Clinical setting and infection control measures
5th Department of Internal medicine of the University Hospital in Bratislava as the largest department in this filed-field in capital city of Slovakia, was repurposed for the treatment of critically ill, COVID-19 positive patients with internal disease co-morbidities, during the peak of second wave of COVID-19 pandemic (December 2020) in the country. From the beginning of December 2020 until the end of April 2021, 395 patients with COVID-19 were hospitalised in two separate isolation wards (ward A and ward B, with a capacity of 40 beds in total) and one isolation ICU (with a capacity of 8 beds) in the department. Enhanced infection control measures were applied for the isolation areas (isolation red zones), including use of personal protective equipment (PPE) with biosafety level 3 (BSL-3).
The general internal surveillance program of drug-resistant bacteria in the department flagged a considerable increase in number of positive VRE cases (all caused by E. faecium) in COVID-19 positive patients in December 2020 (a 60% increase in comparison to the previous three months, (6 positive cases were diagnosed for period of three months between September to November 2020 versus 8 positive diagnosed cases during December 2020), as well as in comparison to the same period in the previous year cases (2 positive diagnosed cases during December 2019 in comparison to 8 positive diagnosed cases in December 2020). However, since most of hospitalized patients were transferred from other clinics, hospitals, or nursing homes, the positive VRE cases were considered possible imported cases rather than part of an in-department outbreak. After a change in antibiotic treatment and physical isolation, these patients were only followed until January 2021 and not included in this study.
In January 2021, we continued to record positive VRE and initiated the epidemiological investigation following the implementation of strategies for terminating the VRE outbreak. These strategies included:
1) implementing a new antibiotic therapy algorithm for treating bacterial super-infections and minimizing inappropriate initiation of antibiotic therapy (specifically, the use of cephalosporins). A combination of Augmentin and Clarithromycin was used for treating community-acquired pneumonia and Tazocin with an aminoglycoside was used for hospital-acquired pneumonia, until antibiotic susceptibility testing results were obtained. For positive cases of VRE, a combination of linezolid with doxycycline was used.
2) We implemented a point prevalence screening on all patients upon admission and once a week thereafter.
3) We intensified the frequency of surface disinfection (from once a day to twice a day every 12 hours) of all surface areas and used instruments in red and green zones (green zone are areas outside of red zones within the ward in which BSL 2 or lower was required). For disinfection, we used Incidin Oxyfoam S (Ecolab, Deutschland GmbH, Monheim am Rhein), Meliseptol Foam pure (B. Braun Melsungen AG, Germany) and Chloramine 0.2%/L solution.
4) Patients diagnosed with VRE were accommodated in same room (designated with a VRE positive sign), which had separated toilets or bedside toilets and higher hygienic precautions, and dedicated examination instruments. Complete isolation of VRE patients (one patient per room) was not possible during the peak of the pandemic. As VRE positive cases continued to emerge throughout February 2021 (although with a mild decrease in number; 7 cases in December 2020, 6 cases in January 2021, and 4 cases in February 2021), we performed environmental sampling for detection of VRE contaminations in red and green zones and all samples from diagnosed patients underwent further genetic examination for epidemiological mapping of transmission routes.
Data and demographical information, inclusion, and exclusion criteria
Data used in this study was gathered from all patients hospitalised in the red zone during a period of 4 months (beginning of January 2021 until the end of April 2021). All patients included in this study had a confirmed infection of SARS-CoV-2 based on a RT-PCR test. We recorded demographical data for each patient including age and gender, and collected information on the duration of hospitalisation, place of sampling, antimicrobial susceptibility, type of acquired infection (HAI, or non-healthcare associated infection, NHAI), whether the patient was dismissed or died during our monitoring, and the duration of hospitalisation before diagnosis of VRE infection. Other than positive COVID-19 PCR test and clinically relevant diagnosis of and VRE (positive samples from the urinary tract, haemoculture, decubitus, or sputum in patients with a high likelihood of bacterial infectious process such as elevated CRP and procalcitonin), no other inclusion or exclusion criteria were used in this study.
Environmental sampling, patient sampling, and testing methods
In accordance with the internal guidelines of the department, sterile swab samples of the tonsils and nose, and samples of sputum and mid-stream urine, were collected for microbiological investigation from all patients upon admission. Other specific microbiological sampling (haemoculture, stool culture, etc.) were performed as deemed necessary by the patient’s physician. After detection of the VRE outbreak and implementation of stricter hygienic measures, we performed anal swab sampling for all patients in red zones at the time of admission and once a week thereafter. Blood sampling was performed using peripheral intravenous puncture. Urine was collected by mid-stream urine sampling or via urinary catheter.
Environmental sampling was conducted twice during February 2020 in both green and red zones. Sampling was performed using sterile swab samples of surfaces and examination instruments We collected a total of 30 environmental samples (20 from the red zone and 10 from the green zone).
Patient samples and environmental samples were processed by conventional methods approved by the Slovak Ministry of Health. Bacterial identification was performed by standard bacterial culture from biological specimens (urine, swabs, etc.) – growth on solid media using Columbia blood agar with 7% sheep blood (EnviroLab), and Uriselect 4 chromogenic agar (EnviroLab) for urine culture (Medirex a.s.). Strain identification was performed by matrix-associated laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) (Bruker, Massachusetts, USA). Antimicrobial susceptibility testing was performed using the disc diffusion test on Mueller-Hinton agar culture and results were interpreted according to the EUCAST guideline (European Committee on Antimicrobial Susceptibility Testing).
DNA extraction, library preparation and sequencing
DNA was extracted from plated samples using the DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany) using a protocol for isolation of gram-positive bacteria following the manufacturer’s instructions. DNA libraries were prepared using Nextera XT DNA Library Prep Kit (Illumina Inc., San Diego, CA, USA) as described in the protocol, followed by library validation using Qubit dsDNA high sensitivity assay (LifeTechnologies, Eugene, OR, USA) and Agilent® HS DNA Chip on Agilent® Technologies 2100 Bioanalyzer (Santa Clara, CA, USA). Finally, DNA libraries were normalized to a concentration of 4 nM, denatured, and sequenced on a MiSeq system (Illumina Inc., San Diego, CA, USA) using MiSeq Reagent kit v3 with paired end of 2x 300 bp reads.
Data analysis
Adapters and low-quality ends of sequenced reads were removed using Trimmomatic version 0.36 (7) based on quality control statistics generated by FastQC version v0.11.5 (8). After trimming, fragments without sufficient length of both reads (>35 bp) were removed from the data set. We used SPAdes version 3.10.1 (9) for de novo assembly of data and BLAST software (Basic Local Alignment Search Tool) version 2.10.1 (10) for the classification of sequences. Computational analyses were written and executed using the SnakeLines framework version 0.11.6 (11, 12). We used Type Genome Server (TYGS) version 281 (13) for genome-based taxonomy analysis and visualization (Figure 1). Sequences were mapped with Burrow-Wheeler Aligner (BWA) version 0.6 (14) against the van genes listed in Table 4. Then we used multilocus Sequence Typing (PubMLST) (15). Finally, the percentage of similarity between de novo assembled samples (Table 2) was determined with BLAST (10).