ABSTRACT
OBJECTIVE: To study the molecular epidemiology of vancomycin-resistant Enterococcus (VRE) colonization and to identify modifiable risk factors among patients with hematologic malignancies. SETTING: A hematology-oncology unit with high prevalence of VRE colonization. PARTICIPANTS: Patients with hematologic malignancies and hematopoietic stem cell transplantation recipients admitted to the hospital. METHODS: Patients underwent weekly surveillance by means of perianal swabs for VRE colonization and, if colonized, were placed in contact isolation. We studied the molecular epidemiology in fecal and blood isolates by pulsed-field gel electrophoresis over a 1-year period. We performed a retrospective case-control study over a 3-year period. Cases were defined as patients colonized by VRE, and controls were defined as patients negative for VRE colonization. Case patients and control patients were matched by admitting service and length of observation time. RESULTS: Molecular genotyping demonstrated the primarily polyclonal nature of VRE isolates. Colonization occurred at a median of 14 days. Colonized patients were characterized by longer hospital admissions. Previous use of ceftazidime was associated with VRE colonization (P < .001), while use of intravenous vancomycin and antibiotics with anaerobic activity did not emerge as a risk factor. There was no association with neutropenia or presence of colonic mucosal disruption, and severity of illness was similar in both groups. CONCLUSION: Molecular studies showed that in the majority of VRE-colonized patients the strains were unique, arguing that VRE acquisition was sporadic rather than resulting from a common source of transmission. Patient-specific factors, including prior antibiotic exposure, rather than breaches in infection control likely predict for risk of fecal VRE colonization.