Professional Documents
Culture Documents
Art 2
Art 2
Review article
Current Concepts
Tuberculosis in Children
Carlos M. Perez-Velez, M.D., and Ben J. Marais, M.D., Ph.D.
W
From Grupo Tuberculosis Valle-Colorado idespread implementation of the strategy of directly ob-
and Clínica León XIII, IPS Universidad de served treatment short course (DOTS) during the 1990s resulted in im-
Antioquia, Medellín, Colombia (C.M.P.-V.);
and the Sydney Emerging Infections and proved global control of tuberculosis.1 However, its effectiveness has been
Biosecurity Institute and Children’s Hos- limited in areas where poverty and infection with the human immunodeficiency
pital at Westmead, University of Sydney, virus (HIV) or drug-resistant tuberculosis are prevalent, and the emphasis on a
Sydney (B.J.M.). Address reprint requests
to Dr. Marais at the Clinical School, Chil- positive sputum smear as the diagnostic criterion actually excludes most children
dren’s Hospital at Westmead, Locked from care.2 Tuberculosis remains a major but often unrecognized cause of disease
Bag 4001, Westmead, Sydney, NSW 2145, and death among children in areas where the disease is endemic3; service delivery
Australia, or at ben.marais@health.nsw
.gov.au. in such areas is hampered by the absence of pragmatic strategies to guide diagnosis
and management. This article provides a brief overview of basic principles, current
N Engl J Med 2012;367:348-61.
DOI: 10.1056/NEJMra1008049
controversies, and recent advances related to the care of children with tuberculosis,
Copyright © 2012 Massachusetts Medical Society. with an emphasis on intrathoracic disease.
The emergence of drug-resistant tuberculosis atic children who have been exposed to persons
poses a major threat to global tuberculosis con- with infectious disease.22 The recent formulation
trol.13 The initial complacency in addressing the of an international consensus on reference stan-
problem was influenced by studies indicating that dards and uniform research methodology should
the acquisition of isoniazid resistance reduced the facilitate progress.23-25
pathogenicity of the strain.14 However, the devel- The sequence of events that follows reinfection
opment of multidrug-resistant tuberculosis (char- (which is common in areas where tuberculosis is
acterized by resistance to isoniazid and rifampin) endemic) remains poorly defined. In cases of re-
in children exposed to persons with infectious current tuberculosis, strain typing makes it pos-
drug-resistant tuberculosis,15 as well as its clonal sible to differentiate relapse from reinfection but
spread in New York City16 and the Russian prison cannot be used to quantify the risk of reinfec-
system,17 has provided clinical evidence of the tion. Composite data analysis suggests that there
transmissibility of multidrug-resistant strains. Ad- is a 79% reduction in the risk of disease progres-
ditional proof was provided by an explosive out- sion among previously infected immunocompe-
break of extensively drug-resistant tuberculosis tent adults as compared with previously uninfected
(multidrug-resistant bacteria with additional re- adults after documented exposure26; however, the
sistance to a fluoroquinolone and a second-line epidemic contribution made by reinfection de-
injectable agent) among patients with HIV infec- pends on the frequency of its occurrence in a
tion in South Africa.18 Although the incidence of particular environment.
drug-resistant tuberculosis among children is un- It is important to differentiate infection from
known, pediatric cases provide a valuable epide- disease, since infection is a common event and
miologic perspective, since they reflect ongoing the approaches to managing the two conditions
transmission within communities. In places where are very different. Disease progression is usually
the rates of drug-resistant tuberculosis in children indicated by persistent, nonremitting symptoms,
have been monitored, the rates among children although the rate of progression is variable.21 In
were similar to those among adults from the the vast majority of cases (>90%), disease occurs
same community.15 The World Health Organiza- within 1 year after the primary infection, with the
tion (WHO) estimated that in 2008, 3.6% of inci- youngest children at greatest risk for progres-
dent tuberculosis cases globally were of the mul- sion. The risk profile is bimodal, with adolescents
tidrug-resistant or extensively drug-resistant type, being at increased risk.21 Exploring the mecha-
which suggests that there was a similar burden nisms underlying the increased risk and the sud-
of this type of disease among children.13 den switch in phenotype toward adult-type cavitary
disease that occurs with the onset of puberty
NAT UR A L HIS T OR Y OF DISE A SE should provide new insights into the immuno-
pathogenesis of tuberculosis.27
An understanding of the natural history of tuber-
culosis is required to appreciate both the variations A PPROACHE S T O DI AGNOSIS
in susceptibility to disease and the diverse spec-
trum of clinical manifestations observed in chil- Children are usually evaluated for tuberculosis
dren. Meticulous descriptions of tuberculosis in after presenting with symptoms or signs sugges-
the literature published before the introduction tive of disease (passive case finding) or as a result
of chemotherapy provide valuable insight into the of contact investigation or routine immigration
sequence of events that follows primary infection screening (active case finding). The clinical pre-
with Mycobacterium tuberculosis (Table 1 and Fig. sentation of children whose infection is detected
1).19,20 An important observation documented in through active case finding differs from that of
these earlier studies was the presence of tran- children whose infection is detected through pas-
sient hilar adenopathy, and even excretion of sive case finding, with the former group often hav-
M. tuberculosis, in children who had never had ing infection but not disease or having disease
progression to disease.21 This finding poses a in a very early phase. Among children in whom
major problem in case definition for studies, such M. tuberculosis infection is detected, young chil-
as vaccine efficacy trials, that use active case- dren and those with recent exposure are at in-
finding strategies in populations of asymptom- creased risk for progression to disease. Knowl-
350
Disease Phase Results on Tuberculin
and Timing Clinical Syndrome Group at Greatest Risk Immunopathogenesis Skin Test and IGRA Manifestations on Imaging
Primary infection
Incubation, 0–6 wk Asymptomatic All ages No adaptive immunity Negative None
Immune conver- Self-limiting symptoms (mild, viral- All ages Acquisition of adaptive immunity Generally positive; Transient hilar or mediastinal lymph-
sion, 1–3 mo like); hypersensitivity reactions infection may be adenopathy detected in 50–70%
(fever, erythema nodosum, lifelong; no test for of cases; transient Ghon focus
phlyctenular conjunctivitis) reinfection usually not detected
Early disease progres-
sion†
2–6 mo Uncomplicated lymph-node disease <10 yr of age Inadequate innate immunity, adap- Generally positive‡ Hilar or mediastinal lymphadenopa-
tive immunity, or both thy without airway involvement;
Ghon focus without cavitation
Progressive Ghon focus <1 yr of age or severely Inadequate innate immunity, adap- Generally positive‡ Ghon focus with visible cavitation
The
Peripheral lymphadenitis (most fre- 1–10 yr of age Inadequate local control Generally positive‡ Ultrasonography usually not needed,
quent extrathoracic disease man- but may reveal matting and adja-
ifestation, usually in the neck) cent soft-tissue edema
Late disease progres-
sion§
Downloaded from nejm.org on August 19, 2015. For personal use only. No other uses without permission.
8–24 mo Adult-type pulmonary disease >10 yr of age, but can Overly aggressive innate immunity, Generally positive‡ Apical cavities in one or both lungs,
(difficult to differentiate among occur in children as adaptive immunity, or both minimal or no lymph-node
primary infection, reactivation, young as 8 yr of age enlargement (previously referred
and reinfection; reactivation may to as postprimary tuberculosis)
occur >20 yr after initial infection)
Current Concepts
* Adapted from Wallgren19 and Lincoln and Sewell.20 Age ranges, risk groups, and timelines are intended to provide general guidance only; children infected with the human immunode-
hydronephrosis, calyceal dilata-
of exposure changes the pretest probability of
ficiency virus are particularly susceptible to tuberculosis and may present with atypical features. For the phases of disease, the times shown are the interval between initial exposure
disease and the positive predictive value of subse-
¶In cases of inadequate local control, manifestations are usually restricted to the local focus of disease, although disease can disseminate from any active focus.
characterizing the symptoms.28 The diversity of
the clinical presentation and the nonspecific na-
ture of most symptoms complicate diagnosis.
Constitutional symptoms often include failure to
Generally positive‡
Generally positive‡
Imaging Studies
In clinical practice, chest radiography is one of
the most useful diagnostic studies. Both frontal
and lateral views should be obtained, since a lat-
eral view assists in the assessment of the medias-
† At least 90% of disease manifestations occur within 12 months after infection.
and the onset of the disease phase. IGRA denotes interferon-γ–release assay.
>5 yr of age
>3 yr
tients with tuberculous meningitis, and it also hensive studies that focus on children.) Micro-
allows superior evaluation of the spine and soft scopical examination of sputum smears is the
tissues.33 cornerstone of diagnosis in most countries, but
its usefulness is limited in young children with
Laboratory Studies paucibacillary disease who are unable to expecto-
Table 2 provides an overview of the laboratory rate. Both the tuberculin skin test and the
examinations used in the diagnosis of tuberculo- interferon-γ release assay fail to differentiate
sis. (See the Supplementary Appendix, available M. tuberculosis infection from active disease. The
with the full text of this article at NEJM.org, for WHO recommends that the assay not be used in
a list of references that includes recent compre- place of the tuberculin skin test,34 although the
copy has a sensitivity that is about 10% ciation of acid-fast bacilli (a particular
greater than that of light microscopy problem when nontuberculous myco-
and a sensitivity that is 26–37% greater bacteria may be present, e.g., in biopsy
in patients coinfected with HIV; speci- specimens from fine-needle aspiration
ficity is similar to that of light micros- of lymph nodes or gastric aspiration);
copy and examination time is shorter no differentiation of viable and dead
bacilli
Detection of M. Solid medium: egg-based (e.g., Diagnosis of tuberculosis, Solid medium: high specificity; capable of Solid medium: in young children with clini-
tuberculosis Löwenstein–Jensen; Ogawa), agar- species identification, phenotypic and genotypic speciation cal diagnosis of intrathoracic tubercu-
growth based (e.g., Middlebrook 7H10 or 7H11), drug-susceptibility test- and drug-susceptibility testing; useful losis, sensitivity is low (about 10–40%),
and thin- or thick-layer plated agar ing, monitoring of treat- in all specimen types except formalin- depending on disease severity and
ment response fixed tissue; can include colorimetric specimen-collection method; slow
indicators of growth, such as nitrate turnaround time (2–12 wk); fastest
reductase assay and colorimetric redox turnaround provided by thin-layer agar
n e w e ng l a n d j o u r na l
ALERT), microscopical (e.g., tubercu formalin-fixed tissue; less dependent solid medium; high per-test cost rela-
losis microscopical observation drug- on operator expertise than other op- tive to other options; slow turnaround
susceptibility test kit), radiometric tions; most automated liquid-culture time (2–8 wk)
(e.g., BACTEC 460), and manometric systems and microscopical observa-
(e.g., VersaTREK) tion tests for drug susceptibility testing
are endorsed by the WHO; in children,
Downloaded from nejm.org on August 19, 2015. For personal use only. No other uses without permission.
sensitivity varies depending on disease
severity and quality of specimen collec-
tion and processing
Detection of DNA Positive or negative smear samples Diagnosis of tuberculosis, Moderate sensitivity (50–85% for negative Drug-susceptibility testing limited with
specific to (sputum, gastric aspirates, and species identification, results on smears and positive results most tests (except with line-probe
M. tuberculosis others) used for diagnosis but not drug-susceptibility test- on culture of respiratory specimens [at assays, which require culture isolates);
for drug-susceptibility testing (e.g., ing, monitoring of treat- higher end of range when at least two no differentiation of viable and dead
Amplified MTD, artus MTB, ment response samples are collected] and >95% for bacilli; high risk of cross-contamination
GenoQuick MTB, GenoType MD, positive results on smears); in chil- in open-tube–based systems; strict
LightCycler MTB, Loopamp MTB, dren, sensitivity of Xpert MTB/RIF is quality control essential; cost remains
RealArt MTB, and TaqMan MTB) 20%, fluorescence microscopy 9%, cul- high, but ongoing innovation and
Positive or negative smear samples ture 25% in one to two IS samples economies of scale should continue
(sputum, gastric aspirates, and from patients with clinical diagnosis of to drive it down; polymerase-chain-
others) used for diagnosis and limited intrathoracic tuberculosis; substantial reaction inhibitors may be problematic
drug-susceptibility testing (e.g., Xpert incremental yield for smear-negative with use of certain specimens
MTB/RIF) disease (33–61%) as compared with
Positive smear samples only (sputum, culture; high specificity (>97%); fully
gastric aspirates, and others) and automated and integrated platforms
culture isolates used for diagnosis and available (e.g., GeneXpert and
expanded drug-susceptibility testing COBAS); GenoType and Xpert MTB/
(e.g., GenoType MTBDRplus, Geno RIF and are endorsed by the WHO; am-
Type MTBDRsl, and INNO-LiPA Rif.TB) plified-MTD and TaqMan MTB ap-
proved by FDA; can be used in all spec-
imen types; rapid turnaround (2–12 hr)
Detection of Lipoarabinomannan, with assay in urine Diagnosis of tuberculosis Sensitivity best in immunocompromised Sensitivity generally very low (about 50%);
mycobacterial (e.g., Clearview TB ELISA and disease patients with HIV infection (CD4 count no differentiation between M. tubercu-
antigens Determine TB-LAM Ag) <50, 85%; CD4 count 50–100, 71%; losis and nontuberculous mycobacteria;
CD4 count 101–150, 56%); high speci- availability of pediatric studies limited,
ficity (about 95%); rapid turnaround some currently under way; current ver-
Current Concepts
Downloaded from nejm.org on August 19, 2015. For personal use only. No other uses without permission.
355
356
Table 2. (Continued.)
Studies of immune Tuberculin skin test (stimulation with in- Identification of M. tuberculo- Tuberculin skin-test sensitivity about 70– Tuberculin skin-test specificity reduced by
response to myco- tradermal injection of purified protein sis infection 80% (lower in immunocompromised BCG vaccination, especially in infants;
bacteria derivative, e.g., Mantoux method) patients); low cost; sensitivity of IGRAs skin-test sensitivity is low in immuno-
IGRA (stimulation with M. tuberculosis– 75–90% (lower in immunocompro- compromised patients and second vis-
specific antigens [e.g., ESAT-6 plus mised patients); specificity unaffected it required after 48–72 hr; IGRA sensi-
CFP-10, with or without TB 7.7], with by BCG vaccination; IGRAs may have tivity low in immunocompromised pa-
The
detection by ELISPOT assay [e.g., higher sensitivity for detection of re- tients; assay not well validated in chil-
T-SPOT.TB] or ELISA [e.g., cent rather than remote infection; test- dren; indeterminate assay results are
QuantiFERON–TB Gold]) ing includes a negative control and re- problematic; IGRA currently too com-
quires only a single visit plex and costly for routine use in set-
tings with limited resources; neither tu-
berculin skin test nor IGRA can differ-
entiate M. tuberculosis infection from
active disease
Antibody tests None None with current tests; advances awaited Values for sensitivity and specificity of cur-
rent commercial serologic tests are
highly variable; not recommended for
the diagnosis of pulmonary or extra
pulmonary tuberculosis
n e w e ng l a n d j o u r na l
fluids: detection of colorimetric method disease, monitoring of in specimens from pleural and pericar- nervous system tuberculosis; validated
biochemical mark- treatment response dial effusions, ascitic fluid, and CSF, for certain body fluids only
ers consistent with especially for isoenzyme ADA2 of ade-
Lysozyme (muramidase), identified with High sensitivity and specificity in speci- Further evaluation in children is needed;
Downloaded from nejm.org on August 19, 2015. For personal use only. No other uses without permission.
turbidimetric method mens from pleural and pericardial effu- validated for certain body fluids only
sions, depending on threshold values
Current Concepts
* Many new tests are in different phases of development and validation; see the Supplementary Appendix for further resources. BCG denotes bacille Calmette–Guérin, CSF cerebrospinal
expensive and often inaccessible in set-
Often provide the initial suggestion of the Value of results depends on skills of opera-
should guide case management (Fig. 2).
fluid, ELISA enzyme-linked immunosorbent assay, ELISPOT enzyme-linked immunosorbent spot, FDA Food and Drug Administration, HIV human immunodeficiency virus, IGRA
The most important variables to consider in
disease management are bacillary load and ana-
No Yes Yes
No Yes No Yes
Immune recovery after the initiation of antiret- a prolonged course may be required, depending
roviral treatment for HIV-coinfected individuals on the degree to which the patient’s immune
or nutritional rehabilitation may unmask sub- system has been compromised and the extent of
clinical disease or induce paradoxical deteriora- disease.49,50
tion, despite adequate treatment for tuberculo-
sis. A finding of IRIS does not indicate treatment PR E V EN T ION A ND C ON T ROL
failure, and treatment should not be interrupted;
patients with severe IRIS may require a course of Transmission of tuberculosis within health care
glucocorticoids. Despite the risk of IRIS, data on facilities is a particular concern in settings where
adults indicate that antiretroviral therapy is most immunologically vulnerable children may be ex-
effective when initiated within 8 weeks after the posed. In hospitals and clinics, careful consider-
start of tuberculosis treatment, or for patients with ation should be given to areas where patients are
severely compromised immune systems, within treated and to air-exchange patterns. It is also im-
2 to 4 weeks after the start of treatment.47 The only portant to recognize that symptomatic parents or
exception would be patients with central nervous caregivers may pose transmission risks.51 Vaccina-
system tuberculosis, in whom IRIS can have dev- tion with bacille Calmette–Guérin (BCG) reduces
astating consequences.48 With HIV-associated the risk of disseminated (miliary) disease and tu-
tuberculosis, treatment should be given daily, and berculosis meningitis in young children but offers
no consistent protection against adult-type tuber- raising the possibility that subclinical tuberculo-
culosis.52 No benefit of BCG vaccination has been sis was present at trial entry. The second trial
established in HIV-infected children, and it is enrolled young infants (3 to 4 months of age)
contraindicated in such children because of the who had been exposed to HIV but had no known
risk of disseminated BCG disease.53 The develop- exposure to tuberculosis.60 The infants were
ment of a safe and effective vaccine remains a top randomly assigned to receive open-label isonia-
priority among global health researchers. zid or placebo; those who were infected with
With good adherence, a 6-month course of HIV also received early antiretroviral therapy. All
isoniazid preventive therapy provides excellent infants were closely monitored for subsequent
protection against tuberculosis disease.54 De- exposure to tuberculosis. The investigators found
spite universal recommendations regarding the no significant difference in the incidence of tu-
provision of preventive therapy and strong evi- berculosis or mortality between the treatment
dence of the greatly increased risk of tuberculosis and placebo groups, suggesting that preexpo-
and the increased mortality among children in sure prophylaxis against tuberculosis has little
close contact with persons who have tuberculo- value if HIV-infected infants are enrolled in
sis,55 the implementation of preventive strategies management programs early, with meticulous
remains poor. Pragmatic solutions are required monitoring for tuberculosis exposure and provi-
to close the pronounced gap between policy and sion of postexposure prophylaxis. However, the
practice.56 Parents are often reluctant to provide value of preexposure prophylaxis in areas where
preventive treatment for an otherwise well child, monitoring for tuberculosis exposure is likely to
and the long duration of preventive therapy is a be poor remains unresolved.49,54
source of further discouragement. One study With the use of isoniazid preventive therapy
showed that a 3-month course of preventive after the completion of tuberculosis treatment in
therapy with isoniazid and rifampin was similar HIV-infected adults, it has been estimated that
in efficacy to a 9-month course of isoniazid 83 recurrences can be prevented for every 1000
alone.57 A regimen of 12 doses of weekly rifapen- cases treated.12 The WHO recommends isonia-
tine and isoniazid has been shown to be effica- zid preventive therapy for 6 to 36 months after
cious in adults,58 but this regimen is not yet the completion of tuberculosis treatment in all
recommended for children younger than 12 years patients with HIV infection, including children
of age because specific data on safety and efficacy who live in areas with a high prevalence of
in this age group are required. The efficacy of tuberculosis. However, the added value of pre-
abbreviated regimens has not been well studied ventive therapy as compared with ongoing
in children with HIV infection. A disadvantage screening for tuberculosis exposure and me-
of the regimens that include rifampin or rifapen- ticulous postexposure prophylaxis has not been
tine is the interactivity of these drugs with the evaluated.
protease inhibitors included in the antiretroviral It is possible to drastically reduce the morbid-
therapy provided for infection with HIV49,50; rif ity and mortality associated with pediatric tuber-
abutin is less reactive, but its use in preventive culosis if case detection is improved and preven-
therapy regimens has not been evaluated. tive therapy and curative treatment are made
Although the value of postexposure prophy- more accessible globally. Many challenges and
laxis is universally acknowledged, the value of research priorities remain (Table S3 in the Sup-
preexposure prophylaxis remains in question. plementary Appendix), but while we await the
Successive randomized, controlled trials of pre- development of new vaccines, better diagnostics,
exposure prophylaxis in children with HIV infec- and shorter treatment regimens, much can be
tion have had contradictory findings. The first of achieved with pragmatic approaches and sensi-
these trials, involving children with minimal ac- ble application of existing tools.
cess to antiretroviral therapy, was discontinued No potential conflict of interest relevant to this article was
because of increased mortality in the placebo reported.
group.59 The reduction in mortality among those Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
receiving isoniazid preventive therapy was con- We thank Mario Perez, M.D., for his valuable assistance with
fined to the first 2 to 3 months of treatment, an earlier version of Figure 1.
References
1. Global tuberculosis control report drug-resistant tuberculosis in the West- nosis of childhood tuberculosis. Int J Tu-
2011. Geneva: World Health Organiza- ern Cape Province of South Africa. Acta berc Lung Dis 2002;6:1038-45.
tion, 2011 (WHO/HTM/TB/2011.16). Paediatr 2006;95:523-8. 30. Edwards DJ, Kitelele F, van Rie A.
2. Marais BJ, Raviglione M, Donald PR, 16. Moss AR, Alland D, Telzak E, et al. A Agreement between clinical scoring sys-
et al. Scale-up of services and research city-wide outbreak of a multiple-drug- tems used for the diagnosis of tuberculo-
priorities for diagnosis, management and resistant strain of Mycobacterium tuber- sis in the HIV era. Int J Tuberc Lung Dis
control of tuberculosis: a call to action. culosis in New York. Int J Tuberc Lung Dis 2007;11:263-9.
Lancet 2010;375:2179-91. 1997;1:115-21. 31. Gie RP. Diagnostic atlas of intratho-
3. Marais BJ, Gupta A, Starke JR, El Sony 17. Drobniewski F, Balabanova Y, Niko- racic tuberculosis in children: a guide for
A. Tuberculosis in women and children. layevsky V, et al. Drug-resistant tubercu- low income countries. Paris: Internation-
Lancet 2010;375:2057-9. losis, clinical virulence, and the domi- al Union Against Tuberculosis and Lung
4. Newton SM, Brent AJ, Anderson S, nance of the Beijing strain family in Disease, 2003.
Whittaker E, Kampmann B. Paediatric tu- Russia. JAMA 2005;293:2726-31. 32. Andronikou S, van Hoenacker FM, de
berculosis. Lancet Infect Dis 2008;8:498- 18. Gandhi NR, Moll A, Sturm AW, et al. Backer AI. Advances in imaging chest tuber-
510. Extensively drug-resistant tuberculosis as culosis: blurring-of-differences-between-
5. Nelson LJ, Wells CD. Global epidemi- a cause of death in patients co-infected children-and-adults. Clin Chest Med 2009;
ology of childhood tuberculosis. Int J Tu- with tuberculosis and HIV in a rural area 30:717-44.
berc Lung Dis 2004;8:636-47. of South Africa. Lancet 2006;368:1575-80. 33. Pienaar M, Andronikou S, van Toorn
6. Menzies HJ, Winston CA, Holtz TH, 19. Wallgren A. The time-table of tuber- R. MRI to demonstrate diagnostic fea-
Cain KP, Mac Kenzie WR. Epidemiology culosis. Tubercle 1948;29:245-51. tures and complications of TBM not seen
of tuberculosis among US- and foreign- 20. Lincoln EM, Sewell EM. Tuberculosis with CT. Childs Nerv Syst 2009;25:941-7.
born children and adolescents in the in children. New York: McGraw-Hill, 1963: 34. Use of tuberculosis interferon-gam-
United States, 1994-2007. Am J Public 1-315. ma release assays (IGRAs) in low- and
Health 2010;100:1724-9. 21. Marais BJ, Gie RP, Schaaf HS, et al. middle-income countries: policy state-
7. Sandgren A, Hollo V, Quinten C, Man- The natural history of childhood intra- ment. Geneva: World Health Organization,
issero D. Childhood tuberculosis in the thoracic tuberculosis: a critical review of 2011.
European Union/European Economic the pre-chemotherapy literature. Int J Tu- 35. Updated guidelines for using inter-
area, 2000 to 2009. Euro Surveill 2011;16: berc Lung Dis 2004;8:392-402. feron gamma release assays to detect My-
pii:19825. 22. Hatherill M, Hanslo M, Hawkridge T, cobacterium tuberculosis infection — United
8. Lawn SD, Bekker LG, Middelkoop K, et al. Structured approaches for the States, 2010. MMWR Recomm Rep 2010;
Myer L, Wood R. Impact of HIV infection screening and diagnosis of tuberculosis 59(RR-5):1-25.
on the epidemiology of tuberculosis in a in a high prevalence region of South Af- 36. Nicol MP, Zar HJ. New specimens and
peri-urban community in South Africa: rica. Bull World Health Organ 2010;88: laboratory diagnostics for childhood pul-
the need for age-specific interventions. 312-20. monary TB: progress and prospects. Pae-
Clin Infect Dis 2006;42:1040-7. 23. Hatherill M, Verver S, Mahomed H. diatr Respir Rev 2011;12:16-21.
9. Cotton MF, Schaaf HS, Lottering G, Consensus statement on diagnostic end 37. Vargas D, Garcia L, Gilman RH, et al.
Wever HL, Coetzee J, Nachman S. Tuber- points for infant tuberculosis vaccine tri- Diagnosis of sputum-scarce HIV-associat-
culosis exposure in HIV-exposed infants als. Clin Infect Dis 2012;54:493-501. ed pulmonary tuberculosis in Lima, Peru.
in a high-prevalence setting. Int J Tuberc 24. Graham SM, Ahmed T, Amanullah F, Lancet 2005;365:150-2.
Lung Dis 2008;12:225-7. et al. Evaluation of tuberculosis diagnos- 38. Perez-Velez CM, Wilches-Luna EC,
10. Hesseling AC, Cotton MF, Jennings T, tics in children: 1. Proposed clinical case Hernández-Sarmiento JM, et al. Prelimi-
et al. High incidence of tuberculosis definitions for classification of intratho- nary results of a comparative yield study
among HIV-infected infants: evidence racic tuberculosis disease: consensus of induced sputum, string test, and gas-
from a South African population-based from an expert panel. J Infect Dis 2012; tric aspirate for the microbiological diag-
study highlights the need for improved 205:Suppl 2:S199-S208. nosis of pulmonary tuberculosis in chil-
tuberculosis control strategies. Clin In- 25. Cuevas LE, Browning R, Bossuyt P, et dren. Am J Respir Crit Care Med 2010;
fect Dis 2009;48:108-14. al. Evaluation of tuberculosis diagnostics 181:A1775.
11. Violari A, Cotton MF, Gibb DM, et al. in children: 2. Methodological issues for 39. Wright CA, Warren RW, Marais BJ.
Early antiretroviral therapy and mortality conducting and reporting research evalu- Fine needle aspiration biopsy: an under-
among HIV-infected infants. N Engl J ations of tuberculosis diagnostics for in- valued diagnostic modality in pediatric
Med 2008;359:2233-44. trathoracic tuberculosis in children: con- mycobacterial disease. Int J Tuberc Lung
12. Guidelines for intensified TB case- sensus from an expert panel. J Infect Dis Dis 2009;13:1467-75.
finding and INH preventive therapy for 2012;205:Suppl 2:S209-S215. 40. Rachow A, Clowes P, Saathoff E, et al.
people living with HIV in resource-con- 26. Andrews JR, Noubary F, Walensky RP, Increased and expedited case detection by
strained settings: annexes. Geneva: World et al. Risk of progression to active tuber- Xpert MTB/RIF assay in childhood tuber-
Health Organization, 2011 (http:// culosis following reinfection with myco- culosis: a prospective cohort study. Clin
whqlibdoc.who.int/publications/ bacterium tuberculosis. Clin Infect Dis Infect Dis 2012;54:1388-96.
2011/9789241500708_annexes_eng.pdf). 2012;54:784-91. 41. Nicol MP, Workman L, Isaacs W, et al.
13. Multidrug and extensively drug-resis- 27. Donald PR, Marais BJ, Barry CE III. Accuracy of the Xpert MTB/RIF test for
tant TB (M/XDR-TB): 2010 Global report The influence of age on the epidemiology the diagnosis of pulmonary tuberculosis
on surveillance and response. Geneva: and pathology of tuberculosis. Lancet in children admitted to hospital in Cape
World Health Organization, 2010 (WHO/ 2010;375:1852-4. Town, South Africa: a descriptive study.
HTM/TB/2010.3). 28. Marais BJ, Gie RP, Hesseling AC, et al. Lancet Infect Dis 2011;11:819-24.
14. Middlebrook G, Cohn ML. Some ob- A refined symptom-based approach to di- 42. Desk-guide for the diagnosis and
servations on the pathogenicity of isonia- agnose pulmonary tuberculosis in chil- management of TB in children. Paris: In-
zid-resistant variants of tubercle bacilli. dren. Pediatrics 2006;118(5):e1350-e1359. ternational Union Against Tuberculosis
Science 1953;118:297-9. 29. Hesseling AC, Schaaf HS, Gie RP, and Lung Disease, 2010 (http://www
15. Schaaf HS, Marais BJ, Hesseling AC, Starke JR, Beyers N. A critical review of .theunion.org/index.php/en/resources/
Gie RP, Beyers N, Donald PR. Childhood diagnostic approaches used in the diag- scientific-publications/tuberculosis/
item/193-desk-guide-for-diagnosis-and- ment of tuberculosis in HIV-infected chil- 55. Gomes VF, Andersen A, Wejse C, et al.
management-of-tb-in-children-). dren: recommendations for a public Impact of tuberculosis exposure at home on
43. WHO rapid advice — treatment of tu- health approach, 2010. Paris: Internation- mortality in children under 5 years of age
berculosis in children. Geneva: World al Union Against Tuberculosis and Lung in Guinea-Bissau. Thorax 2011;66:163-7.
Health Organization, 2010 (WHO/HTM/ Disease, 2010. 56. Hill PC, Rutherford ME, Audas R, van
TB/2010.13). 50. Guidelines for the prevention and Crevel R, Graham SM. Closing the policy-
44. Volmink J, Garner P. Directly ob- treatment of opportunistic infections practice gap in the management of child
served therapy for treating tuberculosis. among HIV-exposed and HIV-infected contacts of tuberculosis in developing
Cochrane Database Syst Rev 2007;4: children. MMWR Morb Mortal Wkly Rep countries. PLoS Med 2011;8(10):e1001105.
CD003343. 2009;58(Early Release):1-166 (http://www 57. Spyridis NP, Spyridis PG, Gelesme A,
45. Schaaf HS, Marais BJ. Management of .cdc.gov/mmwr/preview/mmwrhtml/ et al. The effectiveness of a 9-month regi-
multidrug-resistant tuberculosis in chil- rr58e0826a1.htm). men of isoniazid alone versus 3- and
dren: a survival guide for paediatricians. 51. Muñoz FM, Ong LT, Seavy D, Medina 4-month regimens of isoniazid plus rif
Paediatr Respir Rev 2011;12:31-8. D, Correa A, Starke JR. Tuberculosis ampin for treatment of latent tuberculosis
46. Ettehad D, Schaaf HS, Seddon JA, among adult visitors of children with sus- infection in children: results of an 11-year
Cooke G, Ford N. Treatment outcomes for pected tuberculosis and employees at a randomized study. Clin Infect Dis
children with multidrug-resistant tuber- children’s hospital. Infect Control Hosp 2007;45:715-22.
culosis: a systematic review and meta- Epidemiol 2002;23:568-72. 58. Sterling TR, Villarino ME, Barosimov
analysis. Lancet 2012 February 24 (Epub 52. Trunz BB, Fine P, Dye C. Effect of AS, et al. Three months of rifapentine and
ahead of print). BCG vaccination on childhood tubercu- isoniazid for latent tuberculosis infec-
47. Blanc FX, Sok T, Laureillard D, et al. lous meningitis and miliary tuberculosis tion. N Engl J Med 2011;365:2155-66.
Earlier versus later start of antiretroviral worldwide: a meta-analysis and assess- 59. Zar HJ, Cotton MF, Strauss S, et al.
therapy in HIV-infected adults with tuber- ment of cost-effectiveness. Lancet 2006; Effect of isoniazid prophylaxis on mortal-
culosis. N Engl J Med 2011;365:1471-81. 367:1173-80. ity and incidence of tuberculosis in chil-
48. Török ME, Yen NT, Chau TT, et al. 53. Revised BCG vaccination guideline dren with HIV: randomised controlled
Timing of initiation of antiretroviral ther- for infants at risk of HIV infection. Wkly trial. BMJ 2007;334:136.
apy in human immunodeficiency virus Epidemiol Rec 2007;82:193-6. 60. Madhi SA, Nachman S, Violari A, et
(HIV)-associated tuberculous meningitis. 54. Marais BJ, Ayles H, Graham SM, God- al. Primary isoniazid prophylaxis against
Clin Infect Dis 2011;52:1374-83. frey-Faussett P. Screening and preventive tuberculosis in HIV-exposed children.
49. Guidance for national tuberculosis therapy for tuberculosis. Clin Chest Med N Engl J Med 2011;365:21-31.
and HIV programmes on the manage- 2009;30:827-46. Copyright © 2012 Massachusetts Medical Society.