We evaluated the utility of vancomycin-resistant Enterococcus (VRE) surveillance by varying 2 parameters: admission versus weekly surveillance and perirectal swabbing versus stool sampling.

Prospective, patient-level surveillance program of incident VRE colonization.

Liver transplant surgical intensive care unit (SICU) of a tertiary-care referral medical center with a high prevalence of VRE.

All patients admitted to the SICU from June to August 2015.

We conducted a point-prevalence estimate followed by admission and weekly surveillance by perirectal swabbing and/or stool sampling. Incident colonization was defined as a negative screen followed by positive surveillance. VRE was detected by culture on Remel Spectra VRE chromogenic agar. Microbiologically-confirmed VRE bloodstream infections (BSIs) were tracked for 2 months. Statistical analyses were calculated using the McNemar test, the Fisher exact test, the t test, and the χ2 test.

In total, 91 patients underwent VRE surveillance testing. The point prevalence of VRE colonization was 60.9%; VRE prevalence on admission was 30.1%. Weekly surveillance identified an additional 7 of 28 patients (25.0%) with incident colonization. VRE BSIs were more common in VRE-colonized patients than in noncolonized patients (8 of 43 vs 2 of 48; P=.028). In a direct comparison, perirectal swabs were more sensitive than stool samples in detecting VRE (64 of 67 vs 56 of 67; P=.023). Compliance with perirectal swabbing was 89% (201 of 226) compared to 56% (127 of 226) for stool collection (P≤0.001).

We recommend weekly VRE surveillance over admission-only screening in high-burden units such as liver transplant SICUs. Perirectal swabs had greater collection compliance and sensitivity than stool samples, making them the preferred methodology. Further work may have implications for antimicrobial stewardship and infection control.