Anca Rath (Regensburg / DE), Bärbel Kieninger (Regensburg / DE), Nilufarbayim Mirzaliyeva (Regensburg / DE), Aila Caplunik-Pratsch (Regensburg / DE), Jürgen Fritsch (Regensburg / DE), Stephan Schmid (Regensburg / DE), Patricia Mester (Regensburg / DE), Wulf Schneider-Brachert (Regensburg / DE)
Introduction
Precise outbreak detection is crucial for efficient outbreak management. The German legal outbreak definition implies the nosocomial acquisition of the same bacteria by minimum two epidemiologically linked patients. For vancomycin-resistant enterococci (VRE) the "same bacteria" is typically understood as identical species and van-genotype. In this study, we evaluated how the additional use of whole-genome sequencing (WGS) affects outbreaks identification at an intensive care unit of a tertiary care hospital during January 2020 – February 2021.
Methods
All patients with VRE (E. faecium only) were identified retrospectively. Admission period and time of VRE detection were documented. VRE detection >48h after admission was considered nosocomial. Two outbreak definitions were used: 1. the legal definition as stated above, and 2. the legal definition combined with WGS data. Hereby outbreak confirmation was achieved, if a difference in ≤3 alleles in pairwise comparison by core-genome (cg) multilocus sequence typing (MLST) was detected. WGS (Illumina MiSeq) was performed on the first VRE isolates per patient, and analyzed using SeqSphere+.
Results
41 VRE-positive patients were identified, including 22 nosocomial cases. The median length of stay was 20 days (2-113 days), whereas nosocomial VRE detection occurred after a median of 17 days (3-52 days). According to the legal definition, eight VRE outbreaks involving 2-6 patients were suspected during the study period.
WGS revealed the presence of six different VRE sequence types (ST) - predominantly ST1299, ST80, and ST117 - and 11 complex types (CT). Four of the suspected outbreaks proved to be false as isolates differed by MLST. Moreover, cgMLST only confirmed two outbreaks of the remaining four clusters. These involved two and four (of initially six) patients, respectively.
Conclusion
Only two of eight legally suspected VRE outbreaks could be confirmed using WGS. This proves both a gross overestimation of the frequency of VRE outbreaks, and a clear deficit in precision of outbreak identification on a daily basis without the use of WGS, which lead to wasteful use of outbreak management resources.