[ad_1]
Summary
Background
The burden and influence of health-care associated severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections is unknown. We aimed to examine the use of rapid SARS-CoV-2 sequencing combined with detailed epidemiological analysis to investigate health-care associated SARS-CoV-2 infections and inform infection control measures.
Methods
In this prospective surveillance study, we set up rapid SARS-CoV-2 nanopore sequencing from PCR-positive diagnostic samples collected from our hospital (Cambridge, UK) and a random selection from hospitals in the East of England, enabling sample-to-sequence in less than 24 h. We established a weekly review and reporting system with integration of genomic and epidemiological data to investigate suspected health-care associated COVID-19 cases.
Findings
Between March 13 and April 24, 2020, we collected clinical data and samples from 5613 patients with COVID-19 from across the East of England. We sequenced 1000 samples producing 747 high-quality genomes. We combined epidemiological and genomic analysis of the 299 patients from our hospital and identified 35 clusters of identical viruses involving 159 patients. 92 (58%) of 159 patients had strong epidemiological links and 32 (20%) patients had plausible epidemiological links. These results were fed back to clinical, infection control, and hospital management teams, leading to infection-control interventions and informing patient safety reporting.
Interpretation
We established real-time genomic surveillance of SARS-CoV-2 in a UK hospital and showed the benefit of combined genomic and epidemiological analysis for the investigation of health-care associated COVID-19. This approach enabled us to detect cryptic transmission events and identify opportunities to target infection-control interventions to further reduce health-care associated infections. Our findings have important implications for national public health policy as they enable rapid tracking and investigation of infections in hospital and community settings.
Funding
COVID-19 Genomics UK funded by the Department of Health and Social Care, UK Research and Innovation, and the Wellcome Sanger Institute.
Introduction
originating from an intermediate animal host.
Owing to the error prone nature of the viral replication process, RNA viruses, such as SARS-CoV-2, accumulate mutations over time resulting in sequence diversity. The current mutation rate of SARS-CoV-2 is estimated to be approximately 2·5 nucleotides per month.
Sequencing of SARS-CoV-2 can provide valuable information on virus biology, transmission, and population dynamics.
,
,
,
When linked with detailed epidemiological data and on a timescale of days, genomic data can support epidemiological investigations of potential hospital-acquired infections. On a larger population scale, genomic surveillance of SARS-CoV-2 can inform which lineages of the virus are circulating in the human population, how these change over time as an indicator of the success of control measures, how often new sources of virus are introduced from other geographical areas, and how the virus evolves in response to interventions.
Worldwide, more than 22 000 cases of COVID-19 infection in health-care workers were reported in the WHO situation report from May, 2020, which is likely to be an underestimate.
As the number of community-acquired COVID-19 cases reduces, health-care settings are likely to act as reservoirs of infection. Identifying transmission events in these settings will therefore become increasingly important to manage outbreaks and effectively monitor infection control.
Evidence before this study
Since the emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in December, 2019, the infection has spread worldwide, infecting more than 9·1 million people and causing more than 472 000 deaths as of June 23, 2020. Despite investigation, substantial gaps remain in our understanding of virus biology and transmission. We searched PubMed Central, medRxiv, and bioRxiv with combinations of “SARS-CoV-2”, “genome”, or “genomic” and “hospital-acquired” or “healthcare-associated” for articles published in English from database inception until May 11, 2020, and returned few relevant results. Previous studies have analysed SARS-CoV-2 biology, diversity and evolution, transmission networks, and health-care worker infections. Very few have applied genomic epidemiology to tackle health-care associated infection and, to our knowledge, none have been at the scale of a hospital COVID-19 epidemic comprising hundreds of patients in real-time. Attempts with other pathogens have often been assessed retrospectively, and in a timeframe that was not actionable.
Added value of this study
We present the first report, to our knowledge, applying rapid genome sequencing to systematically investigate SARS-CoV-2 health-care associated infections, integrating genomic and epidemiological data to identify transmission networks and inform targeted infection control interventions. In 6 weeks, we sequenced 1000 genomes, including 70% of all COVID-19 cases tested at our hospital. We uncovered ward outbreaks of hospital-acquired infections and substantial transmission in health-care associated community settings. Genomic analysis identified cryptic transmission that had not been suspected by clinical or infection control teams. These complex transmission networks involved patients and health-care workers and spanned hospital and community health-care settings. By feeding results back to the hospital weekly, the data could be actioned by the infection control team during the course of outbreak investigations.
Implications of all the available evidence
Rapid viral sequencing can contribute to health-care associated infection investigations by uncovering evidence for or against transmission events. As transmission shifts from the community to health-care settings, the use of rapid sequencing integrated with epidemiological investigations can help to reveal complex transmission chains and inform targeted infection control and public health interventions. This strategy could support the new test, track, and trace initiative of the UK Government, enabling a more targeted approach to disease control.
We aimed to examine the use of rapid sequencing of SARS-CoV-2, combined with detailed epidemiological analysis, to investigate health-care associated COVID-19 infections and to inform infection control measures in our hospital.
Methods
Study design and participants
The study was done as part of surveillance for COVID-19 under the auspices of Section 251 of the National Health Service Act 2006. It therefore did not require individual patient consent or ethical approval. The COVID-19 Genomics UK (COG-UK) study protocol was approved by the Public Health England Research Ethics Governance Group.
Procedures
Most samples were from patients with symptomatic COVID-19; two samples were from asymptomatic individuals.
Samples were assigned COG-UK sequencing codes that were integrated back into the master metadata file. For samples that were not sequenced locally, the remaining RNA extract was collected from the diagnostic microbiology laboratory and sent to the Wellcome Sanger Institute (Hinxton, UK) for sequencing as part of the COG-UK consortium. 14 samples sequenced on site were also sent to the institute to compare consistency across sequencing platforms. Each week, clinical metadata and sequencing data were combined and formatted for upload to the Medical Research Council CLIMB system. Data manipulations were done in R (version 3.6.2) with the tidyverse packages (version 1.3.0) installed onto computers within the Trust network.
Amplicon libraries were sequenced using MinION flow cells version 9.4.1 (Oxford Nanopore Technologies, Oxford, UK). Genomes were assembled with reference-based assembly and a bioinformatic pipeline
with 20 × minimum coverage cutoff for any region of the genome and 50·1% cutoff for defining single nucleotide polymorphisms (SNPs).
Analysis
Phylogenetic trees were produced with IQ-TREE
and visualised in Microreact
for weekly hospital reports and the R package ggtree (appendix pp 2–3).
A pairwise SNP distance matrix was produced from the alignment with use of the snp-dists package. Viral lineages
were assigned with the PANGOLIN package, version 1.07.
TableDefinitions of health-care associated COVID-19
Role of the funding source
The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding authors had full access to all the data in the study and had final responsibility for the decision to submit for publication.
Results
(including two asymptomatic individuals). 374 patients with PCR-confirmed COVID-19 were tested at CUH between March 10 and April 24, 2020 (figure 1, appendix pp 7–8). The median age was 64 years (range 0–98) and 233 (62%) of 374 were male. 74 (20%) of 374 patients were admitted to critical care units and 75 (20%) died. Excluding the health-care workers screening samples, 262 (70%) of 374 CUH COVID-19 samples had sequencing data available (figure 1; appendix p 9). 57 (15%) infections were suspected or highly likely to be hospital-acquired, of which 49 (86%) had genome sequences available. A further 32 (9%) admissions were community-acquired but likely to be health-care associated, and nine (2%) were health-care workers (not counting health-care workers identified through screening). The CUH epidemic curve showed that weekly admissions peaked in week 4 (commencing March 30, 2020) and then declined (figure 2). The UK went into full lockdown on March 23, 2020. In the early stages of the epidemic, community-onset and community-acquired infections predominated but the frequency of health-care associated infections increased from March 23 until April 6, 2020, and then declined.
Figure 1Study profile
We prioritised CUH samples for nanopore sequencing on site for quick turnaround to investigate health-care associated infections. Of the 166 samples that did not pass quality control, two were removed as their genomes were less than 29 kb and 164 were removed as they had more than 2990 undefined nucleotides. SARS-CoV-2=severe acute respiratory syndrome coronavirus 2. CUH=Cambridge University Hospitals National Health Service Foundation Trust. *Of 3891 samples assigned for sequencing, 2940 were uploaded to the CLIMB server, as have all of the 1000 genomes sequenced on site, for national COVID-19 Genomics UK analyses. †The 37 health-care workers were identified through a CUH screening programme. In addition, nine self-presented to CUH admission units and are counted as patients (n=46).
Figure 2Epidemic curve of COVID-19 at CUH
Data for 374 patients tested at CUH. Classification of infection: (1) community onset, community associated; (2) community onset, suspected health-care associated; (3) hospital onset, indeterminate health-care associated; (4) hospital onset, suspected health-care associated; (5) hospital onset, health-care associated; (6) health-care worker. These data do not include 37 health-care workers identified through the CUH screening programme, but do include nine health-care workers that self-presented for testing. Note that data for week commencing April 20, 2020, stops at 8 AM on April 24, 2020, so does not include the weekend. CUH=Cambridge University Hospitals National Health Service Foundation Trust.
Most samples in both the East of England and CUH belonged to lineage B.1. There were no lineage A samples, which have mainly been identified in China, the USA, South Korea, and Australia (figure 3, appendix pp 12–13).
Figure 3SARS-CoV-2 lineages identified in the East of England and CUH
Map of the East of England region (A) showing the breakdown of SARS-CoV-2 lineages by collecting hospital site. Lineage B.1 was the most prevalent lineage throughout the region (B). SARS-CoV-2=severe acute respiratory syndrome coronavirus 2. CUH=Cambridge University Hospitals National Health Service Foundation Trust.
Figure 4Phylogenetic tree of SARS-CoV-2 genomes from CUH
Figure 5Epidemiological timeline of ward A cluster
Six patients on ward A were diagnosed with COVID-19. All had been admitted for more than 7 days before their specimen date and were considered likely to have hospital-acquired infections. The date of the first positive sample collection is shown with a red box and patient death date indicated with a solid black bar. Five of the viral genomes (patients A to E) had zero single nucleotide polymorphism differences between them, and one of the viral genomes (patient F) differed by one single nucleotide polymorphism.
In the community cluster, 18 patients were admitted to CUH with COVID-19 between April 5 and April 21, 2020, with genetically identical viruses. Nine patients were residents at a community care home (care home A). A review of medical records revealed that another patient in this genetic cluster worked in care home A and one was a retired nurse who worked in an unknown care home. Three cases were paramedics and two were nurses (who worked in different wards at CUH but lived with paramedics). The final case did not have any discernible epidemiological links with the others. In summary, this investigation revealed a cluster of cases with evidence of linked transmission coming from either the same care home, the ambulance service, or shared accommodation. None of these associations had been detected by clinicians or infection control.
The information from these combined epidemiological and genomic investigations was fed back to the clinical, infection control, and hospital management teams. This information triggered further investigations into patient isolation, ward cleaning procedures, use of PPE, and staff physical distancing behaviour. Health-care associated infections were also assessed in relation to potential harm caused to patients and recorded in the hospital’s patient safety reporting system for follow-up and further action.
Discussion
,
,
,
,
We sought to embed genomic surveillance as part of an active SARS-CoV-2 infection control process in a large UK hospital. A rapid sequencing workflow was established on March 23, 2020, with multiplex PCR-based nanopore sequencing, which has been shown to be effective in a wide range of clinical samples and viral loads.
We aimed to sequence all available positive samples from CUH and a selection from each of the East of England regional hospitals submitted to the diagnostic microbiology laboratory, linking with clinical metadata pulled from the hospital electronic patient records system. We also included 37 samples collected as part of a local health-care worker screening programme.
In 5 weeks, more than 1000 SARS-CoV-2 genomes had been sequenced including most of the CUH samples from this phase of the epidemic. We applied this system to investigate nosocomial and health-care worker COVID-19 cases at CUH, integrating genomics with epidemiological and clinical data.
We examined the diversity in SARS-CoV-2 at CUH and found that overall genetic diversity was low and reflected the pattern seen in the East of England as a whole, with most viruses belonging to the B.1 lineage. We identified a median of eight (range 0–24) SNP differences between viruses, and 4·5% of pairwise comparisons between CUH genomes had zero to one SNP differences. Given the virus’ mutation rate and infectious timeframe, cases might share linked transmission within a few days or weeks if there are fewer than approximately two SNP differences. We applied a more stringent definition of genetic clusters with zero SNP differences, and despite detailed epidemiological data, we found no identifiable connection between 22% of pairwise comparisons within clusters. This finding reflects the low genetic diversity in circulating SARS-CoV-2 during the study, and emphasises the need for in-depth epidemiological analysis to unravel potential transmission networks. The ability of genomics to resolve transmission events might increase as the virus evolves and accumulates greater diversity. Genetics can be used more confidently to rule out transmission—eg, if viruses from two patients suspected as being linked belong to different lineages.
We investigated groups of patients and health-care workers at CUH in response to queries from the clinical and infection control teams. Using ward location data for patients and health-care workers, we analysed epidemiological data to establish if there had been ward-based contact. We compared the genomes of patients and health-care workers in the suspected groups with those of other patients at CUH and in hospitals in the East of England to examine relatedness. This approach enabled us to add supporting evidence or to refute linked transmission between patients and health-care workers (eg, adding confidence to our assessment of whether an infection was hospital or community acquired). Ruling out transmission was useful as a mechanism to monitor and target infection control measures (eg, showing that viruses on the renal ward belonged to a distinct lineage from those in the outpatient dialysis unit). Genomic analysis also enabled us to identify cryptic transmission (ie, additional cases that were not initially suspected to be linked to the original ward or health-care worker clusters, including multiple instances of patients with identical viruses from the same care homes).
,
,
,
we reported results of our investigations to the clinical, infection control, and management teams on a weekly basis, thus enabling them to respond to this information and act accordingly within the timeframe of ongoing ward outbreaks. The genomic data informed reviews of patient placement and isolation procedures, assessment of PPE use, and staff break arrangements, supporting us to better focus efforts at a time of unprecedented demand on infection control teams. Finally, these analyses are being used to inform existing patient safety review processes within our hospital, including investigations related to hospital-onset COVID-19 in which the patient has come to harm.
Our study highlights the importance of understanding SARS-CoV-2 transmission within health-care settings in managing the pandemic. The transmission networks that we identified were complex, involving patients and health-care workers in both hospital and community settings such as care homes, outpatient units, and ambulance services, which have been poorly studied. Of note, we have identified transmission events on nine wards that were considered green (ie, no known patients with COVID-19) at the start of each cluster. Although there were strong epidemiological and genomic associations between cases, the mechanism and direction of transmissions within these clusters are unclear. The role of asymptomatic intermediates, fomites (including PPE), and the environment are not well understood and require further investigation. During the timeframe of this study, several infection-control interventions were implemented across the hospital, as well as national public health measures to reduce community spread. Understanding the interaction between such interventions and nosocomial transmission are complex (especially in the context of the comparably long incubation period for SARS-CoV-2 relative to other respiratory viruses), but essential in enabling health-care providers to safely deliver existing services in the context of a pandemic.
We acknowledge several limitations to our study. Firstly, we were unable to sequence all genomes from samples that were collected during the study period. We might therefore have missed the opportunity to investigate all potential transmission events. 166 (17%) of 1000 sequenced genomes did not pass our quality filtering. This finding reflects the stringent coverage threshold used (90%), the desire not to bias sample selection by sequencing only low Ct samples, and the nature of the diagnostic material used for sequencing. We only had main ward location data for health-care workers and could have overlooked potential epidemiological links with patients with whom they had contact on other wards or within shared communal areas. Due to its low genetic diversity, highly similar genomes could not be used definitively to infer meaningfully linked transmission events without supportive epidemiological data. However, our experience indicates that further investigation of genomic clusters with highly similar genomes can uncover previously unknown epidemiological links. Furthermore, we were able to use rapidly generated genomic data to investigate health-care associated infections within the hospital setting. Similar approaches could be applied in future studies to assess infections in health-care workers and community settings such as care homes. As the practical challenges associated with implementing real-time genome sequencing during epidemics are overcome, unlocking the real power of genomic epidemiology will require its integration with clinical and public health systems to support decision making on local, national, and international scales. This implementation might be of particular benefit in supporting the UK Government’s test, track, and trace initiative, enabling a more targeted approach to disease control.
Contributors
LWM contributed to the data collection, data analysis, data interpretation, figures, and tables. WLH, BW, and CJH contributed to the data collection, data analysis, data interpretation, figures, tables, literature review, and writing. MH, ASJ, MDC, SP, LGC, SLC, FAK, AY, GH, TF, SF, SS, MPW, SB, NB, EM, and TG contributed to the data collection, data analysis, and data interpretation. AP, IR, and MR contributed to the data collection, data analysis, data interpretation, and figures. SJP and GD contributed to the writing of the manuscript. MET and IG designed and supervised the study and contributed to the data collection, data analysis, data interpretation, literature review, and wrote the first draft of the manuscript. All authors reviewed and approved the final manuscript.
Declaration of interests
CJH, SLC, MPW, and SB report grants from Wellcome. GH reports grants from Rotary International. MET reports grants from Academy of Medical Sciences, Health Foundation, grants from Medical Research Council, grants and non-financial support from National Institute of Health Research, during the conduct of the study. MET has published books with Oxford University Press and receives royalty payments from them, personal fees from the Wellcome Sanger Institute, personal fees from University of Cambridge, outside the submitted work. IG reports grants from Wellcome; and grants from Medical Research Council part of UK Research & Innovation. All other authors declare no competing interests.
Acknowledgments
We acknowledge the assistance of the laboratory staff of Public Health England Clinical Microbiology and Public Health Laboratory for processing the diagnostic samples; the clinical teams (infectious diseases, microbiology, virology, infection control) at CUH for their assistance with the investigation of health-care associated infections; and the Wellcome Sanger Institute for sequencing 14 samples included in this study. We are grateful to Richard Smith (Head of Patient Safety, CUH), Lucy Rivett, Dominic Sparkes, Nick K Jones, and Matthew Routledge (for providing health-care worker data), and Anthony Underwood (Centre for Genomic Pathogen Surveillance) for helpful discussions and advice. This work was funded by COVID-19 Genomics UK, which is supported by funding from the Medical Research Council part of UK Research and Innovation, the National Institute of Health Research, and Genome Research, operating as the Wellcome Sanger Institute. It was also supported by the Wellcome (Senior Fellowship 207498/Z/17/Z and ARTIC Network Collaborative Award 206298/B/17/Z to IG, Collaborative Award 204870/Z/16/Z supporting CJH, Senior Research Fellowship to SGB 215515/Z/19/Z, Senior Clinical Research Fellowship 108070/Z/15/Z to MPW), the Academy of Medical Sciences and the Health Foundation (Clinician Scientist Fellowship to MET), and the National Institute for Health Research Cambridge Biomedical Research Centre at the CUH (BW, GD, MET). The views expressed are those of the authors and not necessarily those of the National Health Service, the National Institute of Health Research, or the Department of Health and Social Care.
Supplementary Material
References
- 1.
A pneumonia outbreak associated with a new coronavirus of probable bat origin.
Nature. 2020; 579: 270-273
- 2.
The proximal origin of SARS-CoV-2.
Nat Med. 2020; 26: 450-452
- 3.
Temporal signal and the phylodynamic threshold of SARS-CoV-2.
bioRxiv. 2020; ()
- 4.
Tracking virus outbreaks in the twenty-first century.
Nat Microbiol. 2019; 4: 10-19
- 5.
Towards a genomics-informed, real-time, global pathogen surveillance system.
Nat Rev Genet. 2018; 19: 9-20
- 6.
Clinical and biological insights from viral genome sequencing.
Nat Rev Microbiol. 2017; 15: 183-192
- 7.
Genomic epidemiology reveals multiple introductions of Zika virus into the United States.
Nature. 2017; 546: 401-405
- 8.
Clinical characteristics of 138 hospitalized patients with 2019 novel coronavirus-infected pneumonia in Wuhan, China.
JAMA. 2020; 323: 1061-1069
- 9.
Coronavirus disease 2019 (COVID-19) Situation Report–82.
- 10.
Screening of healthcare workers for SARS-CoV-2 highlights the role of asymptomatic carriage in COVID-19 transmission.
eLife. 2020; 9e58728
- 11.
- 12.
nCoV-2019 sequencing protocol.
- 13.
nCoV-2019 novel coronavirus bioinformatics protocol.
- 14.
MAFFT multiple sequence alignment software version 7: improvements in performance and usability.
Mol Biol Evol. 2013; 30: 772-780
- 15.
IQ-TREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies.
Mol Biol Evol. 2015; 32: 268-274
- 16.
Microreact: visualizing and sharing data for genomic epidemiology and phylogeography.
Microb Genom. 2016; 2e000093
- 17.
ggtree: an R package for visualization and annotation of phylogenetic trees with their covariates and other associated data.
Methods Ecol Evol. 2017; 8: 28-36
- 18.
A dynamic nomenclature proposal for SARS-CoV-2 to assist genomic epidemiology.
bioRxiv. 2020; ()
- 19.
COVID-19 Genomics UK (COG-UK) Consortium Weekly Report #1 23rd March 2020.
- 20.
COVID-19 data.
- 21.
Rapid outbreak sequencing of Ebola virus in Sierra Leone identifies transmission chains linked to sporadic cases.
Virus Evol. 2016; 2vew016
- 22.
Real-time, portable genome sequencing for Ebola surveillance.
Nature. 2016; 530: 228-232
- 23.
Virus genomes reveal factors that spread and sustained the Ebola epidemic.
Nature. 2017; 544: 309-315
- 24.
Whole-genome sequencing of measles virus genotypes H1 and D8 during outbreaks of infection following the 2010 Olympic Winter Games reveals viral transmission routes.
J Infect Dis. 2015; 212: 1574-1578
- 25.
Use of genome sequencing to define institutional influenza outbreaks, Toronto, Ontario, Canada, 2014–15.
Emerg Infect Dis. 2018; 24: 492-497
- 26.
Multiplex PCR method for MinION and Illumina sequencing of Zika and other virus genomes directly from clinical samples.
Nat Protoc. 2017; 12: 1261-1276
- 27.
Rapid whole-genome sequencing for investigation of a neonatal MRSA outbreak.
N Engl J Med. 2012; 366: 2267-2275
- 28.
Whole-genome sequencing for analysis of an outbreak of meticillin-resistant Staphylococcus aureus: a descriptive study.
Lancet Infect Dis. 2013; 13: 130-136
- 29.
Longitudinal genomic surveillance of MRSA in the UK reveals transmission patterns in hospitals and the community.
Sci Transl Med. 2017; 9eaak9745
- 30.
Use of whole-genome sequencing of adenovirus in immunocompromised pediatric patients to identify nosocomial transmission and mixed-genotype infection.
J Infect Dis. 2018; 218: 1261-1271
Article Info
Publication History
Identification
Copyright
© 2020 Elsevier Ltd. All rights reserved.
ScienceDirect
Linked Articles
[ad_2]
Source link