Impact of empiric cefepime therapy on mortality among patients with bloodstream infections due to extended-spectrum lactamase-producing pathogens

Abstract number: P1320

Chopra T., Zhao J., Tansek R., Hatahet D., Hothi J., Chaudhry K., Rahbar H., Chen T., Truong T., Rodriguez V., Bernabela L., Bhargava A., Marchaim D., Alangaden G., Kaye K.

Objective: To study the impact of empiric cefepime for the treatment of blood stream infections (BSIs) due to extended spectrum b-lactamase-producing pathogens (ESBLs).

Methods: A cohort study was conducted at DMC from 1/05 to 12/07 in 5 hospitals. Retrospective chart review was conducted on patients with ESBL-producing Klebsiella pneumoniae and Escherichia coli (ESBLs). Patient variables collected included demographics, comorbid conditions, empiric treatment and in-hospital mortality. Empiric antibiotics were defined as antibiotics initiated during the time spanning from 2 days prior to culture to 3 days after culture. Logistic regression was used to create a prediction model for in-hospital mortality.

Results: 145 ESBL BSIs were identified; 83% were K. pneumoniae and 16.5% were E. coli. 35.2% of the BSI were catheter-related. The mean age of the patients was 66 years, 51% were female and 79.3% were African-American. The in-hospital mortality rate was 35% (n = 51). Cefepime was utilized as an empiric antibiotic in 68 patients (46.9%) and carbapenems in 40 (27.6%) patients. In logistic regression, predictors of in-hospital mortality included: admission to intensive care unit (OR = 2.14, 95% CI 0.96–4.76), central line present prior to positive culture (OR = 2.13, 95% CI 0.70–6.45), presence of a rapidly fatal condition at the time of admission (OR = 5.71, 95% CI 2.36–13.8) and prior hospitalization (OR = 1.75, 95% CI 0.76–4.02). Type of empiric antibiotic was not associated with mortality. In multivariate analysis, receipt of cefepime alone (n = 42) was associated with increased mortality, although this association did not reach statistical significance (OR = 1.58, 95% CI 0.67–3.70). There was a trend between empiric carbapenem therapy and decreased mortality (OR = 0.59, 95% 0.24–1.45). These associations remained unchanged after controlling for hospital and anatomic source of infection. A sub-analysis was performed on 42 patients who were treated with empiric cefepime alone, comparing survival and MIC to cefepime. No association was found between cefepime MIC and mortality in this sub-group.

Conclusion: Empiric use of cefepime alone for BSI due to ESBLs was associated with a trend towards increased mortality. We were not able to demonstrate an association between MIC to cefepime and mortality, but were underpowered to do so. Additional study of the efficacy of cefepime therapy in treating ESBLs should be conducted.

MIC Cefepime (mg/ml)In-hospital mortality [deaths/total (%)]
leqslant R: less-than-or-eq, slant24/12 (33.3%)
41/4 (25%)
81/2 (50%)
geqslant R: gt-or-equal, slanted1610/24 (42%)

Session Details

Date: 10/04/2010
Time: 00:00-00:00
Session name: Abstracts 20th European Congress of Clinical Microbiology and Infectious Diseases
Location: Vienna, Austria, 10 - 13 April 2010
Presentation type:
Back to top