SEARCH

SEARCH BY CITATION

Keywords:

  • Culture;
  • rapid detection;
  • sensitivity;
  • Thin Layer Agar;
  • tuberculosis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Sputum smear microscopy for the diagnosis of tuberculosis (TB) is cheap and simple but its sensitivity is low. Culture on Löwenstein–Jensen (LJ) is more sensitive but it takes a long time to yield results. Thin-Layer Agar (TLA) culture was suggested as an equally sensitive and faster alternative. We evaluated the performance of TLA for diagnosing TB in Jogjakarta, Indonesia. People with suspected TB presenting from July 2010 to July 2011 to two chest clinics of the National TB Control Programme network of Jogjakarta were eligible for inclusion. A sputum sample was sent to the Gadjah Mada University microbiology laboratory for concentration, smearing, Ziehl–Neelsen staining and culture on LJ and TLA. Sensitivity of cultures was evaluated against a composite reference standard (any positive culture). Time to detection of Mycobacteria was recorded. Out of 1414 samples, 164 (12%) were smear positive, 99 (7%) were scanty and 1151 (81%) were negative. On TLA and LJ respectively, 168 (12%) and 149 (11%) samples were positive, 72 (5%) and 32 (2%) were contaminated (κ = 0.64; 95% CI 0.59–0.69, p <0.01). Using the reference standard, 196 (14%) TB cases were identified. The sensitivity of TLA was 0.86 (95% CI 0.80–0.90), significantly higher (p 0.03) than for LJ (0.76; 95% CI 0.69–0.81). The median time to detection in days was significantly shorter (p <0.01) for TLA (12; 95% CI 11–13) than for LJ (44; 95% CI 43–45). TLA is a rapid and sensitive method for the diagnosis of TB. Implementation studies to evaluate the cost-effectiveness and impact of its introduction into programmatic settings are urgently needed.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

With an estimated 8.7 million new cases and 1.4 million deaths in 2011, tuberculosis (TB) remains a serious worldwide health problem and one of the leading causes of death from infectious diseases, especially in developing countries [1]. The current TB control strategy relies mainly on detection and treatment of active TB. In this framework, the use of a performing diagnostic method is crucial. Sputum smear microscopy, the most widely used test, is simple, rapid and inexpensive but it has a low sensitivity under field conditions [2]. Culture is more sensitive and permits species identification and drug susceptibility testing. When performed on solid medium it takes 4–8 weeks, whereas culture on liquid medium is faster, yielding results in less than 2 weeks, but it is more expensive and prone to contamination [2]. Tests based on molecular techniques are costly, and need sophisticated equipment and highly skilled personnel [2]. Recently, emphasis has been put on GeneXpert, a semi-automated molecular assay for the rapid diagnosis of TB and rifampin-resistant TB [3] but the high cost and still unsolved operational concerns limit the feasibility of its extensive roll out for routine TB diagnosis in resource-limited settings [4].

There is an urgent need for simple, accurate, inexpensive and rapid diagnostic tests for TB. Thin-Layer Agar (TLA) has been proposed as an easily implemented technique of culturing and identifying Mycobacteria and has been shown to be rapid and at least equally sensitive as conventional methods [5]. Most of the studies on TLA come from developed countries [6-8] or from Latin America [5, 9-12]. Many included mixed samples [6-8, 11-13], some come from high HIV prevalence settings [14].

We aimed to compare the performance of culture of sputum on TLA with culture on Löwenstein–Jensen (LJ) medium in terms of accuracy, contamination rate and turnaround time, for diagnosing pulmonary TB, in a setting with high TB incidence, low HIV prevalence, and limited resources.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Study setting

The study took place in Jogjakarta municipality, Indonesia, which has a population of around 500 000, an estimated TB incidence of 63/100 000 [15] and 1.9% HIV prevalence among TB patients [16]. The National TB Programme relies on nine hospitals, 18 health centres and two chest clinics. Four health centres perform smear microscopy. The chest clinics perform smear microscopy, chest X-ray and offer voluntary HIV counselling and testing; they are run by general practitioners and they are regularly visited by chest specialists. The clinics accept patients referred from health centres or self-referred.

The Microbiology Laboratory, Faculty of Medicine, Gadjah Mada University is a biosafety level II plus TB diagnostic referral laboratory performing microscopy, Mycobacterium culture using LJ and drug susceptibility testing using proportional methods. It has a biosafety cabinet class two type B2. The study was carried out in the two chest clinics and in the Microbiology Laboratory of Gadjah Mada University and it was endorsed by the Indonesian National TB Programme.

Participants

Participants were included prospectively. All new patients with suspected TB presenting from July 2010 to July 2011 to the chest clinics were eligible. Following national guidelines [17], they were individuals with cough lasting more than 2 weeks, never treated for TB. Exclusion criteria were age <18 years, refusal to participate in the study and inability to produce sputum. In the first phase of the study, non-residence in Jogjakarta municipality was an exclusion criterion, but from January 2011 these patients were included because it allowed for accelerated recruitment without affecting the validity of results.

Training, supervision and study preparation

To allow performing TLA, two candle jars were added to the equipment of the laboratory.

The laboratory staff participated in a 6-day training carried out locally by an expert from the Institute of Tropical Medicine in Antwerp, Belgium, which covered Ziehl–Nielsen staining and reading of sputum smears, decontamination, media preparation, inoculation and reading of TLA. The laboratory referent and one member of staff participated in a 2-week training at the Institute of Tropical Medicine. Both trainings were part of a wider project including other laboratory techniques. Had TLA been the only subject, we estimate that 5 days would have been sufficient.

The staff of the clinics participated in an initial workshop to motivate and to familiarize with the study procedures. A field coordinator conducted supervision to check adherence to the procedures monthly for the first 3 months and subsequently quarterly.

Pre-testing of all procedures was performed 1 week before the study roll-out.

Sample collection at chest clinic and transportation

Following national guidelines, eligible TB suspects produced three sputum samples (spot-morning-spot), after receiving thorough instructions. The first sample of this set was used for the study. In case of failure to produce the first spot sample, the morning sample was used. Samples were collected directly in a sterile tube and, after preparing a direct smear for microscopy reading in the chest clinic, they were sent to the laboratory by a dedicated courier. In case of need to keep samples overnight, they were kept at 4°C.

Sample processing at Gadjah Mada University Laboratory

After decontamination, concentration and re-suspension of the pellet, samples were processed for smear microscopy examination (Ziehl–Nielsen staining) and culture in parallel on LJ and TLA. Samples were processed in batches of a maximum of ten samples. Blinded technicians performed the tests and read the results. When it was not possible to process the samples on the day of receipt, they were kept at 4°C.

Sputum was digested and decontaminated using the sodium hydroxide/N-acetyl-l-cysteine method [18], with a final concentration of 1%; neutralization was performed with sterile phosphate-buffered saline solution. Samples were then concentrated by centrifugation at 4193 g for 15 min, in a Hettich® centrifuge Universal 320R cat 1406 and the pellet was re-suspended in 1 mL phosphate-buffered saline solution.

Smear microscopy examination using the Ziehl–Nielsen method (0.3% carbol-fuchsine, heated) was performed from the re-suspended pellet. Results were reported as negative, scanty and positive (1+ to 3+) according to the International Union Against Tuberculosis and Lung Disease/WHO scale [19].

Thin-layer agar culture was performed as described by Martin and Palomino [20]. For each sample, two plates were inoculated (without spreading), each one with 100 μL of re-suspended pellet. One plate was prepared by means of a reusable 60 × 15 mm Pyrex® Petri dish containing 10 mL Middlebrook agar 7H11 (Difco 0838-17), supplemented with 10% OADC (oleic acid, albumin, dextrose and catalase) (Becton Dickinson, Franklin Lakes, NJ) and piperacillin, trimethoprim and amphotericin B (Sigma Aldrich, St Louis, MO, USA; shelf life of the stock solution: 6 months at −20°C), all at 4 mg/L. The second plate was prepared as above and supplemented with 2% freshly prepared p-nitro-benzoic acid (PNB) solution at 500 μg PNB/mL (final concentration: 10 μg PNB/mL of medium). The inoculum was allowed to be absorbed by the medium overnight under the safety cabinet. Plates were then sealed with parafilm, leaving a space of 1–2 cm, incubated at 37°C in 5% CO2, checked after 24 h for contamination and then examined twice a week up to 6 weeks with a conventional microscope with a low-power objective (10×). A positive culture was identified by the characteristic colony morphology of Mycobacterium tuberculosis growth, taking into account consistency, colony border and cord formation. Non-tuberculous mycobacteria were recognized by their lack of cording and the ability to grow on PNB. Fungal or bacterial contamination was recognized by rapid overgrowth on plates. Further identification of non-tuberculous mycobacteria was performed at the reference laboratory in the Institute of Tropical Medicine. The shelf-life of the TLA plates was 2 months at 4°C.

LJ was prepared following the standard operating procedures recommended by the Indonesian National TB Programme [21], which are in line with the WHO guidelines [22]. For each sample, two tubes containing LJ and one containing LJ and PNB were inoculated with 100 μL of re-suspended pellet. Plates were read twice weekly up to 8 weeks. A single final result was given for each sample.

Quality control for both culture methods was performed using H37Rv reference strain.

Data collection and analysis

Data on the suspect's gender, age, symptoms, examinations and diagnosis were collected. In the laboratory, results of smear, cultures and time to detection of growth were recorded in separate registers by blinded technicians. Data were entered in a database and checked for errors. Statistical analysis was performed using SPSS for Windows v20 and the online software OpenEpi [23]. To evaluate sensitivity, in the absence of a reference standard, we used a composite reference standard [24] composed by bacteriologically proven TB by positive culture in any of the media, as generally done [5, 6, 8, 10, 13, 25]. Exact 95% CI were calculated for proportions. The McNemar test was used for pairwise comparisons of proportions. The degree of agreement between tests was determined by a Kappa concordance analysis. Time to detection of growth was compared using Mann–Whitney U test for unpaired comparisons and Wilcoxon signed rank test for paired comparisons.

Sample size considerations

Based on standard sample size calculation for diagnostic test evaluation [24], assuming a sensitivity of 80% of TLA versus the composite reference standard described above and a precision of ± 5%, specimens from at least 245 TB cases were needed. Based on an expected TB prevalence of 18% among TB suspects, at least 1300 TB suspects had to be included.

Ethics

The ethics committee of Faculty of Medicine, Gadjah Mada University, Institutional Review Board of National Institute of Health Research and Development, Indonesia and the Institute of Tropical Medicine Institutional Review Board approved the study. All participants gave written informed consent.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Of 2836 new TB suspects presenting to the clinics during the study period, 1414 were included. Reasons for non-inclusion were: for 53 (4%), age below inclusion criteria, for 609 (43%), residency outside Jogjakarta municipality, for 760 (53%) the reason was not reported. The first spot sample was collected for all but 73 suspects, whose morning sample was used.

A total of 164 (12%) samples had a 1+ to 3+ smear, 99 (7%) scanty, 1151 (81%) negative. Table 1 shows the results of cultures: the proportion of positive cultures was significantly higher (p 0.04) for TLA than for LJ; on the other hand, the proportion of contaminated cultures was significantly higher (p <0.01) for TLA than for LJ.

Table 1. Sputum culture results in 1414 new tuberculosis suspects in Jogjakarta, Indonesia, July 2010–2011
ResultThin-Layer Agar n (%)Löwenstein–Jensen n (%)
Positive, Mycobacterium tuberculosis168 (12)149 (11)
Positive, Non-tuberculous Mycobacterium1 (0.1)0
Negative1173 (83)1233 (87)
Contaminated72 (5)32 (2)
Total1414 (100)1414 (100)

Only one sample was positive for non-tuberculous mycobacteria, only on TLA. The same sample was positive for M. tuberculosis on LJ and it was excluded from the analysis on sensitivity.

Table 2 shows the pattern and frequency of occurrence of test results. There was complete concordance of the three tests in 81% of samples (in 74% all negative, in 7% all positive). Considering only cultures, 120 (8%) were positive on both methods, 18 (1%) were positive on LJ and negative on TLA, 36 (3%) were positive on TLA and negative on LJ, 10 (1%) were positive on LJ and contaminated on TLA, 12 (1%) were positive on TLA and contaminated on LJ, 1148 (81%) were negative on both methods, 49 (4%) were negative on LJ and contaminated on TLA, 7 (0.5%) were negative on TLA and contaminated on LJ, 13 (1%) were contaminated on both methods. The kappa coefficient was 0.64 (95% CI 0.59–0.69; p <0.01).

Table 2. Pattern and frequency of test results in 1414 new tuberculosis suspects in Jogjakarta, Indonesia, July 2010–2011
Thin Layer AgarLöwenstein-JensenConcentrated Smearn (%)a
  1. +, positive for Mycobacterium tuberculosis; −, negative for Mycobacterium tuberculosis.

  2. a

    Percentages do not add up to 100 because of rounding.

+++92 (7)
++Scanty6 (0.4)
++22 (2)
+Contaminated+8 (1)
+Contaminated4 (0.3)
+ContaminatedScanty0
Contaminated++7 (1)
Contaminated+2 (0.1)
Contaminated+Scanty1 (0.1)
++17 (1)
+Scanty2 (0.1)
+17 (1)
++8 (1)
+Scanty2 (0.1)
+9 (1)
+25 (2)
Scanty81 (6)
1042 (74)
Contaminated+3 (0.2)
ContaminatedScanty5 (0.4)
Contaminated41 (3)
Contaminated+0
ContaminatedScanty0
Contaminated7 (1)
ContaminatedContaminated+4 (0.3)
ContaminatedContaminated7 (1)
ContaminatedContaminatedScanty2 (0.1)
Total1414 (100)

As shown in Table 3, a total of 196 TB cases (14%) were identified using the composite reference standard. The sensitivity of TLA was significantly higher (p 0.03) compared with LJ and, for both methods, it was lower in smear negative samples. Among 163 samples with a 1+ to 3+ smear, 131 (80%) TB cases were identified using the composite reference standard compared with 11 out of 99 (11%) samples with scanty smear and 54 out of 1151 (5%) with negative smear. Considering only the 1322 samples uncontaminated in both cultures, 174 TB cases (13%) were identified using the composite reference standard; the sensitivity of TLA was 0.90 (95% CI 0.84–0.94), significantly higher (p 0.02) compared with LJ (0.79; 95% CI 0.73–0.85); the kappa coefficient was 0.79 (95% CI 0.74–0.85; p <0.01).

Table 3. Sensitivity of Thin Layer Agar and Löwenstein–Jensen stratified by smear results, according to the composite reference standard ‘Any positive culture’, in 1413 new tuberculosis suspects in Jogjakarta, Indonesia, July 2010–2011
Smear (total n)Any culture positive, n TLAa positive, n Sensitivity TLAa (95% CI)LJb positive, n Sensitivity LJb (95% CI)
  1. a

    TLA, Thin-Layer Agar.

  2. b

    LJ, Löwenstein–Jensen.

All (1413)1961680.86 (0.80–0.90)1480.76 (0.69–0.81)
Positive (262)1421250.88 (0.82–0.93)1150.81 (0.74–0.87)
3 + (45)43410.95 (0.85–0.99)350.81 (0.68–0.91)
2 + (36)33300.91 (0.77–0.98)280.85 (0.70–0.94)
1 + (82)55460.84 (0.72–0.92)430.78 (0.66–0.88)
Scanty (99)1180.73 (0.42–0.93)90.82 (0.52–0.97)
Negative (1151)54430.81 (0.67–0.89)330.61 (0.48–0.73)

Most (96%) samples were processed on arrival at the laboratory. Only 57 were processed with a delay (median 4 days). Information on time to detection was available for all positive cultures (Table 4). The median time in days from reception into the laboratory to detection was significantly shorter (p <0.01) for TLA than for LJ. Seventy-five percent of positive samples were detected within 19 days on TLA versus 52 days for LJ. On TLA, the median time to detection was significantly shorter (p 0.01) for smear-positive samples than for smear-negative ones. The results did not change meaningfully when excluding the 57 samples with delayed processing.

Table 4. Median time in days from entry of sample into laboratory to detection on culture on Thin Layer Agar and Löwenstein–Jensen, stratified by smear results, in new tuberculosis suspects in Jogjakarta, Indonesia, July 2010–2011
Method smear result n Median time (95% CI)Interquartile rangep value
  1. a

    Thin Layer Agar versus Löwenstein–Jensen.

  2. b

    Smear negative versus smear positive.

Thin layer agar16812 (11–13)8–19<0.001a
Smear negative4315 (12–19)10–240.01b
Smear positive (1 +  to 3 +  or scanty)12511 (10–13)8–18 
Löwenstein–Jensen14944 (43–45)36–52 
Smear negative3343 (42–45)39–490.84b
Smear positive (1+ to 3+ or scanty)11644 (42–46)36–50 

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

In this prospective study on a large number of sputum samples, TLA proved to be a more rapid and sensitive test than LJ for the detection of M. tuberculosis in people with suspected pulmonary TB.

This is the only study carried out on a consistent number of samples with homogeneous characteristics (only sputa), in an area of low HIV prevalence.

A study limitation is the absence of information on the HIV status of the participants, but we do not expect a high proportion of HIV positives, because a low prevalence of HIV in TB patients (1.9%) was previously found in the same setting [16]. Our results are not applicable to specimens other than sputum and, because we included only new suspects never treated for TB, they may possibly only be extrapolated to such cases.

The proportion of samples with scanty smear was high (38% of the 263 positive smears) and can be only partially explained by the fact that we used concentrated sputum, which increases the sensitivity of microscopy [26]. The possibility of false smear positivity cannot be completely ruled out. The percentage of samples with scanty smear and one or both positive cultures was particularly low (10%). Decontamination can result in killing of bacteria, leading to false negative culture [27] and this might become more evident in particularly sparse bacillary samples, such as those with scanty smear on concentrated sputum. We checked whether errors in the decontamination procedures could have led to ‘overkilling’ of bacteria, but did not find any overt irregularity. Comparisons with other studies are unfortunately impossible: few reported the smear reading scale and none provided specific information on scanty smears.

The sensitivity of TLA was higher than that of LJ. Robledo et al. [5], whose study is the most comparable to ours, made similar observations. A recent systematic review [28] estimated an overall pooled sensitivity of TLA of 0.87 (95% CI 0.79–0.94) for the diagnosis of TB, but it reported considerable heterogeneity in the results. Indeed, the included studies were carried out in varied settings and used diverse methods: most included other samples besides sputum, some included only smear-positive or smear-negative samples, some used both solid and liquid culture as reference standard, some did not use a composite reference standard.

The performance of both cultures was lower in smear negative samples, in particular for LJ. This is in agreement with other studies [5, 8, 14, 29] and was reported also for genetic testing [30] and liquid culture [25].

The contamination rate of TLA in our study was low (5%) and comparable to that in other observations [5, 9, 10, 13]. The reason for some studies reporting a higher contamination rate [9, 14, 25] might be related to the long time lag between sample collection and processing [25], the inclusion of only smear negative samples [14] and/or the use of a lower antibiotics concentration [9, 14] compared with our study.

Concordance between TLA and LJ was good and, excluding contaminated samples, the Kappa coefficient was 0.79. The only previous study reporting this [11], observed a lower concordance (κ = 0.58) for pulmonary samples, but it included other respiratory samples besides sputum.

Similarly with other studies [5-11, 14, 29], time to detection was remarkably shorter for TLA than for LJ and, for TLA, it was shorter in smear-positive versus smear-negative samples.

Besides a good sensitivity and a short time to detection, TLA has other advantages. It does not require additional sophisticated laboratory equipment to that needed to perform conventional culture. In the current study, including decontamination, it had a moderate cost of 4.6 euros per test for reagents and consumables. Importantly, the training needed was short (5 days) and it is easy to gain expertise in microscopic visualization of colonies. Simultaneous inoculation in PNB allows direct identification of the growth without the need for opening the plates and avoids the biosafety risk related to this procedure. Liquid culture such as on Mycobacteria Growth Indicator Tube has a similar time to detection and sensitivity [25] but it lacks most of these advantages. TLA is also used for the rapid detection of drug-resistant tuberculosis, with fast turnaround time, similar to automated commercial liquid culture drug susceptibility tests [31]. Other non-commercial techniques such as Nitrate Reductase Assay and Microscopic Observation Drug Susceptibility, have also been shown to be accurate and rapid yet inexpensive methods both for diagnosis of TB and drug sensitivity testing [28, 31, 32]. GeneXpert is even more rapid to perform but it does not eliminate the need for culture facilities, it requires cumbersome operating conditions and the cost is still prohibitive in most high-incidence settings, especially as first-line diagnostic test [33]. The major bottleneck of TLA is the need for microscopic observation. In our study it took around three minutes to read a plate (Ktut Rentyasti Palupi, personal communication). One reader could handle reading 120 plates/day, which corresponds to an inflow in the laboratory of 30 samples/week, making TLA more suitable for a laboratory with medium to low workload. LJ is less work-intensive than TLA and is cheaper (including decontamination, around three euros for reagents and consumables in our study), but the lower sensitivity and longer turnaround time are a disadvantage.

In conclusion, TLA is a simple, accurate, reasonably priced and rapid method for the diagnosis of pulmonary TB. Implementation studies are urgently needed to evaluate the cost-effectiveness and impact of its introduction into programmatic settings.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

This work received financial support from the Belgian government FWO (Fonds Wetenschappelijk Onderzoek – Vlaanderen). Thank you to the staff of the participating health facilities and to Alonso Soto for his comments on the manuscript. We dedicate this work to the memory of our dear colleague Francine Matthys.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References
  • 1
    World Health Organization. Global tuberculosis control report 2012. Geneva: World Health Organization, 2012.
  • 2
    Nahid P, Kim PS, Evans CA et al. Clinical research and development of tuberculosis diagnostics: moving from silos to synergy. J Infect Dis 2012; 205(suppl 2): S159S168.
  • 3
    Boehme CC, Nabeta P, Hillemann D et al. Rapid molecular detection of tuberculosis and rifampin resistance. N Engl J Med 2010; 363: 10051015.
  • 4
    McNerney R, Maeurer M, Abubakar I et al. Tuberculosis diagnostics and biomarkers: needs, challenges, recent advances, and opportunities. J Infect Dis 2012; 205(suppl 2): S147S158.
  • 5
    Robledo JA, Mejia GI, Morcillo N et al. Evaluation of a rapid culture method for tuberculosis diagnosis: a Latin American multi-center study. Int J Tuberc Lung Dis 2006; 10: 613619.
  • 6
    Welch DF, Guruswamy AP, Sides SJ, Shaw CH, Gilchrist MJ. Timely culture for mycobacteria which utilizes a microcolony method. J Clin Microbiol 1993; 31: 21782184.
  • 7
    Idigoras P, Perez-Trallero E, Alcorta M, Gutierrez C, Munoz-Baroja I. Rapid detection of tuberculous and non-tuberculous mycobacteria by microscopic observation of growth on Middlebrook 7H11 agar. Eur J Clin Microbiol Infect Dis 1995; 14: 610.
  • 8
    Gil-Setas A, Torroba L, Fernandez JL, Martinez-Artola V, Olite J. Evaluation of the MB/BacT system compared with Middlebrook 7H11 and Lowenstein–Jensen media for detection and recovery of mycobacteria from clinical specimens. Clin Microbiol Infect 2004; 10: 224228.
  • 9
    Mejia GI, Castrillon L, Trujillo H, Robledo JA. Microcolony detection in 7H11 thin layer culture is an alternative for rapid diagnosis of Mycobacterium tuberculosis infection. Int J Tuberc Lung Dis 1999; 3: 138142.
  • 10
    Caviedes L, Lee TS, Gilman RH et al. Rapid, efficient detection and drug susceptibility testing of Mycobacterium tuberculosis in sputum by microscopic observation of broth cultures. The Tuberculosis Working Group in Peru. J Clin Microbiol 2000; 38: 12031208.
  • 11
    Mejia GI, Guzman A, Agudelo CA, Trujillo H, Robledo J. Five year experience with thin layer agar medium for rapid diagnosis of tuberculosis. Biomédica 2004; 24(suppl 1): 5259.
  • 12
    Almeida da Silva PE, Wiesel F, Santos Boffo MM et al. Microcolony detection in thin layer culture as an alternative method for rapid detection of Mycobacterium tuberculosis in clinical specimens. Braz J Infect Dis 2007;38:421423.
  • 13
    Satti L, Ikram A, Abbasi S, Malik N, Mirza IA, Martin A. Evaluation of thin-layer agar 7H11 for the isolation of Mycobacterium tuberculosis complex. Int J Tuberc Lung Dis 2010; 14: 13541356.
  • 14
    Martin A, Munga WP, Babu OF et al. Implementation of the thin layer agar method for diagnosis of smear-negative pulmonary tuberculosis in a setting with a high prevalence of human immunodeficiency virus infection in Homa Bay, Kenya. J Clin Microbiol 2009; 47: 26322634.
  • 15
    Soemantri S, Lolong DS, Senewe FE et al. Tuberculosis prevalence survey in Indonesia. Jakarta: National Institute of Health Research and Development Ministry of Health, Republic of Indonesia, 2005.
  • 16
    Mahendradhata Y, Ahmad RA, Kusuma TA et al. Voluntary counseling and testing uptake and HIV prevalence among tuberculosis patients in Jogjakarta, Indonesia. Trans R Soc Trop Med Hyg 2008; 102: 10031010.
  • 17
    Ministry of Health Republic of Indonesia. National guideline for tuberculosis control. Jakarta: Ministry of Health, Republic of Indonesia, 2007.
  • 18
    Kent PT, Kubica GP. Public health mycobacteriology: a guide for level iii laboratory. Atlanta, GA: US Department of Health and Human Services, Public Health Service, Centers for Disease Control, 1985.
  • 19
    International Union Against Tuberculosis and Lung Disease. Technical guide: sputum examination for tuberculosis by direct microscopy in low income countries, 5th edn. Paris: International Union Against Tuberculosis and Lung Disease, 2000.
  • 20
    Martin A, Palomino JC. Procedure manual. Thin Layer Agar (TLA). Microcolony detection. Rapid culture of Mycobacterium tuberculosis. [Online] Available from: http://tbevidence.org/documents/rescentre/sop/Procedure%20manual%20TLA%20version%2006%202012.pdf [updated 1-6-2012; cited 30-10-2012].
  • 21
    Ministry of Health Republic of Indonesia. Guideline for culture, identification and sensitivity test of Mycobacterium tuberculosis from respiratory tract. Jakarta, Indonesia: Ministry of Health Republic of Indonesia, 2010.
  • 22
    World Health Organization. Laboratory Services in tuberculosis control. Geneva: World Health Organization, 1998.
  • 23
    Dean AG, Sullivan KM, Soe MM. OpenEpi: Open Source Epidemiologic Statistics for Public Health, Version 2.3.1. Available from: http://openepi.com [updated 23-6-2011; last accessed 10-03-2013].
  • 24
    Banoo S, Bell D, Bossuyt P et al. Evaluation of diagnostic tests for infectious diseases: general principles. Nat Rev Microbiol 2006; 4(suppl 9): S21S31.
  • 25
    Martin A, Fissette K, Varaine F, Portaels F, Palomino JC. Thin layer agar compared to BACTEC MGIT 960 for early detection of Mycobacterium tuberculosis. J Microbiol Methods 2009; 78: 107108.
  • 26
    Steingart KR, Ng V, Henry M et al. Sputum processing methods to improve the sensitivity of smear microscopy for tuberculosis: a systematic review. Lancet Infect Dis 2006; 6: 664674.
  • 27
    van Deun A, Salim AH, Cooreman E et al. Scanty AFB smears: what's in a name? Int J Tuberc Lung Dis 2004; 8: 816823.
  • 28
    Leung E, Minion J, Benedetti A, Pai M, Menzies D. Microcolony culture techniques for tuberculosis diagnosis: a systematic review. Int J Tuberc Lung Dis 2012; 16: 1623.
  • 29
    Irfan S, Hasan R, Kanji A, Hassan Q, Azam I. Evaluation of a microcolony detection method and phage assay for rapid detection of Mycobacterium tuberculosis in sputum samples. Southeast Asian J Trop Med Public Health 2006; 37: 11871195.
  • 30
    Chang K, Lu W, Wang J et al. Rapid and effective diagnosis of tuberculosis and rifampicin resistance with Xpert MTB/RIF assay: a meta-analysis. J Infect 2012; 64: 580588.
  • 31
    Minion J, Leung E, Menzies D, Pai M. Microscopic-observation drug susceptibility and thin layer agar assays for the detection of drug resistant tuberculosis: a systematic review and meta-analysis. Lancet Infect Dis 2010; 10: 688698.
  • 32
    Martin A, Panaiotov S, Portaels F, Hoffner S, Palomino JC, Angeby K. The nitrate reductase assay for the rapid detection of multidrug resistant M. tuberculosis: a systematic review and meta-analysis. J Antimicrob Chemother 2008; 62: 5665.
  • 33
    Evans CA. GeneXpert—a game-changer for tuberculosis control? PLoS Med 2011; 8: e1001064.