An outbreak report of vancomycin-resistant Enterococcus faecium outbreak in a Dutch general hospital, 2014-2017: successful control with a low impact strategy

Background We describe the control of a large in a in Methods


Abstract Background
We describe the control of a large VRE outbreak in a Dutch general hospital in 2014-2017.

Methods
After the outbreak was identi ed, a screening policy consisting of a single rectal swab culture (with enrichment broth) to identify VRE carriage in high risk patients, was implemented. In addition to screening, measures to improve compliance with standard infection control precautions and enhanced environmental cleaning were implemented to control the outbreak.

Results
Between September, 2014 and February, 2017, 140 patients were identi ed to be colonised by vanA mediated vancomycinresistant Enterococcus faecium (VREfm). Three of these patients developed bacteraemia. AFLP typing showed that the outbreak was caused by a single clone. Extensive environmental contamination was found in multiple wards. Within nine months after the detection of the outbreak no new VRE cases were detected.

Conclusion
We implemented a control strategy based on targeted screening and isolation in combination with implementation of general precautions and environmental cleaning. The strategy was less stringent than the Dutch national guideline for VRE control. This strategy successfully controlled the outbreak, while it was associated with a reduction in the number of isolation days and the number of cultures taken.

Background
In line with the increased global spread of multi-drug resistant microorganisms, the prevalence of vancomycin resistance among enterococci is rising. Whereas vancomycin-resistant enterococci (VRE) were detected only sporadically in the Netherlands before 2012, an increasing number of VRE outbreaks have required considerable resources to contain over the past eight years (1). At present, VRE outbreaks are among the largest and the most frequently reported in the Netherlands (period 2018 -June 2019: VRE 682 patients in 24 outbreaks versus MRSA 93 patients in 20 outbreaks) (2). Due to low virulence of VRE and its ability to survive on hospital environmental surfaces, VRE outbreaks have the potential to become substantial in size before they are detected by routine clinical cultures.
Strategies to prevent VRE transmission include screening of contact patients and isolation precautions of (suspected) VRE carriers (3,4). The Netherlands has a national control strategy of highly resistant micro-organisms including VRE (5). During outbreaks, contact patients (those admitted to the same room or ward as the index patient) are pre-emptively isolated while awaiting test results to prevent further spread. In this context, there is ongoing discussion about the number of rectal swabs to be tested -with culture considered as gold standard -before a VRE suspected patient can be declared VRE negative. The negative predictive value of 1 negative rectal swab is considered insu cient, and the Dutch national guideline advocates taking 3-5 rectal cultures on separate days (5). The combination of late outbreak detection and multiple cultures per contact-patient before VRE carriage can be excluded can result in very large numbers of patients to be isolated and screened, hence it is a substantial burden for hospital infection control departments, lead to signi cant laboratory costs, and exhaust hospital isolation facilities.
We describe the successful control of a VREfm outbreak in a general hospital in The Netherlands, while implementing a screening policy consisting of a single instead of multiple rectal swab culture (with enrichment broth) for excluding carriage in VRE-suspected patients.

Design, setting and participants
We describe the interventions that were implemented to control an outbreak of VREfm in a 364-bed general hospital with approximately 25,000 admissions per year in a setting which is non-endemic for VRE.
The Admiraal De Ruyter Hospital (ADRZ) is a general hospital with four separate locations in south-west of the Netherlands. The hospital has a catchment area of 248 000 inhabitants and supplies 85-90% of the requested hospital care in this area. For tertiary care patients are transferred to the surrounding academic centres. Most patients were admitted to an acute admission unit (AAU), consisting of six single rooms and eight multi-bed room with ve beds each (total 46 beds), before being transferred to their speci c wards (on average after 48 hours). The hospital has four locations but all VREfm colonised patients were detected on the location Goes.
Participants were all patients admitted to the Admiraal De Ruyter Hospital (ADRZ) between September 1, 2014 and February 5, 2017. A case was de ned as any patient infected or colonised with vanA mediated VREfm, multilocus sequence type (MLST) ST117. Patients were categorised into three groups according to their VRE status and potential risk of VRE carriage: (1) VRE carrier: VRE-positive patients who met the case de nition; (2) VRE suspected patients: all patients with prior hospitalisation in the ADRZ hospital location Goes from September 1, 2014. In the beginning of the outbreak it was unclear what the initial source of the outbreak was, therefore patients transferred from nursing homes or rehabilitation centres were also categorised as VRE suspected in the rst phase of the outbreak. Patients with no prior hospitalisation in the ADRZ hospital in the outbreak period, nursing home or rehabilitation centre were categorised as (3) low-risk patients.

Interventions
The control measures could be divided into three phases based on the different screening and isolation policies. Table 1 summarizes the dates, isolation and screening policies per phase and risk group. In Table 2, an overview of the implemented control measures during the outbreak is shown. Healthcare workers wear gown with long sleeves and gloves before each contact with the patient or the patient environment. Rooms were daily cleaned, medical devices were cleaning and disinfected (alcohol 70%) before leaving the room, and after patient discharge, the room is cleaned and disinfected (250 ppm chlorine) Table 2 Overview of the implemented control measures during VREfm outbreak period In the last and third phase (November 14, 2016 -February 5, 2017) the point prevalence survey was changed from a weekly to a monthly screening of all patients hospitalised for more than two days.
From February 5, 2017, only patients hospitalised for more than seven days were screened every week.

Audits, cleaning and education
Cleaning tasks had to be performed by nurses or by dedicated cleaning personnel depending on the objects. During audits it became clear that the tasks had not been de ned clearly and consequently some items were not always cleaned. As an intervention the tasks were speci ed in writing and subsequently the cleaning responsibilities were clearly de ned. Also, damaged hospital equipment and furniture were repaired or replaced, and the cleaning frequency of sanitary facilities was increased. Audits on implementation of infection control measures and cleaning practises were performed, including adenosine triphosphate (ATP) measurements (6) performed by the infection prevention department. The ATP measurements were performed in patient rooms and common areas and to control cleaning after discharge (data not shown). All healthcare workers and cleaning personnel received mandatory training on standard infection control and cleaning policies. The number of alcohol-based hand rub (ABHR) dispensers was increased so ABHR was available at 'point of care' in all wards.

Culturing and typing
Environmental sampling To assess the extent of environmental contamination in general (surveillance) and after cleaning and disinfection of the patient rooms, a range of high touch surfaces and (medical) equipment's were sampled using 10 cm x 10 cm sterile gauzes moistened with sterile saline (7) Microbiology Rectal swabs or feces was collected by nursing staff using the eSwab medium (Copan, Murrieta, USA). In total, 100 µL eSwab transport medium was transferred to a brain heart infusion broth containing 4 mg/L amoxicillin. After an overnight incubation at 35-37 °C, 10 µL of the broth was transferred to a Columbia colistin nalidixic acid -agar with 5% Sheep Blood and vancomycin (6 µg/mL) and grown overnight at 35-37 °C. For all suspected colonies growing on the selective media, species identi cation and susceptibility testing was performed using automated systems (Vitek MS and Vitek 2) (bioMérieux, Marcy l'Etoile, France) and E-test (bioMérieux, Marcy l'Etoile, France).

Molecular methods
Resistance genes were detected using an in house vanA/vanB duplex polymerase chain reaction (Elisabeth-TweeSteden Hospital, Tilburg, The Netherlands).
DNA was extracted using the QIAsymphony DSP virus/pathogen midi kit and pathogen complex 400 protocol of the QIAsymphony Sample Processing (SP) system (Qiagen, Hilden, Germany). Ampli cation reactions were performed in a volume of 25 µL with PCR mastermix (QuantiTect Multiplex PCR NoROX Kit, QIAgen) and 10 µL DNA sample. A multiplex PCR using vanA-, vanB-, and E. faecium-speci c primers and probes (Table 3)

Detection of the outbreak
During the last months of 2014 and in January 2015, VREfm strains were isolated ve times from clinical materials from patients that were admitted in the hospital (Fig. 1). These were initially considered to be incidental ndings without a clear epidemiological link. Retrospectively, they were typed and it was shown that they were strongly related.

Control of the outbreak
No further cases occurred over a three months period and control measures were terminated in February, 2017. Admission and prevalence surveillance cultures were discontinued and all outbreak related 'VRE suspected' labels in the electronic patient system were removed. Furthermore, a hospital-wide VRE rectal screening limited to patients admitted for at least 7 days, was implemented as a standard surveillance form that moment on.

Environmental cultures
In January 2016, environmental cultures were obtained throughout the hospital to assess the extent of environmental contamination. The cultures showed extensive VRE contamination on the surgical, internal, pulmonary and neurology wards (43/80 samples VRE positive; 53.7%). Environmental samples of the AUU, ICU and dialysis department were VRE negative (0/60 samples). (Fig. 2a) In June 2016, environmental screening was repeated on multiple wards (n = 130 samples), and again extensive VRE contamination was found in the surgical ward (19/20 samples, 95,0%) and to a lesser extent on the cardiology ward (4/20 samples; 20.0%). Consequently, stepwise cleaning and disinfection (250 ppm chlorine) of these wards was performed. After cleaning these wards were closed pending VRE negative environmental results. Following a peak in VRE transmission, environmental surveillance was continued and intensi ed: from June 2016, rooms previously occupied by VRE-positive patients were only released after cleaning and negative environmental cultures. Ten percent (74/713 culture) of the room surfaces remained VREfm positive after terminal disinfection. (Fig. 2b) In some cases, VREfm was still detected after two rounds of terminal disinfection on e.g. patient bed, infusion pole and pull-up bar.
Infections during the outbreak period Eight (5.7%) patients developed a VREfm infection, of whom three (2.1%) had bacteraemia. Two of these patients, with extensive co-morbidities, died shortly after detection. One patient, also with extensive co-morbidities (including renal failure, haemodialysis and vascular disease) developed a severe osteomyelitis following a surgical procedure, which eventually led to amputation of her left hand.

Discussion
Here we describe the successful control of a VREfm outbreak in a hospital using a mitigated screening and isolation policy, as compared with the national guidelines. With this approach, within nine months after the detection of the outbreak no new VREfm cases were detected and after twelve months the outbreak was considered fully controlled. Besides the targeted screening and isolation there was an intensive focus on optimisation of environmental cleaning procedures.
In general, there is no consensus on the optimal VRE screening, isolation and surveillance protocol, re ected by the variation in infection control approaches within and between countries (11)(12)(13). Especially, the number of rectal cultures required to consider a patient (known carrier or contact patient) VRE-negative is unclear. Studies show that the sensitivity of a single rectal swab is low, ranging from 42,5-79% (14)(15)(16)(17), and this increases when taking multiple swabs: Pearman et al. showed that on average four rectal swabs, collected on separate days, were needed to detect 95% of carriers compared to 56% with one rectal swab (13). Explanations for the increase in sensitivity when taking multiple rectal swabs include a uctuation in faecal excretion of VRE, and/or the presence of an intestinal transit time after VRE is transmitted (time between transmission event and detectable VRE levels in the faeces). It should be noted that in most of these studies the rectal swabs were inoculated directly on selective media. Addition of a broth enrichment step (as done in our study) increases the yield of a rectal swab culture substantially (18). Lastly, sensitivity may depend on the load of VRE in faeces (17); a high VRE load in faeces also results in a higher sensitivity of a rectal swab, whereby patients with lower faecal loads probably contribute less to transmission.
In 2015, a Dutch guideline was published which recommends taking 3-5 rectal cultures on separate days to reliably exclude carriage in a suspected VRE carrier (5). This guideline makes no speci c recommendation about the timing of the 3-5 separate rectal culture collection, apart from recommending that the last culture is ideally performed at least seven days after the last exposure. During our outbreak, the screening policy consisting of a single rectal swab culture (with enrichment broth) upon re- It should be noted that in our study, for the majority of the cohort of suspected VRE patients the last exposure was > 7 days ago (for many patients even several months), which (in combination with intensive focus on general precautions and environmental cleaning) may have been an important reason for the success of this strategy. In addition, the prevalence among re-admitted patients was < 1%, indicating a low background risk of undetected introductions of VRE in the hospital.
As described we are careful to generalize our ndings to other settings. Our results suggest, however, that the current general Dutch recommendations to take 3-5 cultures to exclude VRE carriage in all exposed patients may be reconsidered for centres with lower complexity of care, especially when the last exposure was > 7 days ago. This lowers the costs and limits the duration of pre-emptive isolation.

Conclusion
To conclude, we describe a large VRE outbreak in a general hospital in The Netherlands, that was successfully controlled, while substantially reducing the number of cultures to be taken and the number of isolation days, and thereby cutting laboratory costs. Ethics approval and consent to participate

Abbreviations
The data of patients used in this study were part of routine clinical practices in the ADRZ hospital and their anonymous use is beyond the scope of the Medical Research Involving Human Subjects Act.

Consent for publication
Not applicable Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests