Multiprimer PCR System Diagnosis of Pulmonary Tuberculosis in Cochabamba, Bolivia

Bolivia has one of the highest incidence rates of tuberculosis (TB) in the Americas. An estimated 15,000 new cases per year are detected (1), which corresponds to an incidence rate of 112 cases per 100,000 population; 1,600 deaths due to TB are reported to occur annually (10). The actual figures are likely to be higher, because similar to that in many other developing countries, routine TB diagnosis in Bolivia is still based primarily on cultures or microscopic examination of specimens taken from individuals suspected of having TB. The drawbacks of these approaches are that (i) neither culture nor microscopy allows for the differentiation of Mycobacterium species; (ii) a large number of Mycobacterium bacilli are required for positive identification (i.e., microscopy has poor sensitivity); (iii) though specific, cultures are time-consuming; and (iv) the sensitivities and specificities of both diagnostic approaches are dependent on the expertise of the laboratory personnel (e.g., with regard to specimen collection, processing, and evaluation). PCR has become a much-used approach for the detection of mycobacteria among cultured strains and in uncultured clinical samples (5, 9), and a range of PCR protocols to detect Mycobacterium tuberculosis have been developed and tested previously (3, 6, 8). However, few protocols have been evaluated, let alone operationalized, in a developing-country setting.

We used a previously published multiprimer system PCR (MS-PCR) protocol (4) to detect Mycobacterium spp. in sputum samples of suspected TB patients attending a regional TB laboratory in Cochabamba, Bolivia. The tested PCR protocol was compared to routine culture and microscopy according to the guidelines of the Bolivian TB Control Program (2); i.e., cultures were grown on Lowenstein-Jensen medium at 37°C for 60 days, and slides were processed with Ziehl-Neelsen stain. For MS-PCR, samples were decontaminated using the Petroff method (7) prior to storage at −20°C. Unlike Del Portillo et al. (4), we extracted DNA from samples by using a commercial DNA extraction kit (DNAzol; Invitrogen, Carlsbad, CA). All samples were amplified with the following primers under previously described conditions (4): PT1 and PT2 (species-specific primers targeting the MTP40 gene), MT1 and MT2 (genus-specific primers targeting the gene coding for the 32-kDa alpha antigen), and IS5 and IS6 (complex-specific primers targeting the IS6110 insertion sequence). Each amplification cycle included negative (no DNA or DNA from an uninfected person and a dog) and positive (water-lysate mixtures of reference strain cultures) controls. Amplification products (10 μl) were visualized under UV light after electrophoresis on 1.5% agarose gels. To avoid cross contamination, separate areas were used for DNA extraction, PCR sample preparation, and amplification; PCR-grade H₂O was used throughout.

Between July and September 2004, of 284 patients attending the regional TB laboratory with suspected pulmonary TB, 63 patients satisfied the study inclusion criteria (i.e., they were >15 years of age, provided nonsalival sputum samples, and had a productive cough for 3 weeks prior to the date of study inclusion). Because of logistical and technical constraints, only 33 of these patients were randomly selected to participate in our study and had sputum samples taken for study purposes. Among the tested study samples, 11 (33%), 7 (21%), and 11 (33%) of 33 tested positive by MS-PCR, microscopy, and culture, respectively. Assuming that culture is the diagnostic “gold standard,” both microscopy and MS-PCR were 100% specific. The sensitivities of MS-PCR and microscopy were 100 and 67%, respectively. The MS-PCR protocol used clearly identified all clinical samples positive for M. tuberculosis, with the characteristic species- and genus-specific bands readily observed in all cases.

For pulmonary TB, PCR has previously been shown to have sensitivity between 77 and 95% and specificity between 95 and 100% (3, 6, 8), the fluctuation being due to whether the samples tested were compared to samples from clinically suspected cases with or without positive microscopy or culture results. Our study shows that an MS-PCR approach can readily be implemented in Bolivia. Based on our results, we intend to carry out further studies to develop a novel diagnostic algorithm including MS-PCR to allow for prompt and cost-effective diagnosis of TB patients in Cochabamba. Such an algorithm could then be scaled up to a national level in order to contribute to improved patient management and better surveillance of the disease in the country.