Clinical Biochemistry 37 (2004) 450 – 455

Childhood tuberculosis and its early diagnosis

James W. Gray *
Department of Microbiology, Birmingham Children’s Hospital, Birmingham B4 6NH, UK

Received 7 November 2003; received in revised form 9 March 2004; accepted 9 March 2004

Available online 13 April 2004

Abstract

Traditional methods for laboratory diagnosis of tuberculosis are unsatisfactory, especially for children, in whose specimens mycobacteria
are usually sparse. Recent changes in tuberculosis epidemiology in developed countries, including a large increase in incidence in children
from certain ethnic minorities, have prompted interest in newer diagnostic methods. Liquid-based culture detection systems offer improved
sensitivity and speed of diagnosis, although the time taken for d etection of growth is still upwards of 1 week. Nucleic acid amplification
techniques offer more rapid results, but perform best on smear-positive samples; sensitivities may be as low as 50% in smear-negative
specimens. Although these newer techniques are widely used in some developed countries, in others, they are not perceived as offering
sufficient benefit to justify their routine use. The diagnostic accuracy of mycobacteriophage and serologic methods is insufficient to justify
their wide use even in developing countries. Despite recent developments, there is still no panacea for diagnosis of childhood tuberculosis.
D 2004 The Canadian Society of Clinical Chemists. All rights reserved.

Keywords: Childhood tuberculosis; Mycobacteriophage; Epidemiology

Introduction 66% of tuberculosis cases in the UK now occur in non-

white ethnic groups [3].
Worldwide, there are an estimated 8 million new cases of Overall, children are believed to account for 5– 15% of the
tuberculosis each year [1]. Not only is the global burden of total worldwide tuberculosis disease burden [4], but they may
tuberculosis rising, but resistance to first-line antitubercu- account for a much higher proportion of cases in developing
lous drugs is also increasing [2]. The highest incidences of countries [4]. The clinical presentation of childhood tuber-
tuberculosis are in sub-Saharan Africa, much of Asia, and in culosis differs in developed and developing countries. In the
some Eastern European countries, where rates typically former, a substantial proportion of cases are asymptomatic
exceed 100 per 100 000 population [2]. Developing coun- and are only diagnosed through contact investigation of
tries account for over 90% of the tuberculosis burden. adults: in a recent Scottish study, only 60% of children
Twenty-two of these countries, including India, Pakistan, treated for tuberculosis had clinical disease [5]. In developing
South Africa, and Brazil, account for 80% of all worldwide countries, there is a greater incidence of symptomatic child-
cases each year, and are designated by the World Health hood tuberculosis [6]. The incidence and severity of child-
Organisation as high-burden countries [2]. In most high- hood tuberculosis in some of these countries is compounded
income developed countries, the overall incidence of tuber- by coexisting diseases, especially malnourishment and HIV
culosis is relatively static at 10 per 100 000 or less [2]. infection [7]. We need to be alert to the possibility that the
However there has been a significant change in the distri- changing epidemiology of tuberculosis in developed coun-
bution of disease in these countries, with ethnic minority tries will lead to an increase in the incidence of childhood
groups accounting for an increasing proportion of cases: tuberculosis, and to a change in clinical presentation with a
greater number of symptomatic cases. In England and Wales,
there is as yet no evidence of a sustained overall increase in
* Department of Microbiology, Birmingham Children’s Hospital,
childhood tuberculosis, with incidences in children aged
Steelhouse Lane, Birmingham B4 6NH, UK. Fax: +44-121-333-9811. under 15 years of 3.3 and 3.6 per 100 000 in 1998 and
E-mail address: [email protected]. 1998. However, these figures mask an increase in the

0009-9120/$ - see front matter D 2004 The Canadian Society of Clinical Chemists. All rights reserved.
doi:10.1016/j.clinbiochem.2004.03.003
 J.W. Gray / Clinical Biochemistry 37 (2004) 450–455 451

incidence of tuberculosis in children of black African ethnic identification, and drug susceptibility testing of M. tubercu-
origin from 15 to 71 per 100 000 over the same period [8]. losis is typically 4 – 6 weeks, although cultures should be
In all countries, the diagnosis of childhood tuberculosis is maintained for up to 12 weeks before being reported as
often made without laboratory confirmation. In developing negative [15]. However, the US Centers for Disease Control
countries, up to 90% of symptomatic cases are diagnosed have recommended that M. tuberculosis should be identified,
solely from clinical history and examination, and tuberculo- and first-line drug susceptibility testing completed, within 30
sis score charts have been widely used to assist in this days of specimen collection [16]. Recent developments in
process [9,10]. However, increasing numbers of malnour- the laboratory diagnosis of tuberculosis have attempted, so
ished and HIV-positive children have compromised the value far with only limited success, to address the poor sensitivity
of such scoring systems [9]. Non-laboratory investigations and slowness of traditional diagnostic methods.
have a limited role in the diagnosis of childhood tuberculo-
sis. Tuberculin skin testing (e.g., the Mantoux test) gives Specimens
negative results in up to 40% of children with tuberculosis,
and radiographic findings are often non-specific [11]. Almost all laboratory techniques for diagnosis of tuber-
Reliable diagnosis of tuberculosis depends on laboratory culosis depend upon detection of mycobacteria. Therefore,
confirmation, but traditional diagnostic methods have long obtaining good quality specimens, containing the highest
been recognised to be less than optimal. In recent years, possible numbers of mycobacteria, is central to the diagno-
several new diagnostic techniques have been developed to sis of tuberculosis. Multiple samples will improve the
improve the speed and reliability of tuberculosis diagnosis. diagnostic yield, and specimen volumes should be as large
However, none of these has proven to be a diagnostic as possible, especially from sites such as CSF where myco-
panacea. Another consideration is that with the decline in bacteria are especially scanty.
incidence of tuberculosis in the indigenous children of many
industrialised countries, clinicians’ expectations of tubercu- Pulmonary tuberculosis
losis testing are changing. Now, the need is often to exclude For childhood pulmonary tuberculosis, the best specimen
tuberculosis from the differential diagnosis in low-risk remains early morning gastric aspirates. However, the
patients, so that drug therapy can confidently be withheld sensitivity is still poor (<50% in most hands, and 33% in
or discontinued. A national audit found that only 3.3% of one recent UK report) [5]. In theory, flexible fibreoptic
primary samples submitted for mycobacterial diagnosis in bronchoscopy might be expected to deliver good quality
the UK produced a positive mycobacterial culture [12], and specimens, however, in practice, it has proved inferior to
in the author’s own laboratory, fewer than 1% of specimens gastric aspirates [13]. However, bronchoscopy can be clin-
investigated for tuberculosis are positive. ically useful, especially in guiding the use of steroid therapy
The purpose of this article is to review the state-of-the-art for children whose chest radiographs are not suggestive of
position on diagnosis of childhood tuberculosis. bronchial involvement [17].

Pleurisy
Microbiological diagnosis of tuberculosis In tuberculous pleurisy biochemical analysis of the
pleural fluid is indicative of a mild exudate, with a protein
The mainstays of diagnosis of tuberculosis remain mi- concentration of 20 – 40 g/l. There are several thousand
croscopy, and culture on solid and liquid media. However, leucocytes/mm3, with an early predominance of polymorphs
neither microscopy nor culture offers good sensitivity. This giving way to a later lymphocytosis. Smears are rarely
is a particular problem for pediatric microbiology laborato- positive, because of the paucity of mycobacteria in the
ries because the bacterial load in childhood tuberculosis is fluid, and for the same reason, cultures are only positive
substantially lower than that in post-primary tuberculosis in in 15 –70% of cases [18,19]. Although ZN-stained smears
adults. The problem of the relative paucity of mycobacteria of pleural biopsy material are positive in fewer than 10% of
in pediatric samples is compounded by the difficulty in cases, histopathological examination reveals granulomas
obtaining good quality clinical samples of satisfactory vol- that are acid fast-positive in over 50% of cases, and cultures
ume, especially from younger children who are unable to are positive in over 90% [18].
produce sputum [13]. Microscopy can provide same-day
preliminary information, but cannot distinguish between Meningitis
Mycobacterium tuberculosis and other Mycobacterium spe- The CSF leukocyte count is usually in the range of 10–
cies. In adults, microscopy has the added advantage of 500/mm3 and lymphocytes usually predominate. The CSF
providing a measure of infectivity, but this observation protein concentration is usually markedly elevated (4– 50 g/
may be less important in pediatrics, where secondary cases l), the highest levels being associated with hydrocephalus
in contacts of children with smear-positive respiratory secre- and spinal block. The CSF glucose concentration is reduced,
tions or gastric aspirates appear to be rare [14]. With typically to 2.2 mmol/l or less [19]. The success of tests for
traditional culture methods, the time taken for isolation, detection of M. tuberculosis depends on both the volume of
452 J.W. Gray / Clinical Biochemistry 37 (2004) 450–455

Table 1 intact cell walls, and then bind to DNA or RNA in a

Typical mean times to detection of mycobacterial growth using conven-
sequence-specific (and therefore species-specific) way. De-
tional culture, the BACTEC 460TB system, and other liquid-based culture
detection systems spite reports of successful use of fluorescent-labelled pep-
tide nucleic acids to distinguish between species directly on
Type of infection Typical mean time (days) to detection of
mycobacterial growth clinical material, this technology has never become widely
used in clinical practice [15].
Conventional BACTEC Other
culture 460 systems
Conventional culture
Smear-positive 20 – 25 9 – 13 9 – 20
tuberculosis M. tuberculosis is a slow-growing and fastidious bacte-
Smear-negative 21 – 33 15 – 22 18 – 36 rium that requires specialised culture media. Traditional
tuberculosis culture is performed using solid or liquid media, including
Non-tuberculous 22 – 40 13 – 17 12 – 25 Lowenstein-Jensen, Kirchner, and the various Middlebrook
mycobacterial
formulations [15]. The average time to detection of growth
infections
on conventional culture media is 2 – 4 weeks, although
cultures are usually maintained for up to 12 weeks before
CSF examined and on the techniques used. In one recent being reported as negative. Although liquid media may be
series, microbiological confirmation was obtained in around faster, manual handling of vials is cumbersome and time-
50% of cases, with liquid culture (positive in 34.5% of consuming. Primary isolates are then subcultured for species
cases) being the most sensitive method, followed by nucleic identification and antibiotic susceptibility testing by pheno-
acid amplification techniques (13.8%) and microscopy typic methods (see later), leading to a further delay of 2 to 3
(1.7%) [20]. Whilst comparative studies may give a fair weeks before a final result is available [15].
estimation of the overall success of microbiological diag-
nosis, they almost certainly underestimate the effectiveness Liquid-based culture mycobacterial detection systems
of individual techniques because smaller sample volumes It is now well established that these techniques offer
are inevitably used for individual tests. improved sensitivity and speed of detection of mycobac-
teria. The first of these was the radiometric BACTEC
Infection at other sites 460TB system (Becton Dickinson, USA), but this has
Good-sized specimens of pus or tissue from the sus- several limitations, including the need for manual loading,
pected site of infection are the preferred specimens. Tissue potential for cross contamination related to invasive reading,
samples should also be referred for histopathological exam- lack of computerised data management, and the accumula-
ination. Comprehensive information on specimen collection tion of low-level radioactive waste. Several non-radiometric
from such sites has been produced by the American Tho- continuous monitoring culture systems have since been
racic Society and Centers for Disease Control [21]. introduced that circumvent mot or all of these drawbacks,
including the MB/BacT (Biomerieux, France), BACTEC
Diagnostic techniques 9000 (Becton Dickinson, USA) and the BACTEC MGIT
960 (Becton Dickinson, USA). The BACTEC MGIT 960 is
Microscopy also available as a manual system for use in smaller
Microscopy for mycobacteria has traditionally used the laboratories whose sample throughput does not justify the
Ziehl-Neelsen stain, which involves staining with carbol high capital cost of monitoring equipment.
fuchsin dye, which is retained by the mycolic acid-rich cell The relative performances of conventional culture, the
wall of mycobacteria after washing with acid alcohol. The BACTEC 460 system, and other systems are outlined in
time taken for microscopic examination can be reduced by Tables 1 and 2 [15,23]. Although there are few specifically
the use of fluorescent microscopy and staining with aur- pediatric-related data on these systems, it might be expected
amine phenol. However, examination of smears stained by
either method is an insensitive technique that can only detect
Table 2
5  103 AFBs/ml [15]. Another significant limitation of Typical recovery rates in mycobacterial infections using conventional
these microscopic techniques is the inability to distinguish culture, the BACTEC 460TB system, and other liquid-based culture
between M. tuberculosis and other Mycobacterium species, detection systems
so that microscopy is unsuitable for examining specimens Type of infection Typical recovery rates (%) as a proportion of all
such as urine that may be contaminated with non-pathogenic culture-positive cases
mycobacteria. Pseudoepidemics of tuberculosis have also Conventional BACTEC Other
been described where due to atypical mycobacteria being culture 460 systems
present in the hospital water supply [22]. Tuberculosis 79 – 95 90 – 100 77 – 97
Peptide nucleic acids are DNA-like molecules in which Non-tuberculous 33 – 60 76 – 85 58 – 80
the sugar phosphate backbone is replaced by a peptide-like mycobacterial
structure. Because they are hydrophobic, they pass across infections
 J.W. Gray / Clinical Biochemistry 37 (2004) 450–455 453

that they would be particularly useful in this setting because tively poor sensitivity, with positive results obtainable in
of the need for a sensitive detection method for M. tuber- 80 – 100% and 50– 95% of smear-positive and smear-nega-
culosis and other Mycobacterium species. Whilst the BAC- tive culture-positive respiratory samples [29]. NAATs may
TEC MGIT 960 system has been reported to offer be even less sensitive with non-pulmonary samples [29].
comparable sensitivity and speed of detection to the BAC- DNA purification before amplification can improve sensi-
TEC 460 in some studies [24], others have found the 460 tivity, but probably at the expense of specificity [30,31]. A
system to be superior for detection of non-tuberculous nested PCR was recently reported to offer improved sensi-
mycobacteria [23]. This is surprising, given the highly tivity for detection of M. tuberculosis in children [32].
enriched formulation of the MGIT medium that presumably Another more straightforward means of improving sensitiv-
accounts for the high contamination rate that has been ity is simply to test multiple samples, presumably because
consistently reported with the MGIT 960 [23 –25]. not all samples will contain detectable nucleic acid [15].
All of these systems are more expensive than conven- Whilst few studies have specifically evaluated the use of
tional culture, and even in developed countries, competition NAATs in pediatrics, a metaanalysis of PCR for diagnosis of
for limited health care resources means that they have not smear-negative tuberculosis concluded that the sensitivity
been adopted universally. Instead, in the UK at least, their and specificity of PCR compared less favourably in the six
use has been largely limited to samples from individual studies that included gastric aspirates [31]. The same meta-
patients in whom there is a particular clinical need for a analysis concluded that NAATs are not consistently accurate
rapid result, for example, in smear-negative clinically sus- enough to be routinely recommended for the diagnosis of
pected cases or where drug resistance is suspected. smear-negative pulmonary tuberculosis, but that they might
be most useful for testing bronchial specimens (as opposed
Species identification and drug susceptibility testing to sputum or gastric aspirates) in highly suspicious TB cases
These final elements in the laboratory diagnosis of [31]. NAATs may also be sometimes useful to confirm M.
tuberculosis are usually centralised at regional or national tuberculosis in smear-positive cases. Whenever and however
reference laboratories. As such, detailed discussion of NAATs are used, it is important that requesting clinicians
developments in this area is outside the scope of this article. have information on local experience of their sensitivities
Briefly, rapid genotypic methods are now widely used for and specificities to assist interpreting the significance of both
speciation, and rapid genotypic and phenotypic methods for positive and negative results.
drug susceptibility testing, of Mycobacterium species [12].
Whilst molecular techniques are the most rapid means of Mycobacteriophage
speciation, at present, their use in drug susceptibility testing Many of the advances in tuberculosis diagnosis described
is largely restricted to screening for rifampicin resistance above are dependent on high capital cost equipment that is
(because most mutations conferring resistance to this drug inaccessible to most high-incidence countries. The FAST-
are confined to a short section of the gene encoding the h PlaqueTB (BIOTEC Laboratories, UK) is a commercial
subunit of bacterial RNA polymerase). The most widely used technique that has been marketed to provide rapid diagnosis
phenotypic methods are based on the liquid-based culture of tuberculosis in settings where there is limited access to
systems. Using these, identification of M. tuberculosis to diagnostic laboratories. It utilises a mycobacteriophage (a
species level and drug susceptibility results are usually virus that specifically infects mycobacteria) to reflect the
available within 3 to 6 and 4 to 12 days, respectively [26]. presence of M. tuberculosis in a specimen within 48 h.
Sensitivities of 58.3% to 87.5% in culture-positive samples
Nucleic acid amplification techniques have been reported [33,34].
The use of nucleic acid amplification techniques (NAATs)
is theoretically highly attractive for the diagnosis of tuber- Serology
culosis because they offer potential for high sensitivity and Despite the availability of purified and recombinant
specificity as well as rapid results. Various NAATs are antigens and monoclonal antibodies for competition assays,
commercially available, including PCR (Amplicor, Roche, serological testing has never been developed satisfactorily
Switzerland); transcription-mediated amplification (Gen- for diagnosis of tuberculosis. A particular problem in
Probe, USA), strand displacement amplification (ProbeTec- pediatrics is that children tend to have lower antibody titres,
SDA, Becton Dickinson, USA), and ligase chain reaction making distinction from natural exposure antibodies more
(LCx, Abbott Systems, USA). However, in practice, these difficult [35]. A number of protein and non-protein myco-
techniques have failed to fulfil some of their theoretical bacterial antigens have been investigated for use in antibody
promise. This compares with chlamydia diagnosis, where detection assays. However, IgA, IgG, and/or IgM responses
NAATs rapidly established themselves as the diagnostic tests to these appear to offer adequate sensitivity (29 – 75%) or
of choice [27]. Since they were first developed, the specific- specificity (50 – 92%), either alone or in multiantigen assays,
ities of NAATs for detection of mycobacteria have improved to be useful as a confirmatory test for childhood tuberculosis
markedly, but remains at less than 100% [28]. However, the [36 –41]. Another problem is that sensitivity is markedly
most disappointing feature of NAATs has been their rela- lower in patients with HIV co-infection [40].
454 J.W. Gray / Clinical Biochemistry 37 (2004) 450–455

Provision of tuberculosis laboratory diagnostic services mens that are negative by microscopy); the possibility of
In 1988, the World Health Organisation published defi- false positives; the additional cost; and the fact that with
nitions of peripheral, intermediate, and central tuberculosis limited sample volumes (as is often the case in pediatrics),
diagnostic laboratories [42]. The former are laboratories that use of multiple tests may actually decrease the overall
are capable of undertaking microscopy on sputum smears. likelihood of obtaining a positive result.
Intermediate laboratories have a greater repertoire of micro- Whilst newer diagnostic techniques for tuberculosis do
scopic techniques, undertake mycobacterial culture, and have some advantages over conventional methods, the
undertake species identification of M. tuberculosis. Central question that has to be addressed by laboratories consider-
laboratories can identify all Mycobacterium species and ing their use is whether the benefits that they offer justify
undertake antibiotic susceptibility testing. These definitions the additional cost. In developed countries, where tubercu-
were intended mainly for use in developing countries where losis is confirmed in fewer than 5% of patients investigated,
the lack of financial and infrastructure resources are major the challenge is to determine evidence-based guidelines for
limiting factors, and although they pre-dated the wide the clinical indications and specimen types at which the
commercial availability of newer diagnostic technologies, newer tests should be targeted. The potential value of these
they probably continue to represent a reasonable standard newer diagnostic techniques in developing countries with a
for most developing countries to aim to attain. high incidence of symptomatic childhood tuberculosis is
In developed countries, where financial constraints are clear, but these countries are unlikely to be able to provide
less (but still) important, there is more scope to develop much of the necessary infrastructure for, or afford the
diagnostic laboratory services according to local epidemio- additional costs of, these tests in the foreseeable future.
logical patterns and clinical need. For example, in areas Advances in the diagnosis of tuberculosis applicable to
with a high incidence of HIV infection or multi-drug- children in high-burden countries remain frustratingly out
resistant M. tuberculosis, the need for rapid turnaround of of reach.
species identification and antimicrobial susceptibility results
may justify local testing rather than referral of isolates to a
distant reference laboratory. The US Centers for Disease
Control [43] and the American Thoracic Society [21] have References
published recommendations for laboratories, which empha-
sise the timeliness of reporting, including the requirement [1] Dye C, Scheele S, Dolin P, Pathania V, Raviglione M. Consensus
statement. Global burden of tuberculosis: estimated incidence, prev-
for a 7 days per week service. As a result, many US alence and mortality by country. WHO Global Surveillance and Mon-
laboratories have adopted the newer technologies to reduce itoring Project. JAMA 1999;282:677 – 86.
turnaround times for identification of M. tuberculosis and [2] WHO Report 2003. Global tuberculosis control. Surveillance, plan-
other important species (e.g., Mycobacterium avium com- ning, finance. Geneva: WHO; 2003.
plex) and for first-line drug susceptibility testing. In the UK, [3] Akhtar M, Antoine D. Preliminary report on tuberculosis cases
reported in 2001 in England, Wales and Northern Ireland. London:
there has been little central guidance on the provision of PHLS CDSC; 2003.
mycobacterium diagnostic services, and it remains very [4] Donald PR. Childhood tuberculosis: out of control? Curr Opin Pulm
difficult to obtain funding to develop local diagnostic Med 2002;8:178 – 82.
services. Most hospital laboratories continue to rely on [5] Bland RM, Corrigan DL, Chapman JP, Paton JY. Tuberculosis in
microscopy and conventional culture, with investigation of children at the Royal Hospital for Sick Children, Glasgow. January
1995 – April 1998. Health Bull (Edinb) 2000;58:102 – 11.
primary isolates centralised on five regional mycobacterium [6] Salazar GE, Schmitz TL, Cama R, et al. Pulmonary tuberculosis in
reference centers [12]. Liquid-based mycobacterial detec- children in a developing country. Pediatrics 2001;108:448 – 53.
tion and NAAT facilities are also mainly provided from [7] Hesseling AC, Schaaf HS, Gie RP, Starke JR, Beyers N. A critical
regional centers and are not as widely used as in the US review of diagnostic approaches used in the diagnosis of childhood
tuberculosis. Int J Tuberc Lung Dis 2002;6:1038 – 45.
[12,15].
[8] Balasegaram S, Watson JM, Rose AM, et al. A decade of change:
tuberculosis in England and Wales 1988 – 1998. Arch Dis Child 2003;
88:772 – 7.
Conclusion [9] van Rheenen P. The use of the paediatric tuberculosis score chart in an
HIV-endemic area. Trop Med Int Health 2002;7:435 – 41.
Despite recent developments, the laboratory diagnosis of [10] Weismuller MM, Graham SM, Claessens NJ, Meijnen S, Salaniponi
FM, Harries AD. Diagnosis of childhood tuberculosis in Malawi: an
childhood tuberculosis remains unsatisfactory [44]. Newer audit of hospital practice. Int J Tuberc Lung Dis 2002;6:432 – 8.
liquid-based detection systems such as the BACTEC and [11] Khan EA, Starke JR. Diagnosis of tuberculosis in children: increased
MB/BacT systems offer improved sensitivity and more rapid need for better methods. Emerg Infect Dis 1995;1:115 – 23.
detection of growth than conventional culture, but the time [12] Drobniewski FA, Watt B, Smith EG, et al. A national audit of the
to reach a positive result is still measured in weeks rather laboratory diagnosis of tuberculosis and other mycobacterial diseases
within the UK. J Clin Pathol 1999;52:334 – 7.
than days. NAATs can provide a same-day or next-day [13] Bibi H, Mosheyev A, Shoseyov D, Feigenbaum D, Kurzbart E, Weil-
result. However, the benefits of this need to be balanced ler Z. Should bronchoscopy be performed in the evaluation of sus-
against the relatively low sensitivity (especially in speci- pected pediatric pulmonary tuberculosis? Chest 2002;122:1604 – 8.
 J.W. Gray / Clinical Biochemistry 37 (2004) 450–455 455

[14] George RH. Tuberculosis in hospital. J Hosp Infect 1984;5:109 – 17. [30] Honore-Bouakline S, Vincensini JP, Giacuzzo V, Lagrange PH, Herr-
[15] Watterson SA, Drobniewski FA. Modern laboratory diagnosis of my- mann JL. Rapid diagnosis of extrapulmonary tuberculosis by PCR:
cobacterial infections. J Clin Pathol 2000;53:727 – 32. influence of sample preparation and DNA extraction. J Clin Microbiol
[16] Tenover FC, Crawford JT, Huebner RE, Geiter LJ, Horsburgh Jr CR, 2003;41:2323 – 9.
Good RC. The resurgence of tuberculosis: is your laboratory ready? [31] Sarmiento OL, Weigle KA, Alexander J, Weber DJ, Miller WC. As-
J Clin Microbiol 1993;31:767 – 70. sessment by meta-analysis of PCR for diagnosis of smear-negative
[17] de Blic J, Azevedo I, Burren CP, Le Bourgeois M, Lallemand D, pulmonary tuberculosis. J Clin Microbiol 2003;41:3233 – 40.
Scheinmann P. The value of flexible bronchoscopy in childhood pul- [32] Montenegro SH, Gilman RH, Sheen P, et al. Improved detection of
monary tuberculosis. Chest 1991;100:688 – 92. Mycobacterium tuberculosis in Peruvian children by use of a hemi-
[18] Hasaneen NA, Zaki ME, Shalaby HM, El-Morsi AS. Polymerase chain nested IS6110 polymerase chain reaction. Clin Infect Dis 2003;36:
reaction of pleural biopsy is a rapid and sensitive method for the 16 – 23.
diagnosis of tuberculous pleural effusion. Chest 2003;124:2105 – 11. [33] Albay A, Kisa O, Baylan O, Doganci L. The evaluation of FASTPla-
[19] Starke JR. Tuberculosis. In: Jenson HB, Baltimore RS, editors. Pedi- queTB test for the rapid diagnosis of tuberculosis. Diagn Microbiol
atric infectious diseases. Principles and practice. Second ed. Philadel- Infect Dis 2003;46:211 – 5.
phia: Saunders; 2002. [34] Marei AM, El-Behedy EM, Mohtady HA, Afify AF. Evaluation of a
[20] Hooker JA, Muhindi DW, Amayo EO, Mc’ligeyo SO, Bhatt KM, rapid bacteriophage-based method for the detection of Mycobacterium
Odhiambo JA. Diagnostic utility of cerebrospinal fluid studies in tuberculosis in clinical samples. J Med Microbiol 2003;52:331 – 5.
patients with clinically suspected tuberculous meningitis. Int J Tuberc [35] Mahadevan S. Clinical utility of serodiagnosis of tuberculosis. Indian
Lung Dis 2003;7:787 – 96. J Pediatr 1997;64:97 – 103.
[21] American Thoracic Society. Diagnostic standards and classification of [36] Gupta S, Bhatia R, Datta KK. Serological diagnosis of childhood
tuberculosis in adults and children. Am J Respir Crit Care Med 2000; tuberculosis by estimation of mycobacterial antigen 60-specific
161:1376 – 95. immunoglobulins in the serum. Tuber Lung Dis 1997;78:21 – 7.
[22] Stine TM, Harris AA, Levin S, Rivera N, Kaplan RL. A pseudoepi- [37] Swaminathan S, Umadevi P, Shantha S, Radhakrishnan A, Datta M.
demic due to atypical mycobacteria in a hospital water supply. JAMA Sero diagnosis of tuberculosis in children using two ELISA kits.
1987;258:809 – 11. Indian J Pediatr 1999;66:837 – 42.
[23] Scarparo C, Piccoli P, Rigon A, Ruggiero G, Ricordi P, Piersimoni [38] Imaz MS, Comini MA, Zerbini E, et al. Evaluation of the diagnostic
C. Evaluation of the BACTEC MGIT 960 in comparison with the value of measuring IgG, IgM and IgA antibodies to the recombinant
BACTEC 460TB for detection and recovery of mycobacteria from 16-kilodalton antigen of Mycobacterium tuberculosis in childhood
clinical specimens. Diagn Microbiol Infect Dis 2002;44:157 – 61. tuberculosis. Int J Tuberc Lung Dis 2001;5:1036 – 43.
[24] Cornfield DB, Gleason Beavis K, Greene JA, Bojak M, Bondi J. [39] Julian E, Matas L, Perez A, Alcaide J, Laneelle MA, Luquin M.
Mycobacterial growth and bacterial contamination in the mycobacte- Serodiagnosis of tuberculosis: comparison of immunoglobulin A
ria growth tube indicator tube and BACTEC 460 culture systems. (IgA) response to sulfolipid I with IgG and IgM responses to 2,3-
J Clin Microbiol 1997;35:2068 – 71. diacyltrehalaose, 2,3,6-triacyltrehalose, and cord factor antigens.
[25] Lee JJ, Suo J, Lin CB, Wang JD, Lin TY, Tsai YC. Comparative J Clin Microbiol 2002;40:3782 – 8.
evaluation of the BACTEC MGIT 960 system with solid medium [40] Perkins MD, Conde MB, Martins M, Kritski AL. Serologic diagnosis
for isolation of mycobacteria. Int J Tuberc Lung Dis 2003;7:569 – 74. of tuberculosis using a simple commercial multiantigen assay. Chest
[26] Banaiee N, Bobadilla-del-Valle M, Riska PF, et al. Rapid identifica- 2003;123:107 – 12.
tion and susceptibility testing of Mycobacterium tuberculosis from [41] Singh K, Dong Y, Hinds L, et al. Combined use of serum and urinary
MGIT cultures with luciferase reporter mycobacteriophages. J Med antibody for diagnosis of tuberculosis. J Infect Dis 2003;188:371 – 7.
Microbiol 2003;52:557 – 61. [42] Narvaiz de Kantor I, Jae Kim S, et al. Laboratory services in tuber-
[27] Johnson RE, Newhall WJ, Papp JR, et al. Screening tests to detect culosis control. Part I: organization and management. Geneva: WHO;
Chlamydia trachomatis and Neisseria gonorrhoeae infections—2002. 1998.
Morb Mortal Wkly Rep, Recomm Rep 2002;51(RR-15):1 – 38. [43] Hinman AR, Hughes JM, Snider DE, Cohen ML. Meeting the chal-
[28] Gomez-Pastrana D. Tuberculosis in children—Is PCR the diagnostic lenge of multidrug-resistant tuberculosis: summary of a conference.
solution? Clin Microbiol Infect 2002;8:541 – 4. Morb Mortal Wkly Rep 1992;41:51 – 7.
[29] Iinuma Y, Senda K, Fujihara N, et al. Comparison of the BDProbeTec [44] Diagnosis of tuberculosis: countdown to new tools. UNDP/World
ET system with the Cobas Amplicor PCR for direct detection of Bank/WHO Special Programme for Research and Training in Tropi-
Mycobacterium tuberculosis in respiratory samples. Eur J Clin Micro- cal Diseases (TDR). Meeting report. Geneva, Switzerland; 29 – 30
biol Infect Dis 2003;22:368 – 71. June 2000.