Clinical Predictors of Bacteremia Outcome After Initial Empirical Antimicrobial Therapy in Patients with Hematological Malignancies: A Retrospective Analysis.

OBJECTIVE

We performed a retrospective analysis to investigate the clinical predictors of bacteremia outcome involving () and after initial empirical antimicrobial therapy among hematological malignancy cases.

METHODS

This retrospective study was conducted between April 2018 and April 2023. All bloodstream infections (BSIs) caused by and in hospitalized hematological malignancy (HM) patients were identified. Data on patient demographics, clinical characteristics, empirical antimicrobial treatment, outcomes and the antimicrobial susceptibility were collected from medical records. Multivariate analyses were utilized to assess the risk factors for all-cause mortality within 28 days and carbapenem resistance. Optimal cutoffs for continuous predictive variables were evaluated by receiver operating characteristic (ROC) curve analysis.

RESULTS

Among 61 individuals diagnosed with bacteremia, 39 cases were caused by bacteremia, while the remaining 22 were identified as bacteremia. Out of these, there were 10 cases of carbapenem-resistant Enterobacteriaceae (CRE) and 12 cases resulted in all-cause mortality within 28 days. Analysis indicated that Pitt score was an independent risk factor for mortality and a cut-off of 2.5 was a reliable predictor with 83.3% sensitivity and 85.7% specificity, respectively. Impaired mental status and elevated body temperature exceeding 38.6°C as well as a procalcitonin (PCT) level over 8.24 ng/mL on the third day (d3) after antimicrobial treatment were identified as independent risk factors for predicting carbapenem resistance.

CONCLUSION

We found that Pitt score with a cut-off of 2.5 was a reliable predictor for mortality within 28 days in HM bacteremia cases. Impaired mental status and elevated temperature exceeding 38.6°C as well as a procalcitonin (PCT) level over 8.24 ng/mL on d3 after antimicrobial treatment were identified as predictive risk factors to carbapenem resistance.