At the start of the SARS-CoV-2 pandemic, a key concern was how SARS-CoV-2 transmission would unfold amongst unvaccinated university students, and whether infections amongst students would impact the surrounding (and potentially more vulnerable) community. Importantly, understanding these dynamics is essential in order to devise effective infection control measures while minimising disruption to teaching, research, and the mental health of students and staff. We sought to address these questions at the University of Cambridge, taking advantage of the high case ascertainment provided by the student SARS-CoV-2 screening programmes and the sequencing capabilities of the COVID-19 Genomics UK Consortium (COG-UK).
By the first university term of the 2020/21 academic year, the University of Cambridge had implemented a cost-effective asymptomatic and symptomatic screening programme; students living in university accommodation could access weekly pooled asymptomatic testing (Figure 1) and all university members (students and staff) could access symptomatic testing. Simultaneously, COG-UK had established community SARS-CoV-2 surveillance sequencing in the United Kingdom, demonstrating the public health utility of studying the genomic epidemiology of SARS-CoV-2 through the investigation of outbreaks in various settings1,2.
The university setting was prime for a genomic epidemiological study:
- No large-scale genomic studies in university settings had been conducted and understanding transmission in this setting was a public health priority.
- The screening programme had consistently high levels of student engagement3,4 where the ascertainment of asymptomatic cases would minimise the limitation of under-sampling and missed transmission chains.
- The relatively closed university environment, and mixing of unvaccinated socially active students where variants of concern had not yet arisen, would enable significant insights into the transmission dynamics of SARS-CoV-2.
- Strong links between the local health protection team, Public Health England (now UKHSA), local hospitals, and the university, would facilitate the integration of detailed epidemiological and contact tracing information with genomic findings to provide more robust insights.
The approach we took was simple at first. A preliminary analysis examined clusters of infections from the early weeks of term5. This early analysis highlighted an unusually large cluster of infections amongst students, with reassuringly few community cases involved. Understanding the origin and subsequent evolution of this cluster appeared critical to understanding the transmission dynamics within the university; especially as it would go on to represent just under 70% of all university cases sequenced across the academic term. Two pieces of information were important to rule out: 1) was this large cluster the result of multiple introductions or a single introduction into the university? 2) was the cluster introduced into the university from the surrounding community?
Epidemiological data proved critical to understanding the large university cluster
Epidemiological information derived from numerous sources including public health agencies, COG-UK, and the University of Cambridge, helped determine the origin and dispersal of the large university cluster. Fortuitously, the lineage representing the large university cluster had been recently sequenced in the UK, distant to the university. The epidemic curve of this lineage in the UK was largely represented by university members after it was introduced into the university - it was not seen in the Cambridgeshire community for three weeks after the start of the student term. Confidence that the lineage had been introduced into the university from outside of Cambridge was bolstered with a time-scaled phylogenetic tree, demonstrating a significant divergence between clade representing the large university and sequences from the surrounding community. Furthermore, we were able to show the cluster was likely to represent a single transmission chain of infections with a locally developed tool, A2B-COVID.
These higher-level insights provided the opportunity to study the drivers of transmission within the large-cluster and therefore the university (the cluster was highly epidemiologically diverse early in its history – comprising the majority of courses, colleges, and year groups). Contact tracing data, provided by the university, local health protection teams, and NHS Test and Trace, was vital to identify the source of dispersal of this cluster across the university - a nightclub venue. Further, this data proved useful again when considering sources for the rise in effective reproduction number in the cluster prior to the November 2020 UK national lockdown.
A notable difference between this cluster and other transmission clusters that did terminate within a few weeks, was the fact that this cluster had been seen across multiple accommodation sites, study courses and study years, making containment a more difficult prospect.
Drivers for transmission were not always external to the university. We noted multiple factors – shared accommodation, courses, and academic years, as important considerations for the development of transmission clusters. These have become an important basis for targeted infection control measures, such as enhanced asymptomatic screening to minimise transmission within the university.
What about student and community transmission?
Through a variety of phylogenetic methods, and examination of clusters generated using the CIVET clustering tool, there was a collective sigh at the demonstration of limited cross-transmission between the university and its surrounding community. These findings were confirmed by examination of contact tracing data. This analysis also highlighted one of the most powerful uses of genomics when assisting outbreak investigations; numerous epidemiologically defined clusters were refuted – a clear benefit of SARS-CoV-2 genomic epidemiology that can be plausibly scaled to national contact tracing efforts.
We did, however, note a possible conduit for transmission between university students and the local community in the form of medical students. Here, we conducted a separate analysis taking advantage of the high case ascertainment in the university and local hospital (the local hospital had been running its own screening programme – from which positive tests were routinely sequenced). Transmission clusters were noted to have a higher representation of medical students when compared to other courses. This finding was not unexpected, given the ongoing face-to-face teaching in clinical settings, but gave weight to the need for prioritisation of medical students for vaccination, as is the case with other healthcare staff. This is now more of an issue in countries where SARS-CoV-2 vaccination programmes are still in their infancy.
Integrated genomic-epidemiology studies of outbreaks are required
Combining detailed epidemiological data with SARS-CoV-2 genomic sequences, ascertained from a well sampled cohort of individuals, led to highly useful public health insights. Close working between public health agencies, COG-UK, and the university, allowed us to reliably determine the predominant sources of transmission inside and outside the university. Genomic information was critical in shining a light on important transmission clusters – this has been echoed in other important settings and highlights the need to maintain an infrastructure for SARS-CoV-2 genomic surveillance. Screening of individuals provided the high case ascertainment required for genomic assessment, early case detection (especially asymptomatic and pre-symptomatic), and isolation. It is important to note that not all higher education facilities have the same infrastructure and differences in transmission patterns of SARS-CoV-2 may exist - this work will hopefully provide a framework for other higher education to replicate now and in future pandemics. Ultimately, the SARS-CoV-2 pandemic has had a significant toll on students and staff in higher education facilities – we need to work hard to ensure that the lessons learnt today are used tomorrow.
- Meredith, L.W., et al. Rapid implementation of SARS-CoV-2 sequencing to investigate cases of health-care associated COVID-19: a prospective genomic surveillance study. Lancet Infect Dis 20, 1263-1271 (2020).
- Hamilton, W.L., et al. Genomic epidemiology of COVID-19 in care homes in the east of England. Elife 10(2021).
- Ben Warne, J.E., Marina Metaxaki, Stewart Fuller, Richard J Samworth, Gillian Weale, Jon Holgate, Vijay Samtani, Craig Brierley, Dinesh Aggarwal, Sarah Hilborne, Mahin Bagheri Kahkeshi, Sarah Berry, Julie Douthwaite, Alexandra L. Orton, Rebecca Clarke, Darius Danaei, Rory Dyer, Rob Glew, Oliver Lambson, Aastha Dahal, Ben Margolis, Edna Murphy, Matthew Russell, Vickie Braithwaite, Kathryn Faulkner, Elizabeth Wright, Afzal Chaudhry, Lenette Mactavous, Ajith Parlikad, University of Cambridge Asymptomatic COVID-19 Screening Programme Consortium, Jane Greatorex, Ian Leslie, Andy Neely, Steve Rees, Ashley Shaw, Martin Vinnell, Linda Sheridan, Emmeline Watkins, Stephen Baker, Ian G. Goodfellow, Paul J. Lehner, Kathleen Liddell, Rob Howes, Michael P. Weekes, Duncan McFarlane, Patrick H. Maxwell, Nicholas J. Matheson. Feasibility and efficacy of mass testing for SARS-CoV-2 in a UK university using swab pooling and PCR. bioRxiv (2021).
- Matheson, N.J., Warne, B., Weekes, M.P. & Maxwell, P.H. Mass testing of university students for covid-19. BMJ375, n2388 (2021).
- Dr Dinesh Aggarwal, D.T.F., Dr Ben Warne, COVID-19 Genomics UK (COG-UK) Consortium and the University of Cambridge Asymptomatic COVID-19 Screening Programme Consortium. Genomic epidemiology of SARS-CoV-2 in the University of Cambridge identifies dynamics of transmission: an interim report, 10 December 2020. https://www.gov.uk/government/publications/genomic-epidemiology-of-sars-cov-2-in-the-university-of-cambridge-identifies-dynamics-of-transmission-an-interim-report-10-december-2020 (2021).