Interfering substances have been reported to inhibit PCR assays for the direct detection of Mycobacterium tuberculosis in clinical specimens. Using an internal control, we determined that 52% of respiratory specimens interfered with our PCR assay. On the basis of these findings, we tried to circumvent the problem by simply diluting prepared sediments. With sediment from a routinely processed sputum known to be inhibitory to PCR, one aliquot was prepared in a routine manner for PCR. Remaining sediment was diluted in phosphate-buffered saline, Middlebrook 7H10 broth, or BACTEC 12B broth; an internal control was added to all reaction mixtures and controls. Internal control was detected only in the sample diluted with BACTEC 12B medium. Components of the BACTEC 12B medium including PANTA reagent (polymyxin B, amphotericin B, nalidixic acid, trimethoprim, and azlocillin), reconstituting fluid, 0.2% glycerol, 0.05% Tween 80, and 0.05% bovine serum albumin (BSA) were tested in a similar manner. Only 0.05% BSA resulted in amplification of the internal control DNA. Varying concentrations of BSA were added to 11 aliquots of a respiratory sediment known to be inhibitory to the PCR. Internal control was detected in all reaction mixtures containing 0.00038 to 0.1% BSA. To determine the ability of BSA to override inhibition, respiratory specimens were run in triplicate: undiluted, diluted 1:2 with BACTEC 12B medium, or diluted with 0.026% BSA. For 21 of 22 inhibitory specimens, BSA was able to override the presence of interfering substances. These data suggest that the presence of BSA in a PCR assay is critical for the direct detection of M. tuberculosis in respiratory specimens.
