 |
| About Bioline | All Journals | Testimonials | Membership | News |
|
||||||
|
||||||
Indian Journal of Medical Microbiology, Vol. 28, No. 4, October-December, 2010, pp. 303-307 Original Article Comparative evaluation of two polymerase chain reactions targeting different genomic regions to detect Mycobacterium tuberculosis in sputum S Sankar1, B Balakrishnan1, B Nandagopal1, K Thangaraju2, S Natarajan3 1 Division of Biomedical Research, Sri Narayani Hospital and Research Centre, Thirumalaikodi, Sripuram, Vellore - 632 055, Tamil Nadu, India Date of Submission: 17-May-2010 Code Number: mb10095 PMID: 20966559 Abstract Purpose: Tuberculosis remains an important health problem all over the world, especially in resource poor settings like India. The Ziehl-Neelsen (ZN) staining of sputum smear is still the method of choice in the diagnosis of tuberculosis in spite of its low sensitivity and specificity. This paper evaluates comparison of two different polymerase chain reaction (PCR) assays with sputum smear findings to detect Mycobacterium tuberculosis. Materials and Methods: A total of 191 sputum samples were collected from 84 patients attending a tertiary care hospital, who were suspected of having pulmonary tuberculosis, were examined by PCR targeting two different genomic regions, namely, TRC 4 by non-nested format and IS6110 insertion element by nested format in comparison to ZN staining of sputum smears. Results: Among the patients tested, 20.24% (Mid-p 95%CI: 31.5-52.4) were smear positive, 7.14% (Mid-p 95%CI: 2.94-14.26) were positive by TRC 4 PCR and 41.67% (Mid-p 95%CI: 12.7-29.8) were positive by IS6110 nested PCR (nPCR). The median age of overall positive cases was 42 years. Among the nPCR positives, the median for age of rural and peri-urban community was 46 and 32 years, respectively. The kappa coefficient between smear findings and TRC4 PCR findings was 0.27 and an agreement of 0.83 was observed (Z = 2.99; one-tailed P = 0.001). TRC 4 PCR picked two unique positives that were negative by smear and IS6110 nPCR. Conclusion: The non-nested TRC 4 PCR showed inability for accurate detection of M. tuberculosis in sputum samples. The study concluded that the nPCR targeting IS6110 is superior and more sensitive than TRC 4 PCR. Keywords: Mycobacterium tuberculosis, polymerase chain reaction, peri-urban, rural, sputum smear Introduction Tuberculosis causes significant morbidity and mortality all over the world. World Health Organization has recognised India as a major hot-spot region for tuberculosis and as the leading cause of death. Factors such as antibiotic resistance and HIV infection compound the severity and spread of the disease rapidly among communities. The current treatment and control regime of revised national tuberculosis control programme (RNTCP) has achieved a marginal reduction in the prevalence of pulmonary tuberculosis. [1] Pulmonary tuberculosis is the most common form and Ziehl-Neelsen (ZN) staining of sputum smear is still the method of choice for diagnosis in most of the hospitals and laboratories. The method is not very sensitive and is prone to problems with microscopy competency of even the trained technician. Radiological examination, tuberculin skin test, and immunological assays also are inadequate in sensitivity and specificity. Definitive identification of Mycobacterium tuberculosis by culture in Lowenstein Jensen (LJ) medium after extensive processing of sputum samples, or automated mycobacterium culture system in automated blood culture system such as BacT/Alert 3D takes weeks to give a result. [2] Polymerase chain reaction (PCR) is a known valuable technique for disease diagnosis from a variety of clinical samples. Only a very few studies exist in the literature on the evaluation of sputum samples by PCR compared to acid fast staining, from tropical developing countries. [3],[4] Quintanilla et al. [5] reported the application of a hemi-nested PCR in smear negative and culture positive sputum samples with a high sensitivity. Gengvinji et al. [6] developed a nested PCR (nPCR) for the detection of pulmonary tuberculosis from sputum specimens by targeting 16S rRNA gene. However, the study was not clinically evaluated. Recently, several reports on the detection of M. tuberculosis from sputum specimens using real-time PCR [7],[8],[ 9] and line probe assay [10] have been reported. These assays, though sensitive and specific than conventional PCR, need expertise and cannot be routinely performed due to lack of specialised equipment and expensive instrumentation even in a tertiary care hospital setting in India. The objective of this prospective study is to compare the efficacy of the conventional PCR targeting TRC 4 and nPCR targeting IS6110 to that of sputum smear findings in identification of M. tuberculosis from sputum samples. Materials and Methods Study group A cross-sectional study was carried out on prospectively recruited, self-reporting patients attending a tertiary care hospital situated in rural area of Vellore district, who were clinically suspected as cases of pulmonary tuberculosis. The study period is from mid-April 2009 to mid-Dec 2009. Patients presenting with highly productive cough, fever of any grade and weight loss were designated as TB suspects and considered under study group. All patients investigated for tuberculosis were under the Revised National Tuberculosis Control Programme (RNTCP), Tamil Nadu, Government of India, for the disease management. Sample collection Sputum samples were collected from the patients in a wide-mouth container by a trained study technician with all safety measures. The first sputum was collected at point of contact with the study physician. Samples were collected for next two consecutive days from each patient. A total of 191 sputum samples were collected from 84 patients for both smear and PCR testing. Sputum samples were not collected in sterile containers, as it was not for culture, but in one-time use disposable containers. Sputum processing The collected sputum samples were processed immediately after the collection. Samples were decontaminated and concentrated by modified Petroff's method. Briefly, the samples were decontaminated by mixing 4% NaOH with an equal volume of sputum sample, mixed and placed in water bath (37°C) for 1 hour with intermittent shaking. The sample was then centrifuged at 3500 g for 30 minutes. The supernatant was discarded and the pellet was resuspended in 1 ml of 1Χ phosphate buffered saline (PBS). To inactivate the viability of the bacteria, the samples were inactivated by 56°C for 30 minutes in dry bath. The inactivation of samples was confirmed by culturing a few representative samples in LJ medium and the inactivation found adequate. The inactivated sample was stored at −20°C until DNA extraction. DNA extraction was performed using QIAamp blood DNA extraction kit (Qiagen, Hilden, Germany) as per manufacturer's instructions. PCR targeting TRC 4 A PCR targeting TRC 4 , a conserved repetitive element in M. tuberculosis complex, was performed using specific primers as described previously. [11] The cycling conditions used were 15 minutes initial denaturation at 95°C, 45 seconds denaturation at 95°C, 1 minute annealing at 57°C, 1 minute extension at 72°C and 10 minutes final extension at 72°C. PCR reagents including Hotstar Taq polymerase, dNTPs and nuclease free water were procured from Qiagen (Hilden, Germany). Negative controls replacing the template with nuclease free water were included in every assay. A DNA extracted from M. tuberculosis strain was used as positive control in every run. Utmost precautions were employed for PCR testing, such as flow through, disposable plastic ware, and gloves, filter blocked tips and dedicated micropipettes, as previously described in standard protocols. [12] PCR targeting IS6110 To compare the PCR results, the same DNA samples were subjected to nPCR using another set of primers targeting IS6110. The nPCR reagents, negative controls and positive control included in every run were similar to that of TRC 4 PCR. The forward and reverse external and internal primers used and the cycling conditions were as described in our previous study. [13] A 5 μl product of the first round formed the template for the second round of the nPCR. The PCRs were carried out in Eppendorf thermal cycler (Mastercycler® personal 5332, Hamburg, Germany). Analysis of amplification products An aliquot of 5 μl amplicon was analysed by gel electrophoresis in 2% agarose (Sigma, St. Louis, MO, USA) prepared in Tris-Borate-EDTA buffer containing 0.5 μg/ml of ethidium bromide (Sigma, MO, USA). The gels were examined in a gel documentation system (Genei, Bangalore, India) for a 173-bp TRC 4 product and 198-bp IS6110 amplification product. Criteria for interpretation The criteria of any one of the two or more samples collected from each patient, positive by smear or PCR, was taken as valid. This is due to the inherent variation that may occur in the distribution of the pathogen in sputum samples, a well-recognised phenomenon. Statistical analysis The data were entered in Excel spreadsheet in MS office 2007. The median values were calculated in excel statistical programme. Kappa coefficient, Chi-square test and test for significance were calculated in EpiInfo 6.04d statistical program (CDC, USA). The centre for disease control (CDC) provides this software online free of cost. Results The sputum smear report was reported to the treating physician in real time. The TRC 4 PCR and IS6110 nPCR were done in batch testing on archived DNA samples. A total of 84 patients were tested, among whom 67 (79.76%) were from rural community and 17 (20.24%) were from peri-urban community. In the rural study population, 52 (77.61%) were males and 15 (22.39%) were females, and in peri-urban study population, 9 (52.94%) were males and 8 (47.06%) were females. A total of 191 samples were collected from 84 patients. The other 61 samples were not included in the study because the patient did not come back for the second or third sample testing or the samples were rejected by the study technician due to presence of saliva rather than sputum. Presence of acid fast bacilli in any one of three consecutive day samples was considered as positive for sputum smear. Similarly, specific amplification in any one of three samples was considered positive for PCR. There was variation in the positive status of the samples from the same patient, but all smear positives were IS6110 nPCR positive. Overall, the sputum smear was positive for 17 cases and the findings were graded according to RNTCP guidelines as 1+, 2+ or 3+. Out of 17 positive cases, 10 patients underwent radiological examination. All of them showed positive signs of tuberculosis, such as opacities in Roentgenograms. Out of 84 patient samples, 17 (20.24%) were positive for smear (Mid-p 95%CI: 31.5-52.4) and 6 (7.14%) were positive for TRC 4 PCR (Mid-p 95%CI: 2.94-14.26) (Yates' corrected χ2 = 6.05; P = 0.14). Among the patients, however, 35 (41.67%; Mid-p 95%CI: 12.7-29.8) were positive by IS6110 nPCR. The overall median of the age of study group was 45.5 years. Among the TRC 4 positives, five were males and one was female; five were from rural community and one was from peri-urban community. The findings in patients are shown in [Table - 1]. The median age of overall positive cases was 42 years. Among the nPCR positives, the median for age of rural and peri-urban community were 46 and 32 years, respectively. The median age of males and females were 48 and 40 years, respectively. [Figure - 1] shows the gel picture showing amplification as 173 bp product of TRC 4 . The kappa coefficient between smear findings and TRC 4 PCR findings was 0.27 and an agreement of 0.83 was observed (Z = 2.99; one-tailed P = 0.001). The kappa coefficient between TRC 4 PCR and IS6110 nPCR was 0.08 and an agreement of 0.60 was observed (Z = 1.29; one-tailed P = 0.098). Accuracy indices were not calculated since the gold standard culture method was not followed. The comparison of proportions between TRC 4 and IS6110 was found to be statistically significant (χ2 = 27.1; P < 0.0001), with TRC 4 detecting a considerably lower number. Discussion In our study, 17 (20.24%) were positive for sputum smear and 6 (7.14%) were positive for TRC 4 PCR. This study is on clinically suspected cases of tuberculosis. The nPCR targeting IS6110 was previously validated on buffy coat preparations from blood in comparison to automated BacT/Alert system. [13] Subsequently, this PCR was also used to test sputum samples and the results were very encouraging. [14] So, in the present study the TRC 4 PCR was compared to the previously established IS6110 nPCR. In this series, when two or more consecutive samples were tested from a patient, the tests were positive in one or more of the samples. But, when the individual sample which was positive in the most sensitive assay (nPCR for IS6110), was negative in the TRC 4 PCR, it was retested twice. This was not done for the smear examination as it was deemed to be less sensitive from previous experience. All TRC 4 PCR positive samples were also repeated twice including the sample which was negative by smear and IS6110 nPCR. The results were highly reproducible. The chest X-rays were done on selected patients who could afford the investigation charges. Interestingly, all the smear positive cases had positive signs of tuberculosis in X-rays. There are totally 42 RNTCP centres in Vellore district and the patients reporting to our centre would be mainly fresh cases. In a previous study by Narayanan et al., [11] CSF samples were evaluated using PCR targeting a novel primer set TRC 4 . In that study, PCR had been compared with non-nested IS6110 and clinical TB meningitis cases. It detected 61 positive cases among 67 clinically suspected meningitis cases and 22 negative cases among 29 negative patients correctly. The study also reported very low false negatives compared to IS6110 and has recommended this test as many Indian isolates of M. tuberculosis lack IS6110. In contrast, in our study, the primer set showed poor sensitivity as it detected only 6 positives among 17 smear positive cases. The IS6110 targeted primers, however, detected 35 positive cases which were also smear positive except one patient. The higher sensitivity of IS6110 compared to TRC 4 could be justified that the IS6110 targeting primers used in our study were in nested format, thus obviously increasing the sensitivity and specificity of the detection of M. tuberculosis. Furthermore, in our study, Hotstar Taq polymerase was used for both TRC 4 and IS6110 PCRs, which is known to enhance the detection of M. tuberculosis especially when tested in patient's samples like sputum, avoiding the appearance of non-specific bands due to false annealing. [15] The use of nPCR format, QIAamp DNA extraction kit and Hotstar Taq polymerase for amplification used in the present study has significantly increased the detection rate. This is in accordance with previous reports. [13] Non-nested TRC 4 PCR used in this study was found to be unsuitable for the detection of M. tuberculosis. Since TRC 4 PCR did not improve in its sensitivity though Hotstar Taq polymerase was used, we suggest that nested format for the same region might improve the detection capability. From our study, it is obvious that nPCR targeting IS6110 can thus be used in spite of its low availability of number of insertion copies. M. tuberculosis is reported to have 1-10 copies of IS6110, though it is predominantly used as a target for PCR. [16] The primers used in this study have been previously validated and evaluated in detection of M. tuberculosis in buffy coat samples in patients with suspected mycobacteremia. The lower limit of detection of the nPCR was 0.1 plasmid per microlitre as determined by titration of cloned plasmid. [13] Djelouagji and Drancourt [17] emphasised that complete inactivation of live bacteria is essential for biosafety. In order to completely inactivate live bacteria prior to DNA extraction and PCR, heat inactivation at 56°C for 30 minutes was performed for all samples. We ensured complete inactivation of live bacilli in the inactivated samples by culturing few representative samples in LJ medium and found it optimal. Hosek et al. [18] reviewed the techniques for mycobacterial DNA isolation from different clinical samples, such as by using N-acetyl l-cysteine (NALC)-NaOH, dithiothreitol (DTT) and SDS with proteinase K followed by centrifugation. Among the methods, lysis extraction method was found to be superior. In our study, NALC-NaOH or DTT was not used for concentration and decontamination; we used Petroff's concentration technique due to its wide use. In our opinion, it seems to be a simple and useful technique. Culture in LJ medium is now replaced with liquid broths and detection by automated blood culture systems such as BacT/alert MB, BACTEC 460 TB system. The detection is either by radiometric or non-radiometric methods. The applicability of these systems is evaluated for specific detection of mycobacteria with clinical samples. [19],[20] However, samples such as sputum are to be extensively processed before inoculation and positive signaling broths are needed to confirm species with further commercial probe assays such as Gen-probe technology, which add substantially to the cost factor and its long-drawn-out time to detection and identification. The study patients, if found positive, were immediately considered for Directly Observed treatment Short-Course Chemotherapy (DOTS) either in our hospital or referred to the Government referral hospital close to the patient's home based on the patient's consent. The study is cross-sectional in nature; these data do not exactly reveal the prevalence of tuberculosis in the population since the patients may also visit other hospital for diagnosis. There are totally 42 RNTCP centres in Vellore district. The total population of Vellore district covered by RNTCP is 3.7 million. In 2009, 3238 patients were diagnosed by RNTCP as smear positive. The suspects examined per 100,000 population per quarter numbered 285, with the percentage of smear positive cases among suspects being 8% in Vellore district. In spite of rigorous detection and control strategy of RNTCP in Vellore district, the prevalence of tuberculosis is underestimated due to insufficient detection technique, i.e., sputum examination by ZN staining. In conclusion, our reports state the inability of TRC 4 primers for accurate detection of M. tuberculosis in patient's sputum samples and found nPCR format targeting IS6110 sequence to be superior and highly sensitive compared to TRC 4 in a previously described non-nested PCR format. Acknowledgement The research work was funded by Division of Biomedical Research, Sri Narayani Hospital and Research Centre, and is gratefully acknowledged. References
Copyright 2010 - Indian Journal of Medical Microbiology The following images related to this document are available:Photo images[mb10095t1.jpg] [mb10095f1.jpg] |
| |||||||||
![[Table - 1]](/showimage?mb/photo/mb10095t1.jpg){kind=link}
![[Figure - 1]](/showimage?mb/photo/mb10095f1.jpg){kind=link}