The oft repeated mantra is that COVID-19 is not flu. And on the night that curfews are announced for pubs, bars and restaurants across the nation, it bears repeating. The differences are legion ranging from symptoms to the age profile of vulnerable populations to the lack of vaccines/treatment and the impact on multiple organ systems. But for now, let us focus on the differences in infectivity. Unlike SARS and MERS, SARS-CoV-2 and influenza are infectious prior to symptoms. Indeed, SARS-CoV-2 may be most infectious for some infected individuals in this asymptomatic stage, and yet never go on to develop symptoms. This observation hints at an underlying feature of the virus that guides our strategy of containment.
Most infections are democratic. Infections are driven by the many not the few. But not COVID-19. Studies have highlighted that only 20% of household members become infected despite long exposure – a very different pattern to what we expect with influenza or noroviruses that spread like wildfire. Contact tracing between January and April in Hong Kong detected 4-7 superspread events and determined that 19% of cases were responsible for 80% of onward transmission. Modelling of super-spreading events in Georgia, USA concluded that 2% of cases were responsible for 20% of all infections. Assimilating the different strands of evidence, Nielsen and Sneppen characterised the crucial role of the superspreaders in COVID-19 and the important dispersion parameter or k-value. Whilst for influenza 80% of all infections might be caused by 40% of initial cases, the figure for COVID-19 is closer to 10%. Superspreader events are key to the relatively high value of R0 for SARS-CoV-2.
As yet we do not know why some individuals develop very high levels of titres of virus prior to symptoms – perhaps characteristics of the individual or the context of the initial infection. But the importance of the superspreaders to the onward propagation of the virus drives our approach to mitigation, and deciphers the intent behind the ubiquitous use of masks, the Rules of Six and early curfews. For a democratic infection, the key focus is to reduce the period of exposure. Lower exposures reduce likelihood of transmission. For a heterogeneous infection like COVID-19, the key focus is reducing diversity of contact. Reducing numbers of contacts reduces the number of individuals who might be infected by a superspreader and the likelihood of a new superspreader. Early closing reduces diversity of contact. Indeed, one publican remarked on the airwaves that a 10pm curfew would reduce his takings by half, giving a clear indication of the level of footfall in that last hour.
Until we are able to identify who and why the superspreaders are, our tools remain blunt. That said, I have to highlight one of the most impressive databases that I have come across (and there have been many over the last 6 months), and express my thanks to the volunteers and their painstaking work. A group at the London School of Hygiene and Tropical Medicine under Dr. Gwen Knight, Dr. Quentin Leclerc and Dr. Fatim Lakha has established a global database that includes 1,500 cluster events up until 11 August linking 190,000 cases. This database provides invaluable information on where and how these cluster events came about, but even this excellent work only represents 1% of the globally confirmed cases, and hence can only be indicative. That said, it is interesting to put the contribution from the hospitality industry into context, so let me end by sharing some analysis on the top 20 categories of cluster events from the database.
