MSACL 2019 EU Abstract

Optimization of MALDI-based Techniques for the Rapid Diagnostics of Sepsis

Miriam Cordovana (1), Thomas Maier (2), Simone Ambretti (1), Markus Kostrzewa (2)
(1) A.O.U. di Bologna Policlinico Sant'Orsola-Malpighi, Bologna, Italy (2) Bruker Daltonik GmbH, Bremen, Germany

Warning: Undefined variable $headshot in /var/www/html/view_abstract/view_abstract_in_program.php on line 704

Miriam Cordovana (Presenter) A.O.U. Policlinico Sant'Orsola-Malpighi

Relevant Financial Disclosures (within past 24 months) Committee/Board/Advisory Board Executive Committee of ESGAI (ESCMID Study Group for Anaerobic Infections) Discussing Products? YES, presenter will present educational content related to products/services of one or more ineligible company(ies) listed above. Conflict mitigation IN PROCESS.

Background
The rapid identification of the causative agent of sepsis is crucial for the patients’ outcome.
MALDI-TOF MS has been shown to be applicable directly to positive blood culture samples.
In this study, we evaluate different MALDI-based approaches for the rapid bacterial identification from blood culture samples. We compared a short subculture-based method with a commercially available lysis-centrifugation method. Further, we optimized the latter to improve its efficacy for challenging-to-identify bacteria.

Methods
In this study, non selected consecutive blood culture samples from the routine were investigated.
N=231 bottles underwent MALDI bacterial identification, in parallel, with short subculture-based protocol (relying on the inoculum of the bacterial pellet obtained from 10 ml of positive blood culture onto chocolate agar, incubated for 1.5 h), and with the commercially available lysis-centrifugation method Sepsityper kit (Bruker Daltonik), relying on the lysis of the blood cells to enable the extraction of the bacterial pellet.
Further N=303 blood cultures underwent the classical and a shortened Sepsityper sample preparation protocol. Another N=3649 samples were processed with a new formulation, to assess performance and robustness of the method.
More, two modified methods were tested to improve the efficiency of identification for Streptococcus pneumoniae (N=21 samples) and Bacteroides fragilis (N=39 samples), two important species for which the standard protocols exhibited a sub-optimal performance.
Results obtained from the positive bottles were compared with the plate subcultures, in terms of correct identification and time-to-reporting.

Results
The lysis-centrifugation method showed a better efficiency than the short subculture method (91.7 % vs 67.3 % correct results), and it enabled to shorten the time-to-report of 2.5-3 hours.
The novel formulation of the kit showed a further improved efficiency (+ 3.6%) in comparison with the old version (from 269/303 (88.8%) to 280/303 (92.4%)) in the comparative study. This test performance was confirmed with the large collection of samples (3289/3649 of identifications – 90.1%).
The dedicated modified protocols enabled to improve the number of correct identifications both for S. pneumoniae (from 19.5% with the shortened method to 71.5%) and B. fragilis (from 35.3% to 89.7%), with the addition of only a 2-minutes step in the protocol.

Discussion
The lysis-centrifugation method proved to be considerably superior than short subculture method for the rapid bacterial identification from positive blood cultures, showing a very good efficacy, and enabling a significantly shorter time-to-report. Further, we showed that with just slight modifications of the protocol, it can be adapted and optimized to succeed in identifying bacteria that are traditionally most difficult to identify directly from the positive blood culture bottles.