Diagnostic approaches in patients with tuberculous pleural effusion.

The conventional bacteriologic methods used for diagnosing pleural tuberculosis are less sensitive and time consuming. The objective of this study was to develop nonbacteriologic methods and to assess their potential utilities for the rapid diagnosis, especially in smear/culture-negative patients. One hundred forty patients with pleural effusion were investigated for tuberculous etiology by bacteriologic methods. Mycobacterium tuberculosis in the pleural fluid specimens was isolated in 11 patients. To establish a tuberculous etiology in the remaining 129 patients, we performed the following assays: a) estimation of tuberculosis-associated glycolipid antigen (TBGL) by a modified indirect enzyme-linked immunosorbent assay (ELISA), b) an immunocytochemical method for the demonstration of TBGL antigen in the Cytospin smears, and c) detection of mycobacterial DNA by polymerase chain reaction (PCR). Estimation of TBGL antigen by ELISA showed 100% specificity and overall 85.5% sensitivity. Immunocytochemistry could be applied only in those samples with adequate number of macrophages. PCR carried sensitivity and specificity of 87% and 93%, respectively. Estimation of TBGL antigen in pleural fluid specimens by ELISA has a definite role in establishing tuberculous etiology, particularly in those patients in whom bacteriologic methods did not demonstrate M. tuberculosis and also in those in whom a distinction between tuberculous and nontuberculous etiology was not possible based on the clinical and radiologic features of the thorax.

Combined use of Amplified Fragment Length Polymorphism and IS6110-RFLP in fingerprinting clinical isolates of Mycobacterium tuberculosis from Kerala, South India.

BACKGROUND: DNA fingerprinting by IS6110-RFLP has shown a high incidence of Mycobacterium tuberculosis isolates having no and low copies of the insertion sequence in Kerala, South India. Amplified Fragment Length Polymorphism (AFLP) would scan the entire genome rather than a few repetitive elements, we thought that this technique would help us in differentiating the large reservoir of isolates from an endemic region. Here we evaluate the ability of Amplified Fragment Length Polymorphism (AFLP) to type clinical isolates. METHODS: Fifty clinical isolates of M. tuberculosis were analysed by conventional radioactive AFLP and IS6110- RFLP. M. bovis, M. bovis BCG and two non tuberculous mycobacteria were also analysed to see species specific differences generated by AFLP. Cluster analysis was performed using the AFLP profile that showed the maximum polymorphism within M. tuberculosis and this was compared to the number of copies of IS6110 insertions. RESULTS: For AFLP, out of ten primer pairs tested, the EO/MC pair generated maximum polymorphism among the clinical isolates of M. tuberculosis. The similarity between the isolates ranged between 88 and 99.5%. Majority (nearly 85%) of the 'low copy' IS6110 isolates clustered together, while the rest clustered irrespective of the copy numbers. AFLP could show rare differences between isolates of M. tuberculosis, M. bovis and M. bovis BCG. The AFLP profiles for non-tuberculous mycobacteria were highly different from those of M. tuberculosis. CONCLUSION: Polymorphism generated by AFLP within the M. tuberculosis species is limited and hence AFLP alone seems to have limited use in fingerprinting the isolates in Kerala. The combined use of AFLP and IS6110-RFLP showed relatively better differentiation of 'high copy' IS6110 isolates, but failed to differentiate the 'low copy' isolates. However, the technique may be efficient in inter-species differentiation, and hence potentially useful in identifying and developing species-specific markers.