%) would have rifampicin resistance.

AUTHORS' CONCLUSIONS

Xpert Ultra and Xpert MTB/RIF may be helpful in diagnosing extrapulmonary tuberculosis. Sensitivity varies across different extrapulmonary specimens: while for most specimens specificity is high, the tests rarely yield a positive result for people without tuberculosis. For tuberculous meningitis, Xpert Ultra had higher sensitivity and lower specificity than Xpert MTB/RIF against culture. Xpert Ultra and Xpert MTB/RIF had similar sensitivity and specificity for rifampicin resistance. Future research should acknowledge the concern associated with culture as a reference standard in paucibacillary specimens and consider ways to address this limitation.

Diagnostic accuracy of Xpert MTB/RIF Ultra for tuberculous meningitis in a clinical practice setting of China

A comparative performance evaluation of the novel Xpert MTB/RIF Ultra (Xpert Ultra) and MTB/RIF Xpert (Xpert) for tuberculous meningitis (TBM) diagnosis was performed. The cerebrospinal fluids of suspected TBM patients were collected consecutively and subjected to smear microscopy, culture, Xpert, and Xpert Ultra. In total, 160 patients were recruited. Xpert Ultra produced a higher sensitivity (45%, 34 of 76) than Xpert (28%, 21 of 76; P = 0.001) and culture (18%, 14 of 76; P < 0.001), respectively. Inclusion of Xpert Ultra outcomes increased the percentage of definite TBM case from 36% (27 of 76) to 51% (39 of 76). Both Xpert Ultra and Xpert accurately identified the one rifampicin (RIF)-resistant and the 5 RIF-sensitive cases defined by phenotypic drug sensitivity test. The specificities of all of the culture, Xpert and Xpert Ultra were 100% (45 of 45). Xpert Ultra outperformed both Xpert and culture for TBM diagnosis, which may speed up the appropriate treatment of patients in clinical practice.

Multidrug-resistant strains of Mycobacterium complex species in Egyptian farm animals, veterinarians, and farm and abattoir workers

Background and Aim:

Mycobacterium tuberculosis complex (MTBC) is a group of mycobacteria that are important human pathogens. Mycobacterium tuberculosis and Mycobacterium bovis cause serious chronic life-threatening disease and also significant economic losses in both production and remedication. Recently, emergence of multidrug-resistant tuberculosis (MDR-TB) complex has generated global recognition of the need for rapid and sensitive diagnosis and development of new treatments. The current study illustrates the isolation/identification of MTBC strains in specimens obtained from cows and humans by conventional and real-time polymerase chain reaction (RT-PCR) techniques. Further, the study assesses sensitivity to antituberculosis drugs in isolated MDR strains.

Materials and Methods:

A total of 1464 samples from cattle (1285 raw milk and 179 lymph node), and 149 human sputum samples, were collected from farms and abattoirs in Delta Egypt. Conventional methods (culture and Ziehl–Neelsen staining) were implemented as were RT-PCR using MTBC universal DNA. The effect of some antituberculosis drugs on obtained isolates was assayed using drug susceptibility proportion and qualitative suspension techniques.

Results:

The MBTC detection rate using the culture method was higher than for Ziehl–Neelsen staining; raw cow milk (2.56 vs. 1.63%), lymph nodes (51.59 vs. 48.04%), and human sputum (5.36 vs. 4.02%). A total of 135 isolates were obtained. Application of RT-PCR detected 138 isolates from the same set of samples. MBTC isolates were resistant to first-line antituberculosis drugs, such as pyrazinamide, isoniazid, rifampicin, and ethambutol by 78.5, 59.3, 40.7, and 31.8%, respectively, and could be highly resistant to kanamycin (82.3%) and amikacin (80.7%). However, isolates remained sensitive to ciprofloxacin (71.1%) and clarithromycin (73.3%) as second-line drugs.

Conclusion:

There is a growing risk for isolation of MDR-TB from raw milk and lymph nodes of field tuberculin positive cattle as well as sputum of veterinarians and workers existed in farms and abattoirs. PCR-based techniques have become the gold standard for the identification of mycobacterial species, showing high efficiency compared to bacteriological and microscopic examination. Application of the first- and second-line antituberculosis drugs in combination could counter the MDR-TB concern once infections are identified.

Advances in the Diagnosis and Management of Tubercular Meningitis in Children

Tubercular meningitis (TBM) continues to be a common cause of neuromorbidity in children. There is no single diagnostic method that can rapidly detect Mycobacterium tuberculosis (M.tb) in TBM patients with high sensitivity and specificity. Newer diagnostic modalities like Xpert/RIF assay and Loop mediated isothermal amplification assay (LAMP) have gained an essential stand in molecular diagnostics due to their high specificity, modest sensitivity in cerebrospinal fluid (CSF) and quick availability of results. Intensified drug regimens using high dose rifampicin, fluoroquinolone and aspirin appear to be useful adjunct therapy but more pediatric clinical trials on large scale are needed to determine their appropriate place in pediatric TBM. The emergence of multi and extreme drug resistant M.tb strains further challenges the standard therapy. In this review authors summarize challenges of the currently used diagnostic methods and treatment for TBM and discuss the recent advances.