omicron variant

How Mild Is Omicron, Really?

Photo: Wang Ying/Xinhua via Getty Images

How mild would Omicron have to be to outweigh its insane rate of spread?

Since the first sequencing of the new variant, and especially since the first suggestions that cases flooding the Gauteng province in South Africa were overwhelmingly mild, epidemiologists have taken care to warn that, even if a new strain proves less severe, the ultimate toll could still be significantly greater than earlier, more pathogenic variants, given Omicron’s high infection rate.

This makes the question of just how virulent Omicron is enormously important. Unfortunately, the early data has been a bit all over the place. Last week, the first major study of South Africa data suggested, overall, a 29 percent reduction in illness severity — a notable drop, but not nearly enough to make up for a rate of spread three times faster than Delta, the fastest-spreading of the earlier variants.

Today, a new analysis suggested both that the hospitalization rate in South Africa may have been reduced by as much as 91 percent, compared to the country’s last wave, and that the death rate among those hospitalized may also have fallen by two-thirds — two points that, taken together, suggest as much as a 25-fold or even 30-fold reduction in mortality. On the same day, a report produced by the Imperial College of London suggested there was no clear indication that, among early British cases, Omicron was any less severe at all — though that analysis was based on just a few dozen cases, and, as is always the case with such preliminary data, the absence of a clear finding is not the same as a definitive finding that the new cases are not more mild.

On Friday, I spoke with the infectious-disease epidemiologist and computational geneticist Francois Balloux, of the University College of London, who calls himself a “militant corona-centrist.” His view, in short: The South African data is no longer all that tentative, and it is very, very encouraging, but that it is still too early to safely project the experience with Omicron in that country onto others, like the U.S.

Let’s talk about what we know right now. There’s been relatively encouraging reports out of South Africa, but then there was this Imperial College report today that was striking a quite different tone, giving a quite different picture. How do you see the basic lay of the land right now? And what do you think is ahead for us in the next month or so?
Well, I think we’re starting to have good data from South Africa. And it looks like the proportion of people hospitalized, given infection, is much lower than in previous waves. Some estimates are as optimistic as 90 percent lower. It looks like, given hospitalization, people are less likely to die, or are less likely to die by respective age groups — a two-thirds reduction of death from infection compared to previous waves. There could be all sorts of reasons why, but that is quite spectacular. If you assume that these numbers are correct, it would mean a 30-times reduction in fatality relative to previous waves, which is really something. That brings us down to seasonal flu.

Do you think those numbers are reliable? You did say, “If you assume these numbers are correct.”
It’s difficult to say. We have no idea about asymptomatic cases, since there is just no kind of surveillance testing in South Africa. Hospitalization is also interesting — hospitalization obviously reflects the severity of illness, the state of the patient, but there are also other things that play a role: psychological, social, political. But if anything, I cannot see why any of those would go in a direction that would make the numbers look better than they really are. There’s always considerable uncertainty, but I find it difficult to imagine that mortality and hospitalization or morbidity is not significantly lower than it was before.

That’s the data for South Africa.

And other countries may have a different experience. 
South Africa has a young, though not super-healthy population —

But that’s been true of the earlier waves as well. 
Yeah. But if we extrapolate to other parts of the world, what does it mean in the U.S. and the U.K.?

It seems to me like it’s one thing if we’re talking about a 30-fold reduction in mortality, and another if we’re talking about even a five-fold reduction in mortality — which on its own would be quite significant. Because a five-fold reduction could be mostly or entirely made up for by transmissibility — a much bigger total number of cases would ultimately produce the same number of severe cases, or more. A 30-fold reduction is really going to just totally change the landscape.
Yes. Even if this wave were globally prevalent everywhere, we still wouldn’t expect 30 times as many cases.

So if that 30-fold reduction actually held, in terms of severe cases and deaths —
We would still expect something much more manageable.

And how safe is that to assume? How safe is that to project those numbers?
Well, the problem is, it’s really difficult to project. I think the South African experience, for instance, would project fairly well in other lower-middle-income countries that have relatively similar demographies, similar rates of vaccination. And they do have a very high rate of protection there — essentially everyone there has been exposed. But that is also the case around other places in the world. I think it can probably be safe to extrapolate South African data to other parts of Africa, some parts of Asia, and some parts of South America.

But I’m not sure, at this stage, what it might tell us about the U.K. and U.S. prospects.

What about the Imperial College report, which suggested no lower severity with Omicron?
Well, it’s based on very little data. That’s because there isn’t that much data at this stage.

Twenty-four cases.
And if you just took the raw data, and you just take number of cases and hospitalizations, you get a four-fold reduction.

But I think here, no evidence really should be understood to mean, not enough evidence yet to say anything in either direction.

And looking at South Africa, do you think that the data that we’re seeing so far is reflecting reduced severity or acquired immunity, either through vaccination or infection? 
It’s a very difficult topic. And a very difficult topic to discuss. And I understand why it’s difficult, because saying anything along the lines of infection-acquired immunization is as robust or more robust than vaccine-acquired immunity — obviously, that might send a really, really wrong message, because irrespective of how robust or durable infection-acquired immunity is, ideally no one should be exposed to the virus without having been vaccinated first. It’s an extraordinarily difficult message. I understand why people are cagey on that.

My understanding is that to this point the data has also been mixed on that. There have been some studies showing infection-acquired immunity was stronger and other studies showing less protection. 
I’m not sure the evidence is that mixed. But I think the evidence that has been put forward, or even published or championed — I would say there’s some bias there.

But it’s complicated. In terms of neutralizing antibodies, I think vaccines, at least in the short term, really are superb. They’re also much more homogenous in the response they produce, which is obviously something important. They’re also much more effective in populations that are most at risk. You can expect that the vaccine is more likely to elicit a decent response in someone who’s elderly, partly immunocompromised, rather than infection.

With infection, it can be anything from something really mild to something very severe, and, obviously, this will affect the robustness of the immmuno-response. For instance, 20 percent of infected people seem to generate hardly any response for all sorts of reasons which are not well understood.

But I also think quite a few people will have to revise their priors about how effective infection-acquired immunity is. There’s been this messaging in some health services that infection-acquired immunity just doesn’t work. I think this data coming out of South Africa does not support that.

So when we look at that data, how much of it do you attribute to that immunity and how much do you attribute to the virulence of the new variant?
I don’t think we can say at this stage. If it were really something like 30 times down, in terms of death, then I think it would be difficult to imagine, though this could be entirely due to additional immunization acquired by the population just within the last few months. That would probably be a bit of a stretch.

Even if the virus circulated very widely in the last wave.
And also vaccination rates are not brilliant, but they’re probably reaching soon 30 percent. The most at-risk population, older people, tend to do better with vaccination. So it’s not 30 percent throughout the population as a whole.

And you can see those numbers therefore as not entirely additive, but complementary, right? So if 70 percent of the country has had exposure, 30 percent of the country’s been vaccinated, you’re probably not all the way to a 100 percent, but you might be to 85 or 90 percent that have some protection. 
Yes, I think so. And obviously, we don’t really know what portion of the population has been exposed to the virus. But I think it’s very, very high, in South Africa. But what does very high mean? Does it mean 80 percent or 99 percent or more? And that will make a huge difference.

Well, it’s something that I’m worried about, being here in America. Because when we saw the Delta wave hitting the U.K., we saw a very different phenomenon than ultimately we saw when it came to the U.S.. You guys had your hospitalizations and deaths dramatically reduced by how well-protected your senior population was, and how concentrated that wave was in the younger population. And in the U.S., the ratio of cases to hospitalizations and deaths wasn’t actually any lower for us in Delta than it was before vaccination. And in thinking about Omicron, I’m worrying that some version of that same phenomenon may play out, where we see really encouraging data in South Africa, pretty encouraging data in the U.K., and then in the U.S., it’s a much bleaker picture.

And a lot there hangs on how much of what we’re seeing has to do with the evolution of the virus into this new variant, and how much of it has to do with acquired immunity — and what difference it makes between having the level of vaccine protection that the U.K. has versus the level the U.S. has. How would you advise that Americans think about those dynamics?
Well, I would definitely recommend everyone get vaccinated.

That anyone gets boosted if they haven’t. But, yes, I tend to agree. You cannot necessarily just take the evidence of South Africa and just extrapolate it across the U.K. and U.S. And, in the U.S., actually, I’m a bit puzzled why —

Me too. I’ve spent a lot of time over the last few months asking people why the case-fatality rate was roughly as high during Delta — when a majority of the country and a vast majority of our seniors were vaccinated — as it was during the winter surge, when hardly anyone was vaccinated. 
I don’t fully understand it. There are obvious big gaps in vaccination coverage in some demographies. It seems sometimes, at least in some states, the people would benefit most from vaccination are the least likely to be vaccinated. But still … I agree. I’ve started going over these figures and it’s not entirely clear to me, really, why actually it has been so bad in the U.S., compared to other places.

Yeah. I can get my head around the fact that there are enough unvaccinated people to produce those raw numbers of deaths — about 1,200 a day right now, and over 1,000 a day now from three months running. But I just would’ve thought that if we’re having a thousand people dying a day, that would’ve just implied — or required — a much larger caseload, since so many more people are protected, so many vulnerable elderly in particular. But in fact, we’re seeing a relatively modest caseload. And that’s where I get a bit confused. I guess it does mean that Delta turns out to have been somewhat significantly more severe. 

But, even so.
Well, I think the ratio of cases to infections or infectious cases varies a lot between countries. And, obviously, countries that have absolute free testing and also now increasingly rely a lot on lateral flow tests. They pick up a large number of the cases. So, I think in the U.K. the assumption is that about half the infections are detected. Only one in four is detected in the U.S.

But even so, I would expect — maybe this is a naïve assumption — that those dynamics are not all that different in the United States of September and October 2021 than they were in January and February 2021. Which means that, while country-to-country comparisons may not be especially useful, comparisons within the U.S. might be more illuminating. 
I think some countries have really ramped up their testing capacity and the number of tests performed by some countries like Denmark and the U.K. is insane. Whereas, I would not say it is insane in the U.S.

Does that imply, then, that one possibility is that what changed in the U.K. between the winter surge and the Delta wave in the U.K. was as much a testing revolution as a vaccination revolution? That while the population was much more protected because of vaccines that it was also just being tested much more aggressively, so that it was both the case that the numerator of severe outcomes was going down, but also that the denominator of the total infections registered as cases was going up? And that we just did not have nearly that same revolution in the U.S.?
Yes. I think it might have as much to do with the denominator than the numerator.

I suspect that the U.K. detects at least twice as many infections as the U.S. does. And it might be even higher, actually.

So how should we think about what to expect here, with Omicron?
Well, it goes back to this fundamental problem. If Omicron is milder, then it will be milder everywhere, irrespective of population or irrespective of the vaccine coverage or prior infection. If it’s not particularly milder itself or only marginally so, and what we are observing here with the South African population is a reflection of the fact that essentially the entire population has been exposed or vaccinated or both, then obviously that’s not portable to any other place. It’s a really major question. And sadly, I don’t think we can extrapolate anything from South Africa, really. We don’t actually even know the vaccine status of the hospitalized.

But the early U.K. data is at least relatively encouraging in that regard. It suggests that the wave in the U.K. is following the broad-strokes outlines of the wave in South Africa — in terms of severity, it’s not like India’s Delta wave or anything like that. 
Well, at the moment, we’re seeing an insane increase of cases. And Denmark, which is the other country which also has really remarkable testing capacity in place — they are seeing fairly similar dynamics. It’s probably happening all over Europe. This idea that it’s happening now in London, and it will happen in the future, is correct. I think it’s really happening many places where we are just starting to have testing capacity in place.

But in terms of severity? 
I think we just have to wait. With the U.K. situation, it’s like, had we talked two weeks ago about South Africa, I would be saying, who can say? I think in the U.K. in two weeks’ time we’d be pretty comfortable saying, these are the numbers and these are the patterns.

Okay. I also wanted to ask about this hypothesis you put forward a few weeks ago, that if we were seeing this feature of viral evolution that we do seem to be seeing with Omicron — that it may reproduce better in the respiratory tract and less efficiently in the lung, that that might be a sign of viral evolution in response to vaccination. You suggested that before we knew that Omicron did exhibit those patterns, and before we had observed any possible sign of reduced severity. Do you still feel that that’s a possible explanation?
At this point it’s a hypothesis, because the study out of Hong Kong was just using pseudo-organs on a chip, but it does look like Omicron might really tend to infect mucosa rather than really infect deep in the lungs or in other organs. Which could make it, if it were the case, less virulent. It could also maybe explain the slightly shorter intergenerational times.

And how might that come about as a result of vaccination? 
Vaccines tend to provide very poor humoral immunity. The cell layers of the airways, they’re not really immunized well.

That’s what we’ve been seeing a lot with breakthroughs — essentially, people didn’t have real infections that affected all organs in their body. They tended to have some very superficial infections in their mucosa.

And you think it’s a sort of plausible explanation for this path of evolution that it emerged in response to vaccination?
It could have contributed. But there are a lot of selective pressures that could have contributed. The standard explanation that the virus doesn’t want to kill its host is a bit too simplistic, but obviously, if you had a strain that can start transmitting well before a host feels the symptom, then it might actually transmit very well, because we are testing so much and obviously people tend to isolate on the first sign of symptoms.

I think that probably creates a strong selective pressure. But vaccines could have contributed as well, because vaccines protect us pretty well, except in our upper respiratory tracts. I think it would be very interesting if that turned out to be the case, though I don’t think we will ever be able to prove it.

More on omicron

See All
How Mild Is Omicron, Really?