Direct Species Identification in Positive Blood Culture Bottles From Patients With Hematologic Malignancies.

ABSTRACT

Background In patients with hematologic malignancies, faster species identification is particularly important in the management of bloodstream infection because of their immunocompromised and neutropenic status. In the present study, we analyzed direct species identification in patients with hematologic malignancies, and the factors that might influence the results of species identification. Methods We performed direct species identification using a Sepsityper® kit (Bruker Corporation, Billerica, Massachusetts, United States) and compared the results with a conventional method in patients with hematologic malignancies. Forty-five positive blood culture bottles containing single microorganisms from 37 patients were analyzed by matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS). And patients' clinical data were compared between the groups with spectral scores at acceptable and unacceptable levels. Results Direct species identification correctly identified 42 of 45 isolates and three were misidentified. While 35 of 45 isolates showed a spectral score ≥1.7 (acceptable identification), 10 isolates had a spectral score <1.7 (unacceptable identification) including three misidentified isolates. The group with a spectral score ≥1.7 had significantly lower white blood cell (p<0.01), neutrophil (p<0.01), and platelet (p<0.01) counts in addition to more frequent central venous (CV) line insertion (p=0.01). Multivariate analysis revealed that pathogen type (gram-positive or negative) and CV line insertion were associated with spectral scores. Conclusion Direct species identification using the Sepsityper kit is an upcoming approach for blood culture bottles, which were flagged as positive even in patients with hematologic malignancies when the spectral score was ≥ 1.7. Our study also indicates that direct identification is more accurate in patients with CV lines, and may be less accurate when gram-positive bacteria are detected.