TEXT

Top Abstract Text References

Mycobacteriology laboratories have been using commercially available nucleic acid amplification test kits (NAATs) directly with processed, acid-fast, smear-positive sputum specimens to help reduce the time to detection of Mycobacterium tuberculosis complex (MTBC). This study tested the usefulness of performing one NAAT, the Gen-Probe Amplified Mycobacterium Tuberculosis Direct (MTD) test, with positive BACTEC 12B broth culture bottles (BACTEC/MTD method) rather than directly with processed sputum. This procedure was designed to offer three advantages over direct testing of sputum. The first advantage was that potential inhibitors of amplification would be diluted by a factor of 1:9 with the 12B medium in the BACTEC bottle, reducing the number of false-negative results. Previously, a dilution of at least 1:2 with 12B medium has been shown to be effective in significantly reducing the effect of inhibitory substances on the performance of another NAAT (4). The second advantage was created by using a BACTEC 12B broth culture that had been inoculated and incubated to obtain a positive growth index (GI) to amplify low numbers of MTBC organisms present in the sputum. By amplifying the amount of nucleic acid target for the MTD, sensitivity should be further improved. The third improvement is that the requirement for growth to obtain a positive GI would reduce false-positive results caused by nonviable organisms in the sputum of some patients, including those under treatment for tuberculosis (TB).

Previous studies (1, 3, 7) evaluated a nucleic acid amplification method with growth from BACTEC 12B bottles. Excellent sensitivity and specificity were obtained, but the NAAT was not attempted at the first clear sign of growth (e.g., a GI of 10).

A total of 239 smear-positive specimens representing 89 different patients were tested in this study. Seventy-three of these specimens were processed in the Microbial Diseases Laboratory (MDL), California Department of Health Services, by the N-acetyl-L-cysteine sodium hydroxide method (6). A total of 166 specimens were received via MDL's BACTECs-by-mail program. In this program, local public health laboratories that do not have a BACTEC 460TB instrument process specimens and mail inoculated, unincubated BACTEC 12B bottles to MDL. For this project, the submitting laboratories informed MDL by telephone when they had sent BACTEC bottles from smear-positive patients, and only those BACTEC bottles were included in the study.

All of the BACTEC bottles for this study were incubated at 37°C for 8 weeks or until positive. Growth was detected with the BACTEC 460TB instrument. The first time a BACTEC bottle inoculated with a smear-positive specimen reached a GI of at least 10, a 0.5-ml aliquot was removed and stored at 20°C until testing. A GI value of 10 was chosen because this level of growth is achieved early, but negative cultures do not usually show a GI value this high. The bottles were then reincubated and subjected to the routine laboratory procedure for detection of mycobacteria and identification by Gen-Probe AccuProbe, high-performance liquid chromatography (HPLC), and/or biochemical techniques, as appropriate.

The Enhanced MTD was performed, according to the manufacturer's protocol, with 450 µl of the frozen broth as the test sample. This method provides greater sensitivity than the original MTD method, which used a sample size of only 50 µl (2, 5).

Of the 239 acid-fast, smear-positive specimens included in this study, 215 showed evidence of growth during the 8-week incubation period. The patients whose specimens were acid-fast, smear positive, and culture negative included those who were undergoing treatment and whose samples were being cultured to monitor therapy. Of the 215 cultures that showed evidence of growth, 179 were positive for MTBC by our routine methods (Table 1). All of these, plus two others, were also positive for MTBC by the MTD test. The two specimens that were positive by the MTD test and not by our routine methods were from patients who were previously culture positive for MTBC. By our routine methods, one of these specimens (patient 1) was overgrown with nonmycobacterial contaminants, and the other (patient 2) was overgrown with an unidentified mycobacterium. Patient 1 had other specimens that were culture positive for M. tuberculosis both before and after the specimen that was overgrown by nonmycobacterial contaminants. No mycobacteria other than M. tuberculosis (MOTT) were cultured from specimens of patient 1. It is therefore considered likely that the positive MTD result from patient 1 was due to the presence of M. tuberculosis in the specimen. The unidentified MOTT organism from patient 2 that grew in the culture of the MTD-positive specimen had an HPLC pattern characterized as "NCP 201." Previous experience with this organism had shown that it can sometimes give a false-positive AccuProbe result when the MTBC probe is used, due to a 16S rRNA sequence that is similar to that of MTBC in the probe region (K. Young, Y. Jang, J. Lopez, and E. Desmond, Abstr. 94th Gen. Meet. Am. Soc. Microbiol. 1994, abstr. U-72, p. 185, 1994). Patient 2 also had other specimens that were culture positive for M. tuberculosis. Thus it is possible that the positive MTD result for patient 2 may have been due to the presence of viable M. tuberculosis that was overgrown in culture by the MOTT or may have been due to a cross-reaction by the MTBC probe used in the MTD procedure with the NCP 201 organism.

The sensitivity of the MTD/BACTEC method was 100% compared with that of culture. The specificity was 100% if a clinical diagnosis of TB at any stage was used as the reference for true positivity or 99% if the culture-proven presence of viable M. tuberculosis was used as the reference for true positivity.

In addition to the cultures that grew MTBC, 34 other cultures of smear-positive specimens showed an increase in GI to 10 or greater. The results are shown in the footnotes to Table 1. These cultures were positive for non-TB mycobacteria or were overgrown by nonmycobacterial contaminants.

The time required for a BACTEC culture of a smear-positive specimen to reach a GI of 10 was studied in two different ways: first, with an accelerated schedule of measurement of GI values (daily reading), and, second, with the schedule of measurement of GI values recommended by the manufacturer (three times per week). In these studies, a specimen was considered to be a diagnostic specimen from an untreated patient if it was collected within 1 week of the first specimen collected from the patient.

A total of 221 smear-positive specimens were tested according to the accelerated reading schedule. Of these, 161 were culture positive for M. tuberculosis; the mean time to achieve a GI of at least 10 was 9 days. For diagnostic specimens from untreated patients, the mean time to achieve a GI of at least 10 was 6 days.

Eighteen cultures from smear-positive specimens were read by the conventional three-times-per-week schedule. Seventeen of the 18 specimens in this portion of the study were diagnostic specimens from untreated patients. With this less frequent reading schedule, the mean time for a culture positive for M. tuberculosis to reach a GI of at least 10 was 7 days. With the small sample size in this group, it was not possible to determine a statistically significant difference between time to detection of growth by the two reading schedules, but the 1-day difference in time to detection would be expected when comparing readings made an average of 2.3 days apart (three-times-weekly reading schedule) to daily readings.

Because aliquots harvested at GI 10 were frozen and tested by the enhanced MTD method in batches, no direct measurement of time savings was made in comparison with the time to results obtained with AccuProbe. In the MDL, the mean time from receipt of a specimen to reporting the presence of MTBC was 19 days when identification testing (by AccuProbe or HPLC) was performed twice weekly. Current identification testing in MDL by AccuProbe or HPLC requires a larger amount of growth (GI at least 500). If twice-weekly testing with MTD were adopted and cultures for identification were harvested at a GI 10 rather than 500, a reduction in turnaround time of 4 to 5 days would be expected. Alternatively, if daily readings of GI values were performed and a diagnostic culture from a new patient were tested by MTD without the delays caused by batching, a mean turnaround time of 6 days for the culture to reach GI 10 plus 1 day for testing could be achieved. Thus for cases with exceptional urgency, a sensitive and accurate culture and identification system with a mean turnaround time of 7 days is possible.

In summary, the data show that MTD testing of BACTEC 12B cultures when they reach a GI of 10 is a sensitive method for detecting M. tuberculosis that is more rapid than conventional methods. Because of the requirement for growth before MTD testing, patients (e.g., those undergoing therapy for TB) who are shedding nonviable M. tuberculosis bacilli should not give a false-positive result. When this method is used to determine whether an acid-fast bacillus infecting a smear-positive patient is M. tuberculosis, the very high sensitivity of BACTEC-MTD should enable the laboratory to confidently rule out M. tuberculosis when the MTD is negative for broth samples inoculated with a smear-positive sample and grown to a GI of 10.