23rd October - the COVID-19 coronavirus compendium

Fatality rates, tocilizumab, and long lasting antibodies

Like Comment
Read the paper

We now know a huge amount about SARS-CoV-2 and the disease it causes, COVID-19. But as with any scientific or medical topic, there is always more to learn. I have been reflecting on whether this compendium needs continuing, or whether the basic science and clinical findings are well enough understood that I can stop. For the moment, I will continue, but reduce my blogging to every two weeks. In part, this is because it is useful for me to stay up to date!

This week we learnt that the infection-fatality rate in New York City was between 0.01% and 19%, depending on age, that tocilizumab doesn't reduce mortality from COVID-19, and that neutralising antibodies can persist for at least 5-7 months, if the serology tests include multiple antigens.

Fatal cases

16 Bayesian models were used to estimate that there were 206,000 extra deaths from COVID-19 across 21 industrialised countries, mainly in Europe. There were around 100 excess deaths per 100,000 people in England and Wales, and in Spain, equivalent to a 37-38% increase in these countries. Other countries had lower mortality rates.

There were 225,000 excess deaths in the US this year, 150,000 of which could be directly attributed to COVID-19. Another study found that the US had 60 deaths per 100,000, more than Australia (3 deaths per 100,000) and Canada (25 deaths per 100,000), and comparable to Italy (59 deaths per 100,000), but less than Belgium (87 deaths per 100,000). If rates were comparable to Australia, there would have been 187,000 fewer deaths in the US.

The infection-fatality rate of COVID-19 in New York City was estimated as 1.39%, although it greatly varied by age, with a rate of 0.1% for those aged 25-44 years, and rising to 19% for those aged over 75 years.

A model of the UK population predicted that at least 200,000 deaths would eventually occur from the virus, regardless of non-pharmaceutical interventions, and in the absence of a vaccine. Lockdown simply delays deaths, and can cause more deaths, unless a vaccination programme is rolled out.


Protease inhibitors were designed which bind to NPro, as well as activity-based probes that can detect protease activity in infected cells of COVID-19 patients.


REGN-COV2, a cocktail of two neutralising antibodies, reduces COVID-19 disease in rhesus macaques and golden hamsters. President Donald Trump was treated with this antibody therapy, but clinical trials in humans are still ongoing.

Two randomised controlled clinical trials found that tocilizumab, an anti–interleukin-6 receptor antibody, did not reduce mortality in COVID-19 patients.

The CR3022 antibody, which binds to the spike protein, was visualised using cryo-EM. A phage library was used to construct neutralising VH antibody fragments that disrupt the interaction between the viral spike and ACE-2.


Neutralising antibodies against the virus last for at least 5-7 months, when testing against multiple antigens, showing the importance of multiple independent serology assays.


A pooling strategy, where positive and negative samples are grouped together and tested at the same time, was developed and trialled in low resource settings in Rwanda and South Africa. Pooled testing is much cheaper and could be used to monitor infection in low and middle income countries, as well as high income countries that also need more testing.


40% of COVID-19 cases in skilled nursing facilities in the US were asymptomatic, showing again that symptoms are not a good predictor of infection.

The individual impact of non-pharmaceutical interventions, such as banning large events and closing workplaces, were assessed, with most shown to be effective at reducing viral transmission. A test and trace programme on the Isle of Wight, which included the use of an App, was found to reduce R, although it does not appear to be as effective in the rest of the UK.

Loss of taste and smell, as recorded in online surveys, was a better predictor of local COVID-19 epidemics than official government data. The outbreak in New Zealand, which was successfully controlled with border closures and testing was described.

A case of re-infection with two different strains of SARS-CoV-2 was described in a man from Nevada.


The inactivated SARS-CoV-2 vaccine, BBIBP-CorV, was tested in humans in a phase 1 / 2 clinical trial and shown to be safe and immunogenic. Two RNA-based vaccines were also tested in a phase 1 clinical trial and found to be safe and immunogenic, and the mRNA-1273 vaccine was tested in non-human primates. Inactivated vaccines have been used for more than 100 years, unlike mRNA vaccines, which have yet to be widely used in humans.

A global survey found that 72% of participants were very or somewhat likely to take a COVID-19 vaccine, with those reporting high levels of trust in government more likely to accept a vaccine.

Risk factors

Black patients in the US were more likely to be hospitalised with COVID-19 than white patients, but were no more likely to have a fatal infection. Racial disparities could not be explained by co-morbidities or socioeconomic status.


SARS-CoV-2 expressing nanoluciferase was engineered, which can be incorporated into rapid anti-viral screening and neutralising assays.

Protein-protein interactions for three coronaviruses found that Tom70 interacts with Orf9b from SARS-CoV and SARS-CoV-2 and so could be a potential drug target.


Two studies showed that neutrophilin is a host factor for SARS-CoV-2 that binds to the spike receptor of the virus.

A structural analysis of human antibodies in complex with the coronavirus spike protein identified four categories of binding, providing a useful framework for future studies. Another study carried out a structural analysis of the full length spike protein.

Two studies showed that a truncated form of ACE-2 is induced by interferons and viruses, not the full-length version, as previously reported. The truncated version does not facilitate viral entry.


There was an 8.8% decrease in global carbon dioxide emissions in the first half of 2020, a larger decrease than during previous recessions or during World War II.


The incidence of pre-term births in the Netherlands went down during the early COVID-19 outbreak in March, although the effects did not persist after that. The mechanism is unclear, but may relate to increased handwashing and social distancing.

Ben Johnson

Magazine Editor, Nature Medicine, Springer Nature

I trained as a virologist, starting with an undergraduate degree in virology from the University of Warwick, UK. My PhD, in influenza virus genetics and immunoevasion, was from Public Health England and the University of Reading, UK, with Maria Zambon and Wendy Barclay. My research interests then moved to smallpox vaccines, viral ion channels and cell adhesion, while a postdoc at Imperial College London with Geoffrey Smith, FRS. I then joined open-access publisher BioMed Central in 2011 as an editor and then associate publisher and was Head of Communities & Engagement at Springer Nature from 2016, running the Nature Research Communities and other online engagement activities for researchers. I joined Nature Medicine in 2021, with responsibility for news and opinion content, and am based in the London office.