We may already have herd immunity – an interview with Professor Sunetra Gupta

Are we already immune to coronavirus? Professor Sunetra Gupta, a theoretical epidemiologist at Oxford University, discusses her recent study on the herd immunity threshold, as well as her views on the social costs of lockdown, the inaccuracy of epidemiological models, and the curtailment of academic debate.

By The Reaction Team. Excerpt:

"A study produced by a team at Oxford University indicated that some parts of the United Kingdom may already have reached herd immunity from coronavirus. A significant fraction of the population, according to the study published last week, may have “innate resistance or cross-protection from exposure to seasonal coronaviruses”, making the proportion vulnerable to coronavirus infection much smaller than previously thought.

The Oxford team is led by Sunetra Gupta, a professor of theoretical epidemiology. In recent months, she has argued that the cost of lockdown will be too high for the poorest in society and questioned the language and quality of debate on the pandemic’s impact. 

Reaction interviewed Professor Gupta about these matters and more, with questions from Maggie Pagano, Alastair Benn and Mutaz Ahmed.

Alastair: In the debate over T-cell immunity or cross-reactivity with coronaviruses, the common-sense view is that exposure to things that are similar does give you some protection, and it seems to be borne out in recent studies.

Yes, exactly. The principle of protection from exposure to related viruses, and indeed any kind of pathogen, is one that we’ve known for a very long time. The very first vaccine we had, which is smallpox, was based on the idea that cowpox protects against smallpox. This idea was already there well in advance of us knowing that smallpox was a virus – and indeed in advance of germ theory having been properly established. So we knew about this cross protection even before we knew that diseases were caused by germs. It’s a very old idea.

In my own studies, beginning with malaria and then later thinking about flu, the role of cross-immunity in protecting against disease seemed to be something that very much needed to be factored into our thinking. Most of the people who die from malaria are children, and they die upon their first exposure, because they have no immunity at that stage. That was one of the first things that struck me when I was working on malaria.

And then later when I was working on flu, it seemed to me a very good way of explaining why the 1918 flu had killed so many people, but why that didn’t seem to be repeating itself, was that it was likely that people hadn’t been exposed to flu. Many people would have not had the flu at all. So then that built up this population of naive immunity in people under the age of thirty who were very badly affected when the pandemic came through.

Having those ideas in mind, when the Covid-19 virus started to spread, I was pretty certain it wouldn’t have a huge, devastating impact in terms of mortality, because we had all these other coronaviruses circulating.

What I didn’t anticipate was that some of our responses to previous exposure to seasonal coronaviruses might actually protect us from infection. It’s one thing to get infected and not ill, but what the new studies are showing is that people are actually fighting off infection. So at an even more basic level, the pre-existing antibodies or T-cell responses against coronaviruses seem to protect against infection, not just the outcome of infection.

Mutaz: When the serological studies were conducted a couple of months ago, antibody rates were very low in the UK. Is that because people weren’t vulnerable to infection, or because once they were infected they had some level of innate immunity?

What we know is that the seropositivity rates in many parts of the world are much lower than we’d expect them to be if we assume that the epidemic has passed through and that people are resistant. If you take a very simple scenario where everyone is susceptible, you’d expect 60-70% of them to have some marker of exposure. And that is not what’s been observed.

I mean, there are areas where it is very high. There’s a paper published recently of a seroprevalence study done in the slums of Buenos Aires, which reports a 50% seroprevalence. And there have been studies from Lodi, Italy, where it is I think 60%.

So there are studies on the high side, but I think one can’t really trust studies on either side completely. One has to take all these measures with a certain degree of caution. What we have here are a range of measures. In cities, it’s typically much higher than in rural areas or areas that are non-urban.

One of the things that’s been done in reporting the seroprevalence, which is not correct, is that they’ve been homogenised. When people say only 5-6% of the UK population has been exposed, that’s not correct. I think very few people would agree that exposure rates in London are less than 20%.

The picture that we’re getting is heterogeneous. But even in hotspots, apart from a few reports, they’re still quite low. So why is that?

One reason might be that lockdown stopped the spread of infection, so it was halted at a stage when, say, 20% of people were immune and the rest of the people were still susceptible to infection. Well, under those circumstances, the easing of lockdown should result in fairly rapid growth of cases. And that’s not something we’re seeing.

So we’ve got those two bits of information. The third bit, the missing piece of the puzzle, is this idea that some people are fully resistant to infection, because they just have really good defences. That could just be part of our innate immunological makeup. It’s also becoming clear that some of the people that have beaten it off have had responses to other coronaviruses which could have played a role.

The other bit of the puzzle is that some people do get infected and they make antibody responses, but those responses die very quickly. So if you’re trying to measure exposure, you won’t get the full picture. Some of the measures of seroprevalence might be underestimates.

We’ve got four pieces of the puzzle, then. If we put them all together, which is what the paper that we published on Friday does, it gives you a theoretical framework that you can use to look at how these bits connect up together.

You can see two things. You can see why the seroprevalence level might be low, and you can also infer that the level of herd immunity needed to stop the thing from exploding again is actually much lower than the figures that are currently being thrown around quite incautiously might suggest."