RESUMEN
BACKGROUND: Currently, many laboratories have selected several different methods for the detection of M. tuberculosis in the sputum. To select efficient method for clinical laboratories among the various methods, we compared the results of several methods. METHODS: Total 72 sputums were examined by the six combinations of stain methods. The samples were constructed as follows on the result of direct smear ZN stain; negatives (26), traces (3), 1+(9), 2+(12), 3+(12) and 4+(10). The true positives were determined after close evaluation of the clinical, radiological and other laboratory findings. RESULTS: The sensitivities and specificities of each methods were as follows; direct smear ZN stain were 83.6% and 100%, direct smear Auramine stain were 90.9% and 100%, centrifugation ZN stain were 94.6% and 100%, centrifugation Auramine stain were 98.2% and 94.1%, cytocentrifugation ZN stain were 96.4% and 100%, cytocentrifugation Auramine stain were 100% and 64.7%, nested PCR were 80% and 94.1% and culture were 67.3% and 100% respectively. CONCLUSIONS: Concentration method by centrifugation is suitable for routine laboratory if enough centrifugal force were engaged. Auramine stain is more suitable staining method than ZN stain in direct smear but not in concentrated smear because it has the potency of false positivity. The PCR assay is thought to be not only a fast, sensitive method but also a specific method for the direct detection of M. tuberculosis in the sputum. The culture method using Ogawa media is specific but not sensitive.