Direct identification of bacteria from BacT/ALERT anaerobic positive blood cultures by MALDI-TOF MS: MALDI Sepsityper kit versus an in-house saponin method for bacterial extraction.

In cases of bacteraemia, a rapid species identification of the causal agent directly from positive blood culture broths could assist clinicians in the timely targeting of empirical antimicrobial therapy. For this purpose, we evaluated the direct identification of micro-organisms from BacT/ALERT (bioMérieux) anaerobic positive blood cultures without charcoal using the Microflex matrix-assisted laser desorption/ionization (MALDI) time of flight MS (Bruker), after bacterial extraction by using two different methods: the MALDI Sepsityper kit (Bruker) and an in-house saponin lysis method. Bruker's recommended criteria for identification were expanded in this study, with acceptance of the species identification when the first three results with the best matches with the MALDI Biotyper database were identical, whatever the scores were. In total, 107 monobacterial cultures and six polymicrobial cultures from 77 different patients were included in this study. Among monomicrobial cultures, we identified up to the species level 67 and 66 % of bacteria with the MALDI Sepsityper kit and the saponin method, respectively. There was no significant difference between the two extraction methods. The direct species identification was particularly inconclusive for Gram-positive bacteria, as only 58 and 52 % of them were identified to the species level with the MALDI Sepsityper kit and the saponin method, respectively. Results for Gram-negative bacilli were better, with 82.5 and 90 % of correct identification to the species level with the MALDI Sepsityper kit and the saponin method, respectively. No misidentifications were given by the direct procedures when compared with identifications provided by the conventional method. Concerning the six polymicrobial blood cultures, whatever the extraction method used, a correct direct identification was only provided for one of the isolated bacteria on solid medium in all cases. The analysis of the time-to-result demonstrated a reduction in the turnaround time for identification ranging from 1 h 06 min to 24 h 44 min, when performing the blood culture direct identification in comparison with the conventional method, whatever the extraction method