Background: Hospital-acquired bloodstream infections (HA-BSI), including catheter-associated bloodstream infections (CABSI), cause preventable harm in haemato-oncology patients but surveillance data are limited.

Methods: We performed a retrospective cohort study using prospectively collected data in a large hospital network in Switzerland from 2017-2022. Incidence, source, and microbiology of HA-BSI were compared between (1) haematology patients with acute leukaemia or allogeneic stem cell transplantation (2) oncology patients with solid tumour or lymphoma, and (3) general medical patients. No routine quinolone prophylaxis was prescribed.

Results: We included 320,058 patient-days and 201,081 catheter-days across two haematology, two oncology and nine non-COVID-19 general medical wards. 669 HA-BSI occurred in 547 individual patients. In haematology patients, HA-BSI incidence was 9.1/1000 patient-days (95% CI 8.2-10.3). 224/299 (75%) of episodes were "unknown/other" source. Low virulence Gram-positive organisms (coagulase-negative staphylococci, viridans Streptococci, enterococci) accounted for 232/378 (61%) HA-BSI organisms and 46/52 (88%) CABSI organisms. Compared to oncology and general medical patients, haematology patients had higher HA-BSI incidence, but a smaller proportion of infections caused by virulent organisms (Gram-negative bacteria, Staphylococcus aureus, p < 0.01).

Conclusions: In haematology patients, HA-BSI are less commonly caused by virulent Gram-negative organisms or Staphylococcus aureus compared to solid tumour and general medical patients, in the absence of quinolone prophylaxis.