We have read the recent report by Brunello et al. (1) about a comparative study between the MB/BacT and the BACTEC 460 TB systems for recovery of mycobacteria. We have carried out a similar study with 500 clinical specimens (89% respiratory and 24.3% nonrespiratory), and we would like to make some comments. First, like Brunello et al., we did not find any statistically significant differences between the two systems, but the rates of recovery of Mycobacterium tuberculosis we obtained were lower than those of Brunello et al. (81.6% versus 99.1% for MB/BacT and 84.4% versus 100% for BACTEC 460 TB) (1). Rohner et al. (2) have reported recovery rates of 89 and 93%, respectively, for the MB/BacT and BACTEC 460 TB systems. With regard to the time to detection, Brunello et al. found an average of 13.2 days for MB/BacT and 9.9 days for BACTEC 460 TB. These results are quite different from our own findings (17.9 days for MB/BacT and 14.5 for BACTEC 460 TB), which are in complete agreement with those of Rohner et al. (2). The three studies are coincident in the time lags for detection by the MB/BacT system versus the BACTEC 460 TB system (3 days on average). Brunello et al. (1) suggest that the better performance of the two systems in their study could be explained by the higher number of smear-positive specimens they tested: 84% versus 64.2% in the study of Rohner et al. (2). (The number of smear-positive specimens in our study was 59%.) However, there must be another reason because they also obtained excellent results for smear-negative specimens (98.1% for MB/BacT and 100% for BACTEC 460 TB), in contrast to our own results (65.3% for both systems). We did not find any explanation for this great difference.