Algorithm for empirical glycopeptide treatment in patients with hematologic malignancies and enterococcus faecium blood stream infection

Introduction / objectives

Enterococcus faecium has become a major cause of nosocomial infections especially in patients with hematologic malignancies. The aim of this study was to determine risk factors in those patients who are at risk of Enterococcus faecium blood stream infection (BSI) and should be considered for empirical treatment.

Methods

Retrospectively demographic, clinical and microbiological data in 33 patients with an E. faecium BSI were compared to 66 control patients during a 5-year period at the hematology ward. Multivariate logistic regression was used to explore the independent risk factors in order to develop a prognostic model to determine the risk of E. faecium BSI.

Results

Significant associations of E. faecium BSI were found with age, hospital stay prior to blood culture, duration of hospitalization 1 year before admission, fever prior to blood culture, severity and duration of neutropenia, CRP (C-reactive protein) at time of blood culture withdrawal, colonization with E. faecium prior to blood culture and diarrhea. E. faecium BSIs were found associated with more severe disease and higher mortality rates. Independent risk factors for E. faecium BSI were colonization with E. faecium 30 days prior to blood culture (OR 3.83; CI 1.1-12.8), fever > 1 day (4.02; 1.3-12.8), hospital stay prior to blood culture > 14 days (4.78; 1.3 -18.0), age > 59 years (5.47; 1.6-18.2) and abdominal pain, diarrhea or neutropenia (5.95; 1.1-31.4).

Conclusion

Using prognostic modeling, risk stratification is possible for development of E. faecium BSI in patients with hematological malignancies. Empirical treatment should be considered in patients who are at high risk.

Disclosure of interest

None declared.