BACKGROUND

Optimal management of Enterococcus faecium bloodstream infection (BSI) is not fully understood.

METHODS

Multicentre retrospective observational study of all consecutive adult (≥18 years old) patients with E. faecium BSI, between January 2016 and December 2023, at two tertiary teaching hospitals in northern Italy. Patients who died within 48 h from BSI onset were excluded. Primary and secondary endpoints were 30 day mortality and persistent E. faecium BSI, respectively. Cox regression and logistic binary analyses were used.

RESULTS

Overall, 391 patients were enrolled: median age was 72 (IQR 61-81) years, 225 were male (57.5%), median Charlson comorbidity index (CCI) was 6 (IQR 4-8) and 94 had immunosuppression (24.0%). BSIs were primary, secondary and device-related in 25.1%, 36.2% and 38.7%, respectively. Vancomycin resistance was found in 30.3%. The appropriate empirical therapy rate was given for 29.1%. All-cause 30 day mortality was 34.3% and the rate of persistent BSI was 18.8%. Variables independently associated with 30 day mortality were immunosuppression (HR 1.638, 95% CI 1.022-2.625, P = 0.040), SOFA (HR 1.205, 95% CI 1.144-1.268, P < 0.001), primary BSI (HR 1.839, 95% CI 1.221-2.770, P = 0.004), source control (HR 0.534, 95% CI 0.260-0.972, P = 0.042) and the performance of follow-up blood cultures (HR 0.403, 95% CI 0.280-0.972, P < 0.001). Factors independently associated with persistent E. faecium BSI were: CCI (OR 1.157, 95% CI 1.030-1.300, P = 0.014), source control not performed (OR 3.275, 95% CI 1.113-9.635, P = 0.031) and teicoplanin (OR 2.023, 95% CI 1.018-4.018, P = 0.044).

CONCLUSIONS

Among modifiable clinical factors in patients with E. faecium BSI, source control and the execution of follow-up blood cultures demonstrated a protective effect on 30 day mortality. Teicoplanin as targeted antibiotic treatment was independently associated with persistent BSI.