Investigating pathways of vancomycin-resistant Enterococcus (VRE) contamination and transmission in intensive care units: a prospective genomic surveillance study

Background: Vancomycin-resistant Enterococcus (VRE) species are common healthcare-associated pathogens that cause difficult-to-treat infections. Whole genome sequencing of patients has revealed a substantial burden of patient-to-patient VRE transmission in hospitals, with patients in intensive care units (ICUs) at particularly high risk of acquisition. However, few studies adequately characterize the pathways of VRE transmission between patients in acute care settings, a necessary step to identify current gaps in infection prevention practices. By harnessing genomic clustering analyses of whole genome sequences of VRE isolates from patients, environmental surfaces, and healthcare providers (HCP) in ICUs, we aim to reconstruct indirect pathways of pathogen movement to identify patterns of VRE spread and opportunities for transmission prevention. Methods and Findings: We collected daily samples (N = 6848) from ICUs in two hospitals over 13 weeks from four main sampling sources: patients, HCP hands, patient rooms, and shared surfaces. Samples were cultured on selective media and sent for whole genome sequencing (WGS). We used genomic thresholds to identify clusters of related VRE isolates and distinguish unrelated isolates. VRE was detected in samples from 20 out of 322 unique occupant-stays (6.22%). VRE isolates were detected from all sampling sources except for shared surfaces. A total of 44 unique VRE isolates were identified, 43 Enterococcus faecium (VREfm) and one Enterococcus faecalis (VREf). Two distinct patterns of VREfm spread were observed: 1) an outbreak setting with observed patient-to-patient transmission and low VRE diversity, and 2) high VRE diversity and pathogen movement between occupant-stays facilitated by persistent HCP and environmental contamination, but no observed transmission events. VRE detection probabilities were not significantly different between occupant-stays in outbreak and non-outbreak settings (OR = 0.63, 95% CI (0.23, 1.83), p = 0.32). However, inclusion of VRE isolated from non-patient samples increased the number of occupant-stays with VRE detection from 6 to 20, a 3.3-fold increase, as compared to patient samples alone. Inclusion of non-patient samples also increased the number of VRE multi-isolate genomic clusters detected by 7-fold. Our findings are limited because sampling was primarily conducted in ICUs. Due to the combination of short ICU stay durations and imperfect test sensitivity, VRE transmission events were probably underdetected. Conclusions: Our findings characterize the complex nature of VRE transmission pathways in ICU settings. Even without an ongoing outbreak, we found substantial evidence of VRE movement between occupant-stays, facilitated by a combination of HCP hands and environmental surfaces. This study highlights the importance of environmental sampling for understanding VRE transmission potential, which is likely to be underestimated using patient sampling alone. We recommend that future studies incorporate follow-up sampling after discharge to better understand the true burden of transmission.