ABSTRACT
OBJECTIVE: To determine risk factors for vancomycin-resistant Enterococcus (VRE) colonization during a hospital outbreak and to evaluate Centers for Disease Control and Prevention (CDC)-recommended control measures. DESIGN: Epidemiological study involving prospective identification of colonization and a case-control study. SETTING: A university hospital. PARTICIPANTS: Patients on eight wards involved in outbreak from late 1994 through early 1995. METHODS: Cases were matched by ward and culture date with up to two controls. Risk factors were evaluated with four multivariate models using conditional logistic regression. The first evaluated proximity to other VRE patients and isolation status. The second evaluated proximity to unisolated VRE cases and three variables independently predictive after adjustment for proximity. The third evaluated seven significant univariate predictors in addition to proximity to unisolated VRE in backward, stepwise logistic regression. The fourth assessed proximity to VRE with all other variables collected, clustered in a principal components analysis. Pulsed-field gel electrophoresis was performed to assess clonality of two outbreak strains. RESULTS: The incidence of transmission declined significantly after CDC guidelines were implemented. Proximity to unisolated VRE cases during the prior week was a significant predictor of acquisition in each of four multivariate models. Other significant risk factors in multivariate models included a history of major trauma and treatment with metronidazole. Pulsed-field gel electrophoresis confirmed the clonality of two outbreak strains. CONCLUSIONS: VRE was transmitted between patients during a hospital epidemic, with proximity to previously unisolated VRE patients being an important risk factor. Weekly surveillance cultures and contact isolation of colonized patients significantly reduced spread
