Performance assessment of two lysis methods for direct identification of yeasts from clinical blood cultures using MALDI-TOF mass spectrometry.

ABSTRACT

In cases of fungal infection of the bloodstream, rapid species identification is crucial to provide adapted therapy and thereby ameliorate patient outcome. Currently, the commercial Sepsityper kit and the sodium-dodecyl sulfate (SDS) method coupled with MALDI-TOF mass spectrometry are the most commonly reported lysis protocols for direct identification of fungi from positive blood culture vials. However, the performance of these two protocols has never been compared on clinical samples. Accordingly, we performed a two-step survey on two distinct panels of clinical positive blood culture vials to identify the most efficient protocol, establish an appropriate log score (LS) cut-off, and validate the best method. We first compared the performance of the Sepsityper and the SDS protocols on 71 clinical samples. For 69 monomicrobial samples, mass spectrometry LS values were significantly higher with the SDS protocol than with the Sepsityper method (P < .0001), especially when the best score of four deposited spots was considered. Next, we established the LS cut-off for accurate identification at 1.7, based on specimen DNA sequence data. Using this LS cut-off, 66 (95.6%) and 46 (66.6%) isolates were correctly identified at the species level with the SDS and the Sepsityper protocols, respectively. In the second arm of the survey, we validated the SDS protocol on an additional panel of 94 clinical samples. Ninety-two (98.9%) of 93 monomicrobial samples were correctly identified at the species level (median LS = 2.061). Overall, our data suggest that the SDS method yields more accurate species identification of yeasts, than the Sepsityper protocol.