Evaluation of a IS6110-Taqman real-time PCR assay to detect Mycobacterium tuberculosis in sputum samples of patients with pulmonary TB

Authors


Correspondence

Laís Ariane de Siqueira Lira, Departament of Imunology, Aggeu Magalhães Research Center - Fiocruz, Recife, Pernambuco, Brazil. E-mail: lais.liraa@gmail.com

Abstract

Aim

Evaluate the IS6110-Taqman system performance in sputum samples from patients with pulmonary tuberculosis from health services in north-eastern Brazil as a diagnostic laboratory tool for pulmonary tuberculosis.

Methods and Results

165 sputum samples from respiratory symptomatic patients were evaluated in the IS6110-TaqMan assay: 66 patients with pulmonary tuberculosis and 99 without TB. When the IS6110-TaqMan assay was evaluated using culture and/or clinical response to the specific treatment as the gold standard, IS6110-TaqMan assay obtained a sensitivity of 87·9% and specificity of 98%. The performance of IS6110-TaqMan assay was also evaluated with the sputum smear microscopy, resulting in a sensitivity of 79·7% and specificity 94·8%.

Conclusions

The IS6110-TaqMan was rapid, sensitive and specific for the diagnosis of pulmonary TB.

Significance and Impact of the Study

IS6110-TaqMan assay is a promising auxiliary tool for the diagnosis of pulmonary TB when used in conjunction with routine laboratory tests, clinical and epidemiological criteria of the patient, thus increasing the sensitivity and specificity of diagnosis.

Introduction

Tuberculosis (TB) is an infectious disease caused by Mycobacterium tuberculosis. It has been estimated that 8·8 million cases of TB occurred worldwide in 2010 alone, resulting in approximately 1·4 million deaths, most of which were in low- and middle-income countries (World Health Organization 2011). Tuberculosis is a major cause of morbidity and mortality in Brazil occupying the 19th place among the 22 countries responsible for 80% of all TB cases in the world, with an incidence rate of 47·2 cases per 100 000 inhabitants. The state of Pernambuco, located in the north east of the country, occupies the 3rd place in incidence, and Recife is the Brazilian capital with the highest rate of TB-specific mortality (Ministry of Health 2012).

The diagnosis of active TB is usually achieved by a combination of clinical and epidemiological features together with laboratory test results. Conventional diagnostic methods have inherent limitations (Khéchine and Drancourt 2011). Sputum smear microscopy shows low sensitivity, and culture is usually time consuming due to the slow growth rate of M. tuberculosis (Chagas et al. 2010; Shinu et al. 2011). This may lead to severe consequences, including false-negative diagnosis, which may increase the risk of M. tuberculosis transmission, TB incidence and resistance to drugs (Helb et al. 2010). TB is a treatable disease, and many fatal cases are associated with delays in diagnosing and starting treatment. Thus, early diagnosis and a prompt start to treatment is a key strategy towards TB management and control (World Health Organization 2011).

The use of molecular biology-based diagnostic tools for detecting and quantifying M. tuberculosis DNA has become increasingly popular in recent years (Helb et al. 2010). Polymerase chain reaction (PCR) assays have proved to be useful in identifying mycobacterial species, detecting resistance to drugs and diagnosing TB, especially in patients with paucibacillary (Portillo-Gómez et al. 2000; Richardson et al. 2009; Helb et al. 2010). Different PCR assays for detecting M. tuberculosis in various clinical samples have been developed in recent years (Rebollo et al. 2006; Cruz et al. 2011). More recently, real-time PCR assays have been used to detect and identify many infectious organisms, including M. tuberculosis (Broccolo et al. 2003; Armand et al. 2011; Yang et al. 2011). The high sensitivity and specificity, along with it taking little time to establish the result and the possibility of quantification, make real-time PCR an attractive auxiliary method for the laboratory diagnosis of TB (Chauhan et al. 2009). The use of such a rapid, precise and specific diagnostic tool has been shown to be promising to evaluate in the pulmonary form of the disease, where roughly 26·7% of cases remain without diagnostic confirmation when traditional methods are used (Mello 2001).

Given the above, this study set out to evaluate the IS6110-Taqman system performance in sputum samples from patients with pulmonary tuberculosis from health services in north eastern Brazil as diagnostic laboratory tool for pulmonary TB.

Material and methods

Study population

Sputum samples were obtained from 165 patients with clinical signs of pulmonary tuberculosis. They were of both sexes, over 15 years old, and had been referred to public clinics in the metropolitan region of Recife, north-eastern Brazil, between December 2010 and January 2012. The diagnosis of TB was based on isolating M. tuberculosis in culture from sputum samples, clinical improvement of the patient after specific treatment and visualization of alcohol-acid resistant bacillus in sputum smear microscopy. All sputum samples were analysed by smear microscopy, culture and real-time PCR with a view to detecting M. tuberculosis, as described below.

Sputum smear microscopy

The presence and quantification of alcohol-acid resistant bacillus in sputum samples were performed by Ziehl–Neelsen staining (Stewart 1953), in the Public Health Laboratory of Recife (Unified Health System), as per the guidelines of the Brazilian Ministry of Health (Ministry of Health 2008).

Specimen processing

The sputum specimens (1–5 ml) were processed in accordance with the modified Petroff's method using 4% NaOH (Kent and Kubica 1985; Ministry of Health 2008).

Culture

Culture was performed in Lowenstein-Jensen (LJ) medium (Liu et al. 1973; Ministry of Health 2008), and mycobacterium species were identified using biochemical tests: selective inhibition by para-nitrobenzoic acid (PNB) and thiophene-2-carboxylic acid hydrazide (TCH), niacin accumulation and heat-stable catalase at 68°C (Ministry of Health 2008).

DNA extraction

DNA was extracted from sputum samples using a QIAamp DNA Mini Kit (Qiagen, Duesseldorf, Germany), as per the manufacturer's instructions. The reference strain of M. tuberculosis (H37Rv) (Cole et al. 1998) was grown in LJ culture medium, and the genomic DNA was extracted and purified using a Genomic Prep™ – Cells and Tissue DNA Isolation Kit (Amersham Biosciences, Piscataway, NJ, USA), as per the manufacturer's instructions.

Plasmid preparation

The genomic DNA of the reference strain of M. tuberculosis (H37Rv) was subjected to amplification by nested PCR, as per Lima et al. (2009) to detect the target IS6110 (Thierry et al. 1990). The amplified target product was subcloned into vector pCR®4 TOPO® using a TOPO TA Cloning Kit® (Invitrogen Life Technologies, Carlsbad, CA, USA), as per the manufacturer's instructions. The recombinant plasmid (pIS6110) was sequenced (ABI Genetic Analyzer 3100, Applied Biosystems, Foster City, CA, USA), and the concentration was measured spectrophotometrically (514 ng μl−1) (Ultraspec 3000, Amersham Pharmacia Biotech, Piscataway, NJ, USA). The resulting product was sequenced to confirm the presence of the cloned target and its similarity to sequences deposited in the GenBank, while any that may have been cross-contaminated with other mycobacteria were discarded, thus yielding 100% specificity, in addition to the complete fragment of M. tuberculosis (data not shown). With this product, a 10-fold dilution series (100 ng to 1 fg) was used to construct a standard curve for each real-time PCR run.

Real-time PCR conditions

The PCR was processed in real time by means of an ABI Prism 7500 Sequence Detection System (Applied Biosystems) using TaqMan-specific probe (and ROX as a passive reference) and TAQM3 oligonucleotide (5′-AGGCGAACCCTGCCCAG-3′) and TAQM4 (5′-GATCGCTGATCCGGCCA-3′), which amplify a target fragment of 122pb (Broccolo et al. 2003). The cycling conditions were performed according to Broccolo et al. (2003), and the results obtained were analysed using an ABI Prism 7500 Sequence Detection System. Milli-Q water negative controls were included in all amplification reactions. All assays were performed in duplicate. The reactions included 1 μl of DNA primers (each of 300 nm), 1 μl of probe (200 nmol l−1), 12·5 μl from a TaqMan® Universal PCR Master Mix kit (Applied Biosystems), 9 μl of DNA template and pure water added up to a final volume of 25 μl.

Additionally, real-time PCR assay targeting the housekeeping gene (β-actin) was performed to assess the quality of the DNA in the sputum samples. PCR conditions were as follows: initial denaturation at 95°C for 10 min, followed by 40 cycles of denaturation at 95°C for 15 s, and annealing and amplification at 60°C for 1 min. PCR reactions included 12·5 μl TaqMan, 2·5 μl of TaqMan® Endogenous Controls (500 nmol l−1 of the probe and 300 nmol l−1 of each primer) and 9 μl of DNA template plus Milli-Q water, for a final volume of 25 μl. Negative controls of Milli-Q water were used in all PCR runs, and assays were performed in duplicate.

Statistical analysis

Culture and/or clinical response to the specific treatment were primarily considered the ‘gold standard’ for performance calculation. Secondarily, the IS6110-TaqMan assay was evaluated with the smear microscopy. The sensitivity and specificity values were calculated to evaluate the system. The Kappa index was used to assess the degree of agreement between tests, and the classification was as follows: Kappa < 0·4 = mild agreement; 0·4 ≤ Kappa < 0·8 = moderate agreement; 0·8 ≤ Kappa < 1 = strong agreement; Kappa = 1 = perfect agreement (Arango 2009). The calculations were performed using spss statistical software for Windows (ver. 18; SPSS, Chicago, IL, USA). A P value of <0·05 was considered significant.

Results

Clinical, epidemiological and laboratory findings

A total of 66 (40%) individuals presented a final diagnosis of pulmonary TB, and 99 (60%) suffered from other respiratory diseases and were classified as the ‘non-TB’ group. Their ages ranged from 17 to 92 years old, their mean age being 41 years old (±14·88). In the group with pulmonary TB (n = 66), 51 subjects were male (77·2%) and 15 female (22·7%). With regard to age, 53 patients were between 17–50 years old, and the age range of the other 13 was from over 50 to 92 years old. 51 (77·2%) patients came from outpatient clinics and 15 (22·7%) from wards of Unified Health System institutions. Forty-seven (71·2%) presented pulmonary parenchyma on X-ray, 7 (10·6%) presented normal X-ray, and X-ray results were not available for 12 (18·2%).

In the group diagnosed as ‘non-TB’, 45 (45·4%) subjects were male and 54 (54·5%) female. Considering the age of the participants, 67 (67·6%) were up to 50 years old and 32 (32·3%) were over 50 years old. As regards from where they had been referred, 87 (87·8%) patients came from outpatient clinics and 12 (12·1%) from public health service wards. Sixteen subjects (16·1%) showed changes that were compatible with other respiratory diseases on X-ray, 78 (78·7%) showed normal X-ray, and 5 (5·0%) did not undergo this examination.

Performance of the IS6110-TaqMan assay with culture and/or clinical response to the specific treatment

All 66 pulmonary TB cases evaluated and the 99 defined as non-TB were confirmed by culture and/or clinical response to the specific treatment. The sensitivity and specificity of IS6110-TaqMan assay in the sputum samples were 87·9% (CI = 80·0–95·8) and 98% (CI = 95·2–100·8), respectively. Culture and/or clinical response to the specific treatment and the IS6110-TaqMan assay showed high agreement between the tests, as demonstrated by the Kappa index of 0·87 (P < 0·0001 Yates corrected) (Table 1).

Table 1. Performance of the IS6110-TaqMan assay on 165 patients suspected of having pulmonary TB with culture and/or clinical response to the specific treatment
Culture and/or clinical response to the specific treatment
IS6110-TaqMan assayTBNon-TBTotalS [95%C.I.]; E [95%C.I.]
Positive58260S = 87·9% (CI = 80·0–95·8)
Negative897105E = 98% (CI = 95·2–100·8)
Total6699165 

Performance of the IS6110-TaqMan assay with the sputum smear microscopy

IS6110-TaqMan assay was also evaluated with the sputum smear microscopy. The sensitivity and specificity of the IS6110-TaqMan assay in the sputum samples were 79·7% (CI = 70·2–89·2) and 94·8% (CI = 90·3–99·2), respectively (Table 2). The smear microscopy and IS6110-TaqMan assay showed good agreement between the tests, as demonstrated by the Kappa index of 0·76 (P < 0·0001).

Table 2. Performance of the IS6110-TaqMan assay on 165 patients suspected of having pulmonary TB with the sputum smear microscopy
Sputum smear microscopy
IS6110-TaqMan assayTBNon-TBTotalS [95%C.I.]; E [95%C.I.]
Positive55560S = 79·7% (CI = 70·2–89·2)
Negative1491105E = 94·8% (CI = 90·3–99·2)
Total6996165 

Discussion

In the clinical routine in Brazil, smear is the most used technique for the diagnosis of pulmonary tuberculosis because it is faster, simple and inexpensive, although it has a limited sensitivity. Despite being considered as a gold standard for the diagnosis of tuberculosis, culture is not widely used in the process of clinical decision to start treatment due to the delay in obtaining the results. Furthermore, culture is performed only in certain cases: when smear is negative, but there is the presence of clinical symptoms compatible with the disease; in patients with a history of prior treatment; in all immunosuppressed patients; in suspected cases of the presence of atypical mycobacteria and with suspected resistance to TB by testing sensitivity to the drugs (Ministry of Health 2008).

When the patient is not identified as having TB by laboratory methods, the clinical symptomatology indicates that treatment should start and TB is confirmed by the therapeutic response. Therefore, this set of criteria would be efficient for confirming the disease. However, it has drawbacks, which means their clinical use is not ideal. Culture takes a long time to produce results and, in many cases, the patient cannot wait 4–8 weeks before starting treatment (Ministry of Health 2011). As clinical and/or smear test results use to be flawed, the physician starts treatment, even in patients not having TB. This may be the cause of resistant strains rising. As a result of clinical and smear test criteria that are flawed, treatment is started, and should the patient not have TB, this can cause the emergence of resistant strains.

From the total 165 sputum samples analysed in the IS6110-TaqMan assay, 60 (36·4%) samples were positive and 105 (63·6%) were negative. The group with pulmonary TB (n = 66), eight clinical samples were negative in the IS6110-TaqMan assay, but they presented positive culture and/or responded to treatment and thus were probably false-negative samples. From these, three had a low number of bacilli (AFB 1 + ) and one smear was negative (Miller et al. 2002; Cleary et al. 2003). In this case, the lack of amplification of the M. tuberculosis DNA using the IS6110-TaqMan assay can be related to the loss of DNA during the extraction process, the presence of inhibitors in the sputum sample (Anjos-filho et al. 2002) or with low copy number of IS6110 target present in the genome of bacillus. The IS6110-TaqMan assay detected the DNA of M. tuberculosis in two samples from patients whose conventional tests (smear and culture) were negative, but were strongly reactive to the Mantoux tuberculin skin test and who had a history of contact with a patient with active tuberculosis. It is likely that these patients had viable circulating bacilli and thus were considered as having TB infection; therefore, they started secondary chemoprophylaxis.

Considering the sputum smear isolated, five of those samples showed to be negative. Four of these samples were positive in the IS6110-TaqMan assay being confirmed as pulmonary TB by culture and, therefore, considered as false negative on sputum smear. One sample showed to negative on sputum smear and culture although positive in the IS6110-TaqMan assay. This patient had a clinical history of M. tuberculosis infection. Analysing the 14 samples with positive sputum smear and negative IS6110-TaqMan assay, we observed that seven samples presented false-positive sputum smear. This false-positive result was confirmed by negative culture test, although the physician started specific treatment even without clinic response elicited by the patient. The performance of IS6110-TaqMan assay was reduced and, therefore, underestimated in face of the results achieved on sputum smear.

In this context in developing countries with high bacillary load, sputum smear is considered diagnostic tool to determine the disease. In this regard, many patients undergo the treatment even not presenting TB which leads to the rise of resistant strains.

Many studies have reported a similar performance of IS6110-TaqMan assay to that found in our study (Broccolo et al. 2003; Lemaître et al. 2004; Armand et al. 2011; Yang et al. 2011). The high specificity found in this study by real-time PCR agrees with the literature where most studies find specificity to be above 95% (Lemaître et al. 2004; Kibiki et al. 2007; Kim et al. 2011; Yang et al. 2011). Molecular methods of DNA amplification proposed for the diagnosis of TB shown to be a more effective approach compared to the traditional methods, thereby facilitating the patient diagnostic and the early establishment of the specific treatment.

The implementation of a real-time PCR system would be important for public health in Brazil, where TB is one of the most prevalent infectious diseases (Hughes et al. 2012), with high rates of morbidity and mortality, especially in patients co-infected with HIV (World Health Organization 2011). It is known that the diagnosis of pulmonary TB in hyperendemic regions should be based on evaluating the clinical, epidemiological, laboratory and therapeutic parameters. However, this study demonstrates that the IS6110-Taqman real-time PCR assay is a promising auxiliary tool for diagnosing active pulmonary TB when used in conjunction with other tests, thus increasing the sensitivity and specificity of the diagnosis.

Acknowledgements

This study was funded by FIOCRUZ/PDTIS, FACEPE and CNPq to whom the authors are grateful. We thank the staff of the Immunoepidemiology laboratory and the professionals involved. We are also grateful to Filipe Dantas Torres for reviewing the article and his contributions and collaboration.

Ancillary