Shifting trends in bloodstream infection-causing microorganisms and their clinical impact in patients with haematologic malignancies in South Korea: A propensity score-matched study

Abstract

Background: This study aimed to identify recent trends in the epidemiology of bloodstream infection (BSI)-causing microorganisms among patients with haematologic malignancies (HMs) between 2011 and 2021, and to determine their impact on patient outcomes. Methods: This retrospective study included 6792 patients with HMs, of whom 1308 (19.3%) developed BSI within 1 y of diagnosis. The incidence of BSI-causing microorganisms was determined, and a propensity score-matched study was performed to identify risk factors for 28-d all-cause mortality in patients with HM. Results: A total of 6792 patients with HMs were enrolled. The cumulative incidence of BSI and neutropenia was significantly higher in the acute myeloid leukaemia and acute lymphoblastic leukaemia groups compared to other groups, and neutropenia and type of HMs were risk factors for the development of BSI. The annual incidence of coagulase-negative staphylococci (CoNS)-BSI decreased significantly ( P < 0.001), whereas Klebsiella pneumoniae- BSI increased ( P = 0.01). Carbapenem nonsusceptibility rates in K. pneumoniae isolates increased from 0.0% to 76.5% ( P < 0.001). BSI caused by K. pneumoniae (adjusted odds ratio 2.17; 95% confidence interval 1.12-4.21) was associated with higher 28-d all-cause mortality compared to that caused by CoNS (adjusted odds ratio 0.86; 95% confidence interval 0.48-1.55). Conclusion: The pathogenic spectrum of BSI-causing bacteria in patients with HMs gradually shifted from Gram-positive to Gram-negative, especially from CoNS to K. pneumoniae . Considering that K. pneumoniae BSI had a significantly higher 28-d mortality rate than CoNS-BSI, this evolving trend could adversely impact the clinical outcomes of patients with HMs. (c) 2024 Elsevier Ltd and International Society of Antimicrobial Chemotherapy. All rights are reserved, including those for text and data mining, AI training, and similar technologies.