TEXT

Top Abstract Text References

The "gold standard" of mycobacterial diagnostic procedures is still cultural detection. The radiometric BACTEC 460TB system (Becton Dickinson, Heidelberg, Germany) in combination with solid media has been the benchmark for sensitivity and speed of cultural detection (8, 11); however, it is labor-intensive, bears the potential risk of cross-contamination, requires special attention regarding radioisotopes, andas manufacture has been haltedwill not be available any more in the future. The BACTEC MGIT 960 system (Becton Dickinson) is a fully automated, nonradiometric, noninvasive device for simultaneous incubation and monitoring of 960 culture tubes. In the culture vials an oxygen-sensitive fluorescent indicator is implemented, responding to the amount of consumed oxygen. Every vial is monitored hourly, and specific algorithms determine the positivity of the vials. Results of several studies with comparison to the BACTEC 460TB system indicated that the MGIT 960 system is a rapid, sensitive, and efficient method for recovery of mycobacteria from clinical specimens (2, 4, 10, 13, 14, 15, 16). However, high percentages of culture-positive specimens (9.2 to 15.1%) and high amounts of smear-positive samples (up to 72%) reported in these studies suggest that they were performed either with a collection of patients with a high incidence of infectious tuberculosis or with selected clinical specimens. One study (16) conducted in a low-incidence laboratory did not use any solid media; contamination rates were high (up to 30%), and specimen numbers were low (two study periods with 859 and 941 samples). In this study we have evaluated the reliability of the MGIT 960 system in comparison to the BACTEC 460TB system and Löwenstein-Jensen (LJ) medium in daily routine diagnostic procedures of a nonspecialized microbiology laboratory where initial screening for the presence of mycobacteria is performed. The laboratory serves a university hospital with units specialized in the treatment of cystic fibrosis patients, human immunodeficiency virus-infected patients, and transplant patients. There is no specialized tuberculosis clinic. Approximately 4,500 clinical specimens are sent annually for mycobacterial culture. Our laboratory adheres to the monitoring procedures laid down by INSTAND (Düsseldorf, Germany) in its voluntary quality control measures (twice-annual checks for sensitivity of microscopic procedures, culture procedures, and differentiation).

During an 8-month study period (August 2000 to March 2001), all specimens (except blood) sent to the mycobacterium laboratory were analyzed in parallel by staining and culture in the BACTEC 460TB system, the MGIT 960 system, and LJ medium. Specimens were processed according to standard protocols (9). Nonsterile specimens were decontaminated by N-acetyl-L-NaOH and concentrated by centrifugation (4,000 × g, 15 min). The sediment was resuspended in 2 ml of phosphate buffer (pH 6.8). Processed specimens were stained by the Ziehl-Neelsen method. Of each specimen 0.5 ml of concentrate was inoculated into the following vials: MGIT 960 tubes, containing Middlebrook 7H9 with the supplement, BBL MGIT OADC (oleic acid, albumin, dextrose, and catalase), and BBL MGIT PANTA (polymyxin B, amphotericin B [AMB], nalidixic acid, trimethoprim [TMP], and azlocillin) as recommended by the manufacturer; and BACTEC 460 12B, containing modified Middlebrook 7H12 with the antibiotic supplement PANTA (polymyxin B [2,000 U/ml], AMB [200 µg/ml], nalidixic acid [800 µg/ml], TMP [200 µg/ml], azlocillin [200 µg/ml]). In parallel, two tubes containing LJ mediumone with glycerol and one without (Biotest, Heidelberg, Germany)were inoculated with 0.2 ml of the same specimen. LJ medium was used without antibiotics (except for respiratory specimens from cystic fibrosis patients, for which LJ medium with the antibiotic supplement PACTpolymyxin B [26 µg/ml], AMB [10 µg/ml], carbenicillin [50 µg/ml], TMP [10 µg/ml]was used). The order of inoculation was random. All specimens were incubated at 37°C (except skin specimens [30°C]) for 8 weeks. MGIT 960 tubes were incubated and automatically monitored in the MGIT 960 instrument. BACTEC 460TB vials were monitored every 2 days during the first week and weekly thereafter. A growth index of 100 was considered positive. LJ cultures were inspected weekly. Nonmycobacterial overgrowth was detected by using blood agar plates. Growth of mycobacteria was verified by microscopy (Ziehl-Neelsen) and subcultivation (LJ medium). Mycobacteria were identified by nucleic acid probes (Gen-Probe, San Diego, Calif.) or by sequencing of the 16S rRNA gene (7) and conventional biochemical tests. Transport time was calculated as the difference between the sampling date and the arrival date in the laboratory, and detection time was calculated as the difference between the date the BACTEC 460TB or MGIT 960 system indicated growth and the arrival date. All statistical results were calculated using Epi Info, version 6 (3). Statistical significances of differences were determined by the ² test, the ²-of-linearity test, or Student's t test, where appropriate, with a P of 0.05 considered to be significant.

From 1,188 patients, a total of 2,624 clinical specimens submitted (respiratory, 55.2%; sterile body sites, 17.1%; urine, 12.3%; gastrointestinal tract, 9.7%; others, 5.7%) were investigated. The number of samples per patient ranged from 1 to 25 (median, 1; mean ± standard deviation [SD], 2.21 ± 2.30).

In all, 4.7% of cultures were positive for mycobacteria and 2.2% were positive for M. tuberculosis complex. There was no statistically significant difference in recovery of M. tuberculosis complex (Table 1), as previously reported (2, 4, 5, 13, 14, 15). Differences in recovery of nontuberculous mycobacteria comparing the BACTEC 460TB and MGIT 960 systems have likewise been reported (5, 15). The recovery rate of M. avium- M. intracellulare in the MGIT 960 system was not significantly different from the recovery rate reported by Hanna et al. (4); still, we were unable to recover more M. avium-M. intracellulare with the MGIT 960 system. Table 2 shows recovery rates in accordance with initial smear results. In contrast to a report on the MB/BacT system (12), the MGIT 960 system was more sensitive in discovering M. tuberculosis complex in smear-negative samples. Still, the combination of the MGIT 960 system and LJ culture would have missed the diagnosis of tuberculosis in one patient, as only a single sputum sample was submitted for culture and the LJ culture remained negative. One smear-positive sample with subsequent growth of M. tuberculosis complex was contaminated in the MGIT 960 system. Tortoli et al. (15) reported on a reduced sensitivity in smear negative samples, especially for M. gordonae recovery, of the MGIT 960 system, similar to our result.

Contamination rates (Table 3) of samples from patients with or without cystic fibrosis for the BACTEC 460TB system and LJ cultures were comparable to those in previous reports (1, 4, 6, 12, 13, 15), as were differences in contamination rates for the BACTEC 460TB and MGIT 960 systems (2, 4, 5, 10, 13, 14, 15, 16). To our knowledge there have been no reports on contamination rates of sputa from cystic fibrosis patients cultured in the MGIT 960 system. Statistically significant differences were found for each of the three systems in different material groups (Table 4). Contamination rates of the MGIT 960 system compared to the BACTEC 460TB system were statistically significantly elevated in contaminated clinical material. Transport time has been discussed (2, 5, 10, 16) as a reason for differing contamination rates. Transport time (mean ± SD) was 0.90 ± 1.17 days (range, 0 to 13; median, 1). Calculation of contamination rates in dependency of transport time revealed no significant differences for the BACTEC 460TB system and LJ medium and no linear relation for the MGIT 960 system (² of linearity, 10.38 [P = 0.016, with linearity being not present when P was <0.05]).

The MGIT 960 system detected mycobacteria significantly earlier than the BACTEC 460TB system (Table 5), especially M. tuberculosis complex and M. avium-M. intracellulare. The results for detection time with the MGIT 960 system are similar to those in previous reports (2, 4, 5, 14, 15, 16). Kanchana et al. (5), who determined the positivity of the BACTEC 460TB system as we did, obtained comparable results for detection time in both systems. In contrast to this, studies (2, 4, 14, 15, 16) using more-intensive schemes to detect positivity in the BACTEC 460TB system were not able to show differences in detection time for the BACTEC 460 and MGIT 960 systems.

In conclusion, the MGIT 960 system is a suitable and fast nonradiometric alternative for M. tuberculosis complex recovery. Contamination rates were associated with underlying disease (cystic fibrosis), the material (gastric aspirate), and the detection system (MGIT 960) but not with transport time. Under our conditions the MGIT 960 system detected mycobacteria significantly earlier than the BACTEC 460TB system, especially M. tuberculosis complex. In the inevitable advent of nonradiometric detection systems (with their systems' inherently higher contamination rates), physicians should, whenever feasible, send respiratory samples (instead of gastric aspirate) and send more repetitive samples for M. tuberculosis complex and nontuberculous mycobacterium recovery.