key: cord-0749472-xmq3n2sy
authors: Lucey, Mary; Macori, Guerrino; Mullane, Niamh; Sutton-Fitzpatrick, Una; Gonzalez, Gabriel; Coughlan, Suzie; Purcell, Aisling; Fenelon, Lynda; Fanning, Séamus; Schaffer, Kirsten
title: Whole-genome sequencing to track SARS-CoV-2 transmission in nosocomial outbreaks
date: 2020-09-19
journal: Clin Infect Dis
DOI: 10.1093/cid/ciaa1433
sha: bc65bfb4273c2556da857028816c58d4761b498b
doc_id: 749472
cord_uid: xmq3n2sy

BACKGROUND: During the first wave of the coronavirus disease 2019 (COVID-19) pandemic, outbreaks of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in healthcare institutions posed a significant problem. Due to limited evidence, guidance on appropriate infection prevention and control (IPC) measures such as the wearing of face masks varied. Here, we applied whole virus genome sequencing (WvGS) to analyse transmission routes of SARS-CoV-2 in hospital-acquired (HA) COVID-19. METHODS: An investigation was undertaken for all HA cases of COVID-19 from March to April 2020. Fifty SARS-CoV-2 samples were analysed by WvGS and their phylogenetic relationship established. RESULTS: WvGS identified transmission events previously undetected by epidemiological analysis and provided evidence for SARS-CoV-2 transmission between healthcare workers (HCW) and patients and among HCW themselves. The majority of HA COVID-19 cases occurred in patients highly dependent on nursing care, suggesting the likely route of transmission was by close contact or droplet, rather than aerosol, transmission. Mortality among HA COVID-19 infections was recorded as 33%. CONCLUSIONS: This study provides evidence that SARS-CoV-2 transmission occurs from symptomatic and asymptomatic HCWs to patients. Interventions including comprehensive screening of HCWs for COVID-19 symptoms, PCR testing of asymptomatic HCWs upon identification of HA cases and implementation of universal use of surgical masks for all clinical care is indicated to prevent viral transmission. Our study highlights the importance of close collaboration between guidance bodies and frontline IPC experts for developing control measures in an emergency pandemic situation caused by a virus with undefined transmission modus.

The coronavirus disease 2019 caused by the novel virus SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) was first identified in Wuhan, Hubei, China following an unusual cluster of pneumonia in the region. 1 The risk of SARS-CoV-2 transmission within healthcare institutions has been well described. 2, 3 China, the USA and several European countries have reported high rates of COVID-19 infection amongst healthcare workers (HCWs). [3] [4] [5] In Ireland, the Health Protection Surveillance Centre reported 8,215 positive cases amongst HCWs (up to the 1 st of July), amounting to 32% of all of cases diagnosed nationally. Over 1,687 outbreaks/clusters have occurred in Ireland secondary to local transmission, of which hospital outbreaks comprise 10% . 6 Whole genome sequencing (WGS) has been utilized for the investigation of outbreaks of multi-drug resistant bacteria such as carbapenemase-producing Enterobacterales and other clinically significant pathogens. [7] [8] [9] [10] The WGS amplicon strategy has been used to obtain high-quality consensus sequences of SARS-CoV-2. 11, 12 Application of this approach, allows for higher resolution analysis, enhancing our understanding of transmission patterns, whilst guiding policies designed to prevent disease spread.

At the start of the pandemic, recommendations on usage of personal protective equipment (PPE), particularly masks, varied between countries . 13 Initial infection prevention and control guidance recommended the use of surgical masks only for HCWs involved in the care of COVID-19 patients. 14 Environmental contamination with SARS-CoV-2 in hospital was perceived to be lower risk than person-to-person transmission 15 Due to fears regarding PPE shortages, mask use was not advised when caring for asymptomatic persons or for HCWs in nonclinical areas. 16 In this report, we apply the amplicon-based Illumina sequencing protocol to the investigation of nosocomial outbreaks of SARS-CoV-2 and demonstrate how knowledge of transmission networks can better inform IPC

A c c e p t e d M a n u s c r i p t 4

This study was undertaken in a tertiary referral centre. As per national policy, a hospital-acquired (HA) COVID-19 case was defined as a patient admitted without symptoms suggestive of COVID-19 who had SARS-CoV-2 RNA detected at least seven days after admission. 17 In order to apply greater precision to our inclusion criteria we further categorised patients by length of admission greater than 14 days which took into consideration the full incubation period of the virus. Symptoms suggestive of COVID-19 were fever, cough, myalgia, fatigue, headache, anosmia and diarrhoea.

HA COVID-19 cases were identified by laboratory records from March 7 th to May 10 th , 2020. An outbreak was classified as two or more cases of HA COVID-19 on a ward. Visiting of patients was restricted to compassionate care only from March 13 th . Masks, gowns and gloves were worn when assessing suspected or confirmed cases of COVID-19. Positive cases were transferred to a designated COVID ward. Close contacts were isolated with droplet precautions for 14 days.

During March 2020, HCWs were advised by occupational medicine to remain home and attend community testing services if they experienced fever and cough or shortness of breath. On March 24 th , in-house SARS-CoV-2 PCR for HCWs was established; the screening criteria for COVID-19 were expanded on April 6 th to include sore throat, headache, myalgia, fever, cough, anosmia or ageusia. Epidemiological criteria collected on SARS-CoV-2 infected HCWs were date of PCR test and the ward they had been working on.

A close contact was defined as a HCW or patient who spent more than 15 minutes face-to-face within 2 metres of a confirmed case or patients who shared a multi-bedded room with a confirmed case for more than 2 hours. PCR Viral RNA was purified from nasopharyngeal swabs using the MagNA Pure 96 system (Roche Diagnostics). Complementary DNA was obtained from isolated RNA through reverse transcription and multiplex PCR according to the protocol provided by the Artic Network initiative. 18 Libraries were prepared using the NEBNext Ultra II kit (New England Biolabs) and sequenced on an Illumina MiSeq using 300-cycle v2 reagent kits (Illumina). Bowtie 2 was used for aligning the sequencing reads to the reference genome for SARS-CoV-2 (GenBank number, MN908947.3) and SAMtools for manipulating the alignments. 19 SNPs were used to define clusters and a median-joining network was generated including these data from this study and an additional 1,000 strains collected from GISAID available on May 22 nd . Clade annotation was included for the Pangolin, GISAID and NextStrain systems. 20,21

Ethical approval Approval was granted by the institution"s ethics committee.

The first COVID-19 patient presented to our hospital on March 7 th . In Ireland the number of reported cases increased steadily from February 29 th , plateauing at the end of March, with a sustained peak until mid-April ( Figure S1 ). Consistent with this, the number of COVID-19 cases admitted daily to our hospital peaked on A c c e p t e d M a n u s c r i p t 6 A fourth outbreak occurred on medical wards 6 and 8 between the 2 nd and 9 th of April. The index case on ward 8 was originally considered a single hospital-acquired case, but when two additional patients were diagnosed simultaneously three days later on the neighbouring ward 6, all other inpatients on ward 6 and ward 8 were tested. All patients were in single en-suite rooms. The two wards shared nursing teams and clinical teams and several staff members working on the ward had been diagnosed with COVID-19. Six additional SARS-CoV-2 infected patients were detected by screening.

At the same time two outbreaks occurred on surgical wards 7 and adjacent ward 9 involving 7 patients. Similar to the medical wards, the first two patients became symptomatic and additional cases were detected following testing of all the other inpatients: Five patients were in single en-suite rooms and two patients were sharing a room. At the time no obvious epidemiological link with the concomitant outbreak on the medical ward could be established.

In response to the increase in HA cases the occupational health testing criteria were changed. A concern that COVID-19 positive staff members experiencing milder symptoms, such as headache and sore throat, were being missed, led to a decision to broaden the testing criteria to include any other symptom possibly indicating SARS- corresponding HCWs were sequenced to completion. The average sequence quality per samples was > 99% for 46 samples, and between 92 and 94% for 4 samples (Table S1 ). Sequenced SARS-CoV-2 with tiles average quality < 89% were excluded in the study. Phylogenetic analysis identified six independent groups of which clusters 1-3 were related to 39 patients. The combination of the phylogenetic cluster data (Figure 2) , the ward, sampling dates, and classification as HCW or patient, uncovered a hitherto unrecognised association among samples distributed over time (Figure 3) . In a global context, cases clustered in distinct areas of the phylogenetic tree suggesting a diverse genetic origin of the virus (Figure 4 ).

Cluster 1 contained the majority of specimens associated with the outbreak on ward 1 at the end of April ( Figure   2 ). This provides evidence of HCW-patient transmission as 5 HCWs were positive prior to the ward screening on April 29 th . Many patients on this ward did not have direct contact with one another due to immobility. Cluster 3 as shown on the tree illustrates the outbreaks on wards 6/8 and 7/9, which were considered distinct at the time as these wards are geographically separate employing different HCWs. The phylogenetic tree reveals indistinguishable strains on these wards. One of the HCWs identified in this cluster was a "floater". "Floaters" are responsible for the personal care of highly dependent patients and are assigned to different wards on a daily basis. Two "floaters" working between these 4 wards tested positive, providing an epidemiological link between outbreaks. A positive HCW in this cluster had no direct patient contact indicating that transmission may occur in non-clinical areas between HCWs.

All three clusters stem from one main branch and are broadly the same hospital strain with minor sequence variations per ward. The fourth and fifth clusters are suggestive of different strains, originating from distinct branches. A patient in the fourth cluster was identified epidemiologically as part of the ward 1 outbreak on April Clinical characteristics are described in Table 1 . The majority of patients required assistance with mobility (65%) and self-care (77%). The crude mortality rate amongst HA cases was calculated at 33%. As the majority had a ceiling of care established, (i.e. not for cardiopulmonary resuscitation or mechanical ventilation) only 3 patients were admitted to ICU. Overall, 14 patients (27%) died from complications of COVID-19.

HA COVID-19 presents an ongoing threat to vulnerable patients with a high level of dependency and comorbidity. To prevent hospital outbreaks of COVID-19 it is essential to identify modes of viral transmission in order to implement effective IPC measures.

Our study highlights for the first time the analytical strength of WvGS specifically for the investigation of hospital outbreaks of SARS-CoV-2 and how these data can facilitate improved IPC practices.

A study of nosocomial infection amongst HCWs and their families was undertaken in Wuhan. Nine separate clusters were identified and WvGS was utilised to sequence 12 strains revealing a high level of dissemination across transmission chains. Only HCW strains were sequenced and therefore genomic links between HCW and patients were not established. 22 WvGS accurately identified dissemination routes of SARS-CoV-2 across Australia. Provisional contact tracing suggested an epidemiological link between shared HCWs and 54 cases of COVID-19. However, phylogenetic analysis revealed that the cases were divided into 4 distinct clusters with 15 cases linked through WvGS to a social event unrelated to the health-care setting. 23 Unlike other studies published to date, we sought to use WvGS to enhance our knowledge of hospital transmission and to provide supportive evidence for IPC interventions introduced during the outbreak. 24 Through WvGS, links between outbreaks were established and points of both HCW-to-patient and HCW-to-HCW transmission identified. The majority of HA clusters were broadly linked to the same strain suggesting a high level of intra-hospital transmission. Prior to sequencing, clear epidemiological links between cases were often not apparent. WvGS data suggested the majority of HA cases were spread via HCWs as many patients A c c e p t e d M a n u s c r i p t 9 were either immobile or in single en-suite rooms. Furthermore, phylogenetic analysis highlighted the potential of WvGS to distinguish separate strains introduced into the hospital.

A number of changes brought about a decline in HA cases starting with a change to staff testing allowing testing of HCW with mild viral symptoms. This was the first step in minimising HCW-patient transmission.

At the start of the pandemic, decisions around PPE were particularly influenced by a lack of supply supporting the use of masks only for activity involving COVID-19 patients. 16 On the 31 st of March the European Centre for Disease Prevention and Control stated that the use of surgical masks should be considered for all HCWs. 25 Despite an increasing number of hospitals reporting outbreaks, national guidance advised using masks only for the care of COVID-19 patients until the end of April. 26 By April 13 th , five separate ward outbreaks had occurred in our centre, with a significant number of patients epidemiologically linked to positive staff contacts. In addition, clear evidence of pre-symptomatic transmission reported in the literature added to concerns regarding asymptomatic spread of infection between HCWs and patients. 27 Masks were mandated for all HCW-patient interactions in clinical areas and no HA cases were diagnosed for ten days. On June 5 th the WHO issued guidance supporting the continuous use of masks by HCWs in areas with community transmission or large-scale outbreaks. 28 Despite the introduction of universal mask use in clinical care, an outbreak occurred at the end of April.

Restricted space in non-clinical areas posed a challenge for social distancing. The reported removal of masks upon entry made it a likely point of transmission between staff. In addition, we described an agitated wandering patient positive for COVID-19. A number of studies highlighted the phenomenon of a "superspreader". [29] [30] [31] It is possible that this patient amplified the transmission of COVID-19 via close contact with HCWs. This particular outbreak suggests that masks should be worn at all times, including in administrative areas and particularly where social distancing cannot be observed. The transmission of SARS-CoV-2 via environmental contamination has been hypothesised as a potential source of hospital outbreaks. 32 Transmission via contaminated surfaces or other fomites to hands of HCW or to patients was felt to be unlikely in this study, as hand hygiene and environmental cleaning compliance remained above 90% in regular audits.Our study identified common characteristics amongst patients across all hospital disciplines with HA COVID-19. Most of the hospital outbreaks described to date are confined to specific patient cohorts. 33, 34 The average age was higher than studies describing both community and HA cases. 35 Hence a large proportion of patients required assistance with mobility and self-care, thereby putting them at greater risk of infection from asymptomatic staff.

A c c e p t e d M a n u s c r i p t 10 The overall mortality rate of 33% amongst HA cases is consistent with other studies describing hospital outbreaks and is not a reflection of overall COVID-19 deaths but rather the frailty of this patient population. 32 A nosocomial outbreak study of SARS-CoV-2 in a London hospital assessed the impact of various IPC measures.

The study reported a reduction in HA cases once PPE (gloves, gowns and surgical masks), were introduced for all hospital interactions. Unlike our study, the majority of cases were attributed to patient-patient transmission in a shared ward. The study acknowledged that in the absence of sequencing data, all putative transmission routes were merely inferred. 32 The majority of HCW-patient exposure occurred before the introduction of universal surgical masks. A minority of patients constituted contacts of one another in multi-bedded rooms. However not all patients in six-bedded rooms became infected. This suggests that the majority of transmission is spread via droplets and close contact, as patients with high levels of dependency were shown to be at greater risk than independent patients occupying the same room. Outside of aerosol generating procedures, it is likely that aerosolization occurs in rare circumstances, such as in superspreading events described above. The infectivity of SARS-CoV-2 containing aerosols has been demonstrated in tissue-culture assays. 36 This may provide an explanation for the extent of the outbreak which occurred at the end of April.

The focus of our epidemiological analysis was to identify specific patterns of SARS-CoV-2 transmission within our hospital. We demonstrate the role of WvGS for the purpose of outbreak investigation and show how sequencing informs our understanding of transmission patterns. Our study suggests that in nosocomial outbreaks, apart from aerosol generating procedures, the majority of transmission is spread via droplets and 

This work is based on research supported by University College Dublin. We would like to thank staff on all the medical and surgical wards and the occupational health department who helped in collecting epidemiological information. We would also like to thank all members of the microbiology department for providing SARS-CoV-2 PCR and for helping to retrieve specimens for sequencing, Dr O"Donnell and the staff of the health protection surveillance centre for providing Irish epidemiology data and Prof C.Taylor for editing the manuscript.

This study was funded by University College Dublin. 

Initial Cluster of Novel Coronavirus (2019-nCoV) Infections in Wuhan, China is Consistent with Substantial Human-to-Human Transmission

Transmission of COVID-19 to Health Care Personnel During Exposures to a Hospitalized Patient

Death from Covid-19 of 23 Health Care Workers in China

European Centre for Disease Prevention and Control. Coronavirus disease 2019 (COVID19) in the EU/EEA and the UK -ninth update

Epidemiology of COVID-19 in Ireland daily reports

Genomics for molecular epidemiology and detecting transmission of carbapenemase-producing Enterobacterales in Victoria

Molecular characterisation of multi-drug resistant Escherichia coli of bovine origin

Whole-genome sequencing for analysis of an outbreak of meticillin-resistant Staphylococcus aureus: a descriptive study. The Lancet infectious diseases

Molecular epidemiology of methicillin-resistant and methicillinsusceptible Staphylococcus aureus in the ovine dairy chain and in farm-related humans

nCoV-2019 sequencing protocol

Spread of SARS-CoV-2 in the Icelandic population

Rationale use of face masks in the COVID-19 pandemic

Prevention and management of COVID-19 in community residential and inpatient facilities. Health Protection Surveillance Centre

Surfaces and equipment contamination by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in the emergency department at a university hospital

The sequence alignment/map format and SAMtools

Nextstrain: real-time tracking of pathogen evolution

Nosocomial outbreak of COVID-19 pneumonia in Wuhan, China

Tracking the COVID-19 pandemic in Australia using Genomics

Association Between Universal Masking in a Health Care System and SARS-CoV-2 Positivity Among Health Care Workers

European Centre for Disease Prevention and Control. Infection prevention and control preparedness for COVID-19 in healthcare settings. Second update 31 st

Guidance on the use of surgical masks when healthcare is being provided in the context of the COVID-19 pandemic. Health Protection and Surveillance

Temporal dynamics in viral shedding and transmissibility of COVID-19

Advise on the use of masks in the context of COVID-19. Interim guidance 5 th

Superspreading and the effect of individual variation on disease emergence

Transmission of SARS-CoV-2 by inhalation of respiratory aerosol in the Skagit Valley Chorale superspreading event

Estimating the overdispersion in COVID-19 transmission using outbreak sizes outside China

Nosocomial transmission of COVID-19: a retrospective study of 66 hospital-acquired cases in a London teaching hospital

First reported nosocomial outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in a pediatric dialysis unit

Clinical and Transmission Characteristics of Covid-19 -A Retrospective Study of 25 Cases from a Single Thoracic Surgery Department

Clinical Characteristics of 138 Hospitalized Patients with 2019 Novel Coronavirus-Infected Pneumonia in Wuhan, China

Aerosol and Surface Stability of SARS-CoV-2 as Compared with SARS-CoV-1

A c c e p t e d M a n u s c r i p t 13 M a n u s c r i p t 15 Neurological conditions including epilepsy, multiple sclerosis, spinal cord injury and Parkinson"s disease A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t