HTB

The tuberculosis diagnostics pipeline

2016 Pipeline Report coverErica Lessem, TAG

Introduction

Over the past decade, several new options for improving TB diagnosis have become available. The past year saw considerably more progress than the previous one. However, the reality of how most TB is diagnosed – or not – remains largely unchanged. The world failed to detect 3.6 million of the estimated 9.6 million new cases of TB in 2014. [1]

Sputum smear microscopy – which misses over half of TB cases [2] and gives no indication of drug susceptibility to guide appropriate treatment – is still the diagnostic standard in most of the world, despite the availability of the far more sensitive GeneXpert MTB/RIF for six years. [3] Late in 2015, the World Health Organization (WHO) approved Alere’s TB lipoarabinomannan (LAM) test – a very affordable, simple, rapid, noninvasive, point-of-care (POC) rule-in test for people with HIV with very low CD4 counts – but no country has begun to implement it yet. New versions of line probe assays – Hain’s MTBDRplus and MTBDRsl and a product from Nipro – received WHO recommendation, facilitating rapid drug susceptibility testing (DST), but the world is still a long way from universal DST, with an estimated 59 percent of cases of multidrug-resistant TB (MDR-TB) undetected. [4]

Research and development (R&D) for TB diagnostics features some promising developments since last year (see table 1). Improvements on nucleic acid amplification tests (NAATs) such as GeneXpert Omni and Ultra and Molbio’s TrueNAT are being validated, positioning them for possible WHO recommendation. Further upstream, encouraging research into gene sets that can predict active TB disease and reliably distinguish it from latent TB and other infections may eventually underpin new blood tests (currently, there is no effective serological test for active TB). Incremental advances are being made to improve detection of pediatric TB (see Extending quality).

Yet, overall, with a mere US$65 million in 2014 funding out of an estimated annual need of $340 million, [5] the pipeline for evidence-based new diagnostics has largely remained stagnant (see table 2). Dismayingly, some companies continue to move forward with marketing for their products when the data are unavailable or scream that they should not, as is the case with Epistem, which is marketing GeneDrive in India despite the test’s having flopped in studies.

With use of poor tests predominating, poor uptake of good extant options, poor evidence bases to support the introduction of new tests, and poor funding to support the development of better tests, it’s no wonder we’ve made little headway in diagnosing TB.

Table 1. 2016 Tuberculosis Diagnostics Pipeline: Products in Later-Stage Development or on Track for Evaluation by the WHO with New Published Data or Policy Updates Since the 2015 Pipeline Report
Test Type Sponsor Status Comments
Molecular/NAAT
BD MAX MTB assay qPCR for MTB in automated BD MAX BD In 16 M. tuberculosis samples, 100% sensitivity, 97.1% specificity [6]
Genedrive MTB/RIF Portable RT-PCR for MTB + RIF resistance Epistem Worse sensitivity than smear [!] in 2016 study [7] Marketed in India
GenoType MTBDRplus Line probe assay for RIF + INH resistance Hain Lifescience WHO now recommends based on FIND evaluation [8] WHO guidance pending
GenoType MTBDRsl Line probe assay for FQ + SLID resistance Hain Lifescience WHO now recommends [9] FIND’s multicountry evaluation of MTBDRsl version 2.0 from 2015 still unpublished
MeltPro Closed-tube RT-PCR Zeesan Biotech New study from China of 2,057 smear-positive TB patients shows sensitivity of detecting resistance to rifampin 94.2%, isoniazid 84.9%, ofloxacin 83.3%, amikacin 75.0%, kanamycin 63.5% [10]
NTM+MDRTB Detection Kit 2 Line probe assay for RIF + INH resistance Nipro WHO now recommends based on FIND evaluation [11] WHO guidance pending
RealTime MTB/ TB MDx m2000 Automated RT-PCR for MTB; can be added to HIV RNA platform Abbott Sensitivity 100%, 95% CI: 98.6–99.9 in smear-positive samples, similar to GeneXpert MTB/RIF [12]
Truenat MTB Chip-based NAAT with RT-PCR on handheld device for MTB Molbio Diagnostics, Bigtec Labs FIND and ICMR studies underway
Xpert MTB/RIF Ultra Next-generation cartridge-based detection of MTB + RIF resistance Cepheid FIND study results anticipated end 2016
Xpert Omni Single-cartridge mobile platform that can use single MTB/RIF or Ultra cartridge Cepheid FIND study pending but delayed
Xpert XDR NAAT Cepheid FIND study anticipated 2018
Antibody/Antigen Detection
Determine TB LAM Ag Urine dipstick for TB LAM protein Alere WHO recommended use in people with HIV with CD4 count <100 [13]

CI: confidence interval
FLD: first-line drugs (isoniazid, rifampin, ethambutol, pyrazinamide)
FQ: fluoroquinolone
ICMR: Indian Council of Medical Research
INH: isoniazid
LAM: lipoarabinomannan
MDR-TB: multidrug-resistant tuberculosis
MTB: Mycobacterium tuberculosis
NAAT: nucleic-acid amplification test
qPCR: quantitative polymerase chain reaction
RIF: rifampin
RT-PCR: real-time polymerase chain reaction SLID: second-line injectable drug (e.g., amikacin, capreomycin, or kanamycin) WHO: World Health Organization

Table 2. Later-Stage or Marketed TB Diagnostic Test Candidates with No New Published Evaluation Data
Test Type Sponsor Status
Molecular/NAAT
EasyNAT Isothermal DNA amplification/lateral flow to detect MTB Ustar No new data since poor in Tanzanian field study [14]
FluoroType MTB Semiautomated direct MTB detection; PCR in a closed system; results in 3 hours Hain Lifescience Marketed
FluoroType MTB RNA MTB RNA for monitoring of anti-TB therapy Hain Lifescience No published data
GeneChip RT-PCR for RIF + INH DR CapitalBio Marketed
LATE-PCR with Lights-On/Lights-Off Probes + PrimeSafe Single-tube PCR to detect MTB, resistance to INH, RIF, EMB, SLID Hain Lifescience/Brandeis University, Stellenbosch University No published data
LiPA pyrazinamide Line probe assay for PZA resistance Nipro Marketed
PureLamp Manual NAAT by loop-mediated isothermal amplification for MTB detection Eiken WHO June 2015 review impeded by data quality issues; guidelines development group reexamined evidence January 2016; recommendations expected June 2016
REBA MTB-MDR Line probe assay for RIF + INH resistance YD Diagnostics No new data; marketed
REBA MTB-XDR Line probe assay for FQ + SLID DR YD Diagnostics No new data; marketed
TREK Sensititre MYCOTB MIC plate Dry microdilution plate to detect MICs for FLD + SLD (except PZA) TREK Diagnostic Systems, Thermo Fisher Scientific No new evaluation data but used in study in Cameroon [15]
TRC Rapid MTB Automated rapid rRNA to detect MTB Tosoh No new data
Volatile Organic Compounds
Giant African pouched rats (Cricetomys gambianus) Trained sniffer rats to detect MTB in sputum Apopo Foundation No new data
Automated Imaging
CAD 4TB Digital CXR for TB screening Delft Imaging Systems Marketed; in 2016 WHO to review available evidence on computer-aided radiographic TB detection and organize a scoping meeting to determine research needs and if guidelines should be developed [16]
Antibody/Antigen Detection
MBio Array System POC cartridge to measure ~57 simultaneous MTB antigen-antibody reactions MBio Diagnostics,FIND No new data

CXR: chest x-ray
DR: drug resistance
EMB: ethambutol
FLD: first-line drugs (isoniazid, rifampin, ethambutol, pyrazinamide)
FQ: fluoroquinolone
INH: isoniazid
MDR-TB: multidrug-resistant TB
MIC: minimum inhibitory concentration
MTB: Mycobacterium tuberculosis
NAAT: nucleic-acid amplification test
PCR: polymerase chain reaction POC: point of care
PZA: pyrazinamide
RIF: rifampin
RT-PCR: real-time polymerase chain reaction
SLD: second-line drug
SLID: second-line injectable drug (e.g., amikacin, capreomycin, or kanamycin)
TB: tuberculosis
WHO: World Health Organization

Trot, Trot to Market

Perhaps the most exciting advance in TB diagnostics came in late 2015, when the WHO recommended Alere’s Determine LAM Ag simple urine dipstick test for ruling in TB in people with HIV with CD4 counts below 100/mm3 or who are seriously ill. [17] The test’s imperfect sensitivity (56% pooled sensitivity based on five studies of people with CD4 counts below 100/mm3) means that a negative test must still be followed up with other testing to rule out TB. [18] However, given the extremely high mortality of people with TB and HIV (TB is thought to be the cause of death in nearly 40% of HIV-positive patients, half of which is undiagnosed) [19] and challenges in diagnosing TB in people with low CD4 counts, having an inexpensive ($2.26 per test), simple, and noninvasive test to use in this very high-risk population is a major advance. In fact, the LAM test is the first TB test to ever demonstrate a mortality benefit in a randomized controlled clinical trial: among 578 people with HIV in hospitals in South Africa, Tanzania, Zambia, and Zimbabwe, using LAM was associated with an absolute reduction of all-cause mortality at eight weeks of 4% (95% confidence interval [CI]: 1%–7%) from 25% to 21%, and a relative risk reduction of 17% (95% CI: 4%–28%). [20] This difference appeared to be attributable to the test’s allowing earlier initiation (by one day on average) of anti-TB therapy. [21] Countries with large burdens of TB/HIV, including many countries in sub-Saharan Africa, should roll out LAM testing immediately, along with proper accompanying training to ensure the test is used only in the recommended population and that follow-up tests are done as necessary. [22] Whoever ends up with the rights to the test – Abbott is trying to pull out of a putative acquisition of Alere – should ensure its continued manufacture as well as marketing. [23] Further upstream, funding from the Global Health Innovative Technology Fund (GHIT Fund, which is itself funded by the Japanese government, pharmaceutical companies, the Bill & Melinda Gates Foundation, and the Wellcome Trust) to FIND and Fujifilm will support the development of what is hoped to be a more sensitive LAM test. [24, 25, 26]

Other policy advances can help improve timely detection of drug-resistant TB. In late 2015, the WHO extended its 2008 guidance, which recommended the use of the Hain version 1 line probe assay (LPA), to recommend the use of two alternative LPAs with the capability to detect TB and rifampin resistance: the Hain version 2 LPA (also called the GenoType MTBDRplus assay) and the Nipro Assay. [27] The WHO still does not recommend using LPAs on smear-negative samples. In 2014 and 2015, FIND conducted a cross-sectional noninferiority study to compare the accuracy of these two tests to that of Hain Version 1 assay, evaluating their performance both on clinical isolates and on sputum specimens from people with pulmonary TB; both tests showed comparable performance in detecting Mycobacterium tuberculosis (the bacterium that causes TB infection and disease) and rifampin resistance in smear-positive samples: on clinical isolates, sensitivity and specificity compared with the phenotypic reference standard for Hain V1, HainV2, and Nipro were 90.3%/98.5%, 90.3%/98.5% and 92.0%/98.5%, respectively, for detection of rifampin resistance and 89.1%/99.4%, 89.1%/99.4%, and 89.6%/100.0%, respectively, for detecting isoniazid resistance. In sputum testing, sensitivity and specificity were 97.1%/97.1%, 98.2%/97.8%, and 96.5%/97.5% for rifampin resistance and 94.4%/96.4%, 95.4%/98.8%, and 94.9%/97.6% for isoniazid resistance. [28] The WHO will update its guidance on LPAs later in 2016. While this certainly reflects progress, Hain and Nipro launched these assays in 2011; it’s taken five years to optimize and fully evaluate them.

In May 2016, the WHO also recommended and issued guidance on Hain’s MTBDRsl, an LPA capable of detecting resistance to fluoroquinolones and second-line injectables. [29] LPAs can produce results in 24–48 hours, much quicker than the two weeks that liquid culture or two to three months that solid culture take. As such, the MTBDRsl LPA can guide appropriate treatment selection. The announcement of the WHO’s MTBDRsl recommendation accompanied its recommendation of the shortened or “modified Bangladesh” regimen, whose introduction the test can help facilitate, as the shortened (9- to 12-month) regimen is not suitable for fluoroquinolone- or injectable-resistant TB (pre-XDR-TB; see Tuberculosis Treatment). [30] Countries and donors must scale up the introduction of this test and work with Hain to further reduce the price. FIND negotiated a public sector price of €7.50 (approximately $10) per test strip in 138 countries; however, the total cost of running a test (which requires other laboratory supplies) can result in costs of $20–$30. The test equipment itself can cost $8,000–$40,000, depending on its size and whether it automatically reads results or not. [31]

Newer iterations of GeneXpert are moving closer to market. Recent investments may make it more suitable for use in a variety of settings and more sensitive. GeneXpert Omni, a smaller and more rugged single-cartridge version of the test that is dustproof and runs on batteries could be a point-of-care test for TB. The test device’s anticipated cost is $2,895. [32] Cepheid claims that another new product, the GeneXpert Ultra cartridge, is more sensitive than the MTB/RIF, approximating the sensitivity of culture, and has a shorter processing time. [33] FIND is currently validating both the Ultra cartridge and Omni platform: Ultra results are expected at the end of 2016; Omni results have been further delayed. If studies show them to indeed be as promising as the company claims, the WHO will issue recommendations and formulate guidance accordingly. Data on the use of Ultra in smear-negative specimens are expected in 2017, which could inform recommendations on whether Ultra can be used to replace culture. Cartridge prices are expected to remain consistently high at $9.99, as even though MTB/RIF sales volumes have increased, [34] those profits are said (by Cepheid) to have been reinvested in R&D. In 2017, Cepheid plans to release the XDR assay, designed to genotype resistance to isoniazid, fluoroquinolones, and second-line injectables when MTB/RIF (or Ultra) indicates rifampin-resistant TB, though no peer-reviewed data yet exist on this product. [35] A FIND evaluation is expected in 2018.

Molbio’s TrueNAT, an Indian GeneXpert competitor that has been on the market since 2013, is finally being validated by outside parties (FIND and the Indian Council on Medical Research). [36] A recent study compared TrueNAT to Xpert MTB/RIF on 274 patient specimens, using culture as the reference standard. The assays had similar sensitivity on sputum-smear-positive samples: TrueNAT had 99% sensitivity (95% CI: 94.2%–99.95%) versus MTB/RIF’s 100% (95% CI: 96.5–100.0%, respectively). With sputum-smear-negative, culture-positive samples, the sensitivity of the TrueNAT was 86.2% (95% CI: 74.1%–93.4%) as compared with 90.1% (95% CI: 88.7%–94.35%) for Xpert MTB/RIF. [37] The cartridge-based sample preparation extraction tool, TruePrep – which is rugged and portable – costs $7,000, and each assay costs $14; the public sector will receive a further discount. [38]

Other products without the data to back them up continue to be marketed by unscrupulous manufacturers. Epistem’s Genedrive performed dismally in a recent clinical study of 336 participants: sensitivity was 45.4% (95%CI: 35.2%–55.8%) versus 91.8% (95% CI: 84.4%–96.4%) for Xpert MTB/RIF and 77.3% (95% CI: 67.7%–85.2%) for smear microscopy. In smear-negative cases, sensitivity of GeneDrive was 0% (95% CI: 0, 15.4) versus 68.2% (95% CI: 45.1%–86.1%) for Xpert. [39] Yet just after those data were published, Epistem announced full commercial launch of Genedrive TB tests in India, claiming it “enables early detection of TB and antibiotic resistance without need for central laboratory facilities.” [40] The regulatory authority in India should ban the marketing of this test, and private providers should be extremely wary and not waste patients’ time, money, and effort by subjecting them to it (see box).

Extending Quality and Affordability to the Private Sector

Many countries, including 12 of the 22 countries with the highest TB burdens (India, Pakistan, the Philippines, Bangladesh, Afghanistan, Kenya, Uganda, Vietnam, Indonesia, Myanmar, Nigeria, and Cambodia) have large private-sector markets for TB diagnosis and care. [41] Ensuring access to affordable, quality diagnosis is both critical and challenging. The use of unvalidated tests, or using tests off-label, can endanger patients and their communities and, at best, wastes their money and time. Important tests, such as GeneXpert, are not commercially available in the private sector in Burma, Cambodia, Indonesia, Nigeria, Uganda, or Vietnam. [42] But even when good tests are available in the private sector, patients pay dearly, as concessional prices are normally available only to the public sector, and some private practitioners are aiming to maximize profit. For example, in Afghanistan, Bangladesh, India, Kenya, Pakistan, and the Philippines, GeneXpert is available, but the average price charged by private laboratories is $68.73 (range $30.26–$155.44). [43] Private diagnosis without case notification to the public program also impedes getting a true picture of local, national, and global TB epidemiology.

In India, where about half of patients seek TB diagnosis and care in the private sector, the Indian government and other actors have taken steps to mitigate the inappropriate use of TB diagnostics, such as banning the use of serological tests and discouraging the use of Quantiferon TB Gold (a test for latent TB that was being inappropriately marketed and used in India to screen for active TB). [44,45] Efforts are underway to ensure best TB diagnostic practices among private-sector providers in India at an affordable price. The Initiative for Promoting Affordable and Quality TB Tests (IPAQT) offers WHO- recommended TB diagnostics to laboratories who agree to pass on these price reductions to patients by agreeing to a maximum ceiling price, participating in quality assurance programs, and notifying cases to the public program. [46]

To date, IPAQT has involved 116 laboratories across India, with over 250,000 presumptive TB cases tested, and volumes climbing. IPAQT has notified 23,000 cases in five cities under a pilot program and plans to further involve more decentralized laboratories and to streamline notification, as well as to look into expanding into cross-disease diagnostic support, such as including the HIV1 viral load GeneXpert cartridge in the IPAQT framework. Other countries with robust private-sector activity in TB should follow suit. The Clinton Health Access Initiative’s Nigeria team recently conducted an analysis to determine feasibility there.

In Development

Back in the lab, a promising development came from a team at Stanford University that identified a three-gene set indicative of active TB (GBP5, DUSP3, and KLF2) in whole blood across eight data sets containing over 1,000 samples from both adult and pediatric patients in ten countries. This gene signature could accurately separate people with active TB from healthy controls, from people with latent TB, and from people with other diseases. HIV status, bacillus Calmette-Guérin (BCG) vaccination, and drug resistance did not confound expression of the three-gene set. The gene set may be of use in monitoring treatment, as its expression increases with disease severity and decreases with time of treatment, though this must be validated prospectively. Further validation and development are required. [47]

Researchers from the University of Washington and the University of Cape Town have received funding to further study a simple oral swab to test for TB DNA, and it detected TB well in 18 of 20 patients in a proof-of-concept study (90.0% sensitivity compared with GeneXpert MTB/RIF; 95% CI: 66.9%–98.2%). [48, 49]

Several researchers are exploring the value of immune activation markers such as C-reactive protein (CRP) to indicate active TB disease or to identify good responses to TB therapy. One small study pre- and post-treatment in Gambia showed that CRP levels showed the most significant decrease by two months of treatment (P < .0001), whereas two other markers, beta2 microglobulin and neopterin, showed little change by two months but a significant decrease by six months of treatment (P = .0002 and P < .0001, respectively). [50] A larger prospective study identified a seven-marker biosignature including CRP as well as transthyretin, interferon-γ, complement factor H, apolipoprotein-A1, inducible protein 10, and serum amyloid A that identified TB disease in the test set (N = 210) with a sensitivity of 93.8% (95% CI: 84.0%–98.0%) and a specificity of 73.3% (95% CI: 65.2%–80.1%), regardless of HIV infection status. [51] CRP may be of particular interest for pediatric development (see Tuberculosis Diagnostics Research for Children below).

DST research saw some developments. Development of a rapid colorimetric method for detection of resistance to pyrazinamide – one of the most important drugs to treat TB, for which DST development remains challenging due to the enormous number of resistance-associated mutations in the M. tuberculosis pncA gene – using a dye called 5-cyano-2,3-ditolyl tetrazolium chloride (CTC) progressed when initial testing in a small test of 50 isolates showed DST results could be available in four to six days with 97.1% sensitivity and 81.3% specificity, in comparison with liquid culture. [52] Unfortunately, research that could underpin DST development for bedaquiline has not yet been as successful – examining the 12 cases who had developed over fourfold increases in bedaquiline minimum inhibitory concentrations (MICs) in study C209 revealed that all had M. tuberculosis with mutations in the Rv0678 gene, but there was no correlation between MIC change and treatment response, making it unclear what might be a clinically meaningful breakpoint. [53] Developing DST for bedaquiline will be important, as resistance to it has already started to develop. [54]

Recommendations

With new products moving forward and interesting new leads to pursue, we are pleased to report progress with the TB diagnostics pipeline. But the world is still far from ensuring all those with TB get appropriate diagnostics. Both R&D and access need dramatic infusions of funding and political will. In particular:

  • National governments and donors must substantially increase funding for TB programs to allow for best diagnostic practices. This includes the widespread scale-up of NAAT to supplant microscopy, universal DST using liquid culture or LPAs, digital X-ray, and the rapid adoption of LAM testing in areas with high HIV burdens.
  • National governments, donors, and the private sector must invest far more in TB R&D to advance better tests, including those for children. This should include a commitment to rapidly and rigorously evaluating new technologies and to publishing peer-reviewed results. Greater resources are necessary to achieve the requirements set out in Target Product Profiles (TPPs). [55]
  • National governments and donors should work closely with the nonprofit and private sectors to ensure only quality and affordable tests are used. In countries with large proportions of care-seeking in the private sector, access to appropriate diagnostics is extremely limited and can be catastrophically expensive. Good programs such as IPAQT to address this exist and should be expanded and replicated.
  • Developers must commit to timely and rigorous validations of their tests prior to marketing, and health and regulatory authorities and private practitioners should hold them accountable for doing so. Epistem and other companies who market ineffective or as-yet-unproven tests must cease doing so immediately. National governments should ban the import and use of inappropriate tests and enforce those bans. Those working in TB globally should call to task companies such as Epistem that inappropriately market them.

Tuberculosis Diagnostics Research for Children

By Lindsay McKenna

The ability to confirm the presence of tuberculosis (TB) bacteria in the body (microbiological diagnosis) underpins much of the existing technology and paradigm for diagnosing TB in adults, but this approach is problematic for children. An estimated 60 percent of children with TB go undiagnosed: in 2014, national TB programs reported 358,521 cases of TB among children to the World Health Organization (WHO), [56] yet credible models estimate that one million cases of incident TB occur among children each year. [57] TB is also likely a major unrecognized co-morbidity or cause of illness and death among children affected by pneumonia, meningitis, HIV, and malnutrition. [58] Children with the disease have fewer TB bacteria in their bodies (paucibacillary disease), difficulty producing sputum, and high rates of extrapulmonary TB. As a result, diagnosis is often empirical (presumed, rather than confirmed) and based on a combination of clinical and epidemiologic information.

The gold standard for diagnosing TB, microbiological confirmation using culture, is only obtained in 15–20% of children with clinically diagnosed TB disease. [59] Compared with culture, Xpert MTB/RIF has 62% pooled sensitivity when performed on induced or expectorated sputum (36% more sensitive than smear microscopy) and 66% pooled sensitivity when performed on gastric aspirate or lavage (44% more sensitive than smear microscopy). [60] Xpert MTB/RIF sensitivity in culture-negative children clinically diagnosed with TB is just 2% for induced or expectorated sputum. [61] The WHO recommends Xpert MTB/RIF as the initial diagnostic test in children suspected of having multidrug-resistant TB (MDR-TB) or HIV-associated TB or, where resources allow, as the initial diagnostic test in all children suspected of having TB. [62] These recommendations apply to both pulmonary and extrapulmonary specimens, with the exception of stool, urine, and blood, given the lack of data for the utility of Xpert MTB/RIF for these specimen types. [63] It is important to note that a majority of studies evaluating Xpert MTB/RIF’s performance have been conducted at higher-level health facilities, where it is likely that sicker children with higher rates of smear-positive TB are present for evaluation. How Xpert MTB/RIF performs among children in an outpatient setting, and with different levels of TB disease severity, has not been well studied. That said, while Xpert MTB/RIF is superior to smear microscopy and helps provide rapid confirmation of disease, it should not be used as a rule-out test for TB in children: clinical evaluation remains important in diagnosing TB in children.

Research is ongoing to determine the most feasible and sensitive combinations of tests and specimen types (including urine and stool) for diagnosing TB in children [64]; to compare the performance of smear, culture, and Xpert at baseline and during treatment; and to optimize specimen sample collection and processing to improve diagnostic yields in children. In addition to efforts to optimize existing tools, new technologies in the pipeline might also improve our ability to detect TB in children in the future. The WHO recently recommended molecular line probe assays (LPAs) for the rapid detection of resistance to second-line TB drugs, including fluoroquinolones and injectable agents in children with confirmed rifampin-resistant TB or MDR-TB, based on extrapolation from data in adults. [65] Xpert MTB/RIF Ultra, currently under evaluation in adults, is expected to have increased sensitivity and ability to detect paucibacillary TB disease, which is common in young children. Xpert XDR is expected to improve our ability to quickly diagnose resistance to first- and second-line TB drugs. These advances, though extremely important, are incremental. To radically improve diagnosis of all forms of pediatric TB, a rapid biomarker-based test that does not rely on sputum and can be used at the point of care is necessary. [66]

The discovery and validation of biomarkers for TB diagnosis and treatment monitoring in children is an urgent research priority. A blueprint for pediatric TB biomarker identification and development, resulting from a 2014 U.S. National Institutes of Health (NIH)-convened workshop, is a call to action. The blueprint identifies critical research needs, including enhancing the detection of pathogen biomarkers and identifying host biomarkers, and calls for collaboration to advance the field. [67] Efforts by an NIH-organized working group are underway to harmonize pediatric biorepositories (specimen collection methods and clinical data collection) to optimize their use for the future discovery and development of TB biomarkers in children.

Compared with adults, children have increased risk of progression from infection to active disease. [68] While it is possible to diagnose TB infection in children using biomarker-based tuberculin skin testing (TST) and interferon-gamma release assays (IGRAs), these tests have shortcomings, including their inability to differentiate between TB infection and disease, cross-reactivity with other mycobacteria, particularly for TST, and increased false-negative tests among immune-compromised children. The ability to differentiate between infection and disease and to identify children at increased risk of progression to active disease would improve feasibility and make more efficient the targeted provision of preventive therapy to child contacts of TB patients in high-burden settings.

Efforts to identify and validate biomarkers of TB disease and risk of progression in adults and children are ongoing and, if proven, will greatly improve the reliability and ease of TB diagnosis. However, drug-susceptibility tests rely on microbiological samples. Where a microbiological sample cannot be obtained, biomarkers that enable treatment monitoring in children could provide an interesting opportunity to improve access to appropriate treatment. Select pediatric biomarker research highlights are presented below.

Pediatric Biomarker Research Highlights

LAM

Because both children and people with HIV tend to have higher rates of extrapulmonary TB, it was expected that the lateral flow urine LAM assay (currently recommended in HIV-positive adults who have low CD4 counts or are seriously ill) would work well in children, too. However, the LAM test demonstrated poor sensitivity (48.3%) and specificity (60.8%) compared with culture in HIV-positive and HIV-negative children with TB. [69] The WHO recommendation for LAM in people with CD4 counts <100 (or with advanced HIV disease) does extend to children, based on the generalization of data from adults, while acknowledging very limited data in children.

C-reactive protein

C-reactive protein (CRP), a nonspecific marker of inflammation detectable in blood and measurable with existing assays at the point of care, has shown potential for screening for TB disease and indicating response to TB treatment in adults (see In Development above). In one study, CRP demonstrated 98% sensitivity and 59% specificity for TB among South African adults with smear-negative, culture-positive TB with or without HIV. [70] A study to identify the expression patterns of biomarkers in the plasma of HIV-negative children in India with pulmonary and extrapulmonary TB compared with healthy controls found that children with active TB showed significantly elevated levels of CRP. These findings are not surprising, as a detectable difference in inflammation can be expected when comparing healthy and sick children. As such, CRP should be evaluated as a marker of active TB and for use in TB diagnostic algorithms in larger pediatric cohorts, inclusive of children with latent TB and other pulmonary infections and HIV. [71]

TAM-TB

Encouragingly, a novel T-cell activation marker-tuberculosis assay (TAM-TB) demonstrated 83.3% sensitivity and 96.8% specificity among children with TB symptoms compared with culture. The pediatric cohort (N = 113) in this prospective proof-of-concept study included HIV-positive and HIV-negative children 6 months to16 years old. The combined use of the TAM-TB assay and Xpert MTB/RIF demonstrated 94% sensitivity compared with culture. TAM-TB is a rapid blood-based test with the potential to improve the detection of active TB in children; further refinement and testing, especially in HIV-positive children with low CD4 cell counts, are necessary. [72]

RNA expression signatures

A genome-wide analysis of RNA expression in blood among children undergoing evaluation for TB, including those with HIV, identified a 51-transcript signature capable of distinguishing TB from other diseases and from latent TB infection. The 51-transcript signature demonstrated 82.9% sensitivity and 83.6% specificity for culture-confirmed TB (Xpert MTB/RIF sensitivity was 54.3%). For culture-negative TB where children are deemed to have highly probable, probable, or possible TB, the 51-transcript signature had an estimated sensitivity of 62.5–82.3%, 42.1–80.8%, and 35.3–79.6%, respectively (estimated sensitivity for Xpert MTB/RIF was 25%–35.7%, 5.3%–13.3%, and 0%, respectively). The 51-transcript signature distinguished TB from latent infection with a sensitivity of 94% and a specificity of 100%. The 51-transcript signature identified higher proportions of culture-confirmed and culture-negative cases of TB than Xpert MTB/RIF; however, innovation is needed to translate transcriptional signatures into diagnostic tools for resource-poor settings – current methods used to detect RNA transcripts are complex and costly. [73]

Gene expression signatures

A multicohort analysis of data sets available in two public gene expression microarray repositories identified a three-gene signature (GBP5, DUSP3, and KLF2) capable of diagnosing active TB in adults and children, irrespective of bacillus Calmette-Guérin (BCG) vaccination or HIV status. The TB score derived from the three-gene signature demonstrated 85 percent sensitivity and 93% specificity in a cohort made up of both healthy people and those with active TB, 80% sensitivity and 86% specificity in a cohort made up of people with latent or active TB, and 81% sensitivity and 74% specificity in a cohort made up of people with other diseases or active TB. The TB score showed a significant decreasing trend with progression of treatment, suggesting its potential as a biomarker of clinical response to treatment. However, TB scores in children with culture-negative TB were significantly lower than those in children with culture-positive TB. The TB score demonstrated 86% sensitivity and specificity for latent TB versus culture-positive active TB in children. [74] A blood-based test offers much advantage over what currently exists, but the ability to detect TB in culture-negative children is extremely important, limiting the potential utility of this three-gene signature for children with culture-negative TB, who account for 80% of children with TB.

Recommendations for Pediatric TB Diagnostics

Much work remains to develop novel diagnostic technologies to accurately detect TB infection and disease; to predict disease progression in healthy, infected children; and to monitor treatment in children. The few tests that have been validated and recommended for use in children, including Xpert MTB/RIF, are sub-optimal and underutilized, partly due to difficulties with specimen collection. At the same time, efforts to identify children at risk for TB – especially within maternal and child health programs, where sick children often first present for care – and referral systems and decentralized capacity to diagnose childhood TB, clinically or with available tools, are urgently needed. We also need:

  • To validate tests in adults and children in parallel to expedite access to improved diagnostic technologies for children. These evaluations should include a variety of sample types in children with and without HIV and should assess age-related performance;
  • Increased investments in research to discover and validate biomarkers and innovation to translate these biomarkers into simple and affordable tests that can rapidly and accurately diagnose TB, monitor treatment, and predict disease progression in children. In 2014, less than $2.3 million and $2.8 million was spent globally on research and development for pediatric TB diagnostics and basic science, respectively; [75]
  • To establish and support harmonized and collaborative pediatric biorepositories important for biomarker discovery and development;
  • To support and create networks of sites that support field evaluation of new diagnostics and can pool data to more rapidly demonstrate the impact of new tools;
  • To scale up and decentralize the use of existing technologies and strategies to diagnose pediatric TB infection and disease, especially within maternal and child health programs; and
  • To train health care workers to improve their ability and confidence to clinically diagnose children with TB when tests are unavailable or come back negative.

Acknowledgments

Erica Lessem thanks Mark Harrington for his editing, Rachel Schiff for her support for references, and the contributions of many researchers, developers, and technical experts that made this chapter possible. Lindsay McKenna sends special thanks to Dr. Anne Detjen, Dr. Jennifer Cohn, and Dr. Devasena Gnanashanmugam for their thoughtful reviews and comments on the pediatric section.

References

  1. World Health Organization. Global tuberculosis report 2015. Geneva: World Health Organization; 2015. Available from:
    http://www.who.int/tb/publications/global_report/en/. (Accessed 2016 May 17)
  2. Steingart KR, Ng V, Henry M, et al. Sputum processing methods to improve the sensitivity of smear microscopy for tuberculosis: a systematic review. Lancet Infect Dis [Internet]. 2006 Oct;6(10):664–74. doi: 10.1016/S1473-3099(06)70602-8.
  3. World Health Organization. WHO endorses new rapid tuberculosis test. Geneva: World Health Organization; 2010 December 8. Available from:
    http://www.who.int/mediacentre/news/releases/2010/tb_test_20101208/en/. (Accessed 2016 May 17)
  4. World Health Organization. Global tuberculosis report 2015.
  5. Frick, M. 2015 report on tuberculosis research funding trends, 2005–2014: a decade of data. New York: Treatment Action Group; 2015. Available from:
    http://www.treatmentactiongroup.org/tbrd2015. (Accessed 2016 May 15)
  6. Rocchetti TT, Silbert S, Gostnell A, Kubasek C, Widen R. Validation of a multiplex real-time PCR to detect Mycobacterium spp., Mycobacterium tuberculosis complex and Mycobacterium avium complex directly from clinical samples using the BD MAX™ open system. J Clin Microbiol. 2016 Mar 23. doi: 10.1128/JCM.00241-16. [Epub ahead of print]
  7. Shenai S, Armstrong DT, Valli E, et al. Analytical and clinical evaluation of the Epistem Genedrive assay for detection of Mycobacterium tuberculosis. J Clin Microbiol [Internet]. 2016 Apr;54(4):1051–7. doi: 10.1128/JCM.02847-15. Epistem Holdings PLC (Press Release). Epistem begins full commercial launch for Genedrive® tuberculosis tests in India. 2016 April 18. Available from:
    http://www.publicnow.com/view/F٢٩A٩F٨DA٩E٠AEA8714A469AE1DA1B25BBE01F0B?2016-04-18-11:00:36+01:00-xxx4802. (Accessed 2016 May 18)
  8. Nathavitharana RR, Hillemann D, Schumacher SG, et al. A multi-center non-inferiority evaluation of Hain GenoType MTBDRplus Version 2 and Nipro NTM+MDRTB line probe assays for the diagnosis of rifampin and isoniazid resistance. J Clin Microbiol. 2016 Apr 13. doi: 10.1128/JCM.00251-16. [Epub ahead of print]
  9. World Health Organization. The use of molecular line probe assays for the detection of resistance to second-line antituberculosis drugs: policy guidance. Geneva: World Health Organization; 2016. Available from:
    http://www.who.int/tb/WHOPolicyStatementSLLPA.pdf?ua=1. (Accessed 2016 May 18)
  10. Pang Y, Dong H, Tan Y, et al. Rapid diagnosis of MDR and XDR tuberculosis with the MeltPro TB assay in China. Sci Rep. 2016 May 6;6:25330. doi: 10.1038/srep25330.
  11. Nathavitharana RR, et al. Evaluation of line probe assays.
  12. Wang SF, Ou XC, Li Q, Zheng HW, Wang YF, Zhao YL. The Abbott RealTime MTB assay and the Cepheid GeneXpert assay show comparable performance for the detection of Mycobacterium tuberculosis in sputum specimens. Int J Infect Dis. 2016 Apr;45:78–80. doi: 10.1016/j.ijid.2016.02.024.
  13. World Health Organization. The use of lateral flow urine lipoarabinomannan assay (LF-LAM) for the diagnosis and screening of active tuberculosis in people living with HIV. Geneva: World Health Organization; 2015. Available from:
    http://www.who.int/tb/areas-of- work/laboratory/policy_statement_lam_web.pdf. (Accessed 2016 March 1)
  14. Mhimbira FA, Bholla M, Sasamalo M, et al. Detection of Mycobacterium tuberculosis by EasyNAT diagnostic kit in sputum samples from Tanzania. J Clin Microbiol. 2015 Apr;53(4):1342–4. doi: 10.1128/JCM.03037-14.
  15. Jubulis J, Dionne K, Osterhout G, et al. Mycobacterium tuberculosis resistance in pulmonary TB patients in Cameroon: a phenotypic susceptibility assay. Int J Tuberc Lung Dis. 2015 Jul;19(7):823–7. doi: 10.5588/ijtld.14.0527.
  16. World Health Organization. Chest radiography for tuberculosis detection. Geneva: World Health Organization, 2016. Available from:
    http://www.who.int/tb/publications/Radiography_TB_factsheet.pdf. (Accessed 2016 May 30)
  17. World Health Organization. Use of lipoarabinomannan assay.
  18. Shah M, Hanrahan C, Wang ZY, et al. Lateral flow urine lipoarabinomannan assay for detecting active tuberculosis in HIV-positive adults. Cochrane Database Syst Rev. 2016 May 10;5:CD011420. doi: 10.1002/14651858.CD011420.pub2.
  19. Dheda K, Barry CE 3rd, Maartens G. Tuberculosis. Lancet. 2016 Mar 19;387(10024):1211–26. doi: 10.1016/S0140-6736(15)00151-8. Erratum in: Lancet. 2016 Mar 19;387(10024):1162.
  20. Peter JG, Zijenah LS, Chanda D, et al. Effect on mortality of point-of-care, urine-based lipoarabinomannan testing to guide tuberculosis treatment initiation in HIV-positive hospital inpatients: a pragmatic, parallel-group, multicountry, open-label, randomised controlled trial. Lancet. 2016 Mar 19;387(10024):1187–97. doi: 10.1016/S0140-6736(15)01092-2.
  21. Ibid.
  22. AIDS & Rights Alliance for Southern Africa, Asia Pacific network of people living with HIV, Delhi Network of Positive People, Global TB Community Advisory Board, International Treatment Preparedness Coalition-South Asia, Nagaland Users Network, Section 27, SWIFT Response Project, TB PROOF, Treatment Action Campaign, Treatment Action Group. RE: Urgent need to roll out LAM testing for TB in some people with HIV in South Africa. 2016 March 10. Available from:
    http://www.tbonline.info/media/uploads/documents/lam_rsa_letter_from_community.pdf. (Accessed 2016 May 18)
  23. Khan N and Cortez M. “Alere Says It Rejected Abbott’s Request to End Acquisition.” Bloomberg [Internet]. 2016 April 29. Available from:
    http://www.bloomberg.com/news/articles/2016-04-29/alere-says-it-rejected-abbott-request-to-end-5-8-billion-deal. (Accessed 2016 May 29)
  24. FIND. FIND awarded GHIT grant for support on new TB test. Geneva: FIND; 2016 April 22. Available from:
    http://www.finddx.org/news/find-awarded-ghit-grant-support-new-tb-test/. (Accessed 2016 May 18)
  25. Global Health Innovative Technology Fund. Highly sensitive POC TB-LAM rapid diagnostic test. Tokyo: Global Health Innovative Technology Fund; 2015. Available from:
    https://www.ghitfund.org/impact/portfolio/awpdetail/detail/78. (Accessed 2016 May 18)
  26. Global Health Innovative Technology Fund. Partners. Tokyo: Global Health Innovative Technology Fund. Available from:
    https://www.ghitfund.org/about/orgnization/partners. (Accessed 2016 May 29)
  27. World Health Organization. Global TB programme. Geneva: World Health Organization; 2015 November 6. Available from:
    http://us8.campaign-archive2.com/?u=f093a7c38a3780cd9504f8d9d&id=7d9334b3d5&e=56695a1113. (Accessed 2016 May 18)
  28. Nathavitharana RR, et al. Evaluation of line probe assays.
  29. World Health Organization. Molecular line probe assays.
  30. World Health Organization. Rapid diagnostic test and shorter, cheaper treatment signal new hope for multi-drug resistant tuberculosis patients. Geneva: World Health Organization; 2016 May 12. Available from:
    http://www.who.int/mediacentre/news/releases/2016/multidrug-resistant-tuberculosis/en/. (Accessed 2016 May 18)
  31. World Health Organization. Tuberculosis diagnostics: molecular line probe assay for the detection of resistance to second-line anti-TB drugs (SL-LPA). Geneva: World Health Organization; 2016. Available from:
    http://www.who.int/tb/Factsheet_SLLPAfinal.pdf?ua=1. (Accessed 2016 May 18)
  32. UNITAID and the World Health Organization. 2015 tuberculosis diagnostics technology and market landscape, 4th ed. Geneva: World Health Organization; 2015. Available from:
    http://unitaid.org/images/marketdynamics/publications/Tuberculosis_diagnostics_technology_and_market_landscape_4th_edition_Oct_2015.pdf. (Accessed 2016 May 18)
  33. FIND. FIND launches multi-centre study of Xpert Ultra for TB & drug resistance. Geneva: FIND; 2016 March 24. Available from:
    http://www.finddx.org/events/find-ultra-study/. (Accessed 2016 May 18)
  34. World Health Organization. Global tuberculosis report 2015.
  35. UNITAID and the World Health Organization. 2015 tuberculosis diagnostics landscape.
  36. Nair C. Overview of technological platforms: Molbio. Oral presentation at: Critical Path to TB Drug Regimens Workshop; 2016 April 4–7; Washington, D.C.
  37. Nikam C, Kazi M, Nair C, et al. Evaluation of the Indian TrueNAT micro RT-PCR device with GeneXpert for case detection of pulmonary tuberculosis. Int J Mycobacteriol. 2014 Sep;3(3):205–10. doi: 10.1016/j.ijmyco.2014.04.003.
  38. Ibid.
  39. Shenai S, et al. Evaluation of Epistem Genedrive assay.
  40. Epistem Holdings PLC. Genedrive® tuberculosis tests in India.
  41. Puri L, Oghor C, Denkinger CM, Pai M. Xpert MTB/RIF for tuberculosis testing: access and price in highly privatised health markets. Lancet Glob Health. 2016 Feb;4(2):e94–5. doi: 10.1016/S2214-109X(15)00269-7.
  42. Ibid.
  43. Ibid.
  44. Specter M. “How banning a TB test will save lives in India.” New Yorker [Internet]. 2012 June 19. Available from:
    http://www.newyorker.com/news/news-desk/how-banning-a-tb-test-will-save-lives-in-india. (Accessed 2016 May 18)
  45. Pai M, DAS J. Management of tuberculosis in India: time for a deeper dive into quality. Natl Med J India [Internet]. 2013 Mar–Apr [cited 2016 May 18];26(2):65–8. Available from:
    http://www.nmji.in/advt/26-2-Editorial.pdf.
  46. Puri L, et al. Xpert MTB/RIF for tuberculosis testing.
  47. Sweeney TE, Braviak L, Tato CM, Khatri P. Genome-wide expression for diagnosis of pulmonary tuberculosis: a multicohort analysis. Lancet Respir Med. 2016 Mar;4(3):213–24. doi: 10.1016/S2213-2600(16)00048-5.
  48. Sharpe E. “Oral swabs for TB diagnosis undergo testing.” Newsbeat: UW Health Sciences [Internet]. 2016 April 6. Available from:
    http://hsnewsbeat.uw.edu/story/oral-swabs-tb-diagnosis-undergo-testing. (Accessed 2016 May 19)
  49. Wood RC, Luabeya AK, Weigel KM, et al. Detection of Mycobacterium tuberculosis DNA on the oral mucosa of tuberculosis patients. Sci Rep. 2015 Mar 2;5:8668. doi: 10.1038/srep08668.
  50. Mendy J, Togun T, Owolabi O, Donkor S, Ota MO, Sutherland JS. C-reactive protein, Neopterin and Beta2 microglobulin levels pre and post TB treatment in The Gambia. BMC Infect Dis. 2016 Mar 8;16(1):115. doi: 10.1186/s12879-016-1447-9.
  51. Chegou NN, Sutherland JS, Malherbe S, et al; AE-TBC consortium. Diagnostic performance of a seven-marker serum protein biosignature for the diagnosis of active TB disease in African primary healthcare clinic attendees with signs and symptoms suggestive of TB. Thorax. 2016 May 4. doi: 10.1136/thoraxjnl-2015-207999. [Epub ahead of print]
  52. Li H, Zhou LP, Luo J, Yu JP, Yang H, Wei HP. Rapid colorimetric pyrazinamide susceptibility testing of Mycobacterium tuberculosis. Int J Tuberc Lung Dis. 2016 Apr;20(4):462–7. doi: 10.5588/ijtld.15.0745.
  53. Andries K, Coeck N, Villellas C, et al. Efflux-pump-mediated acquired resistance to bedaquiline in a clinical trial (Abstract 341-04). Paper presented at: 46th Union World Conference on Lung Health; 2015 December 2–6; Cape Town, South Africa.
  54. Bloemberg GV, Keller PM, Stucki D, et al. Acquired resistance to bedaquiline and delamanid in therapy for tuberculosis. N Engl J Med. 2015 November 12;373(25):1986–1988. doi: 10.1056/NEJMc1505196.
  55. Denkinger C, Kik S, Casenghi M. Meeting report. High-priority target product profiles for new tuberculosis diagnostics: report of a consensus meeting. 28–29 April 2014, Geneva, Switzerland. Geneva: World Health Organization; 2014. Available from:
    http://apps.who.int/iris/bitstream/10665/135617/1/WHO_HTM_TB_2014.18_eng.pdf.
  56. World Health Organization. Global tuberculosis report 2015. Geneva: World Health Organization; 2015. Available from:
    http://www.who.int/tb/publications/global_report/en/.
  57. Dodd PJ, Gardiner E, Coghlan R, Seddon JA. Burden of childhood tuberculosis in 22 high-burden countries: a mathematical modeling study. Lancet Glob Health. 2014 Jul 9; 2(8):e453–9. doi: 10.1016/S2214-109X(14)70245-1.
  58. Graham SM, Sismanidis C, Menzies HJ, et al. Importance of tuberculosis control to address child survival. Lancet. 2014 May 3;383(9928):1605–7. doi: 10.1016/S0140-6736(14)60420-7..
  59. Marais BJ, Hesseling AC, Gie RP, et al. The bacteriologic yield in children with intrathoracic tuberculosis. Clin Infect Dis 2006; 42: e69–71.
  60. Detjen AK, DiNardo AR, Leyden J, et al. Xpert MTB/RIF assay for the diagnosis of pulmonary tuberculosis in children: a systematic review and meta-analysis. Lancet Respir Med. Jun 2015;3(6):451–61. doi: 10.1016/S2213-2600(15)00095-8.
  61. Ibid.
  62. World Health Organization. Policy Update: Xpert MTB/RIF assay for the diagnosis of pulmonary and extrapulmonary TB in adults and children. Geneva: World Health Organization; 2013. Available from:
    http://www.who.int/tb/laboratory/xpert_launchupdate/en/.
  63. Ibid.
  64. Marcy O, Ung V, Goyet S, et al. Performance of Xpert MTB/RIF and alternative specimen collection methods for diagnosis of tuberculosis in HIV-infected children. Clin Infect Dis. 2016 May 1;62(9):1161–8. doi: 10.1093/cid/ciw036
  65. World Health Organization. Policy Guidance: The use of molecular line probe assays for the detection of resistance to second-line anti-tuberculosis drugs. Geneva: World Health Organization; 2016. Available from:
    http://www.who.int/tb/areas-of-work/laboratory/policy_statements/en/.
  66. World Health Organization. High-priority target product profiles for new tuberculosis diagnostics: report of a consensus meeting. Geneva: World Health Organization; 2014. Available from:
    http://www.who.int/tb/publications/tpp_report/en/.
  67. Nicol MP, Gnanashanmugam D, Browning R, et al. A blueprint to address research gaps in the development of biomarkers for pediatric tuberculosis. CID. 2015;61(3):S164–72.
  68. Newton SM, Brent AJ, Anderson S. Paediatric tuberculosis. Lancet Infect Dis. 2008; 8:498–510. doi: 10.1016/S1473-3099(08)70182-8.
  69. Nicol MP, Allen V, Workman L, et al. Urine lipoarabinomannan testing for diagnosis of pulmonary tuberculosis in children: a prospective study. Lancet Glob Health. 2014 Apr 8; 2:e278–84. doi: 10.1016/S2214-109X(14)70195-0.
  70. Wilson D, Badri M, Maartens G. Performance of serum C-reactive protein as a screening test for smear-negative tuberculosis in an ambulatory high HIV prevalence population. PLoS ONE. Jan 2011;6(1):e15248. doi:10.1371/journal.pone.0015248.
  71. Kumar PN, Anuradha R, Andrade BB, et al. Circulating biomarkers of pulmonary and extrapulmonary tuberculosis in children. Clin Vaccine Immunol. May 2013;20(5):704–711.
  72. Portevin D, Moukambi F, Clowes P, et al. Assessment of the novel T-cell activation marker-tuberculosis assay for diagnosis of active tuberculosis in children: a prospective proof-of-concept study. Lancet Infect Dis. 2014 Sep 1;14:931–38. doi: 10.1016/S1473-3099(14)70884-9.
  73. Anderson ST, Kaforou M, Phil M, et al. Diagnosis of childhood tuberculosis and host RNA expression in Africa. N Engl J Med. 2014 May 1; 370(18):1712–23. doi: 10.1056/NEJMoa1303657.
  74. Sweeney TE, Braviak L, Tato CM, Khatri P. Genome-wide expression for diagnosis of pulmonary tuberculosis.
  75. Frick M. 2015 report on tuberculosis research funding.

Links to other websites are current at date of posting but not maintained.