The Diagnosis of Tuberculosis

ESPID REPORTS AND REVIEWS

CONTENTS The Diagnosis of Tuberculosis

EDITORIAL BOARD Co-Editors: Delane Shingadia and Irja Lutsar Board Members David Burgner (Melbourne, Australia) Nicol Ritz (Basel, Switzerland) Tobias Tenenbaum (Mannhein, Germany) Luisa Galli (Rome, Italy) Ira Shah (Mumbai, India) Marc Terbruegge (Southampton, UK) Christiana Nascimento-Carvalho Matthew Snape (Oxford, UK) Marceline van Furth (Amsterdam, (Bahia, Brazil) George Syrogiannopoulos The Netherlands) Ville Peltola (Turku, Finland) (Larissa, Greece) Anne Vergison (Brussels, Belgium)

The Diagnosis of Tuberculosis

Delane Shingadia, MPH, MRCP, FRCPCH

Abstract: Childhood tuberculosis accounts for a ESTABLISHED DIAGNOSTIC in place. Nasopharyngeal aspiration (NPA) significant proportion of the global tuberculosis METHODS has also been used to obtain respiratory sam- disease burden. However, tuberculosis in children ples, as the passage of a nasal cannula may is difficult to diagnose, because disease tends to be Microscopy and Culture elicit a cough reflex. The culture yield from paucibacillary and sputum samples are often not Microscopic examination of respira- NPA (19/64; 30%) was similar to that of easy to obtain. The diagnosis of tuberculosis in tory samples for acid-fast bacilli using the gastric aspirates (24/64; 38%) among Peru- 8 children is traditionally based on chest radiogra- Ziehl-Neelsen and fluorochrome stains, such vian children. However, subsequent studies phy, tuberculin skin testing, and mycobacterial as the auramine and rhodamine, have been have shown relatively poor yields from staining/culture from appropriate samples. Newer the standard and rapid diagnostic tools for NPA samples compared with gastric aspi- 1,2 9,10 diagnostic strategies have included improved bacte- tuberculosis (TB) diagnosis. Recent ad- rate. Since young children tend to swal- vances in light-emitting diode (LED) tech- riologic and molecular methods, as well as new low their sputum rather than expectorate it, nology have widened the applicability of methods for sample collection from children. Re- mycobacterial culture of stool has been fluorescent microscopy.3 In adults and older considered as an indirect way of analysis cently, immune-based diagnostics, such as the inter- children, sputum samples are often obtained of respiratory secretions. However, studies feron-gamma release assays, have been introduced with sensitivity from 34% to 80%.4 In in children have shown relatively poor re- for clinical use. These tests do not offer substantial younger children, who are unable to produce covery from stool, making this an insensi- improvements in sensitivity over tuberculin skin test- sputum samples, alternative methods of ob- tive method for mycobacterial culture. Fur- ing for the diagnosis of active disease but may be taining respiratory samples, such as gastric thermore, the major drawback of stool useful in excluding false-positive tuberculin skin aspirates, are often used. However, micro- culture is the need for stringent decontam- tests. Further research is needed to develop better scopic yields may be Ͻ20% in children with ination procedures to prevent overgrowth diagnostic tests for tuberculosis in children. probable TB.5 The detection rates on micros- of normal bowel flora, which may also kill Key Words: tuberculosis, diagnosis, child copy from other extrapulmonary samples, or inhibit growth of mycobacteria further such as cerebrospinal fluid, are even lower reducing the sensitivity.11 (Pediatr Infect Dis J 2012;31: 302–305) because of the paucibacillary nature of dis- Another novel method of sampling ease at these sites. swallowed respiratory secretions is the string Mycobacterial culture of respiratory test. The string test was developed for the samples has provided a more useful method diagnosis of intestinal parasites such as giar- of diagnosis in children with suspected pul- diasis. This test involves swallowing a gela- monary TB. Three consecutive daily morn- tin capsule containing a coiled nylon string, ing gastric aspirates yield M. tuberculosis in which unravels as the capsule descends into 30% to 50% of cases and may be as high as the stomach. After 4 hours, the string is From the Department of Infectious Diseases, Great 6 Ormond Street Hospital, Great Ormond Street, 70% in infants. Recently, sputum induction withdrawn and cultured for mycobacteria. London, United Kingdom. using nebulized hypertonic (3%–5%) saline Although this test appears to have a better The author has no funding or conflicts of interest to has been used safely and effectively in young culture yield than sputum induction in adults disclose. children. The culture yield from a single with HIV infection (9% vs. 5%), it has not Address for correspondence: Delane Shingadia, MPH, MRCP, FRCPCH, Department of Infectious Dis- induced sputum sample has been shown to been studied in children other than a feasi- eases, Great Ormond Street Hospital, Great Or- be equivalent to that of 3 cumulative gastric bility study where it appears to have been mond Street, London WC1N 3JH, United King- lavage samples.7 There are, however, some well tolerated.12,13 Furthermore, it may be dom. E-mail: [email protected]. concerns regarding the risk of nosocomial of limited use in younger children who will Copyright © 2012 by Lippincott Williams & Wilkins ISSN: 0891-3668/12/3103-0302 transmission following sputum induction if be unable to swallow the capsule in the DOI: 10.1097/INF.0b013e318249f26d adequate infection control procedures are not first place. The ESPID Reports and Reviews of Pediatric Infectious Diseases series topics, authors and contents are chosen and approved independently by the Editorial Board of ESPID.

302 | www.pidj.com The Pediatric Infectious Disease Journal • Volume 31, Number 3, March 2012 The Pediatric Infectious Disease Journal • Volume 31, Number 3, March 2012 ESPID Reports and Reviews

The culture yield from other body flu- cination and Ͼ15 mm for those with BCG and require a laboratory infrastructure simi- ids or tissues from children with extrapulmo- vaccination history.22 The US guidelines use lar to that required for standard cultures. nary TB is usually Ͻ50%.14,15 In children a risk categorization based on epidemiologic There is currently only 1 commercially avail- with palpable peripheral lymphadenopathy, and clinical factors: Ͼ5 mm (close contacts, able kit, the FASTPlaque-TB (Biotec Labo- fine needle aspiration and culture is a very TB disease, immunosuppression), Ͼ10 mm ratories, Ipswich, Suffolk, United Kingdom) useful adjunct to culture of respiratory (increased risk of disseminated disease or assay, which can be used directly on sputum specimens and may have a higher yield increased exposure to TB disease), and Ͼ15 samples for diagnosis. A variant of this than such culture (sensitivity 60.8% vs. mm (children Ͼ4 years of age with no risk assay, the FASTPlaque-Response kit is de- 39.2%, respectively).16 factors).23 signed to detect rifampicin resistance in Recently, automated liquid culture TST is prone to both false-negative sputum specimens, which has been used as systems with continuous monitoring for my- and false-positive results. Up to 10% to 15% a reliable marker for multidrug-resistant cobacterial growth (such as BD BACTEC of otherwise immunocompetent children TB. However, no information exists on the MGIT system or Biomerrieux BacT/ALERT with culture-documented TB do not initially utility of these tests in the diagnosis of 3D) have been a significant advance over show TST reactivity.14 Host factors, such as childhood TB. traditional solid culture (Lowenstein-Jensen young age, poor nutrition, immunosuppres- The potential of a gas sensor array media). In adult studies, these tests offer sion, other viral infections (such as measles, electronic “nose” (E-nose) to detect different improved sensitivity (88% vs. 76%) and re- varicella, and influenza), recent TB infec- Mycobacterium species in the headspaces of duced detection time (13.2 vs. 25.8 days) tion, and disseminated TB diseases, can fur- cultures and sputum samples is another in- compared with solid media.17 It is likely that ther decrease TST reactivity. False-positive novative approach that is currently under these findings can be extrapolated to children TST results may also occur following BCG development. The array uses 14 sensors to with TB, although there is a paucity of pe- vaccination and exposure to environmental profile a “smell” by assessing the change in diatric data. Despite their higher cost and the nontuberculous mycobacteria.24 Skin reac- each sensor’s electrical properties when ex- laboratory infrastructure required, liquid cul- tivity can be boosted, probably through an- posed to a specific odor mixture. In a recent ture has been recommended for all culture in tigenic stimulation, by serial testing with study using sputum samples from adult TB resource-rich settings.18 TST in many children and adults who re- patients and non-TB patients, the E-nose had Newer culture-based methods, such as ceived BCG.25 sensitivity of 68% and specificity of 69%.29 TK medium, use multiple dye indicators for Further research is still required to improve the early detection of mycobacterial growth Radiology sensitivity and specificity as well as its po- with the naked eye. The simple colorimetric Chest radiography is used widely for tential in the diagnosis of childhood TB. system reduces turnaround times, but their the detection of pulmonary TB, including accuracy and robustness in field conditions detection of hilar lymphadenopathy, lung pa- Molecular Diagnostics and Rapid have not been reported. The Microscopic renchymal changes, and miliary TB. Cavi- Resistance Testing Observation Drug Susceptibility assay uses tarydisease is uncommon in younger chil- Nucleic acid amplification tests an inverted light microscope to rapidly detect dren but is often seen in adolescents, who (NAATs) for the detection of mycobacterial mycobacterial growth in liquid growth me- may develop adult-type postprimary dis- DNA or RNA are increasingly being devel- dia. It is an inexpensive method that has ease.26 Computed tomography imaging has oped for clinical use. These tests are theoret- demonstrated excellent performance under been useful in demonstrating early pulmo- ically highly sensitive, able to detect very field conditions (in both adults and children), nary disease, such as cavitation, and intratho- low copy numbers of nucleic acid, rapid, not being more sensitive than standard liquid racic hilar lymphadenopathy.27 Central ner- requiring biosafety level 3 facilities and are broth or solid culture media systems.9,19,20 vous system disease, such as TB meningitis relatively easy to automate. Commercial The test is not widely available at present. or tuberculoma, may also be identified on NAATs have been extensively evaluated in computed tomography imaging, where men- adults showing high specificity (85%– Tuberculin Skin Test ingeal enhancement may be seen with con- 98%), high sensitivity for smear-positive A positive tuberculin skin test (TST) trast. Magnetic resonance imaging has been TB (pooled estimate 96%) but poorer sensi- reaction has been used as a hallmark of found to be useful for musculoskeletal TB, tivity for smear-negative TB (pooled esti- infection with M. tuberculosis, occurring particularly involving bones and joints.28 mate 66%).11 Sensitivity estimates are gen- within 3 to 6 weeks, but occasionally up to 3 erally also lower in most paucibacillary months, and remaining positive lifelong, ADVANCES IN DIAGNOSIS forms of disease, including extrapulmonary, even after treatment.21 which represents most of childhood TB The Mantoux test is the standard TST Novel Culture Systems and cases. Their performance in children has not currently in use and involves the intradermal Detection Methods been thoroughly evaluated; however, limited injection of 2 standardized tuberculin units Bacteriophage-based assays use bac- studies to date suggest that their performance of purified protein derivative solution. Sub- teriophage viruses to infect and detect the in children is likely to be similar to that in sequent induration, rather than erythema, is presence of viable M. tuberculosis in clinical smear-negative adults because of the pauc- measured in millimeters after 48 to 72 hours. samples and culture isolates. Two main ap- ibacillary nature of TB in children.3 In some countries, such as the United King- proaches have been developed: (1) to detect There have been several recent evalu- dom, with low TB incidence a TST is re- the presence of mycobacteria using either ations of NAATs performed on nonrespira- garded as positive with induration of Ͼ5mm phage amplification and (2) to detect light tory samples to diagnosis respiratory disease. in those without prior Bacille Calmette- produced by luciferase reporter phages after One study has reported the presence of small Gue´rin (BCG) vaccination and Ͼ15 mm for their infection of live M. tuberculosis. When fragments of M. tuberculosis IS6110 DNA in those who have received BCG vaccination. the assays detect M. tuberculosis in drug-free urine (so-called transrenal DNA or tr-DNA) The World Health Organization (WHO) samples, but fail to detect M. tuberculosis in of 34 of 43 adults with TB but not in healthy guidelines differ slightly in that a positive drug-containing samples, the strains are clas- controls.30 However, other studies have TST is regarded as positive with induration sified as drug susceptible. In general, phage shown wide variations in performance (7%– Ͼ10 mm for those without prior BCG vac- assays have a turn around time of 2 to 3 days 100% sensitivity), and there are no data on

the performance of these tests in children. A linked immunosorbent assay and enzyme- tinguished active TB from LTBI (88%–98% urinary test that could serve as a rapid and linked immunospot assay, respectively.35 sensitivity, 100% specificity).43 easy diagnostic test has advantages in the IGRAs have been studied in both pediatric population. LTBI and active disease in different geo- Novel Detection Methods NAATs have also been used for the graphic settings. As there is no gold standard A recent innovative approach that has been explored is the urinary detection of rapid detection of rifampicin resistance di- for LTBI, exposure gradients have been used lipoarabinomannan (LAM). LAM is a rectly from sputum samples. The Xpert and comparison made with TST. Overall, 17.5-kD glycolipid component of the outer MTB/RIF is a cartridge-based, automated although evidence is limited, the results cell wall of mycobacteria. LAM is heat sta- diagnostic test that is rapid and simple to use show that IGRAs have modest predictive and correctly identified 98% of bacteria that ble, cleared by the kidney, and detectable in value, perhaps of the same magnitude as urine. As a bacterial product, it has the the- were resistant to rifampicin in a large study TST. For the diagnosis of LTBI, there is a 31 oretical potential to discriminate active TB in adults. In December 2010, WHO en- high agreement between the IGRAs but dorsed this test for use in TB endemic coun- from latent infection, the former having much discordance (mostly TST-positive/ higher quantities of bacteria. The sensitivity tries and declared it a major milestone for IGRA-negative) between the IGRA test and global TB diagnosis. of urinary LAM in adults varies widely TST. The high specificity of IGRAs may be 44,45 As mentioned earlier in the text, (44%–67%). Higher estimates have useful in reducing the number of low-risk young children swallow their sputum, and been reported in HIV-coinfected patients children who receive preventative therapy.36 thus DNA of M. tuberculosis may be de- with advanced immunosuppression, presum- In low TB incidence settings, there tected in stool. At present, there are limited ably because of higher bacterial burden and was higher specificity of IGRA (100% and increased frequency of disseminated dis- data in children, although in several small 46 studies, the sensitivity appears low (Ͻ40%) 98% for Quantiferon-TB (QFT). and T-Spot, ease. At present, there are limited data for compared with adults (sensitivity 86%).32,33 respectively) than TST (58%) in children urinary LAM in children. Real-time polymerase chain reaction with TB disease However, in children with has increasingly become available for clini- nontuberculous mycobacteria (NTM) and CONCLUSION cal use with the advantage of lower cross- other respiratory infections, TST had 100% Advances in the diagnosis of child- 37 contamination and as well as the ability to sensitivity compared with 93% for IGRA. hood TB in the past decade have included the identify rifampicin resistance. The rpoB Two studies in children with TB disease identification of alternative specimen types gene of M. tuberculosis accounts for Ͼ95% from the United Kingdom have shown lower as well as improvement in smear microscopy of rifampicin resistance, and because rifam- sensitivity of IGRA compared with TST and liquid culture systems. A number of picin resistance is usually accompanied by Ͼ15 mm (83%, 80%, and 58% for TST, novel and exciting methods have been iden- isoniazid resistance (monoresistance is rare), QFT, and T-spot and 82%, 78%, and 66% for tified for diagnosis of adult TB, such as this test is used as a marker for multidrug- TST, QFT, and T-spot, respectively).38,39 In integrated real-time polymerase chain reac- resistant TB. both studies, the sensitivities increased to tion detection systems, urine LAM, and test- Line probe assays (LPAs) are NAATs Ͼ90% when the combined IGRA and TST ing for volatile organic compounds in breath. that simultaneously detect infection with M. result was used to diagnose definite TB. However, many of these novel diagnostics tuberculosis and amplify regions of drug Overall for active disease, IGRAs have sub- have not been studied in children, and further resistance. LPAs use strip technology, optimal sensitivity and therefore cannot be research in this area is greatly needed. whereby amplified DNA is applied to strips used in isolation to rule out TB disease in REFERENCES containing probes specific for M. tuberculo- children. Although IGRAs may be used to 1. Ba F, Rieder H. A comparison of fluorescence sis, isoniazid, and rifampicin resistance. The help support a diagnosis of TB, in combina- microscopy with the Ziehl-Neelsen technique in WHO has endorsed LPAs for culture and tion with the TST and other investigations, the examination of sputum for acid-fast bacilli. smear-positive clinical specimens as part of a they should not be a substitute, or obviate the Int J Tuberc Lung Dis. 1999;3:1101–1105. larger commitment to target and implement need, for appropriate specimen collection. 2. Steingart K, Henry M, Ng V, et al. Fluorescence 34 new technology in high-burden countries. Furthermore, IGRA cannot distinguish be- versus conventional sputum smear microscopy for tuberculosis: a systematic review. Lancet In- Immunodiagnosis tween LTBI and active disease, similar to the fect Dis. 2006;6:570–581. TST. Further work is needed to determine Because of the limitations of TST, 3. Marais B, Brittle W, Painczyk K, et al. Use of the added value of IGRAs, beyond conven- light-emitting diode fluorescence microscopy to particularly cross-reactivity with BCG im- tional tests such as microbiology and chest detect acid-fast bacilli in sputum. Clin Infect Dis. munization and environmental mycobacte- radiographs for the diagnosis of active TB 2008;47:203–207. ria, newer diagnostic tests have been the disease. 4. American Thoracic Society. Diagnostic standards developed based on in vitro T-cell-based and classification of tuberculosis in adults and interferon-␥ release assays (IGRA), which The other significant problem encoun- children 1999. Am J Respir Crit Care Med. 2000; measure interferon-␥ production in re- tered with IGRA testing has been the risk of 161:1376–1395. sponse to stimulation to TB-specific anti- indeterminate tests, particularly in younger 5. Strumpf I, Tsang A, Syre J. Re-evaluation of sputum staining for the diagnosis of pulmonary gens (ESAT-6, CFP10 and in QuantiFERON children and the immunocompromised indi- 40 tuberculosis. Am Rev Respir Dis. 1979;119:599– TB Gold TB 7.7). These antigens are present viduals. Rates of indeterminates appear to 602. be generally higher in Quantiferon TB in M. tuberculosis complex but absent from 6. Vallejo J, Ong L, Starke J. Clinical features, all strains of Mycobacterium bovis BCG, and GOLD than enzyme-linked immunospot as- diagnosis and treatment of tuberculosis in infants. almost all environmental mycobacteria. Two say in young children and immunosupressed Pediatrics. 1994;94:1–7. IGRAs—the QuantiFERON-TB Gold assay individuals.41,42 7. Zar H, Hanslo D, Apolles P, et al. Induced spu- (Cellestis Limited, Carnegie, Victoria, Aus- Another immune-based approach has tum versus gastric lavage for microbiological confirmation if pulmonary tuberculosis in infants tralia) and the T SPOT-TB assay (Oxford been the measurement of the immune re- and young children: a prospective study. Lancet. Immunotec, Oxford, United Kingdom)—are sponse to transdermal application of M. tu- 2005;365:130–134. currently available. Both tests measure inter- berculosis MPB-64 antigen. In pilot studies, 8. Franchi L, Cama R, Gilman R, et al. Detection of feron-␥ release from T cells using enzyme- the MPB-64 skin patch test successfully dis- Mycobacterium tuberculosis in nasopharyngeal

aspirate samples in children. Lancet. 1998;352: 22. World Health Organization. Treatment of tuber- cobacterium tuberculosis infection in healthy ur- 1681–1682. culosis: guidelines for national programmes. 5th ban Indians. J Infect Dis. 2001;183:469–477. 9. Oberhelman R, Soto-Castellares G, Caviedes L, ed. Geneva: World Health Organization; 1997. 36. Ling D, Zwerling A, Steingart K, et al. et al. Improved recovery of Mycobacterium tu- 23. American Academy of Pediatrics. Tuberculosis. Immune-based diagnostics for TB in children: berculosis from children using the microscopic In: K. PL, editor. Red Book: 2009 Report of the what is the evidence? Paediatr Respir Rev. 2011; observation drug susceptibility method. Pediat- Committee on Infectious Diseases. 26th ed. Elk 12:9–15. rics. 2006;118:e100–e106. Grove Village, IL: American Academy of Pedi- 37. Detjen A, Keil T, Roll S, et al. Interferon-gamma 10. Al-Aghbari N, Al-Sonboli N, Yassin M, et al. atrics; 2009. release assays improve the diagnosis of tubercu- Multiple sampling in one day to optimize smear 24. Larsson L, Bentzon M, Lind A, et al. Sensitivity losis and non-tuberculous mycobacterial disease microscopy in children with tuberculosis in Ye- to sensitins and tuberculin in Swedish Children. in children in a country with a low incidence of men. PLoS One. 2009;4:e5140. Part 5: a study of school children in an inland tuberculosis. Clin Infect Dis. 2007;45:322–328. 11. Nicol M, Zar H. New Specimens and laboratory rural area. Tubercle Lung Dis. 1993;74:371–376. 38. Kampmann B, Whittaker E, Williams A, et al. diagnostics for childhood pulmonary TB: prog- 25. Sepulveda R, Burr C, Ferrer X, et al. Booster Interferon-gamma release assays do not identify ress and prospects. Paediatr Respir Rev. 2011;12: effect of tuberculin testing in healthy 6-yearpold more children with active tuberculosis that the 16–21. school children vaccinated with bacille Calmette- tuberculin skin test. Eur Respir J. 2009;33:1374– 12. Vargas D, Garcia L, Gilman R, et al. Diagnosis of Guerin at birth in Santiago, Chile. Pediatr Infect 1382. sputum-scarce HIV-associated pulmonary tuber- Dis J. 1988;7:578–582. 39. Bamford A, Crook A, Clark J, et al. Comparison culosis in Lima, Peru. Lancet. 2005;365:150– 26. Khan E, Starke J. Diagnosis of tuberculosis in of interferon-gamma release assays and tubercu- 152. children: increased need for better methods. lin skin test in predicting active tuberculosis in children in the UK: a paediatric TB network 13. Chow F, Espiritu N, Gilman R, et al. La cuerda Emerg Infect Dis. 1995;1:115–123. study. Arch Dis Child. 2010;95:180–186. dulce-a tolerability and acceptability study of a 27. Andronikou S, Joseph E, Lucas S, et al. CT novel approach to specimen collection for diag- scanning for the detection of tuberculous medi- 40. Haustein T, Ridout D, Hartley J, et al. The like- nosis of paediatric pulmonary tuberculosis. BMC astinal and hilar lymphadenopathy in children. lihood of an indeterminate test result from a Infect Dis. 2006;6:67. Pediatr Radiol. 2004;34:232–236. whole-blood interferon gamma release assay for the diagnosis of mycobacterium tuberculosis in- 14. Starke J, Taylor-Watts K. Tuberculosis in the 28. De Backer A, Mortele K, Vanhoenacker F, et al. fection in children correlates with age and im- pediatric population of Houston, Texas. Pediat- Imaging of extraspinal musculoskeletal tubercu- mune status. Pediatr Infect Dis. 2009;28:669– rics. 1989;84:28–35. losis. Eur J Radiol. 2006;57:119–130. 673. 15. Teo S, Riordan A, Alfaham M, et al. Tuberculosis 29. Kolk A, Hoelscher M, Maboko L, et al. 41. Bergamini B, Losi M, Vaienti F, et al. in the United Kingdom and Republic of Ireland. Electronic-nose technology using sputum sam- Performance of commercial blood tests for the Arch Dis Child. 2009;94:263–267. ples in diagnosis of patients with tuberculosis. diagnosis of latent tuberculosis infection in chil- J Clin Microbiol. 2010;48:4235–4238. 16. Wright C, Hesseling AC, Bamford C, et al. Fine dren and adolescents. Pediatrics. 2009;123: needle aspiration biopsy: a first-line diagnostic 30. Cannas A, Goletti D, Girardi E, et al. e419–e424. Mycobacterium tuberculosis DNA detection in procedure in paediatric tuberculosis suspects with 42. Diel R, Loddenkemper R, Nienhaus A. peripheral tuberculosis suspects with peripheral soluble fraction of urine from pulmonary tuber- Evidence-based comparison of commercial inter- lymphadenopathy. Int J Tuberc Lung Dis. 2009; culosis patients. Int J Tuberc Lung Dis. 2008;12: feron-gamma release assays for detecting active 13:1373–1379. 146–151. TB: a meta-analysis. Chest. 2010;137:952–968. 17. Cruciani M, Scarparo C, Malena M, et al. 31. Boehme C, Nabeta P, Hilleman D, et al. Rapid 43. Nakamura R, Einck L, Velmonte M, et al. Meta-analysis of BACTEC MGIT 960 and BACTEC molecular detection of tuberculosis and rifampin Detection of active tuberculosis by an MPB-64 460 TB, with or without solid media, for detection of resistance. N Engl J Med. 2010;363:1005–1015. transdermal patch: a field study. Scand J Infect mycobacteria. J Clin Microbiol. 2004;42:2321–2325. 32. Wolf H, Mendez M, Gilman R, et al. Diagnosis of Dis. 2001;33:405–407. 18. National Institute of Clinical Excellence. Tuber- pediatric tuberculosis by stool PCR. Am J Trop 44. Boehme C, Molokova E, Minja F, et al. Detection culosis: Clinical diagnosis and management of Med Hyg. 2008;79:893–898. of mycobacterial lipoarabinomannan with an an- tuberculosis, and measures for its prevention and 33. Cordova J, Shiloh R, Gilman R, et al. Evaluation tigen-capture ELISA in unprocessed urine of control: Royal College of Physicians of London; of molecular tools for detection and drug suscep- Tanzanian patients with suspected tuberculosis. 2006. tibility testing of Mycobacterium tuberculosis in Trans R Soc Trop Med Hyg. 2005;99:893–900. 19. Moore D, Evans C, Gilman R, et al. Microscopic- stool specimens from patients with pulmonary 45. Mutetwa R, Boehme C, Dimairo M, et al. observation drug-susceptibility assay for the diag- tuberculosis. J Clin Microbiol. 2010;48:1820– Diagnostic accuracy of commercial urinary li- nosis of TB. N Engl J Med. 2006;355:1539–1550. 1826. poarabinomannan detection in African tuberculo- 20. Arias M, Mello F, Pavon A, et al. Clinical eval- 34. Molecular line probe assays for rapid screening sis suspects and patients. Int J Tuberc Lung Dis. uation of the microscopic-observation drug-sus- of patients at risk of multidrug-resistant tubercu- 2009;13:1253–1259. ceptibility assay for detection of tuberculosis. losis (MDR-TB). Geneva: World Health Organi- 46. Lawn S, Edwards D, Kranzer K, et al. Urine Clin Infect Dis. 2007;44:674–680. zation; 2008. lipoarabinomannan assay for tuberculosis screen- 21. Hsu K. Tuberculin reaction in children treated 35. Lalvani A, Nagvenkar P, Udwadia Z, et al. ing before antiretroviral therapy, diagnostic yield with isoniazid. Am J Dis Child. 1983;137:1090– Enumeration of T cells specific for RDi-encoded and association with immune reconstitution dis- 1092. antigens suggest a high prevalence of latent my- ease. AIDS. 2009;23:1875–1880.