Complex Routes of Nosocomial Vancomycin-Resistant Enterococcus faecium Transmission Revealed by Genome Sequencing

Abstract

Background: Vancomycin-resistant Enterococcus faecium (VREfm) is a leading cause of nosocomial infection. Here, we describe the utility of whole-genome sequencing in defining nosocomial VREfm transmission.

Methods: A retrospective study at a single hospital in the United Kingdom identified 342 patients with E. faecium bloodstream infection over 7 years. Of these, 293 patients had a stored isolate and formed the basis for the study. The first stored isolate from each case was sequenced (200 VREfm [197 vanA, 2 vanB, and 1 isolate containing both vanA and vanB], 93 vancomycin-susceptible E. faecium) and epidemiological data were collected. Genomes were also available for E. faecium associated with bloodstream infections in 15 patients in neighboring hospitals, and 456 patients across the United Kingdom and Ireland.

Results: The majority of infections in the 293 patients were hospital-acquired (n = 249) or healthcare-associated (n = 42). Phylogenetic analysis showed that 291 of 293 isolates resided in a hospital-associated clade that contained numerous discrete clusters of closely related isolates, indicative of multiple introductions into the hospital followed by clonal expansion associated with transmission. Fine-scale analysis of 6 exemplar phylogenetic clusters containing isolates from 93 patients (32%) identified complex transmission routes that spanned numerous wards and years, extending beyond the detection of conventional infection control. These contained both vancomycin-resistant and -susceptible isolates. We also identified closely related isolates from patients at Cambridge University Hospitals NHS Foundation Trust and regional and national hospitals, suggesting interhospital transmission.

Conclusions: These findings provide important insights for infection control practice and signpost areas for interventions. We conclude that sequencing represents a powerful tool for the enhanced surveillance and control of nosocomial E. faecium transmission and infection.

Keywords: Enterococcus faecium; genome sequencing.; infection control; transmission; vancomycin resistant.

Figure 1.

Phylogenetic analyses of Enterococcus faecium associated with bloodstream infection in 284 patients at Cambridge University Hospitals NHS Foundation Trust demonstrate numerous independent clades associated with nosocomial transmission. A, Maximum likelihood tree of 284 isolates. Colored dots indicate presence of vanA (red), vanB (yellow), vanA and vanB (orange), or absence of van genes (blue). Clusters 1–6 are used as examples for detailed analyses, and show the medical specialty for each case at the time of bacteremia (blue outer ring). Asterisks indicate isolates in cluster 3 that were not classified as closely related to other isolates in the cluster based on single-nucleotide polymorphism (SNP) differences. Scale bar indicates 84 SNPs. B, Pairwise SNP comparison for the 284 isolates shows the closest genetic match for each isolate. Red, vancomycin-resistant E. faecium (VREfm); blue, vancomycin-susceptible E. faecium (VSEfm). C, Number of isolates per year for clusters 1–6 (n = 93), starting from the earliest isolation date (day zero) in each cluster.

Figure 2.

Complex patterns of Enterococcus faecium transmission. The combination of genetic E. faecium clusters and patient location and movement data over time revealed several patterns of spread. Admission timeline for each patient is shown, with patient ID on the y-axis. Blocks indicate duration of hospital stay. The color of each block is unique to a specific ward or unit. Black dots denote a bloodstream infection, the isolate from which was sequenced. A, Patients in cluster 6 involved in a single ward transmission. The first 3 cases on ward A (purple) had overlapping admission dates indicative of an outbreak due to direct or indirect spread. The fourth case occurred many months later. B, Isolates in cluster 4 involved in transmission within and between 2 different wards (blue and orange). C, Isolates in cluster 2 from numerous wards.

Figure 3.

Genetic relatedness between Enterococcus faecium from Cambridge University Hospitals NHS Foundation Trust (CUH) patients and a collection from the UK and Ireland. Maximum likelihood tree based on single-nucleotide polymorphisms (SNPs) in the core genome of isolates from the clonal expansion of clade A. Outer ring indicates isolates from CUH (red), 15 isolates from patients with bacteremia in 2 hospitals close to CUH (yellow), and 456 isolates from patients with bacteremia in the UK and Ireland (dark blue). Clusters 1–6 are also indicated (clockwise from 1 o'clock: cluster 5, light blue; cluster 3, dark green; cluster 2, red; cluster 4, purple; cluster 6, light green; cluster 1, dark blue). Scale bar indicates approximately 100 SNPs.