|Home | About | Journals | Submit | Contact Us | Français|
Nucleic acid amplification tests have improved tuberculosis diagnostics considerably. This study evaluates a new amplification test, the GenoType Mycobacteria Direct (GTMD) test, for detection of the Mycobacterium tuberculosis complex, Mycobacterium avium, Mycobacterium intracellulare, Mycobacterium kansasii, and Mycobacterium malmoense directly in 61 sputum samples. Thirty (49.2%) samples were auramine smear positive, and 31 (50.8%) were smear negative. The GTMD results were compared to the Gen-Probe Amplified M. tuberculosis Direct (MTD) test results, using culturing and sequencing of the 16S rRNA gene as reference methods. The GTMD test could identify 28 of 29 samples containing the M. tuberculosis complex and was negative in a sputum sample containing M. intracellulare. The overall sensitivity and specificity results were 93.3% and 90.0% for the GTMD test, respectively, and 93.1% and 93.5% for the MTD test, respectively. The GTMD test is rapid and can be easily included in routine clinical laboratories for the direct detection of the M. tuberculosis complex in smear-positive sputum samples as an adjunct to microscopy and culture. Further studies are needed to evaluate the performance of the GTMD test for the detection of atypical mycobacteria.
Worldwide, tuberculosis (TB) is a major cause of illness and death. WHO estimates that in 2006, 9.2 million new cases and 1.7 million deaths occurred from TB globally (25), and the incidence is increasing. The emergence of multidrug-resistant TB, and recently also extensively drug-resistant TB, and the human immunodeficiency virus-TB coinfection are further worsening the situation, and effort to accelerate progress in global TB control is needed. Important factors for TB control are increased case detection and treatment success rates (25). The slow growth of most pathogenic mycobacteria results in diagnosis and treatment delay and has stimulated the development of nucleic acid amplification (NAA) tests for identification of mycobacteria directly in clinical specimens. NAA tests provide test results within 1 day. In general, the specificity result for NAA tests ranges from 95% to 100% (1, 12, 16, 23), but the sensitivity result, especially for acid-fast bacillus (AFB) smear-negative samples, varies greatly, from 33 to 96% (1, 12, 16, 23). For AFB smear-positive respiratory specimens, the sensitivity level is approximately 95%.
Two direct systems approved by the United States Food and Drug Administration (FDA) for detection of pulmonary TB are commercially available, as follows: the Amplicor Mycobacterium tuberculosis test (Roche Diagnostic Systems, Indianapolis, IN) and the Gen-Probe Amplified M. tuberculosis Direct test (MTD test; Gen-Probe, San Diego, CA). Both tests use the 16S rRNA gene as the target amplification gene. The 16S rRNA gene represents a stable property of microorganisms and is widely used as the target for identifying mycobacterium species. Several studies have confirmed an excellent test proficiency (sensitivity and specificity levels of more than 95%) in AFB smear-positive sputum samples but a reduced sensitivity level (82 to 85%) when applied on AFB smear-negative samples (1, 16, 23, 24). Thus, their use was limited to respiratory smear-positive samples from untreated patients. An enhanced version of the MTD test was later approved for use in both smear-positive and smear-negative specimens (5). A novel, commercially available NAA test for diagnosis of TB directly in patient specimens which has not yet been FDA approved is the BD ProbeTec ET test (Becton Dickinson Diagnostic Systems, Sparks, MD). The test is based on strand-displacement amplification of target sequences in IS6110 and the 16S rRNA gene and has a sensitivity level of 90 to 100% and a specificity level of 92% in smear-positive sputum samples (16). To make the NAA tests more rapid, robust, and applicable in laboratories without substantial technical infrastructure, the following novel NAA tests have been developed: the loop-mediated isothermal amplification (LAMP) test (Eiken Chemical Co., Ltd., Tokyo, Japan) (2, 3), the GeneXpert system (Cepheid, Sunnyvale, CA) (9), and the gold nanoparticle probes assay (21). Simple sample processing, amplification, and detection steps make these NAA tests more applicable in low-income countries with high incidence of TB. However, data on test proficiencies are limited so far. Ongoing studies will show if these rapid molecular tests can be alternatives to the conventional TB diagnostic tests.
Recently, a new DNA strip test for detection of mycobacteria directly in smear-positive and smear-negative respiratory samples has been developed. The GenoType Mycobacteria Direct (GTMD) test (Hain Lifescience GmbH, Nehren, Germany) is based on nucleic acid sequence-based amplification and amplifies single-stranded nucleic acids from the 23S rRNA gene in an isothermal reaction. The biotinylated amplified DNA product is hybridized to specific oligonucleotide probes immobilized on the strip. The GTMD test detects members of the M. tuberculosis complex (MTC), Mycobacterium avium, Mycobacterium intracellulare, Mycobacterium kansasii, and Mycobacterium malmoense directly from decontaminated respiratory specimens, and the result is available within 1 day. Few studies have previously evaluated the GTMD test (7, 15, 20).
The aim of this study was to evaluate the performance of the GTMD test and compare that test to the MTD test. Thus, the GTMD and MTD tests were evaluated for sensitivity and specificity using 61 respiratory specimens from patients suspected to suffer from pulmonary TB. Amplification and sequencing of the 16S rRNA gene of strains isolated from specimen culture (solid and automated liquid media) were used as reference methods.
Sputum samples from 61 patients attending the outpatient department of the Lala Ram Sarup Institute of TB and Respiratory Diseases, New Delhi, India, for whom pulmonary TB was suspected, were included in the study. Forty-six of the 61 (75.4%) patients were male, and the mean age was 38 years. The sputum samples were decontaminated by the N-acetyl l-cysteine (NALC)-NaOH method (13), and the processed sediments were aliquoted into two tubes as follows: 500 μl for the GTMD test mixed with 300 μl inhibition removal buffer (see below) and 450 μl for the MTD test. The sediments were stored at −70°C until the amplification tests were performed. The remainder sediments were inoculated onto two Lowenstein Jensen (LJ) slants and were also inoculated into one MB/BacT bottle and applied on a slide for auramine-rhodamine fluorochrome staining.
A 0.2-ml portion of the processed sediment was inoculated onto two LJ solid agar tubes and incubated for 6 weeks at 36 ± 1°C. LJ slants were inspected weekly for growth, and acid fastness from suspect colonies was confirmed by Ziehl-Neelsen (ZN) staining. Nonmycobacterial colonies on LJ agar was further identified by cultivation on blood agar and biochemical testing.
Culture using the MB/BacT system was undertaken as previously described (18). MB/BacT antibiotic supplement (0.5 ml) was added to the MB/BacT system bottles containing 10 ml of modified Middlebrook 7H9 broth enriched with casein, bovine serum albumin, and catalase shortly before specimen inoculation (0.5 ml). MB/BacT bottles were incubated at 37°C and automatically monitored continuously for 6 weeks. Bottles contaminated by nonmycobacterial microorganisms were decontaminated a second time using the NALC-NaOH procedure and reincubated in a new MB/BacT system bottle. Samples for ZN staining were drawn from positive MB/BacT bottles (growth index, ≥100) and subcultured onto LJ agar. Acid fastness from suspect colonies on LJ agar was confirmed by ZN staining, and the species were identified by amplification and sequencing of the 16S rRNA gene.
The Gen-Probe MTD test (Gen-Probe Inc., San Diego, CA) was performed and interpreted according to the instructions supplied by the manufacturer (amplified M. Tuberculosis direct test for in vitro diagnostic use—50 test kit revised package insert; Gen-Probe, San Diego, CA). Positive and negative amplification controls were included in every run. The positive control was prepared from a 104 to 105 dilution of a 1 McFarland nephelometric standard suspension of M. tuberculosis ATCC 27294, and the negative control was made from a similarly prepared suspension of an M. avium isolate. A cutoff value of ≥500,000 relative light units (RLU) was used for positive specimens. The MTD test was repeated when a test value in the indeterminate range (30,000 to 499,999 RLU) appeared. A repeat MTD result of >30,000 RLU was considered as positive test result (amplified M. Tuberculosis direct test for in vitro diagnostic use—50 test kit revised package insert; Gen-Probe Inc., San Diego, CA).
The GTMD test procedure is divided into three parts: RNA isolation using a magnetic beads capture method, amplification based on the nucleic acid sequence-based amplification technique, and reverse hybridization of the amplified products with specific probes which are immobilized as parallel lines on a membrane strip. The test was performed according to the manufacturer's instructions.
Discrepant results in the MTD or GTMD test were retested in both tests evaluated. Test-negative specimens, which turned out to be positive when frozen aliquots were retested, were considered false negatives according to their first test, while test-positive specimens, which turned out to be negative when retested, were considered false positives. Specimens showing internal amplification control (IAC) inhibition by the GTMD test were also retested.
AFB smear-positive colonies from the LJ slants were identified at the species level by amplification and sequencing of the 16S rRNA gene, as previously described (11). Briefly, DNA was extracted from cells in 3- to 4-week-old mycobacterial cultures, using a boiling technique. A 1-μl loopful of cells was suspended in 200 μl TE buffer at pH 8.0 (10 mM Tris-Cl, 1 mM EDTA) and heat killed by incubation at 95°C for 20 min. The supernatant containing the extracted DNA was collected by centrifugation at 12,000 rpm for 7 min. The 16S rRNA gene was amplified from each isolate using the following universal 16S rRNA primers: pA (forward, 5′ AGA GTT TGA TCC TGG CTC AG 3′) (6) and pI (reverse, 5′ TGC ACA CAG GCC ACA AGG GA 3′) (11). The expected size of the PCR product was 1,019 bp, which included the full length of the 16S rRNA gene as well as upstream and downstream sequences.
PCR amplifications were carried out in a GeneAmp PCR system 9700 thermocycler (PerkinElmer, Applied Biosystems, CA). Each reaction mixture (50 μl) contained 2.5 μl of the crude DNA extraction, 2.5 μl of each PCR primer, 25 μl HotStarTaq mastermix kit (Qiagen, Germany), and 17.5 μl H2O. The reaction mixtures were subject to 15 min at 95°C, followed by 30 cycles of 40 s at 94°C, 60 s at 58°C, 40 s at 72°C, and terminated by 15 min at 72°C. Successful gene amplifications were confirmed by the Bioanalyzer DNA 1000 chip kit (Agilent Technologies).
PCR products from each strain were purified with the QIAquick PCR purification kit (Qiagen GmbH, Hilden, Germany), according to the manufacturer's instructions. The purified PCR product was subjected to a sequencing reaction using a BigDye Terminator cycle sequencing kit. The gene amplicons were sequenced using the pB primer (TAA CAC ATG CAA GTC GAA CG) (6). The amplification profile consisted of an initial 5 min of denaturation at 96°C, 30 cycles of 96°C for 30 s, 50°C for 15 s, 60°C for 4 min, and elongation time of 72°C for 15 min. Nucleotide sequences were analyzed using CHROMAS software (Technelysium Ltd).
The 61 sputum samples included in the study were evaluated by auramine smear examination, culture on LJ agar and in the MB/BacT liquid system, and by the MTD and GTMD tests. Sequencing of the 16S rRNA gene from DNA from positive cultures (LJ agar or MB/BacT) was considered the reference method. Thirty (49.2%) samples were auramine smear positive, and 31 (50.8%) were smear negative. One of the 30 auramine smear-positive samples was positive for the MTC in the MTD and GTMD tests but was contaminated by a nonmycobacterial strain in culture. Repeated NALC-NaOH decontamination procedures did not result in a positive mycobacterial culture, and the sample was excluded from further analysis.
Thirty-four out of 60 samples provided a positive mycobacterial culture result, while 26 samples were negative. Twenty-seven of the 29 auramine smear-positive samples showed growth of a single mycobacterial species (26 MTC isolates and one Mycobacterium fortuitum isolate), one sample contained two different mycobacterial species (one MTC isolate and one Mycobacterium simiae isolate), and one sample was culture negative. Six of the 31 auramine smear-negative samples were mycobacterial culture positive on solid or liquid medium; all contained a single mycobacterial isolate. Thus, a total of 35 mycobacterial isolates were isolated from 34 mycobacterial growth-positive samples, as follows: MTC (n = 29), M. simiae (n = 2), M. flavescens (n = 2), M. intracellulare (n = 1), and M. fortuitum (n = 1). All isolates were identified by 16S rRNA amplification and sequencing.
Of the 60 respiratory samples included in the analysis of test proficiency, 54 samples had concordant results in the MTD and GTMD tests, and 6 had discordant results, using 16S rRNA sequencing as the reference method. Twenty-seven samples were MTC positive in the MTD test, and isolates from cultures were identified as the MTC by sequencing (Table (Table1).1). Twenty-nine samples were negative in the MTD test and MTC culture negative. Discrepant test results were found for four samples (Table (Table2),2), with three of them being from smear-positive samples. The overall sensitivity level of the MTD test was 93.1%, and the specificity level was 93.5% (Table (Table11).
Twenty-eight of the 60 samples included were MTC positive in the GTMD test, and the isolates from culture were identified as the MTC by sequencing. Twenty-seven samples were negative in the GTMD test and culture negative (Table (Table1).1). Discrepant test results were found for five samples (Table (Table3),3), three of which were smear-negative samples. Using 16S rRNA amplification and sequencing as reference methods, the sensitivity and specificity results of the GTMD test were 93.3% and 90.0%, respectively (Table (Table1).1). One sample had an absent amplification control band, together with absence of any mycobacterium-specific bands on the GTMD strip, and the test result was not valid. A repeat GTMD test was valid, and the test was interpreted to be MTC positive.
Six of the 60 samples studied contained nontuberculous mycobacteria. The MTD and the GTMD tests were both positive in the M. fortuitum-containing specimen (MTD test, 701,822 RLU) and identified the isolate as the MTC (specimen number 4) (Tables (Tables22 and and3).3). The GTMD test was negative when retesting specimen number 4, but the MTD test showed two subsequent indeterminate results (59,863 and 58,011 RLU), interpreted as a positive test result (amplified M. tuberculosis direct test for in vitro diagnostic use—revised package insert for 50-test kit; Gen-Probe Inc., San Diego, CA). The GTMD test did not detect any of the five remaining nontuberculous mycobacteria, whereas the MTD test reported an indeterminate value (56,226 RLU) in a sample containing M. flavescens. A repeat MTD test was negative (3,446 RLU), and the sample was interpreted as being negative. The GTMD test contains an M. intracellulare-specific probe but was negative when an M. intracellulare-containing sample was tested (sample number 56) (Table (Table33).
One of the main challenges to the effective control of the TB epidemic is the time-consuming diagnosis and drug susceptibility testing by culture. Acid-fast microscopy provides results within a couple of hours but requires a 5,000- to 10,000-bacilli/ml specimen for a positive result and is unable to distinguish M. tuberculosis from nontuberculous mycobacteria. Culture performs better than microscopy, with a sensitivity level of 80 to 85% and a specificity level of approximately 98% (1). NAA tests for detection and identification of bacilli directly in pulmonary specimens represent the most rapid laboratory methods for detection of M. tuberculosis and can provide test results within 1 day. A new, direct NAA test for detection of mycobacteria, the GTMD test, has recently been developed. The GTMD test is an isothermal amplification test which has the potential to identify the MTC and four atypical mycobacteria directly in clinical samples.
In this study, the GTMD and the MTD tests were evaluated by testing 61 sputum samples from patients suspected of suffering from pulmonary TB. Cultivation on LJ agar and in the MB/BacT liquid system, followed by amplification and sequencing of the 16S rRNA gene, was used as the reference method. Both the GTMD test and the MTD test had sensitivities of approximately 93% when testing on a battery of mixed smear-positive and smear-negative specimens. The specificity of the MTD test (93.5%) was somewhat higher than that of the GTMD test (90.0%). Three studies have previously evaluated the performance of the GTMD test. In 2006, Franco-Álvarez de Luna et al. (7) tested a total of 134 respiratory and extrapulmonary samples from 65 patients suspected to suffer from TB or another mycobacteriosis. The specimens were a mixture of smear positive and smear negative. The GTMD test had sensitivity and specificity results of 92% and 100%, respectively, using the COBAS Amplicor M. tuberculosis test (Roche Diagnostic Systems) as the reference (7). Seagar et al. reported a sensitivity level of 80.5% and a specificity level of 75% for detection of the MTC, M. avium, M. intracellulare, M. kansasii, and M. malmoense with the GTMD test, using identification of cultured isolates as the reference (20). The GTMD test failed to identify eight atypical mycobacterial isolates (seven M. avium isolates and one M. kansasii isolate) (20). All 22 members of the MTC were detected. Neonakis et al. (15) obtained a sensitivity level of 93.7% and a specificity level of 100% for the GTMD test when 147 samples from 132 patients highly suspicious for TB were analyzed (125 respiratory and 22 nonrespiratory specimens). The sensitivity and specificity results of the MTD test were 89.6% and 100%, respectively (15). Reference tests were cultures in BacT/Alert 3D (Organon Teknika, Durham, NC) and on LJ agar. Among the 147 samples, the GTMD test identified 7 samples positive for atypical mycobacteria (3 M. intracellulare- and 4 M. avium-positive samples), which were confirmed by culture and the GenoType CM and AS assays (Hain Lifescience). The study did not mention the number of specimens containing atypical mycobacteria that the GTMD test failed to identify. The test performance for the GTMD test in our study was slightly reduced compared to the previously reported test performances of the LAMP test (2, 3) and the gold nanoparticle probes assay (21). However, test proficiency data on the LAMP test and the gold nanoparticle probes assay are limited, and for GeneXpert, data are not yet available. The LAMP and GeneXpert tests are currently evaluated with a larger number of clinical samples.
Three specimens were interpreted as false-positive test results in our study (specimen numbers 4, 37, and 51). Both the MTD and the GTMD tests were MTC positive for specimen number 4. However, according to the sequencing results, the specimen contained M. fortuitum and Rhodococcus equi. The false-positive result can be explained by a cross-reaction between the MTC-specific probes and M. fortuitum or R. equi or by overgrowth of a possible M. tuberculosis isolate in the specimen by M. fortuitum and R. equi, resulting in a false-negative test result for M. tuberculosis. Retesting the M. fortuitum isolate supported the possibility of a cross-reaction in the MTD test (positive test result) but not in the GTMD test. False-positive MTD results have previously been reported due to cross-reaction with isolates of M. avium, M. gastri, M. terrae, M. kansasii, and M. celatum (amplified M. Tuberculosis direct test for in vitro diagnostic use—50 test kit revised package insert; Gen-Probe Inc., San Diego, CA) (4, 8, 10, 11), but to our knowledge, there has been no previous report of an M. fortuitum cross-reaction. Specimen numbers 37 and 51 were also interpreted as being false positive for the MTC. These specimens were bacteriologically sterile on culture. The false-positive results could have been due to an accidental mycobacterial contamination between specimens during the test procedure or detection of nonviable bacilli in the NAA test if the patients received anti-TB treatment and were in a noncultivable organism shedding period (14) when the samples were collected. However, none of the patients reported any history of anti-TB therapy. The reference test used in this study was culture, which can give false-negative results (1). The lack of a satisfactory diagnostic gold standard remains one of the biggest obstacles for evaluating new diagnostic tests. The true accuracy of NAA tests evaluated may be higher than reported, since culture is used as the reference standard (12, 22).
There were three specimens with false-negative test results in the study. False-negative test results can be caused by the presence of inhibitors of enzymatic amplification, a suboptimal target extraction, a small number of mycobacteria, or mycobacteria that are unevenly distributed in the test suspension (17). The GTMD test contains an IAC, which monitors the amplification and detection step by using a second target and excludes the possibility of enzymatic inhibitors. Only one sample had an absent amplification control band on the test strip in our study, but upon retesting the specimen, the IAC was positive. In NAA tests without an IAC, a negative test result must be considered carefully because of the potential for false-negative results.
Studies have previously shown that the sensitivities in NAA tests are lower for smear-negative samples than for smear-positive samples (1, 12, 19, 23) and that the sensitivities for NAA tests in smear-negative samples are too low and variable to be clinically helpful (19). There were only two auramine smear-negative, culture-positive samples in this study (specimen numbers 39 and 40). Of these, one could be detected by the MTD and GTMD tests. Only one study has previously evaluated the GTMD test with smear-negative samples (7), but they have not reported the results for smear-negative samples separately. A larger study with more smear-negative, culture-positive samples would be required to evaluate the GTMD test performance on smear-negative samples.
One of the main limitations for the implementation of NAA tests in TB epidemic areas is the high costs of NAA tests. The costs for testing one sputum sample were EUR 38.42/55.33 U.S. dollars (USD) for the MTD test and EUR 44.32/63.83 USD for the GTMD test (current prices obtained from distributors in Norway). The COBAS TaqMan M. tuberculosis test (has replaced the Amplicor Mycobacterium tuberculosis test in Europe; Roche Diagnostic Systems) has somewhat lower costs than the MTD and GTMD tests (EUR 25.39/36.56 USD per test performed). The LAMP test is reported to be a rapid and cost-effective alternative; turnaround time is 1 to 1.5 h, and the cost is approximately 1 USD per test performed (2, 3). The reported costs do not include costs associated with quality assurance, staff, workload, training, equipment, machines, building facilities, etc. Moreover, costs associated with hospitalization and isolation, and socioeconomic costs of patients suspected to suffer from TB in the community, have to be taken into consideration when new diagnostics tests are evaluated. The NAA tests provide test results within 1 day rather than weeks and months by culture, and rapid diagnosis and initiation of anti-TB treatment may limit the time needed for isolation and hospitalization, as well as the time of transmission of the bacilli to noninfected persons and the progression of TB to more serious disease stages.
In summary, the GTMD test had excellent test proficiency for detection of the MTC, interpretation of results was simple, test results were available in 5 h, and the test could be easily included in the routine work schedule of a diagnostic laboratory. Moreover, the GTMD test includes an IAC to identify sample inhibitors. A disadvantage of using NAA tests is the considerably high cost for each specimen evaluated, which reduces the applicability of these tests in low-income countries. The GTMD test can be recommended for use in routine clinical laboratories for rapid detection of the MTC directly from smear-positive sputum samples as an adjunct to smears and cultures. Further studies are needed to evaluate the performance of the GTMD test in detection of atypical mycobacteria.
We thank Grete Hopland for support in the laboratory. GTMD test kits for this study were kindly provided by Hain Lifescience.
The study was supported financially by the Research Council of Norway and the Western Norway Regional Health Authority.
Published ahead of print on 30 September 2009.