PCR Enhances Acid-Fast Bacillus Stain-Based Rapid Detection of Mycobacterium tuberculosis

Rapid detection of the Mycobacterium tuberculosis complex is important in patient management in terms of initiating appropriate antimycobacterial therapy as well as controlling the spread of this pathogen. Although mycobacterial culture continues to be a valuable diagnostic tool, the results are not rapid. The Centers for Disease Control and Prevention as well as the American Thoracic Society recommends obtaining mycobacterial cultures of at least three sputum specimens from patients in whom tuberculosis is suspected. In contrast to the increased time required for mycobacterial culture, acid-fast bacillus (AFB) staining generally provides rapid evidence for the presence of mycobacteria in a clinical specimen. AFB staining is used by most health care facilities to determine when patients with suspected tuberculosis should be isolated. However, AFB smears will be positive in only 60% of tuberculosis cases (7). Moreover, a recent study found that only 8.6% of patients placed in isolation proved to have tuberculosis yet 19% of patients with pulmonary tuberculosis were not isolated on the first day after hospital admission (9).

Although AFB staining and mycobacterial culture are still the primary diagnostic tests in most laboratories, molecular tests based on nucleic acid amplification techniques have been available for the detection of the M. tuberculosis complex in a variety of clinical specimens (2, 4, 6). We used a colorimetric microtiter plate PCR-enzyme immunoassay (PCR-EIA) (8) targeting two specific genes for the rapid detection of the M. tuberculosis complex. One primer set (IS-1, 5′-CCT GCG AGC GTA GGC GTC GG-3′, and IS-2, 5′-CTC GTC CAG CGC CGC TTC GG-3′) was designed to target IS6110, an insertion sequence within the chromosome of M. tuberculosis (2). Another primer set (MPB-1, 5′-GAG GAG TTG AAA GGC ACC GAT-3′, and MPB-2, 5′-CGC GTC TGG GCG ATG TA-3′) was designed to target an M. tuberculosis gene which encodes the MPB64 protein (6). Two M. tuberculosis complex-specific 5′-biotinylated capture probes (IS-p, 5′-GGC GAA CCC TGC CCA GGT CGA CAC-3′, and MPB-p, 5′-CGG CCA GGC GTG CCA GAT TC-3′) were incorporated into a colorimetric signal amplification procedure to detect and confirm the amplification products as previously described (8). A similar format was used separately to amplify the human β-actin gene to assure the quality of extracted DNA samples.

During a study period from December 2001 until January 2002, a total of 782 respiratory specimens, mainly sputa, were sent to the Diagnostic Laboratory Services, Inc., in Honolulu, Hawaii, for AFB staining and mycobacterial culture. Standard sample decontamination with N-acetyl-l-cysteine-NaOH (2% final concentration) and concentration by centrifugation were performed as described previously (5, 10). The volume of the final sediment was approximately 2 ml. Part of the sediment was used for the AFB smear, which was then stained using auramine-rhodamine fluorochrome staining (10). Part of the sediment was inoculated onto Middlebrook 7H10 solid medium and into BACTEC 12B broth bottles that were monitored for growth using the BACTEC 460TB instrument (Becton Dickinson, Sparks, Md.). Of these specimens, all 15 M. tuberculosis culture-positive specimens and 60 M. tuberculosis culture-negative specimens (which represented 7.8% of all collected M. tuberculosis-negative specimens) were included in the study. Of 75 specimens, 26 (34.7%) grew one or more types of mycobacteria. Of the positive mycobacterial cultures, 15 had the M. tuberculosis complex, 5 had the M. avium complex, 2 had M. simiae, 2 had M. chelonae, 1 had M. fortuitum, and 1 had the M. avium complex and M. chelonae. Nine of 15 specimens which grew the M. tuberculosis complex upon culture were positive by AFB staining, giving a sensitivity of 60%. A specimen that grew the M. avium complex upon culture was also positive by AFB staining, giving a specificity of 98%. The remaining specimens, having grown non-M. tuberculosis complex mycobacterial species, were negative by AFB staining.

Nucleic acid was then extracted directly from 0.2 ml of N-acetyl-l-cysteine-NaOH-decontaminated and concentrated sediments by using IsoQuick (Orca Research, Inc., Bothell, Wash.) according to the manufacturer's instructions, as previously described (8). Extracted DNA was resuspended in 25 μl of water, and 10 μl of extracts was included in a PCR master mixture for PCR amplification (8). A housekeeping gene, human β-actin, was detectable in all DNA samples, indicating that these DNA samples were free of amplification inhibitors. All extracted DNA samples were tested for the M. tuberculosis complex by PCR-EIA targeting both IS6110 and MPB64 genes. M. tuberculosis complex-specific DNA was detected in 13 of 15 specimens that were culture positive for the M. tuberculosis complex, giving a sensitivity of 87%. Only two culture-positive specimens were PCR-EIA negative; both were also negative by AFB staining. All 60 respiratory specimens that were negative by culture were also negative by PCR-EIA, giving a specificity of 100%. These data indicated that PCR-EIA provides more rapid detection of the M. tuberculosis complex in clinical specimens than does AFB staining, with sensitivity improving from 60 to 87% and specificity improving from 98 to 100%.

While our study population in Hawaii still has a relatively high incidence of tuberculosis, other hospitals now face a change in tuberculosis epidemiology. Tuberculosis has become a relatively uncommon disease in the continental United States in recent years. As a result, physicians have less experience with the disease and may have difficulty recognizing it, a circumstance potentially leading to misdiagnosis and spread of the infection (3). As the incidence of tuberculosis declines, more resources will be used inappropriately for patients whose culture specimens grow mycobacteria other than M. tuberculosis. A recent study prospectively assessed the management of patients with suspected tuberculosis in an area with a high prevalence of mycobacteria other than M. tuberculosis and a low incidence of tuberculosis. The high prevalence of mycobacteria other than M. tuberculosis in specimens has a substantial impact on the management of suspected tuberculosis. Clinicians' assessments were sensitive for tuberculosis but had poor predictive value. Owing to a high rate of isolation of mycobacteria other than M. tuberculosis, AFB staining was a weak predictor of tuberculosis, with a positive predictive value of only 33% (1). Therefore, development of additional techniques to enhance the early diagnostic predictive values given by AFB staining will be of benefit in addressing this problem.

Although several nucleic acid amplification-based M. tuberculosis complex detection kits are commercially available, they are relatively expensive and the procedure is time-consuming (4). PCR-EIA as described in this report is simple, rapid, and user friendly. The whole procedure, including specimen processing, DNA extraction, PCR amplification, and amplicon identification, can be completed within an 8-h shift. An additional signal amplification step included in the microtiter plate system enhances test sensitivity, making this test an alternative to the AFB smear test in the clinical setting for tuberculosis patient management.