A retrospective cohort study of antibiotic exposure and vancomycin-resistant Enterococcus recolonization.

OBJECTIVE: In the National Institutes of Health (NIH) Clinical Center, patients colonized or infected with vancomycin-resistant Enterococcus (VRE) are placed in contact isolation until they are deemed "decolonized," defined as having 3 consecutive perirectal swabs negative for VRE. Some decolonized patients later develop recurrent growth of VRE from surveillance or clinical cultures (ie, "recolonized"), although that finding may represent recrudescence or new acquisition of VRE. We describe the dynamics of VRE colonization and infection and their relationship to receipt of antibiotics. METHODS: In this retrospective cohort study of patients at the National Institutes of Health Clinical Center, baseline characteristics were collected via chart review. Antibiotic exposure and hospital days were calculated as proportions of VRE decolonized days. Using survival analysis, we assessed the relationship between antibiotic exposure and time to VRE recolonization in a subcohort analysis of 72 decolonized patients. RESULTS: In total, 350 patients were either colonized or infected with VRE. Among polymerase chain reaction (PCR)-positive, culture (Cx)-negative (PCR+/Cx-) patients, PCR had a 39% positive predictive value for colonization. Colonization with VRE was significantly associated with VRE infection. Among 72 patients who met decolonization criteria, 21 (29%) subsequently became recolonized. VRE recolonization was 4.3 (P = .001) and 2.0 (P = .22) times higher in patients with proportions of antibiotic days and antianaerobic antibiotic days above the median, respectively. CONCLUSION: Colonization is associated with clinical VRE infection and increased mortality. Despite negative perirectal cultures, re-exposure to antibiotics increases the risk of VRE recolonization.