Interhospital transmission of vancomycin-resistant Enterococcus faecium in Aomori, Japan

Spread of vancomycin-resistant Enterococcus (VRE) is a global concern as a significant cause of healthcare-associated infections. A series of VRE faecium (VREf) outbreaks caused by clonal propagation due to interhospital transmission occurred in six general hospitals in Aomori prefecture, Japan. The number of patients with VREf was obtained from thirty seven hospitals participating in the local network of Aomori prefecture. Thirteen hospitals performed active screening tests for VRE. Whole genome sequencing analysis was performed. The total number of cases with VREf amounted to 500 in fourteen hospitals in Aomori from Jan 2018 to April 2021. It took more than three years for the frequency of detection of VRE to return to pre-outbreak levels. The duration and size of outbreaks differed between hospitals according to the countermeasures available at each hospital. Whole genome sequencing analysis indicated vanA-type VREf ST1421 for most samples from six hospitals. This was the first multi-jurisdictional outbreak of VREf sequence type 1421 in Japan. In addition to strict infection control measures, continuous monitoring of VRE detection in local medical regions and smooth and immediate communication among hospitals are required to prevent VREf outbreaks.


Introduction
Vancomycin (VCM)-resistant enterococci (VRE) are listed by the Centers for Disease Control and Prevention (CDC) of the United States of America (US) as antibiotic-resistant bacteria that present "serious threats" to human health [1]. Spread of VRE is a global concern as a significant cause of healthcare-associated infections [2]. According to data collected by European Centre for Disease Prevention and Control (ECDC) from the European Antimicrobial Resistance Surveillance Network (EARS-Net) and the Central Asian and European Surveillance of Antimicrobial Resistance (CAESAR), significantly increasing trends for vancomycin resistance in E. faecium were observed for the period 2016-2020, similar to the previously reported trends for 2015-2019 when the United Kingdom was included [3].
Healthcare-associated VRE infections have been reported in Japanese hospitals since the beginning of this century [4]. As emergence of VRE is uncommon in Japan, even a single case at one hospital is considered an outbreak. A series of vanA-positive E. faecium outbreaks

Data collection
The number of patients sampled for all bacteriological tests in thirty seven hospitals participating in the Aomori Infection Control Network (AICON) was obtained through the Microbiological Information Network Aomori (MINA). AICON is a local infection control network managed by the Infection Control Center at Hirosaki University Hospital. AICON covers approximately 50% of all hospitals including all general hospitals in Aomori prefecture and two laboratory centers. Participating hospitals can share information about infection control through the network. MINA is a bacteriological database run by AICON. The microbiological test results obtained at each hospital were uploaded via an information network and stored in web servers located and managed at the Infection Control Center of Hirosaki University Hospital. The information processing systems for the database was supplied by KD-ICONS (Tokyo, Japan).
The number of patients with VREf detected by active screening was obtained from thirteen hospitals participating in AICON, all of which performed active screening tests for VRE. These data were also included in the database of MINA. Screening examination of stool samples and/or rectal swabs in all patients who had a preceding hospitalization and in new admissions was formally performed between April 2019 and March 2021 according to the protocol of each hospital and the informed consent. Because the expense of screening was burdensome for the hospitals, it was performed at the discretion of each hospital. In all general hospitals and in hospitals at which an outbreak occurred, simultaneous screening examination was carried out for all inpatients at least twice and new admissions during a certain period (for 1-2 years). At other hospitals, a single simultaneous screening for all inpatients and/or screening for at least new high-risk admissions was carried out.
Data regarding the official proportion of VREf on E. faecium isolated from all samples in all 47 prefectures of Japan were obtained from Japan Nosocomial Infection Surveillance (JANIS), which is a national surveillance program organized by the Ministry of Health, Labour and Welfare of Japan. Its clinical laboratory division collects information regarding the number of nosocomial infections and antimicrobial-resistant bacteria in Japanese hospitals. As of January 2020, JANIS covered 2,418 hospitals, including > 80% of hospitals with 500 beds or more [5,6]. JANIS member hospitals are required to submit surveillance data on a regular basis. National data of drug-resistant bacteria in Japan are publicly available on the JANIS website [7]. VRE was defined following the standard provided by the infectious diseases control law (Act on the Prevention of Infectious Diseases and Medical Care for Patients with Infectious Diseases [8]) as Enterococcus spp. resistant to vancomycin (VCM) with MIC of ≥ 16 μg/mL in the microdilution method, which is the definition used by JANIS since 2007, and which corresponds to > 4 μg/mL as defined in the international guidelines.
For economic reasons, pulsed-field gel electrophoresis (PFGE) and genome typing could not be applied to all isolates. Accordingly, these analyses were applied to 20 samples from five general hospitals where outbreaks occurred and from Hirosaki University Hospital, which were able to preserve and provide samples during the series of VRE outbreaks in Aomori prefecture.

Identification of VRE
VRE was identified by the clinical laboratory at each affected hospital according to the CLSI (Clinical and Laboratory Standards Institute). Active screening for VRE was done using stool samples and/or rectal swabs. Prepared fecal samples were inoculated onto VRE selective agar (Becton Dickinson and Company, Franklin Lakes, NJ, USA). Identification of the isolates and drug susceptibility tests were performed by MicroScan WalkAway using Combo panel PC1J (Dade MicroScan, Sacramento, CA, USA). The Biotyper matrix-assisted laser desorption/ionization system (Bruker Daltonics GmbH and Co., Bremen, Germany) in combination with MicroScan WalkAway and PM1J panel was utilized. Several samples isolated in the early phase of outbreaks at each hospital were sent to the Aomori Prefectural Public Health and Environment Center for detection of vanA and vanB genes using multiplex PCR (polymerase chain reaction).

Whole genome sequencing
Genomic DNA was extracted with QIAamp DNA purification kit (QIAGEN, Hilden, Germany) according to the manufacturer's instructions. DNA libraries were prepared for sequencing with Enzymatics 5X WGS reagents (BioStream Co., Ltd, Tokyo, Japan) and then pooled. For Illumina sequencing, paired-end sequencing (2 × 300 bp) was performed using the MiSeq reagent kit v3 on the MiSeq platform. Raw reads were assembled using Shovill v1.0.9 [9] with the default settings. Multilocus sequence typing (MLST) determination was performed using

Outbreak report Setting
Aomori prefecture has six major medical regions: Hachinohe, Kamitosan, Aomori, Tsugaru, Seihokugo, and Shimokita ( Fig. 1). The Hachinohe, Aomori and Tsugaru regions each have a tertiary general hospital with more than 500 beds. There are several medium-sized general hospitals in these six medical regions. During the period from January 2018 to April 2021, outbreaks of VREf which were strongly suspected as interhospital transmission occurred in 14 hospitals in Aomori prefecture, indicated as A-N in chronological order according to the date of the first VREf-positive case identified at each hospital ( Fig. 1), and included two tertiary general hospitals. Six hospitals (A, B, C, D, E, and G) had insidious intra-institutional spread of VREf (Fig. 2), and required long-lasting strict countermeasures to end the outbreak. Hospitals B, C, D, and G had infection control teams (ICTs) including full-time staff and belonged to the AICON at the time of onset of the outbreaks, while Hospitals A and E did not belong to AICON and did not have any infection control specialist (Table 2). Before the outbreak, AICON had regularly held lectures, workshops and meetings about standard precautions and environmental improvements, resistant bacteria, and proper use of antimicrobial. After the onset of the outbreaks, the executive office of AICON discussed the outbreaks with member hospitals and each local public health center in Aomori prefecture, and shared the outbreak data. By the request of the executive office of AICON and local public health centers, an infection control doctor (ICD), infection control nurse (ICN) and

Index case (Hospital A)
Hospital A is a small hospital with < 100 beds in Hachinohe city and has a close relationship with Hospitals B and C in the same medical region (Fig. 1). The first VREf was detected at this hospital in pus from skin in January 2018, followed by a second case three weeks later.

Hospital B
Following the outbreak at hospital A, VREf outbreaks occurred in five general hospitals (B, C, D, E, and G) from 2018 to 2020 (Fig. 2). Hospital B, a general hospital with > 400 beds, received the report that one patient transferred from Hospital B to Hospital A was a VREf carrier from the health center of Hachinohe region in  were evaluated and encouraged. The number of VREfpositive patients returned to zero in January 2020.

Hospital G
Hospital G is a medium-sized general hospital in Kamitosan region, situated slightly separate from Hachinohe city. Patients are sometimes transferred from and referred from general hospitals B, C, and D (Fig. 1). The first VREf-positive patient was found in February 2019, and cases with VREf were found sporadically during periods of VREf outbreaks at other hospitals. The local health center of Kamitosan detected sporadic cases of VREfpositive patients in hospitals in their area of responsibility. Following the recommendation of the local health center, the first screening for VRE was carried out for hospitalized patients in December 2019, and found five VREf-positive patients. In addition to the reinforcement of routine infection control activities, VRE screening of patients at hospitalization was started after the first screening. The outbreak ended within six months.

Other hospitals
From 2019, AICON recommended that participating hospitals should screen for VRE at hospitalization, which was conducted at a total of 13 hospitals: 12 general hospitals and 1 special hospital. A small number of VREf-positive patients were detected by this screening, although it was difficult to discriminate between minimal intrahospital transmission and individual VRE carriages.
No new VRE-positive cases were detected after May 2021.

Epidemiological summary
The total number of cases with VREf amounted to 500 in 14 hospitals in Aomori prefecture from Jan 2018 to April 2021. Table 1

Apr-Sep Oct-Mar Apr-Sep Oct-Mar Apr-Sep Oct-Mar Apr-Sep Oct-Mar Apr-Sep
All samples (Cases) 30 2021. The rate of VREf on active screening samples was highest (0.11%) in the six months between April and September 2019 (Table 1).

Changes in the proportion of VREf on E. faecium in all 47
prefectures of Japan Figure 3 shows changes in the proportion of VREf on E. faecium in all 47 prefectures of Japan from January 2016 to December 2020, obtained from the JANIS surveillance data (number of cases) including all samples. These outbreaks at multiple hospitals in the same period caused a sharp peak in the proportion of VREf. The remarkable size of the area under the curve representing Aomori prefecture suggests the epidemiological importance of the series of VREf outbreaks in this prefecture.

Analysis of VRE strains using PFGE
The right in Fig. 4 lists six source hospitals (Hospitals B, C, D, E, and G and Hirosaki University Hospital) and sampling year for 20 isolates that were tested by PFGE analysis. VRE strains showed identical or very close patterns in PFGE analysis. Lanes 1, 3, 6, 10, 11, 12, 15, 16, and 18 were judged as identical strains. Lanes 2, 4, and 19 were also identical. The former and the latter were judged as closely related with more than 85% similarity, which indicates interhospital clonal spread of VRE in this series of outbreaks. Figure 5 shows the phylogenetic tree based on the core genome SNPs of the same 20 isolates. Whole genome sequencing analysis indicated vanA-type VREf ST1421 for most samples except samples 7 and 8 from hospital D. These were identified as other vanB-type strains, although sample 6, also from Hospital D, was identified as vanA-type VREf ST1421, the same as the strains from Hospitals B, C, E, and G.   ; however, the frequency of detection of VRE was still low in Japan as of 2019 [14]. Using the ECDC surveillance data as a measure of bloodstream infection (BSI) incidence could underestimate the true value in countries with low blood culture frequency [15]. The data show that resistance to vancomycin in E. faecium varied substantially among countries in Europe and Central Asia. In 2020, resistance percentages < 1% were reported by 7/38 (18%) countries reporting data on VREf, whereas percentages ≥ 25% were found in 13/38 (34%), and 4/38 (11%) reported resistance percentages ≥ 50% [3]. Unfortunately, JANIS does not hold data of the resistance percentage of E. faecium in blood culture. However, the resistance percentage in all samples in Japan remained at a level around 1.0-1.5% except the period from January to July in 2019, when the outbreaks and active screenings in Aomori prefecture were at peak levels (Fig. 3). The resistance percentage of E. faecium in blood culture in Aomori was 0% for several years until March 2018, but was 29.4% for the six months from October 2018 to March 2019 and 22.5% for the year from October 2018 (Table 1). These rates are exceptionally high for Japan. It is not easy to prevent VRE spread without active screening because patients infected (colonized) with VRE are usually asymptomatic. By the time an index VRE carrier is found by chance, several asymptomatic patients in proximity to the index case can be found by spot surveillance, as was the case in hospitals in the present series. It takes months for colonized patients to be free from VRE [16]. Furthermore, colonized patients show no obvious signs of infection but can spread VRE through person-toperson contact during hospitalization.

Countermeasures to VRE in six hospitals (A, B, C, D, E, and G)
Christiansen et al. [17] reported that one hundred sixty-nine cases in 23 wards were colonized with a single strain of VREf in a major Australian hospital. The following interventions were introduced: (1) Formation of a VRE executive group; (2) Rapid laboratory identification; (3) Screening of all inpatients with isolation of carriers and cohorting of contacts; (4) Environmental screening and increased cleaning; (5) Electronic flagging of the medical records of contacts; and (6) Antibiotic restrictions. Using these interventions, the outbreak was terminated within only 3 months, but at a cost of 2.7 million Australian dollars (1.9 million US dollars). In a 1600bed hospital in Singapore, eradication of a hospital-wide outbreak comprising 151 cases required a coordinated strategy comprising: (1) Formation of a VRE task force; (2) Hospital-wide screening; (3) Isolation of carriers; (4) Physical segregation of contacts; (5) Surveillance of highrisk groups; (6) Increased cleaning; (7) Electronic tagging of VRE status; and (8) Education and audits [18]. Furthermore, Fournier et al. reported that 45 repeat outbreaks of VREf occurred, comprising 533 cases, between 2004 and 2010 in a 23,000-bed multi-hospital institution in France [19]. During that period, a multidrug-resistant bacteria control program was implemented, including the following measures: (1) Rapid reporting to ICT; (2) Stopping transfers of cases and contact patients to other units or to any other hospitals; (3) Particular attention to daily cleaning of VRE patient environments; (4) Extended screening of contact patients already discharged or transferred from the involved unit; (5) Maintained screening of all contact cases until the outbreak was controlled; (6) Identification of discharged VRE patients and contact patients in case of readmission; and (7) Cohorting patients in three distinct areas with dedicated nursing staff: "VRE patients" section, "Contact patients" section and "VREfree patients" section for newly admitted patients. The number of cases per outbreak was significantly lower after implementation of this program.
These strategies can be summarized as follows.

< Strategies for infection prevention and control of VRE infections >
(1) Formation of a VRE executive ICT and rapid reporting to ICT (2) Rapid laboratory identification of VRE (3) Rapid and repeated hospital-wide screenings (4) Cohorting patients with dedicated staff into sections: "VRE patients"; "Contact patients"; and "VRE-free patients" (5) Stopping transfers of VRE patients and contact patients to other units or to any other hospitals (6) Extended and maintained screening of contact patients already discharged or transferred until the outbreak is controlled (7) Flagging of medical records for identifying discharged VRE patients and contact patients in case of readmission (8) Environmental screening and increased cleaning. (9) Antimicrobial stewardship.
In addition to the above, the common understanding of the need for screening, patient selection, and evaluation of the clinical and economic benefits based on the frequency of detection of VRE should be shared by cooperating hospitals within medical regions. Interhospital patient transfer is commonly required to improve patient management; however, transferring patients can promote interhospital transfer of multidrug-resistant organisms (MDROs). Therefore, "Information sharing using a local network for infection control" could be added as the 10th strategy for preventing MDROs such as VRE.
We investigated these 10 items in the six hospitals that had experienced intra-institutional spread of VRE (Table 2). Countermeasures to "contact patients" were totally insufficient, and identification of VRE and hospital-wide screening were delayed, particularly in two small hospitals (A and E) that did not belong to AICON at that time. Hospitals A, B, C, and E are in the same medical region and have close relationships with each other (Fig. 1). Interhospital transmission of VREf is strongly suspected to have occurred from the above reasons. It is difficult to define an interhospital transmission, but comparison between the consistently low frequency of detection of VRE in Japan (Fig. 3) and the remarkably high proportion of VRE on E. faecium isolated from blood culture and in all samples from Aomori prefecture (Table 1), which is very uncommon in Japan, indicate interhospital transmission. Interhospital VRE transmissions have been reported in Australia [20], Taiwan [21], and Kyoto (Japan) [22]. If VRE is detected in a hospital, the information should be shared by other hospitals in the medical region to prevent interhospital transmission. In their systemic review, Urlich et al. (2017) [23] reported that the duration of VRE outbreak in hematology and oncology departments ranged from 1 to 36 months, with a mean value of 11 months. Strong countermeasures including active VRE screening are necessary to shorten the duration of VRE outbreaks [24][25][26].
Increases in medical costs due to spread of VRE is a matter of concern in countries such as the United States [27], Germany [28], France [29], China [30] and Canada [31]. The cost-saving effect has been shown in a detailed analysis of the costs associated with implementation of a strict policy for controlling spread during outbreaks, including active screening for highly resistant microorganisms [32]. Hospital administrators and infection control specialists need to discuss the costs and benefits of VRE screening to achieve efficient and effective countermeasures based on the medical conditions in each area [33]. PFGE and whole genome sequencing analysis clearly indicated that the multi-jurisdictional outbreaks were caused by vanA-type VREf ST1421, whereas 2 of 20 isolates showed different patterns of PFGE and were identified as other vanB-type strains by whole genome sequencing. The vanA and vanB types could generally be distinguished by susceptibility, but confirmation by genetic analysis was useful. The ST1421 clone lacks the pstS housekeeping gene of the MLST allele [34]. pstSnull VREf clones have recently emerged worldwide and regional spread has been reported in numerous countries, including Australia [34], the UK [35], Scotland [36], Denmark [37] and Korea [38]. To the best of our knowledge, this is the first report of a VRE outbreak by pstSnull VREf ST1421 in Japan.

Study limitation
Data assessed in the present study were obtained from the MINA database, which targeted all clinical samples from approximately 50% of all hospitals (including all general hospitals) and laboratory centers belonging to AICON in Aomori prefecture. However, data from small sized hospitals and clinics which do not make it their routine work to collect bacteriological samples were not included. Therefore, it is possible that some VRE carriers could have been missed.
Furthermore, because our study was limited regarding clinical data from VRE-positive patients, we could not investigate the burden of the infectious disease as compared to asymptomatic carriage. That would have provided more useful information, as symptomatic/infected individuals would likely accumulate higher associated medical costs.

Conclusion
Interhospital transmission of VREf ST1421 occurred in Aomori prefecture despite a prefecture-wide infection control network that provides data of microbiological test results uploaded from 37 hospitals in the prefecture. This was the first multi-jurisdictional outbreak of VREf sequence type 1421 in Japan. In addition to strict infection control measures and a system of continuous monitoring of MDROs, sustained interest in the frequency of detection of VRE in local medical regions and smooth and immediate communication among hospitals are required to prevent VRE outbreaks.