Preparation of mycobacteria-containing artificial sputum for TB panel ...

INT J TUBERC LUNG DIS 10(8):899–905 © 2006 The Union

Preparation of mycobacteria-containing artificial sputum for TB panel testing and microscopy of sputum smears H. Yamada,* S. Mitarai,† L. Aguiman,‡ H. Matsumoto,§ A. Fujiki § * Pathology and † Bacteriology Divisions, Mycobacterium Reference Center, Research Institute of Tuberculosis, Japan Anti-Tuberculosis Association, Tokyo, Japan; ‡ Cebu Regional Tuberculosis Reference Laboratory, Center for Health Development, Cebu, Philippines; § Research Division, Research Institute of Tuberculosis, Japan Anti-Tuberculosis Association, Tokyo, Japan SUMMARY S E T T I N G : Panel testing, blinded cross rechecking and on-site evaluation are the three methods for external quality assessment (EQA) of acid-fast bacilli (AFB) smear microscopy. Panel testing can provide data on laboratory capabilities prior to implementing a rechecking programme, assess the current status of laboratory performance and detect problems associated with diagnostic performance. Thus far, two methods for preparing panel test slides have been reported: these use real AFB-positive and -negative sputum treated with sodium hydroxide (NaOH) or N-acetyl-L-cysteine (NALC). O B J E C T I V E : To evaluate the above methods and to develop a new method to prepare panel test slides with artificial sputum. D E S I G N : Panel test slides were prepared using the NaOH and NALC methods. New artificial sputum preparation

methods were developed and examined using a cultured monocyte cell line, cultured avirulent mycobacteria and methylcellulose or polyacrylamide gel as substrate. Smears prepared by the four methods were compared. R E S U L T S : Panel test slides prepared with NaOH and NALC methods are not macroscopically or microscopically similar to real smears. Our new artificial sputum is similar to real sputum in viscosity and macroscopic and microscopic appearance; it is also consistent in panel positivity grades. C O N C L U S I O N : The artificial sputum described here could contribute to the EQA and training in tuberculosis laboratories or microscopy centres. K E Y W O R D S : panel test; artificial sputum; viscosity; external quality assessment; consistency

THE WORLD HEALTH ORGANIZATION (WHO) has reported that approximately 9 million people develop tuberculosis (TB) annually worldwide and that approximately 4 million cases are smear-positive.1 Direct sputum smear microscopy is the most effective method for detecting TB cases. For effective TB detection, it is essential to assure the quality of smear microscopy.2–4 Panel testing, blinded rechecking and on-site evaluation are the three components of external quality assessment (EQA) of acid-fast bacilli (AFB) smear microscopy. Panel testing can provide preliminary data on laboratory capabilities prior to a blinded rechecking programme.5 In practice, stained and unstained smear slides must be sent to peripheral laboratories to evaluate the ability of laboratory technicians to stain and/or read smears, although panel testing does not serve to assess routine laboratory performance. Two procedures for preparing panel test slides have been reported that use AFB-positive and -negative sputum treated with sodium hydroxide (NaOH) or

N-acetyl-L-cysteine (NALC).5 Although these procedures are simple and rapid, they have several disadvantages. First, large amounts of AFB-negative sputum are needed for dilution of AFB-positive sputum, but AFB-negative sputum is difficult to collect. Second, although highly positive AFB sputum is required, it presents a biohazard. Third, slides prepared with AFBpositive sputum are not standardised relative to grading controls. Finally, the appearance of smear slides prepared from NaOH and NALC-treated sputum are not macroscopically or microscopically similar to real smears. A new improved method is therefore needed to prepare sputum smear slides for panel testing. In this study, slides were prepared with readily available materials, and provided consistent AFB positivity grades. We also developed a method to prepare artificial sputum from cultured cells and avirulent mycobacteria using polyacrylamide gel. This artificial sputum should contribute to the EQA and training of technicians in smear examination.

Correspondence to: Hiroyuki Yamada, Bacteriology Division, Mycobacterium Reference Center, Research Institute of Tuberculosis, Japan Anti-Tuberculosis Association, 3-1-24, Matsuyama, Kiyose, Tokyo, 204-8533, Japan. Tel: (81) 424 93 5072. Fax: (81) 424 92 4600. e-mail: [email protected] Article submitted 15 December 2005. Final version accepted 3 April 2006.

900

The International Journal of Tuberculosis and Lung Disease

MATERIALS AND METHODS Reagents NaOH, formaldehyde and sodium citrate dihydrate were obtained from Kanto Chemical Co Ltd (Tokyo, Japan). Ammonium peroxodisulfate and N,N,N,Ntetramethylethylenediamine (TEMED) were obtained from Wako Pure Chemical Industries (Osaka, Japan). NALC was obtained from Merck (Darmstadt, Germany). RPMI1640 and foetal bovine serum were obtained from Invitrogen Corporation (Grand Island, NY, USA). Acrylamide and N,N-methylenebisacrylamide were obtained from Acros Organics (Morris Plains, NJ, USA). Sputum AFB-positive and -negative sputum samples were collected at Fukujuji Hospital and National Tokyo Hospital in Tokyo, with patients’ informed consent. The samples were stored at 30C without cryopreservative for several months. Preparation of panel test slides with real sputum Panel test slides were prepared using both the NaOH and NALC methods described by Aziz et al.5 The procedures are summarised in Figure 1. Briefly, 3 ml of AFB-positive and -negative sputum were placed in separate 50 ml centrifugation tubes and fixed with three drops (about 50 l) absolute formaldehyde for 1 h at room temperature; 1 ml 4% NaOH or 3 ml of a solution of 2% NALC and 2.9% sodium citrate was then added to the sample. For NaOH treatment, distilled water was added, to a total volume of 20 ml. The sample was mixed well and incubated at 55–60C for 30 min. Distilled water was then added to bring the

total volume to 40 ml, and the sample was mixed by inversion and centrifuged at 3000 g for 20 min at room temperature. After centrifugation, the supernatant was decanted carefully, and 0.5–1 ml of distilled water was added to resuspend the pellet. To evaluate the AFB positivity of the sample, smears were prepared using an ordinary microbiological loop. Preparation of cells and mycobacteria for artificial sputum To obtain cells for artificial sputum, the monocyte cell line THP-1 (American Type Culture Collection, Manassas, VA, USA) was cultured in 150 cm2 flasks containing 90 ml RPMI 1640 supplemented with 10% foetal bovine serum, 2 mm L-glutamine, penicillin (100 U/ml), and streptomycin (100 g/ml). Ten culture flasks were prepared to harvest a sufficient amount of cells. The cells were cultured at 37C in 5% CO2 for 3 weeks. Cells were then harvested by centrifugation and washed three times with 0.1 M phosphate buffer (PB) (pH 7.4) that had been filtered with MillexGS (pore size 0.2 m, Millipore Corporation, Bedford, MA, USA). After the last wash, the cell pellet was briefly dried by inverting the centrifugation tubes on a paper towel for 5 min. To obtain mycobacteria for artificial sputum, Mycobacterium bovis bacille Calmette-Guérin (BCG) Pasteur strain was grown in Middlebrook 7H9 broth for up to 2 weeks, with optical density (OD) being monitored at 530 nm. When the OD of a culture was 0.7, the broth was filtered and cell clumps dispersed by passage through an Acrodisk filter (#4650: pore size 5.0 m, Pall Corporation, Cornwall, UK) and the turbidity of the dispersed bacteria was measured. The filtrate was cultured on Ogawa media for 3–4 weeks and the number of colonies was counted. Experiments using AFB solutions containing 107 cfu (colony forming units)/ml (i.e., 105–107 cfu/ml) did not result in 3 smears (10 AFB/visual field) in at least 20 fields. An AFB concentration 107 cfu/ml was therefore required to obtain distinct 3 smears. To obtain an AFB solution for preparing 2 smears, a filtrate containing 107 cfu/ml was diluted with PB to 5 106 cfu/ml, to produce smears containing approximately 5 AFB/field, or 500 AFB/100 fields. An AFB solution for 1 smear (10–99 AFB/100 fields) was prepared by 10-fold dilution of the 2 smear solution (1–10 AFB/field in at least 50 fields) with PB, and an AFB solution for smears (1–9 AFB/100 fields, i.e., doubtful smears), was made by 10-fold dilution of the 1 smear solution. A 0.5 ml sample of each AFB grade solution was added to a dried cell pellet and mixed. Preparation of artificial sputum with methylcellulose or polyacrylamide

Figure 1 Smear preparation procedure using the NaOH and NALC methods for panel test slides as described in reference 5. AFB acid-fast bacilli; ZN Ziehl-Neelsen; RT room temperature; NaOH sodium hydroxide; NALC N-acetyl-L-cysteine.

Methylcellulose method The methylcellulose (MC) method was originally described by de Kantor et al. for the preparation of samples for proficiency testing to evaluate laboratory

Artificial sputum for panel test slide

901

Figure 2 Smear preparation protocols for the A) MC and B) PAM methods to make the artificial sputum smears used as panel test slides. PB 0.1 M phosphate buffer (pH 7.4); cfu colony forming units; AFB acid-fast bacilli; MC methylcellulose; PAM polyacrylamide.

staff in mycobacterial culture and counting colonies.6 We modified this protocol to make artificial sputum by mixing a 2% MC solution without emulsified egg with a sample of THP-1 cells and mycobacteria. Briefly, 2% MC solution was prepared by dissolving MC in distilled water; 0.5 ml of this solution was immediately added to a cell-mycobacteria mixture and mixed thoroughly. The final concentration of MC then became 1% and the final product was incubated at room temperature for about 1 h (Figure 2A).

Polyacrylamide method For the polyacrylamide (PAM) method, a polyacrylamide stock was prepared by mixing 22.2 g acrylamide, 0.6 g N,N’-methylenebisacrylamide and distilled water to 100 ml. The stock was stored at 4C, protected from light. PAM was made from 1.75 ml polyacrylamide stock, 6.25 ml distilled water, 2 ml TBE buffer (pH 8.0), and 100 l 10% ammonium peroxodisulfate. After mixing well, 8 l of TEMED, a polymerisation accelerator, was added and the mixture was mixed; 0.5 ml of this mixture was immediately added to a cell-mycobacteria sample and mixed thoroughly. The remainder of the PAM mixture containing TEMED was kept for several minutes to confirm proper polymerisation. The mixture containing THP-1 cells, mycobacteria and PAM was left for 1–2 days after final mixing at room temperature on the laboratory table in ordinary room lighting until it became viscous (Figure 2B). Smear examinations Panel test smears of both real and artificial sputum were prepared with a microbiological loop and stained with Ziehl-Neelsen (ZN) and auramine-rhodamine

fluorescent (AO) stains to evaluate the staining properties and AFB contents of these preparations. Observations of ZN and AO stained smears were performed at magnifications of respectively 1000 and 400. The size, evenness, thickness and staining of the smears were scored according to the standard described by Fujiki.7

RESULTS Although the preparation using the NaOH and NALC methods was simple and rapid, the final product after treatment with these chemicals was watery and not viscous, and the sedimented cell debris was powdery (Figure 3A and B). The appearance of these artificial sputum smears was so thin and sparse that it was easy to distinguish them from real smears (Figures 3A, B and E). Furthermore, the NaOH and NALC smears were easier to examine than actual specimens because the slides were clear, with no fibrous substances (Figure 4). Artificial sputum preparation by the MC method was easy, but the final mixture of MC, THP-1 cells and mycobacteria was not as viscous as real sputum and macroscopically the smears were clearly distinguishable from those with real sputum (Figure 3C). However, the mixture of THP-1 cells, mycobacteria and PAM was as viscous as real sputum, and macroscopically the smear preparations after ZN staining could not be distinguished from real smears (Figure 3D). The staining properties of the smears were also compared. The results of microscopic observation of the slides are shown in Figure 4. Although the NaOH and NALC slides contained cells and AFB, there was less fibrous substance compared to real sputum smears. On the other hand, there was more fibrous substance

902


Figure 3 Macroscopic appearances of sample solutions and smears prepared by the A) NaOH method, B) NALC method, C) MC method, D) PAM method, and by E) a smear directly prepared from the actual sputum specimen. NaOH sodium hydroxide; NALC N-acetyl-L-cysteine; MC methylcellulose; PAM polyacrylamide. These images may be viewed online in colour at: http:// www.ingentaconnect.com/content/iuatld/ijtld

Figure 4 Microscopic appearance of smears prepared by the four methods used in these studies. Note that only the PAM smear had fibrous contents as well as cells and AFB. NaOH sodium hydroxide; NALC N-acetyl-L-cysteine; MC methylcellulose; PAM polyacrylamide; AFB acid-fast bacilli. These images may be viewed online in colour at: http://www.ingentaconnect. com/content/iuatld/ijtld


903

Figure 5 Comparison of ZN and fluorescent staining in real sputum and artificial sputum prepared by the PAM method. ZN Ziehl-Neelsen; AFB acid-fast bacilli; PAM polyacrylamide. This image may be viewed online in colour at http://www.ingentaconnect.com/content/iuatld/ijtld

in MC and PAM smears, as well as cells and AFB. In particular, staining of smears by the PAM method was satisfactory and comparable to actual sputum smears, not only for ZN staining but also for fluorescent staining with auramine O and rhodamine B (Figure 5). Panel test slides must include slides with different positivity grades, i.e., 3, 2, 1, and negative. To provide these grades, AFB was cultured in Middlebrook 7H9 broth for 2 weeks followed by filtration with an Acrodisk syringe filter (5.0 m pore size, Pall Corp, UK) to disperse mycobacterial clumps. The filtration procedure gave well-dispersed mycobacteria samples. Dilution of this sample to 5 106 cfu/ml Table 1

gave a 2 mycobacterial sample. Samples for 1 and smears were prepared using 10-fold and 100fold dilutions of the 2 sample, respectively. The procedures and properties of the samples produced by the NaOH, NALC, MC and PAM methods are summarised and compared in Table 1. The positivity of each grade of PAM smears was highly consistent when 100 visual fields per smear were examined (Table 2).

DISCUSSION One of the main reasons for the rare use of panel testing is the difficulty in obtaining appropriate prepara-

Comparison of the properties of artificial sputum prepared by the NaOH, NALC, MC and PAM methods NaOH

NALC

MC

AFB 2, 3 and negative sputum Virulent clinical strain in sputum Materials: low Smear: none Difficult

AFB 2, 3 and negative sputum Virulent clinical strain in sputum Materials: low Smear: none Difficult

Cultured cells and MC

100 50 100 80 Mainly cells

100 50 100 80 Mainly cells

Viscosity General similarity to sputum Reproducibility Quality control of panel test slides

No Less Low Difficult

Suitability as a panel test slide

Low

Materials Source of mycobacteria Biohazard level Adjustment of bacterial contents Macroscopic appearance (%)* Size Thickness Evenness Staining Microscopic appearance

PAM gel

Cultured avirulent strain None

Cultured cells and PAM Cultured avirulent strain None

Easy

Easy

No Less Low Difficult

100 80 80 80 Cells and mucus-like component Medium Less Medium Easy

100 100 100 100 Cells and mucus-like component High High High Easy

Low

Low

High

* Scored according to the standard described by Fujiki.7 NaOH sodium hydroxide; NALC N-acetyl-L-cysteine; MC methylcellulose; PAM polyacrylamide; AFB acid-fast bacilli.

904 Table 2


Consistency in smear slides within each grade

Grade*

1

2

3

Mean, n SD

3.5 (23) 1.6

46.2 (25) 18.9

520.7 (24) 162.1

1000 (19) NC

Average number of AFB per 100 fields in smears observed by four trained technicians, except for the 3 smears in which counting was stopped when 1000 AFB were seen. * See text for definitions. SD standard deviation; NC not calculated; AFB acid-fast bacilli.

tions.2 This study aimed at resolving these difficulties using readily available materials to develop an artificial sputum preparation. The NaOH and NALC methods were simple and easy, but they required real AFB (TB bacilli) positive sputum,5 which contain viable M. tuberculosis and may be hazardous before fixation. Another difficulty with these methods is that they require collection of a large amount of AFB-negative sputum to dilute AFBpositive sputum to make the smear positivity grades. In addition, although the properties of the smears prepared by these methods had the same number of mycobacteria as real smears, the processed samples had lost viscosity and had little fibrous substance. It was also not easy to obtain reproducibly similar smears (data not shown). There were some advantages in the MC and PAM methods. They do not present a biohazard because only cultured avirulent AFB is used. These methods are also convenient, as smears can be prepared within 3 weeks after a panel is requested. Furthermore, the smear properties are highly reproducible because the number of AFB can be easily adjusted. Methylcellulose was originally introduced to make artificial specimens for proficiency testing of mycobacterial cultures.6 In this study, we tried to make artificial sputum for panel test slides with MC. The mixture of MC, THP-1 cells and mycobacteria provided a viscous solution, but the smears were not macroscopically or microscopically similar to real sputum. The PAM method showed several additional advantages. First, the components could be mixed thoroughly under non-viscous conditions, because the viscosity was produced by photopolymerisation. Second, the macroscopic appearance of the smears before and after ZN staining and their microscopic appearance were similar to those of real sputum smears, in which both cells and fibrous substances are in the background. Third, smears prepared with this method could be examined by both ZN and fluorescent staining. The toxicity of acrylamide (monomer) is well known. However, technicians do not have to use special protective devices besides gloves and masks, which are used in routine smear preparation, and most steps for preparing PAM can be performed at a biosafety cabinet. Furthermore, the mixture of acyrylamide and methylenbisacrylamide, the main components of PAM, can be prepared in volumes of more than 100 ml and

stored at 4C for at least several months. Only 1.75 ml of this mixture is used to make 10 ml of PAM, which is sufficient to prepare artificial sputum for 2000 smears. The risk of exposing technicians to the toxic materials can therefore be infrequent. It is not necessary to prepare PAM from the initial materials at the peripheral laboratory. Because the mixture of acrylamide and methylenbisacrylamide can be stored for several months at 4C, the WHO Supranational Reference Laboratory would be able to send the prepared mixture to each national and/or peripheral laboratory. This procedure is simple and would be less harmful for peripheral laboratory technicians. Alternatively, another solution would be for the PAM mixture to be made at the Supranational or National Reference Laboratory, and to send it to the peripheral laboratories so that only the smears are prepared by the peripheral laboratory technicians. This scenario would present the lowest risk of toxicity. PAM- and MC-based smears are stable in heating for fixation and chemicals for staining, and can be stored for at least 1 year at room temperature before and after staining. However, as MC-based artificial sputum comprises natural materials such as cellulose, cells and BCG Pasteur with no preservative reagent, its shelf life may be shorter than that of PAM-based smears. Another important issue is the consistency of the slides. Toman reported that on average 106 bacilli/ml sputum was required to observe one AFB per field,8 which is equivalent to a 2 smear. On the other hand, it has already been demonstrated that there are approximately 104 AFB/ml when one AFB is detected by observing 100 fields under an oil-immersion lens, which is equivalent to a smear. In the present study, we used a mycobacterial preparation of 107 cfu/ml to prepare 3 smears, and 10-fold serial dilutions of the 3 preparation to produce 2, 1 and samples containing 106, 105 and 104 mycobacteria/ml, respectively. These values are in agreement with the report by Toman.8 Finally, our artificial slides showed high consistency in AFB numbers in each grade, which was attributable to the simplicity and reproducibility of both the materials and procedures. In conclusion, a novel form of artificial sputum has been developed using a simple, reproducible procedure, and, as real AFB-positive and -negative sputum is not required, with no biosafety problems. Smears made using this method could be used not only for panel testing, but also as a training tool for laboratory technicians. Acknowledgements This work was supported by a Grant for International Health Cooperation Research (15-6) from the Ministry of Health, Labour and Welfare. We thank Drs H Ogata, Fukijuji Hospital (JATA), and Y Kawabe, National Hospital Organization Tokyo Hospital, for their generous cooperation.


References 1 World Health Organization. WHO media center, Fact sheet no. 104, 2005. Geneva, Switzerland: WHO, 2005. http://www.who. int/mediacentre/factsheets/fs104/en/index.html Accessed May 2006. 2 Lan N T N, Wells C D, Binkin N J, Becerra J E, Linh P D, Cô N V. Quality control of smear microscopy for acid-fast bacilli: the case for blinded re-reading. Int J Tuberc Lung Dis 1999; 3: 55–61. 3 Martinez-Guarneros A, Balandrano-Campos S, Solano-Ceh M A, et al. Implementation of proficiency testing in conjunction with a rechecking system for external quality assurance in tuberculosis laboratories in Mexico. Int J Tuberc Lung Dis 2003; 7: 516–521. 4 Tuberculosis Division. Tuberculosis bacteriology—priorities and

5

6

7 8

905

indications in high prevalence countries: position of the technical staff of the Tuberculosis Division of the International Union Against Tuberculosis and Lung Disease. Int J Tuberc Lung Dis 2005; 9: 355–361. Aziz, M A, Ba F, Becx-Bleumink M, et al. External quality assessment for AFB smear microscopy. Washington, DC, USA: Association for Public Health Laboratories, 2002: pp 1–111. de Kantor I N, Kim S J, Frieden T R, et al. Laboratory services in tuberculosis control. Organization and management. Part I. Geneva, Switzerland: WHO, 1998: pp 42–43. Fujiki A. AFB microscopy training. Tokyo, Japan: The Research Institute of Tuberculosis, 2005: pp 16–23. Toman K. Tuberculosis case-finding and chemotherapy. Questions and answers. Geneva, Switzerland: WHO, 1979: pp 6–7.

RÉSUMÉ

Les tests groupés, le recontrôle croisé, aveugle et l’évaluation sur place sont les trois méthodes de contrôle externe de qualité (EQA) pour l’examen microscopique des frottis à la recherche des bacilles acidorésistants (BAAR). Les tests groupés devraient fournir des données sur la capacité des laboratoires avant de mettre en œuvre un programme de recontrôle, évaluer l’état présent des performances du laboratoire et détecter les problèmes associés à cette performance de diagnostic. A ce jour, on a signalé deux méthodes pour la préparation de tests groupés. Celles-ci utilisaient des crachats réels positifs et négatifs pour les BAAR, traités soit au NaOH, soit à la N-acétyl-L-cystéine (NALC). O B J E C T I F : Evaluer ces méthodes et développer une nouvelle méthode pour préparer des lames de tests groupés avec des crachats artificiels. S C H É M A : Les lames de tests groupés ont été préparées par la méthode au NaOH et à la NALC. Les méthodes CONTEXTE :

nouvelles de préparation de crachats artificiels ont été développées et examinées en utilisant une lignée cellulaire monocytaire en culture, des mycobactéries avirulentes cultivées et comme substrat de la méthylcellulose ou un gel de polyacrylamide. Les frottis préparés par les quatre méthodes ont été comparés. R É S U L T A T S : Les lames de tests groupées préparées par les méthodes au NaOH et à la NALC ne sont similaires ni macroscopiquement ni microscopiquement aux frottis réels. Notre nouveau crachat artificiel est similaire aux crachats réels en matière de viscosité et d’apparence macroscopique et microscopique. Il est également cohérent dans les degrés de positivité au sein du groupe. C O N C L U S I O N : Le crachat artificiel décrit ici pourrait contribuer à l’évaluation externe de qualité et à la formation dans les laboratoires de tuberculose (TB) ou dans les centres de microscopie.

RESUMEN M A R C O D E R E F E R E N C I A : Las baciloscopias en serie, la verificación cruzada con anonimato y la apreciación en el laboratorio son los tres métodos de evaluación externa de la calidad del examen microscópico para bacilos acidorresistentes (BAAR). Las pruebas en serie pueden suministrar datos sobre las capacidades del laboratorio previas a la aplicación del programa de verificación, evalúan el desempeño actual del laboratorio y detectan problemas asociados con el rendimiento diagnóstico. Hasta el presente, se han descrito dos métodos de preparación de los portaobjetos para las pruebas en serie. Estos métodos emplean verdaderas muestras de esputo, positivas o negativos para BAAR, tratadas con NaOH o NALC. O B J E T I V O : Evaluar estos métodos y establecer un nuevo sistema con el fin de preparar los portaobjetos para las pruebas en serie con esputo artificial. D I S E Ñ O : Los portaobjetos para las pruebas en serie se

prepararon por los métodos del NaOH y la NALC. Se concibieron y evaluaron nuevos métodos de preparación con esputo artificial utilizando una línea celular de monocitos, micobacterias avirulentas cultivadas y metilcelulosa o gel de poliacrilamida como sustratos. Se compararon los frotis preparados con los cuatro métodos. R E S U L T A D O S : La apariencia macroscópica y microscópica de los portaobjetos utilizados actualmente para las pruebas en serie es diferente a la apariencia de los frotis clínicos. El nuevo esputo artificial presenta la misma viscosidad del esputo real, el mismo aspecto macroscópico y microscópico y concuerda también con respecto al grado de positividad de las series. C O N C L U S I Ó N : El esputo artificial descrito en este artículo podría contribuir a la evaluación externa de la calidad y al adiestramiento en laboratorios de tuberculosis o en centros de examen microscópico.