Association of Inappropriate Initial Antibiotic Treatment with 30-Day Mortality in Hematological Malignancies Patients with Klebsiella pneumoniae Bloodstream Infection: A Single-Centre Retrospective Study

Abstract

Abstract Background Bloodstream infections caused by multidrug-resistant organisms such as Klebsiella pneumoniae are a significant challenge in managing hematological malignancies. This study aims to identify predictors of inappropriate antibiotic therapy and evaluate its impact on mortality in patients with hematological malignancies experiencing Klebsiella pneumoniae infections. Methods A retrospective analysis was conducted at a single center from January 2017 to December 2020, focusing on 182 patients with hematological malignancies who developed Klebsiella pneumoniae bloodstream infections. We compared the 30-day mortality rates between patients receiving appropriate and inappropriate antibiotic treatments, including the effectiveness of both single-drug and combination therapies. Kaplan-Meier survival analysis and multivariate logistic and Cox regression were used to identify factors influencing mortality risk. Results The 30-day all-cause mortality rate was 30.2% for all patients. The 30-day all-cause mortality rates were 77.2% and 8.8% in patients who received inappropriate initial treatment and appropriate initial treatment (P < 0.001). Inappropriate initial treatment significantly influenced mortality and was a key predictor of 30-day mortality, along with septic shock and previous ICU stays. Conclusions Infections with multidrug-resistant organisms considerably increase mortality in patients with hematological malignancies. A history of antibiotic use is a significant factor in the development of resistance in Klebsiella pneumoniae infections. Customizing treatment based on local drug resistance patterns and patient characteristics can reduce resistance and mortality. The alarming mortality risk in patients with carbapenem-resistant infections underscores the importance of rapid and effective treatment initiation, suggesting a potential role for predictive models in early therapeutic interventions.