| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
 |
ABSTRACT |
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In developing countries the detection of TB-L with traditional diagnostic tools is a major challenge. Direct smears are widely used but lack sensitivity and specificity. Culture takes at least two weeks, is not always available in countries endemic for TB, and lacks sensitivity in extra-pulmonary TB. Subsequent drug sensitivity testing takes 2–3 additional weeks.5 The diagnosis of TB-L is routinely performed by lymph node cytology or biopsy histology. However, histology and cytology lack specificity due to the difficulty of distinguishing other granulomatous pathologies in the absence of acid-fast bacilli, and the process of biopsy is invasive.
The PCR has been evaluated for more than a decade for the diagnosis of pulmonary and extrapulmonary TB, including TB-L.6–9 Its development has required extensive troubleshooting to avoid false-positive and false-negative results, and reproducibility between laboratories. However, it is now a recognized approach in the routine laboratory setting and commercial kits are available to complement conventional diagnostic assays.10,11 The PCR is routinely used in specimens from the site of infection, such as sputum for pulmonary TB, cerebrospinal fluid for tuberculous meningitis, and lymph node material for TB-L. An alternative approach is the detection of M. tuberculosis or M. avium DNA with a PCR in peripheral blood mononuclear cells (PBMC-PCR) of patients with pulmonary TB.12–17 The hypothesis is that patients with active infection harbor mycobacterial DNA in peripheral scavenging cells such as macrophages.13,15 In TB patients coinfected with human immunodeficiency virus (HIV), the PCR has sensitivities of 95–100% and specificities of 89–100%.12–14,17 In case series with immunocompetent and immunosuppressed patients (HIV negative), the performance of the assay is highly variable, ranging from 33% to 87% depending on the study.15,16,18–20 The use of the PBMC-PCR for the diagnosis of extrapulmonary TB is still incipient but holds promise.16,21
Our study was performed at the Aga Khan University, with an affiliated 650 bed modern, private hospital, the Aga Khan Hospital, in Karachi, Pakistan. It has one of the most modern and well-equipped clinical laboratories in the country and provides rapid and timely histologic and microbiologic diagnostic services throughout Pakistan. One of the daily challenges in the clinical laboratory is to provide an accurate diagnosis of TB-L, the most common form of extrapulmonary TB. In this study, we sought to develop and evaluate a PBMC-PCR assay in patients with a clinical suspicion of TB-L.
|
MATERIALS AND METHODS |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Sampling techniques and specimen handling. The FNAC was performed with the palpable lymph nodes of suspected cases by an experienced pathologist as an outpatient procedure. Drops of the aspirate material were immediately expressed and spread on glass slides, fixed, and stained with Papanicolaou stain, hematoxylin and eosin, and Ziehl-Neelsen stain. The material remaining in the syringe was used directly for M. tuberculosis culture or a lymph node PCR (LN-PCR). Similarly, the biopsy specimens from control patients were processed for histologic analysis, direct smear, M. tuberculosis culture, and LN-PCR. For histology, an aliquot of the specimen was fixed in 10% buffered formalin, and 8-mm sections processed in an automated tissue processor for 12 hours, embedded in paraffin, and stained with hematoxylin and eosin and Ziehl-Neelsen stain. A section of the biopsy was homogenized used for M. tuberculosis culture and LN-PCR analysis. Cytology smears and tissue sections stained with Papanicolaou stain, hematoxylin and eosin, and Ziehl-Neelsen stain were screened by experienced cytotechnologists and results were recorded as suggestive TB-L if there was typical caseating necrosis in the absence of acid fast bacilli or positive TB-L if acid fast bacilli were detected by Ziehl-Neelsen stain. Culture for M. tuberculosis was performed using the Bactec 460 radiometric system (Becton Dickinson, Franklin Lakes, NJ) by incubation at 37°C for six weeks or until a positive reaction was observed. A 10-mL blood specimen was collected in heparinized tubes from all patients and controls for PBMC-PCR analysis. The buffy coat from the heparinized blood was separated using Ficoll Hypaque (density 1.007 grams/ml; Sigma, St. Louis, MO) and PBMCs were collected for PCR.
PCR assays. Extraction of DNA from PBMCs, lymph node FNAC samples, and biopsy specimens was performed as described previously.22 Essentially, cell walls were lysed with lysozyme, followed by proteinase K digestion and sodium dodecyl sulfate treatment of proteins. Proteins and macromolecules were precipitated using NaCl and hexadecyltrimethlyammonium bromide/NaCl solutions. Nucleic acids were recovered from the aqueous phase after extraction with chloroform and isoamyl alcohol. The DNA was precipitated overnight with isopropanol at -20°C. The pellet was washed with ethanol, reconstituted in 50 µL of water, and 10 µL were used in the PCR. Amplification of the DNA was performed with the primers INS-1 5'-CGTGAGGGCATCGAG-GTGGC-3'and INS-2 5'-GCGTAGGCGTCGGTGA-CAAA-3', which amplify a 245-basepair segment of the insertion element IS6110.23 Because some of the M. tuberculosis isolates in Pakistan do not carry this insertion element,24 a second set of primers targeting a 439-basepair fragment of the 65-kD heat shock protein from all mycobacterial species was also tested in all specimens. The primer sequences were TB11 5'-ACCAACGATGGTGTGTCCAT-3'and TB12 5'-CTTGTCGAACCGCATACCCT-3'.25 The PCR was performed using 10 mM Tris, 50 mM KCl, 1.5 mM of each dNTP, 0.4 µM of each primer, and 2.5 units of Taq polymerase. Cycling conditions were denaturation at 94°C for two minutes, annealing at 65°C for two minutes, and extension at 72°C for three minutes, for a total of 40 cycles. The amplified products were subjected to electrophoresis on a 1.5% agarose gel and a positive PCR was evaluated by the detection of the expected amplicons after staining with 0.5 µg/ml of ethidium bromide.
|
RESULTS |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Using a clinical diagnosis of TB-L as a reference, individual analysis of each assay indicated, in order of sensitivity, that the combined positive and suggestive cytology was positive in 77% of the cases, including 17 (35%) of 48 suggestive (granuloma with negative direct smear for M. tuberculosis) and 20 (42%) of 48 positive (granuloma with positive direct smear for M. tuberculosis) cytology (Figure 1
). The second highest sensitive assay was the PBMC-PCR, which was positive in 30 patients (62.5%), with identical results with either the IS6110 and hsp65 targets. Finally, the LN-PCR was positive in 16 (33%) cases, and M. tuberculosis was isolated from 13 (27%) (Figure 1). In the clinical setting, a positive culture is the gold standard for diagnosis (Figures 1 and 2). The next most reliable laboratory results are the combination of culture and direct smear (reported as positive cytology), which increased the sensitivity to 56% (Figure 2). Culture and smear are routinely complemented with cytology (positive and suggestive), increasing the overall sensitivity to 81% in this study. However, 12 (25%) of these cases had a suggestive cytology as the only laboratory evidence for TB-L. The additional contribution of the PBMC-PCR to the routine combination of assays confirmed the suggestive cytology result in nine of these 12 cases, and identified three new patients who were negative by all other assays. Thus, with the addition of the PBMC-PCR, there were 39 (81.25%) positive and 3 (6.25%) suggestive cases, increasing the overall sensitivity to 87.5% (Figure 2). The two assays that provided the highest combined sensitivity were the PBMC-PCR and cytology, which identified 85% (41 of 48) of all clinically suspected cases of TB-L. These two assays were concordant in 68.8% of the cases (26 positive and 7 negative). One of the eight control, non-TB-L patients (based on clinical criteria) had a positive result in the LN-PCR. The results of all other assays were negative in the control patients.
|
|
|
DISCUSSION |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The onset of TB-L is insidious. Patients present with enlarged, usually painless lymph nodes, most commonly affecting the cervical region. Systemic symptoms such as fever, weight loss, and malaise may or may not be present at an early stage. Advanced disease presents additionally with cold abscess with or without sinus formation. Prompt diagnosis and effective treatment leads to reduced morbidity and mortality. To obtain early and accurate diagnosis, rapid laboratory techniques are required. Classic methods based on the detection of M. tuberculosis at the site of infection have low sensitivity. Our results using these established methods bear this out and fall within the range of those reported in the literature. Ziehl-Neelsen staining is rapid and inexpensive, but requires more than 10,000 organisms/mL, cannot be used to distinguish between various members of mycobacteria, and its sensitivity ranges between 10% and 50% depending on the study.7,8,26,27 In our study, acid-fast bacilli were detected in 42% of the cases (reported as positive TB-L cytology). Culture is the gold standard, but takes several weeks, requires 10–100 organisms per sample, and reports from different studies detect in fine needle aspirates or biopsy specimens between 9% and 70% positive in the clinically suspected TB-L cases.7,8,26,27 In our study, 27% of the fine needle aspirates from probable TB-L patients were positive. The recognized poor sensitivity of conventional microbiologic assays7,8 is routinely complemented with additional non-specific assays. In FNAC or histology of surgical lymph node biopsy specimens, diagnosis is based on the presence of granulomas, central necrosis (suggestive TB-L), and if possible, demonstration of acid-fast bacilli by staining of tissue sections (positive TB-L). However, histology lacks specificity since granulomas with necrosis are seen in several infectious and non-infectious conditions, and false-negative Ziehl-Neelsen smears are common (presumably 35% in our study assuming all suggestive cytologies were indeed TB-L). Histology and cytology have a sensitivity of approximately 32–59%, depending on the report,6,8,27 and were 77% positive (combined positive and suggestive) in our hands.
The PCR has gained importance in the clinical diagnosis of M. tuberculosis, and is now an important support in the clinical diagnosis of TB in developed countries.11 Despite its considerable demands for technical skills and equipment, it can also be performed successfully in reference laboratories in developing countries, such as the modern Aga Khan clinical laboratory in Karachi, Pakistan.24 Given the availability of a thermal cycler, the rest of the procedure has a cost similar to other routine assays for TB-L diagnosis. Additional advantages of PCR over conventional methods include its high sensitivity, performance in few hours, and depending on the assay design, ability to differentiate between M. tuberculosis complex and mycobacterial species other than TB, and identification of gene mutations associated with drug resistance.28,29 In the past decade, experimental studies describing lymph node PCRs from fine needle aspirates or biopsy specimens have consistently shown improved sensitivity (61–78% depending on the study) when compared with conventional microbiologic methods.6–8 In the present study, the LN-PCR was positive in 33% of the aspirates from patients. This is a lower sensitivity compared with other reports for TB-L patients, perhaps due to the small amount of aspirate remaining after splitting the specimen for the additional microbiologic and cytologic assays.
In one study, comparative analysis showed biopsies had higher sensitivity than aspirates, presumably due to their larger size.8 However, in the routine clinical setting, aspirates are preferred because their collection is less invasive. At the Aga Khan Hospital, aspirates were routinely performed for TB-L diagnosis, while biopsies were performed for histologic analysis of patients with suspected malignancies. This difference in specimen collection between patients and controls was done because aspirates cannot be performed on solid tumors, and it was not justified to perform biopsies, a more invasive procedure, on patients with caseous lymph nodes. If this aspect of our experimental design affected our results, one would expect biopsies to provide higher sensitivity for M. tuberculosis detection given their larger size compared with aspirates. One of the biopsy samples gave a positive LN-PCR result in the control specimens, but the interpretation of this unexpected result is unclear.
The working hypothesis for the PBMC-PCR is that peripheral macrophages have scavenged M. tuberculosis genomic material from the infected sites, or even whole bacilli. The detection of M. tuberculosis DNA, but our inability to obtain culture-positive blood from the same individuals, suggested low-level bacteremia undetectable by culture, or most likely, amplification of DNA remnants from non-viable mycobacteria, presumably inside phagocytic vacuoles from peripheral macrophages.
We had difficulty obtaining a larger number of lymph node and blood specimens from patients who did not have a clinical suspicion of TB-L, and those sampled were significantly older than the TB patients. To control for this age difference, blood samples for the PBMC-PCR were collected from 40 younger, normal, healthy adults, all of whom were negative. Even though a tuberculin skin test was not performed, the fact that the controls from a population that is highly endemic for TB were negative suggested that a positive PBMC-PCR result is usually associated with active TB infection. Likewise, the HIV or AIDS status of the patients or controls was not available. This infection is thought to be essential in increasing the sensitivity of blood-based PCR assays for pulmonary TB.12–14,17 However, it is unlikely that the patients had AIDS because none presented signs or symptoms suggestive of AIDS, and the prevalence of HIV (0.007%) and AIDS (1,710 cases/140 million population) is extremely low in Pakistan.30 These data suggested that the PBMC-PCR can provide a high sensitivity for the diagnosis of TB-L patients in non-AIDS populations.
Our results, combined with findings from a previous study,16,21 encourage us to test the PBMC-PCR in the diagnosis of other forms of extrapulmonary TB where the infections are usually paucibacillary, have an extensive differential diagnosis, and access to specimens requires an invasive procedure. As the PCR technology improves, new advances allowing real time assays, high sensitivity and specificity, and high throughput at low cost open up improved possibilities for diagnosis.
Received April 16, 2003. Accepted for publication July 14, 2003.
Acknowledgments: We thank the field staff of the Urban Health Program for their help during data collection, and the staff of the research laboratories of Aga Khan University for their participation in the study.
Financial support: This project was supported in part by the Aga Khan Health Services of Pakistan.
Authors’ addresses: Shaper Mirza, Department of Microbiology, University of Alabama, Birmingham, Bevill Building, 845 19th South Street, Birmingham, AL 35294, Telephone: 205-934-1923, Fax: 205-934-0605. Blanca I. Restrepo, Joseph B. McCormick, and Susan P. Fisher-Hoch, University of Texas Health Science Center Houston School of Public Health-Brownsville, The University of Texas at Brownsville, 80 Fort Brown SPH, 2nd Floor Room 2.202, Brownsville, TX 78520.
Reprint requests: Blanca I. Restrepo, University of Texas Health Science Center Houston School of Public Health, The University of Texas at Brownsville, 80 Fort Brown, SPH 2nd Floor, Room 2.202 Brownsville, TX 78520, Telephone: 956-279-3841, Fax: 956-554-5152, E-mail: birestrepo{at}utb.edu.
|
REFERENCES |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  P. Kumar, K. Nath, B. Rath, M. K. Sen, P. Vishalakshi, D. S. Chauhan, V. M. Katoch, S. Singh, S. Tyagi, V. Sreenivas, et al. Visual Format for Detection of Mycobacterium tuberculosis and M. bovis in Clinical Samples Using Molecular Beacons J. Mol. Diagn., September 1, 2009; 11(5): 430 - 438. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
