A nationwide, multicenter, case-control study comparing risk factors, treatment, and outcome for vancomycin-resistant and -susceptible enterococcal bacteremia
Sm. Bhavnani et al., A nationwide, multicenter, case-control study comparing risk factors, treatment, and outcome for vancomycin-resistant and -susceptible enterococcal bacteremia, DIAG MICR I, 36(3), 2000, pp. 145-158
National Nosocomial Resistance Surveillance Group participants from 22 hosp
itals across the United States reviewed medical records for hospitalized pa
tients with vancomycin-resistant enterococcal (VRE) or vancomycin-susceptib
le enterococcal (VSE) bacteremia to identify risk factors associated with t
he acquisition of VRE bacteremia, describe genetic traits of VRE strains, a
nd identify factors predictive of clinical outcome. VRE cases were matched
to VSE controls within each institution. Multiple logistic regression (LR)
and classification and regression tree (CART) analysis were used to probe f
or factors associated with VRE bacteremia and clinical outcome. A total of
150 matched-pairs of VRE cases and VSE controls were collected from 1995 to
1997. Using LR, the following were found to be highly associated with VRE
bacteremia: history of AIDS, positive HIV status, or drug abuse (OR 9.58);
prior exposure with parenteral vancomycin (OR 8.37); and liver transplant h
istory (OR 6.75). CART analysis revealed that isolation of Enterococcus fae
cium, prior vancomycin exposure, and serum creatinine values greater than o
r equal to 1.1 mg/dl were predictors of VRE bacteremia. Greater proportions
of clinical failure (60% versus 40%, P < 0.001) and all-cause mortality (5
2% versus 27%, P < 0.001) were seen in patients with VRE versus VSE bactere
mia. Results from both LR and CART indicated that patients with persisting
enterococcal bacteremia, intubation at baseline, higher APACHE II scores, a
nd VRE bacteremia were at greater risk for poor outcome. (C) 2000 Elsevier
Science Inc. All rights reserved.