Evaluation of the MB/BacT system compared with Middlebrook 7H11 and Lowenstein-Jensen media for detection and recovery of mycobacteria from clinical specimens.

MB/BacT vials (an automated system), Middlebrook 7H11 agar plates with the microcolony method for growth detection, and conventional Lowenstein-Jensen egg-based medium were assessed for their rates of recovery, time to detection and ease in aiding the identification of mycobacterial isolates. Of the 2101 consecutive, non-selected specimens for tuberculosis analysed in the laboratory, 158 (7.5%) yielded growth of 159 mycobacteria on at least one of the three media, comprising 111 Mycobacterium tuberculosis and 48 non-tuberculous mycobacteria. Of the 111 specimens positive for M. tuberculosis, 100 isolates were recovered from MB/BacT vials, 99 from M7H11 agar plates, and 86 from Lowenstein-Jensen tubes. The combination of MB/BacT and M7H11 recovered 110 isolates of M. tuberculosis, 107 from M7H11 and Lowenstein-Jensen, and 105 from MB/BacT plus Lowenstein-Jensen. The average time to detection of M. tuberculosis was 11 days with M7H11 using the microcolony method, 16 days with MB/BacT, and 19.5 days with the conventional Lowenstein-Jensen method. The MB/BacT vials recovered greater numbers of non-tuberculous mycobacteria than either of the other two media, but these were considered mostly to be saprophytes. It was concluded that a combination of media was better than a single medium. None of the media combinations showed statistical differences. Use of M7H11 with the microcolony method, which shortens the detection time for mycobacteria and facilitates the detection of mixed cultures, together with the Lowenstein-Jensen tube, was a useful combination.