Comparison of the MB/BacT and BACTEC MGIT 960 system for recovery of mycobacteria from clinical specimens

A total of 543 specimens were cultured in parallel with the MB/BacT and BAC TEC MGIT 960 systems and on the conventional solid media. Mycobacteria were identified from 95 (17.5%) specimens, including 63(66.3%) Mycobacterium tu berculosis and 32 (33.7%) nontuberculous mycobacteria. The recovery rates f or the MB/BacT, MGIT 960, and solid media were 91.6, 87.4, and 54.7%, respe ctively, for all mycobacteria; the recovery rates were 93.6, 88.9, and 63.4 %, respectively, for M. tuberculosis complex alone, and 87.5, 84.4, and 37. 5%, respectively, for all nontuberculous mycobacteria. The mean times to de tection of all mycobacteria by individual systems were 13.9, 8.7, 31.7 days for the MB/BacT, MGIT 960 and solid media, respectively, 13.9, 9.3, 32.9 d ays for M. tuberculosis alone, and 14.1, 8.1, 27.2 days for all nontubercul ous mycobacteria. The contamination rates of the MB/BacT and MGIT 960 were 10.2 and 5.4%, respectively. With regard to detection times and recovery ra tes, both automated systems are superior to the conventional media tall p < 0.005). As compared to the MB/BacT, the MGIT 960 detected mycobacterial gr owth more rapidly (p < 0.001), and had a lower contamination rate (p = 0.00 3); however, there was no statistically significant difference in recovery rates between these two systems. These results indicate that both MGIT 960 and MB/BacT systems are rapid, sensitive, and efficient methods for the rec overy of mycobacteria from clinical specimens. (C) 2000 Elsevier Science In c. All rights reserved.