Longitudinal genomic surveillance of a UK intensive care unit shows a lack of patient colonisation by multi-drug-resistant Gram-negative bacterial pathogens

Abstract

Vulnerable patients in an intensive care unit (ICU) setting are at high risk of infection from bacteria including gut-colonising Escherichia coli and Klebsiella species. Complex ICU procedures often depend on successful antimicrobial treatment, underscoring the importance of understanding the extent of patient colonisation by multi-drug-resistant organisms (MDROs) in large UK ICUs. Previous work on ICUs globally uncovered high rates of colonisation by transmission of MDROs, but the situation in UK ICUs is less understood. Here, we investigated the diversity and antibiotic resistance gene (ARG) carriage of bacteria present in one of the largest UK ICUs at the Queen Elizabeth Hospital Birmingham (QEHB), focusing primarily on E. coli as both a widespread commensal and a globally disseminated multi-drug-resistant pathogen. Samples were taken during highly restrictive coronavirus disease 2019 (COVID-19) control measures from May to December 2021. Whole-genome and metagenomic sequencing were used to detect and report strain-level colonisation of patients, focusing on E. coli sequence types (STs), their colonisation dynamics and antimicrobial resistance gene carriage. We found a lack of multi-drug resistance (MDR) in the QEHB. Only one carbapenemase-producing organism was isolated, a Citrobacter carrying bla _KPC-2. There was no evidence supporting the spread of this strain, and there was little evidence overall of nosocomial acquisition or circulation of colonising E. coli. Whilst 22 different E. coli STs were identified, only 1 strain of the pandemic ST131 lineage was isolated. This ST131 strain was non-MDR and was found to be a clade A strain, associated with low levels of antibiotic resistance. Overall, the QEHB ICU had very low levels of pandemic or MDR strains, a result that may be influenced in part by the strict COVID-19 control measures in place at the time. Employing some of these infection prevention and control measures where reasonable in all ICUs might therefore assist in maintaining low levels of nosocomial MDR.

Keywords: AMR; E. coli; ICU.

Fig. 1.. Workflow of the study showing the number of patients recruited and the number of isolates whole-genome sequenced.

Fig. 2.. Timeline of colonising species seen during an ICU patient stay. Patients were in the ICU at different times, but for illustrative purposes, the timeline is the inpatient stay day number. In cases where the same ST is present multiple times in the same patient, they are shown as the same colour, but these colours cannot be compared between patients. Different shapes represent different species. A gradient of colour indicates different STs of the same species. Each ST is on a separate line. Isolates of the same ST are linked by a solid line where colonisation is not interrupted by another ST. In cases where an ST is interrupted by another ST, a dashed line is used. In cases where species aside from E. coli, Klebsiella, Enterobacter and Citrobacter were identified, an asterisk is used to signify these colonising isolates.

Fig. 3.. (a) ST breakdown for all E. coli isolates (n=107) and (b) the resistance gene profile for all colonising E. coli across all patients (n=21). The number of patients carrying the ST is stated on top of each bar in (a). Presence (navy) and absence (grey) of resistance genes are displayed alongside E. coli ST and patient number in (b).

Fig. 4.. Colonising isolate timeline displaying patients who acquired E. coli during their stay after having no detectable E. coli in their baseline first stool sample. All species confirmed using WGS are displayed alongside species presumptively identified using culture plates only. Antibiotics prescribed during a patient stay are displayed below. Patients were in the ICU at differing times, but for illustrative purposes, the timeline is the inpatient stay day number. In cases where the same ST is present multiple times in the same patient, they are shown in the same colour, but these colours cannot be compared between patients. Isolate links are demonstrated with a solid line where this ST strain is not interrupted by another ST, and where this ST strain is interrupted by another ST, a dashed line is used.

Fig. 5.. Colonising isolate timelines where E. coli colonisation was lost during inpatient stay. All species confirmed using WGS are displayed alongside species presumptively identified using culture plates only. Antibiotics prescribed during a patient stay are displayed below. Patients were in the ICU at differing times, but for illustrative purposes, the timeline is the inpatient stay day number. In cases where the same ST is present multiple times in the same patient, they are shown as the same colour, but these colours cannot be compared between patients. Isolate links are demonstrated with a solid line where this ST strain is not interrupted by another ST, and where this ST strain is interrupted by another ST, a dashed line is used.