Dr Benjamin Parcell Consultant in Medical Microbiology NRS Career Research Fellow

1. Clinical Perspectives in Integrating Whole Genome Sequencing into the Investigation of Healthcare and Public Health Outbreaks Dr Benjamin Parcell Consultant in Medical Microbiology NRS Career Research Fellow NHS Tayside, Ninewells Hospital and Medical SchoolDundee

2. Contents 1. Establishing a WGS service 2. Results • Bacteria and suspected outbreaks • Clinical benefits of translating genomics into clinical practice • WGS can provide results with greater granularity than routine typing methods • Genomic analysis can enhance the detection of ‘alert organisms’ • WGS could replace the need for multiple tests • WGS can be utilised to investigate new resistance mechanisms • Genomic analysis can be used to rule out outbreaks 3. When, what and how much should we sequence? 4. WGS clinical decision aid 5. The challenges and solutions for going forward 6.Recommendations so far...

3. Definitions A healthcare infection incident may be: An exceptional infection episode • A single case of any serious illness which has major implications for others (patients, staff and/or visitors), the organisation or wider public health e.g. infectious diseases of high consequence such as VHF or XDR-TB. A healthcare associated infection outbreak • Two or more linked cases with the same infectious agent associated with the same healthcare setting over a specified time period. or • A higher than expected number of cases of HAI in a given healthcare area over a specified time period.1 Local surveillance systems should be set to have a trigger/ threshold that prompts IPC action.

4. Consequences of outbreaks • Significant risk to patient safety • Result in discomfort, anxiety, pain, disability, death. • Costly and time consuming • Disruption due to ward closures or cancellation of procedures. • Patients/Staff/Relatives may also have to undergo screening tests • Nasal, perineal, throat and rectal swabs taken to determine the extent of transmission and risk of infection.

5. What happens when an outbreak is suspected? • Problem Assessment Group (PAG) or Incident Management Team (IMT) meeting is convened. • These are multidisciplinary meetings that allow planning of IPC measures and discussion of further testing required to identify cases. • BUT determining whether transmission has occurred is a major challengefor Infection Prevention & Control Teams (IPCTs) for the following reasons: • Difficult to establish epidemiological links between patients and the environment. • Typing results do not always have sufficient granularity or robustness to unequivocally define strains and confirm that transmission has taken place. • Results may not be available until after the outbreak is over.

6. WGS • Major developments in WGS in terms of ease of use, cost and turnaround time. • It has been used previously to investigate the following outbreaks to name a few: • Mycobacterium tuberculosis • Vibrio cholerae • E. coli 0104 • MRSA • Neisseria meningitidis • Legionella2,3,4,5,6,7,8,9 • It now has the capacity to allow outbreaks to be detected in real-time, enabling rapid implementation of targeted IPC measures. It may also rule out transmission and prevent the need for beds to close, minimizing disruption to clinical services. • Limited data on how best to establish a WGS service for real-time outbreak investigation in clinical settings.

7. Setting up a WGS service for outbreak investigation • Relevant reference laboratories for routine typing: • Antimicrobial Resistance and Healthcare Associated Infections Reference Unit (AMRHAI), Public Health England, Colindale • Scottish MRSA Reference Laboratory, Glasgow • Scottish AMR Satellite Reference Laboratory, Glasgow Clinical specimens collected during routine care First processed by standard microbiology methods for microorganism ID and antibiotic susceptibility testing: Dundee Microbiology (NHS Tayside) Aberdeen Microbiology (NHS Grampian) Bacterial isolates sent in parallel Infection Group School of Medicine, University of St Andrews for WGS

8. Alert Organisms • Health Protection Scotland Guidance10

9. Statistical process charts (SPC) charts • Show data chronologically and show natural or unnatural variation. • Centre line shows the average number of alert organisms per month. • Trigger line (trigger) represents a warning limit. • Upper control limit is the limit of natural variation and any results above this is unnatural variation and out of statistical control.11

11. Infection Group WGS referral form • A referral form was developed to send bacterial isolates to the Infection Group for WGS. • This captures bacterial, clinical and epidemiological information.

12. Results –bacteria and suspected outbreaks • Over 400 bacterial isolates sequenced so far.. • Community healthcare outbreaks included: MRSA , E. coli hospital outbreaks, Group A Streptococcus and E. coli in care homes, optrA gene positive linezolid resistant E. faecalis, borderline oxacillin-resistant Staphylococcus aureus(BORSA) in the community and secondary care. • Secondary care outbreaks included: Listeriamonocytogenes, Pseudomonas aeruginosa, Enterobactersppand Klebsiella spp. in neonatal intensive care units (NICU) and adult ICU, maternity ward Group A Streptococcus, orthopaedic vancomycin-resistant Enterococcusfaecium (VREfm), respiratory clinic P.aeruginosa and renal ward carbapenemase-producing Enterobacteriaceae (CPE) outbreaks.

13. Results: Clinical benefits of translating genomics into clinical practice • WGS can provide results with greater granularity than routine typing methods • Genomic analysis can enhance the detection of ‘alert organisms’ • WGS could replace the need for multiple tests • WGS can be utilised to investigate new resistance mechanisms • Genomic analysis can be used to rule out outbreaks Initial findings presented as a poster at HIS conference, Liverpool, 2018

14. WGS provided greater granularity than routine typing

15. WGS provided greater granularity than routine typing Example 1: Two patients with invasive Listeria monocytogenesinfection. • Routine typing = serotype 4/ clonal complex 1 • 1 in 6 invasive Listeria isolates are clonal complex 1. • Both isolates underwent WGS and were mapped to the reference chromosome of strain F2365 and SNPs called against this, found to be indistinguishable and therefore highly likely to be epidemiologically linked. • Prompted repeated hospital kitchen inspections which identified the handling of salads/meat did not meet national recommendations. Catering facilities were temporary closed until remedial action was undertaken.

16. WGS provided greater granularity than routine typing Example 2:Application of WGS to 5 neonatal intensive care unit (NICU) suspected outbreaks : • 1x Pseudomonas aeruginosa(3 patients) • 2x Klebsiellapneumoniae(2 patients , 4 patients) • 1x Klebsiellaoxytoca(7 patients) • 1x Enterobacterasubriae(2 patients) • Overall WGS results provided greater detail of the diversity amongst isolates within patients and the environmental niches sampled. • WGS results for the suspected NICU P. aeruginosaoutbreak were available before conventional reference laboratory results (10 vs 12 days). • It was not possible to conventionally type the K.oxytocaoutbreak isolates by pulse-field gel electrophroresis (PFGE) as the DNA was found to degrade however WGS revealed most K.oxytocaisolates were ST178 and closely related within 2 SNPs difference. Oral e-poster presentation ECCMID 2019 Amsterdam

17. Genomic analysis enhanced the detection of ‘alert organisms’

18. Genomic analysis enhanced the detection of ‘alert organisms’ Example 1 - WGS applied to the investigation of a VREfm outbreak • Two patients were identified to be VREfm positive in urine cultures (both admitted to same orthopaedic rehabilitation ward, identical antibiograms). • Further screening samples taken – one patient was colonised with 2 strains of VREfm (one from rectal swab and another from urine) each strain related to different outbreak clusters. • Over a two year period eleven further positive patients identified on a Surgical High Dependency Unit (SHDU) and a Renal ward. Typing by PFGE identified 5 clusters in total (three ST80 clusters, one ST64 culture and one ST203 cluster). • WGS revealed there was only one ST80 cluster in total (subtle interlinking between the ST80 isolates rather than numerous discrete clusters).

19. Genomic analysis enhanced the detection of ‘alert organisms’ • Example 1 (cont.) • WGS identified two vancomycin sensitive Enterococcusfaecium (VSEfm) isolates from two separate patients (previously identified during a separate VSEfm outbreak in ICU the year before) were related to the ST64 cluster. The four isolates were differentiated by only 21 SNP sites suggesting a common course. • The investigation also revealed that interhospital transmission had occurred between local hospitals and also a regional hospital carrying out renal transplants. Our results support findings from: • A retrospective study in the UK in which Raven et al applied WGS to the genetic characterisation of 293 VREfm bacteraemia isolates: • Majority of isolates were hospital-acquired and over 50% of isolates were highly related. • In 32% of cases transmissions occurred over years/across wards. • There was a mixture of vancomycin-resistant and –susceptible antibiotic profiles -due to isolates having lost or gained transposons carrying vancomycin resistance gene (vanA).12

20. WGS can replace the need for multiple tests

21. WGS can replace the need for multiple tests • WGS can be used to streamline outbreak investigations • Replace unnecessary routine laboratory testing • Potentially cut down transport of isolates to various reference laboratories Example 1: • Variable number tandem repeat (VNTR) typing and pulsed-field gel electrophoresis (PFGE) were both required for confirming transmission in a P.aeruginosa ICU outbreak. • When WGS was applied no other typing method was required. Parcell BJ, Oravcova K, Pinjeiro M, Holden MTG, Phillips G, Turton JF, Gillespie SH. Pseudomonas aeruginosa intensive care unit outbreak: winnowing of transmissions with molecular and genomic typing. J Hosp Infect 2018; 98(3):282-288.

22. WGS – one step for P.aeruginosaICU outbreak confirmation Parcell BJ, Oravcova K, Pinheiro M, Holden MTG, Phillipls G, Turton JF, Gillespie SH.Pseudomonas aeruginosa intensive care unit outbreak: winnowing of transmissions with molecular and genomic typing. J Hosp Infect 2018; 98(3):282-288.

23. WGS can replace the need for multiple tests • WGS could streamline testing during CPE outbreak investigations. Example 2- Outbreak involving three renal patients infected with blaKPCpositive ST258 Klebsiella pneumonia WGS could have replaced: • Multiplex polymerase chain reaction (PCR) assays in Ninewells Hospital, Dundee and the Glasgow Satellite Laboratory • PCR assays identifying different capsular types, regulators of mucoviscosity (rmpA and rmpA2 and wcaG) • VNTR analysis in Colindale. • In this outbreak WGS supported transmission and revealed that one of samples was mixed with an E.coliwhich had not been identified on routine testing. • Potential for cost savings Presented as a poster at ECCMID 2019 Amsterdam

24. WGS can replace the need for multiple tests Integration of WGS into CPE diagnostic workflow

25. WGS can be utilised to investigate new resistance mechanisms

26. WGS can be utilised to investigate new resistance mechanisms • WGS can determine movement of different resistance elements. Example 1 - WGS identified three patients from NHS Grampian to have Enterococcus faecalis with new plasmid-mediated optrAgene. • WGS can highlight the risk that linezolid resistance can transfer therefore providing support to IPCTs that patients require isolation to prevent onwards transmission. Example 2- WGS findings can be used to inform the development of PCR based screening which is particularly useful when phenotypic methods or antibiotic susceptibility testing is not accurate e.g. BORSA.

27. Genomic analysis can be used to rule out outbreaks

28. Genomic analysis can be used to rule out outbreaks • Outbreak group meetings can be halted and targeted measures discontinued freeing hospital services from closure and disruption e.g. extra cleaning regimes, increased domestic staff input. • Of value in outbreaks with high consequences e.g. NICU/ICU/Maternity Example - Group A Streptococcus identified from neonatal umbilical swabs samples were taken a day apart on a maternity ward. Patients were both treated with antibiotics. • The two Group A Streptococcuswere different (ST101 and ST382) both differing in all 7 alleles providing reassurance that transmission had not occurred.

29. Staff attending PAGs/IMTs • WGS may save staff time if outbreaks can be refuted by WGS. • This could equate to 38 hours of staff time (including the minute takers time) each time a 2 hour meeting is held.

30. When, what and how much should we sequence? BORSA optrAlinezolid resistant E.faecalis VRE/VSE NICU K.oxytoca ICU P.aeruginosa CPE Group A Strep MRSA Gram negative ESBL producing bacteria

31. WGS Clinical Decision Aid – where is the value of WGS?

32. Overview of issues and solutions

33. Recommendations so far... • WGS is the ultimate typing tool enabling IPCTs to carry out outbreak investigations efficiently and effectively. We have identified that: • It can assist in the formulation of case definitions and support or refute hypotheses in relation to lines of transmission, rule out outbreaks, minimising disruption to healthcare services. • It is particularly useful in outbreaks involving rare organisms or when conventional typing is unable to unequivocally show whether isolates are linked or not e.g. BORSA, Klebsiellaoxytoca • WGS is useful as a first line tool for the investigation of new resistance mechanisms such as the optrA resistance. • The IPCT should consider WGS of sensitive enterococci in VRE outbreaks and consider asking for a variety of samples and repeat samples as patients may carry more than one VRE strain. • WGS findings can be used to inform the development of polymerase chain reaction (PCR) based screening e.g. BORSA • Real-time sequencing could be used to streamline clinical microbiology services by reducing unnecessary testing. • Implementing WGS as a standard of care in real-time would be a major advance in day-to-day IPC practice.

34. Acknowledgements and thanks • NHS Microbiology/Infection Prevention and Control /Management NHS Tayside and NHS Grampian • Jane F. Turton, Antimicrobial Resistance and Healthcare Associated Infections Reference Unit (AMRHAI), Public Health England, Colindale • National Reference Laboratories (Antimicrobial Resistance and Healthcare Associated Infections Reference Unit (AMRHAI), Public Health England, Colindale, Scottish MRSA. Reference Laboratory and the Scottish AMR Satellite Reference Laboratory, Glasgow • Matthew T.G. Holden • Stephen H. Gillespie • Kerry Pettigrew • Katarina Oravcova • Martin McHugh • Prince Agyirey-Kwakye • Miguel Pinheiro • Wai-Lum Sung Graphic Designer at the University of Aberdeen

35. Funding for Sequencing • Bioinformatics and Computational Biology analyses were supported by theUniversity of St Andrews Bioinformatics Unit which is funded by a Wellcome TrustISSF award [grant 097831/Z/11/Z]. • BP, KO, MP, MTGH, GP and SHG are funded by the Chief Scientist Office through the Scottish Infection Research Network, a part of the SHAIPI consortium.

36. References • National Infection Prevention and Control Manual. Chapter 3 - Healthcare Infection Incidents, Outbreaks and Data Exceedance • Mellmann A, Harmsen D, Cummings CA et al. Prospective genomic characterization of the German enterohemorrhagic Escherichia coli O104:H4 outbreak by rapid next generation sequencing technology. PLoS One 2011; 6: e22751 • Ko¨ser CU, Holden MTG, Ellington MJ et al. A neonatal MRSA outbreak investigation using rapid whole genome sequencing. N Engl J Med 2012; 366: 2267–75. • Harris SR, Cartwright EJ, Torok ME et al. Whole-genome sequencing for analysis of an outbreak of meticillin-resistant Staphylococcus aureus: a descriptive study. Lancet Infect Dis 2013;13(2)130-6 • Vogel U, Szczepanowski R, Claus H et al. Ion torrent personal genome machine sequencing for genomic typing of Neisseriameningitidis for rapid determination of multiple layers of typing information. J ClinMicrobiol 2012;50(6):1889-94 • Reuters S, Harrison TG, Ko¨ser CU et al. A pilot study of rapid whole-genome sequencing for the investigation of a Legionella outbreak. BMJ Open 2013 2013 Jan 9;3(1). pii: e002175. doi: 10.1136/bmjopen-2012-002175. Print 2013. • Walker TM, Ip CL, Harrell RH et al. Whole-genome sequencing to delineate Mycobacterium tuberculosis outbreaks: a retrospective observational study. Lancet Infect Dis 2013;13(2):137-46. • Chin C-S, Sorenson J, Harris JB, et al. The origin of the Haitian cholera outbreak strain. N Engl J Med 2011;364:33e42. • Marra MA, Jones SJ, Astell CR, et al. The genome sequence of the SARS-associated coronavirus. Science 2003;300:1399e1404 • Appendix 13 – NHSScotland Alert organism/Condition list. Part of the National Infection Prevention and Control Manual (NIPCM), available at: http://www.nipcm.hps.scot.nhs.uk/ Produced by: Health Protection Scotland, June 2017 • Some information on Statistical Process Control (SPC) c charts that may be useful for clinical teams . Health Protection Scotland • Raven KE, Gouliouris T, Brodrick H, Coll F, Brown NM, Reynolds R, et al. Complex Routes of NosocomialVancomycin-Resistant Enterococcusfaecium Transmission Revealed by Genome Sequencing Clin Infect Dis. 2017 Apr 1;64(7):886-893. doi: 10.1093/cid/ciw872.