Microbiology and predictors of mortality in haematological malignancy patients with gram-negative bacterial bloodstream infections

Abstract

Background: Bloodstream infection (BSI) in haematological malignancy patients caused by gram-negative bacteria (GNB) poses a clinical challenge, which is exacerbated by the increased dissemination of carbapenem-resistant GNB (CRGNB). In this study, we investigated the prevalence and factors for GNB BSI and mortality in this population. Methods: This retrospective study included haematological malignancy patients who developed GNB BSI between 2015 and 2023 at a tertiary teaching hospital in southern China. Risk factors for CRGNB BSI and mortality of GNB BSI were identified by using multivariate logistic analyses. The patients were randomly divided into training and validation cohorts at a ratio of 7:3 to establish the model of 30-day mortality. C-indices, calibration plots, and decision curve analyses were generated to evaluate the model. A nomogram of the model was established. Results: Among the 351 patients with GNB BSIs, acute myeloid leukaemia (51.3%) was the most common. Escherichia coli (28.8%) and Klebsiella pneumoniae (29.7%) were the most common pathogens of GNB BSI and CRGNB BSI, respectively. The risk factors for CRGNB BSI were chronic liver disease, previous exposure to carbapenems, a platelet count < 30×10⁹/l and an albumin concentration < 30 g/l before BSI. The model for 30-day mortality of GNB BSI included neutropenia and an albumin concentration < 30 g/l before BSI, as well as septic shock and mechanical ventilation after BSI. The C-indices were 0.942 and 0.931 in the training and validation cohorts, respectively. The calibration plots and decision curves indicated that the model had good performance. Conclusions: The identified factors allow for the stratification of patients at greatest risk for CRGNB BSI and poor prognosis for GNB BSI, which could help in facilitating timely effective intervention.