We’re still sorting out the answer in terms of severity, where the early returns suggest that while Omicron likely produces less severe illness overall — perhaps 70 percent less — the reduction inherent to the strain may be smaller than has been observed so far in South Africa. Among other things, this suggests that in the U.K., Europe, and the United States, the ultimate toll in terms of hospitalization and death may be larger than in South Africa — as we may already be seeing hints of in the London hospitalization data. But there are also reasons to believe the models may be dramatically overestimating the ultimate size of the wave, and perhaps reason to believe that, thanks to its immune-escape properties, the Omicron wave may be qualitatively different than those produced by earlier strains — that hypothesis is the subject of the interview below.
In the meantime, the equation of alarm is quite straightforward: Even a strain that’s half as virulent can cause much more damage if it is also spreading several times faster. “Hope is not an effective strategy for dealing with a pandemic,” Harvard epidemiologist William Hanage wrote earlier this week. “Let me be clear,” he continued. “I’m not stating definitively that Omicron has some grim future in store for us. I’m saying that there are red flashing warning signs, that we underestimate this virus at our peril and that even the best-case scenario is still bad.”
But for those used to interpreting pandemic news in terms of “good news” and “bad,” the new variant also raises some disorienting questions about future immunity and the course of the pandemic to come. That is in part because, to this point, the Omicron wave has been defined primarily by breakthrough infections and reinfection, rather than new illnesses among the unvaccinated and unprotected. Three-quarters of early U.S. Omicron cases were among the vaccinated or already infected; in Denmark, with some of the world’s best variant surveillance, the share was the same; in other parts of the world, the figure has been 80 percent or higher. If the new variant spread rapidly through the susceptible population, it would mean huge spikes in hospitalizations and deaths, even at much lower levels of severity. But, in theory, at least, it is also those unprotected who would benefit most from exposure to a “mild” strain, acquiring immunity from an encounter with a relatively less-punishing form of the virus. At the moment, however, it does not appear that Omicron is operating that way. Instead, its astonishingly rapid growth appears to be powered by spread among the already protected.
There is so much we still don’t know about Omicron, but since the variant was first identified last month, epidemiologists cautioned to be careful about attributing rapid case growth to either immune evasion or increased transmissibility, in part because, especially before detailed data emerged, the two effects could produce basically the same picture at the population level. But while an early consensus formed that the variant was inarguably more transmissible than Delta, a few newer data points suggest, at least, a more complicated picture. In Denmark, it appears not only that previous infection or vaccination offers less protection against infection with Omicron, but that the new variant is in fact spreading faster among those considered “safe” than those we reflexively consider “vulnerable.” And according to a report published Tuesday by the U.K.’s Office for National Statistics, “Those who have received three doses of a vaccine and test positive for COVID-19 are more likely to be infected with infections compatible with the Omicron variant compared with those who are unvaccinated.”
This doesn’t mean that Omicron prefers the vaccinated, or that the vaccinated are more vulnerable to infection than the unvaccinated, but that the immune evasiveness of the new variant gives it a much bigger relative competitive advantage over Delta among those with prior immune protection than among the unprotected, among whom it may not be more transmissible than Delta at all. What we’ve observed as explosive initial growth may primarily or entirely reflect Omicron’s prolific ability to create breakthrough cases and reinfections, which are overwhelmingly mild. (Even with Omicron, the vaccinated and especially the boosted still appear quite well-protected against severe disease.)
These are just a couple of data points, compiled in the very early — and not necessarily representative — stages of the new wave. The course may well shift as Omicron spreads further, changing the picture and composition of the wave overall. But the early data does suggest the possibility that, as virologist Trevor Bedford warned several weeks ago, Delta and Omicron may be sufficiently distinct that the growth of Omicron doesn’t mean the end of Delta, but that the two ongoing waves are running basically in parallel.
And if the two variants are competing for hosts in two separate populations, it may only be among the vaccinated or previously infected that the new variant has a significant advantage — and that among the vulnerable unvaccinated, Delta may be able to very well hold its own or even outlast Omicron, continuing to infect people and cause severe illness and death after the Omicron wave has crashed. This hypothesis may help explain both why Omicron is producing a significant reduction in severe outcomes, at least for now, and why the ongoing Delta wave that preceded it does not yet appear to be receding all that quickly. It also suggests a relatively pessimistic outcome, even if Omicron proves to be relatively mild: that the new variant might burn through the country, infecting huge numbers of people, without giving them much additional immune protection going forward — against future variants or against the Delta strain that is still killing 1,300 Americans a day on average.
Last week, prompted by Omicron, President Joe Biden offered a stern warning to the unvaccinated: “For unvaccinated,” he said, “we are looking at a winter of severe illness and death,” adding that “the only real protection is to get your shot.” But while the Omicron wave is in its early stages, the dynamics of spread could well change, and the variant inevitably represents a real risk of severe illness to those without any immune protection, it also seems possible that the wave will spread, at first at least, preferentially in those with that immune protection.
On Tuesday, I spoke with epidemiologist Deepti Gurdasani of the Queen Mary University of London about how incomplete our understanding still is of the Omicron wave currently crashing over us.
I wanted to start with this strange-seeming feature of the European data we have so far, which is that Omicron seems to be doing better among the vaccinated than the unvaccinated. How can that be?
Well, first of all, I would say vaccination still protects. That analysis doesn’t mean that vaccinated people are more likely to be infected. What it means is that if a vaccinated person gets infected, it’s more likely to be Omicron. Among the vaccinated and previously infected, Omicron has a higher advantage compared to Delta.
And the majority of people in most of western Europe are now vaccinated, so it’s not surprising at all that you’re going to see the majority of people who are infected with Omicron are vaccinated, or that it’s particularly spreading among people vaccinated.
Because the actual protection offered against infection with Omicron by previous infection or vaccination is actually quite small. There is good protection against severe disease, we think, but against infection —
It’s actually near zero. With two doses of AstraZeneca, it is zero. For two doses of Pfizer, it is only 30 percent. So it’s not surprising that a lot of vaccinated people are getting infected. And the majority of the population of Denmark is vaccinated. So it’s not surprising that the majority of people getting infected are vaccinated.
But it’s not just that the majority of people getting infected are vaccinated, it’s that — at least in the early going — a larger share of Omicron cases are vaccinated than the share of vaccinated people in the population as a whole. In other words, it’s not just that Omicron has eliminated protection from vaccination. It seems, at the population level at least, that vaccination confers almost a negative advantage — it makes you slightly more likely to get Omicron compared to an unvaccinated person. How is that?
Let me start from the beginning. Overall, Omicron has an advantage over Delta. The reasons for that advantage could be multiple. We don’t currently know the exact parameters, but we know, for example, that it’s better at escaping immunity, whether it’s previous immunity from infection or immunity from vaccination, even immunity from boosters. So in a population that’s vaccinated, Omicron will have an advantage over Delta.
But if the advantage over Delta is all immune escape, it means it will perform relatively better in the vaccinated, which is a population it can absolutely dominate, as compared to the unvaccinated population, where it may have a much smaller advantage. Is that right?
That’s the second big question that I think a lot of us are trying to answer is. We know about immune escape. But is Omicron also more transmissible than Delta? By that, I mean, if the population was fully susceptible, would it still have an advantage over Delta? And that’s something we don’t know yet.
But once you know the relative growth, then the other parameters are fixed. If there’s high escape, then it can’t also be a lot more transmissible than Delta based on the parameters we know — the spread we’ve observed. If there’s low escape, then it’s possible it can be slightly more transmissible, but not a lot more transmissible than Delta.
A week or so there was a lot of confidence that it was more transmissible. It sounds like you’re saying that it’s much murkier there than we thought.
We’re just not seeing that rapid replacement of Delta by Omicron.
Right; in a lot of places, at least, it’s like two separate waves.
If you remember with Alpha and with Delta, each of them rapidly replaced previous variants. Here, now, Omicron is growing very rapidly, but it seems to be growing on top of Delta, rather than replacing it, which suggests that they might be occupying different spaces. In some individuals, perhaps Omicron has less of an advantage. Delta may even have an advantage. And if Delta has a small advantage in unvaccinated and unexposed people, what you’ll see is that Delta will continue to circulate in people who are susceptible to both. And if Omicron’s major advantage comes from immune evasion, which is what it looks like, then it’ll have a huge advantage over Delta but mainly in vaccinated groups and those with previous infection.
And does that alone explain how Omicron is doing better among the vaccinated than among the population as a whole? Or are there other factors? Chance? Population structure and social networks, where people behave like the people they interact with?
It may also reflect the fact that Omicron is proportionately less represented among children, who are unvaccinated. I think that is one of the possibilities. Of course, you can’t rule out the fact that the spread of Omicron sort of started in Europe among travelers because it was imported, which is essentially a particular demographic. That’s the sort of thing you can only know when sufficient time has passed, whether it’s a sort of founder effect, about who was impacted first and who their contacts were. But initial dynamics are always going to be influenced heavily by where the import happened. Once that’s over, then you can start commenting on whether it is spreading uniformly across groups. To me, it does look like it’s not spreading uniformly, but I can’t rule out that part of that is still an early spread effect from 20- to 30-year-olds, who are still dominating the spread of Omicron right now.
Then you have to look at external factors: Are schools closed in Denmark, for instance, or have they closed recently or were they open when the Omicron spread started? So I can’t give a definitive answer on all that. To me, what’s more compelling is the lack of replacement of Delta. I mean, we went into Omicron in England with a massive Delta wave. And if it had a huge advantage over Delta in susceptible people, we would’ve expected it to outcompete Delta very, very quickly given its rate of growth.
Which we aren’t, exactly.
And that suggests that while it does have a huge advantage over Delta in some groups, it may not have a massive advantage in other groups.
This is not new. It just sounds new. Delta was more able to escape immune protection than Alpha, so it was more able to spread around vaccinated people than Alpha. The advantage just wasn’t as big.
Here, the advantage seems enormous.
It almost feels like they’re operating in two different immunological compartments, and that the cross-immunity between Omicron and Delta is very low because of that high level of divergence. Whereas with Delta and Alpha, they were kind of in the same immunological compartment, but with some escape.
Let’s focus on the lack of replacement: that we’re basically seeing two waves unfolding at once. First of all, is that true everywhere you’re looking? Because I’ve seen some data that suggests in some places, there has been a more total displacement.
I think that’s because they started with very, very small Delta waves. South Africa had hardly any cases of Delta when they got Omicron there.
Could the “distinct niche” hypothesis help explain why we seem to be seeing such fast peaks with these Omicron waves?
I’ve been wondering about this as well. If the advantage is more limited to populations that are previously immune or vaccinated, that would sort of limit the population within which it can spread. But having said that, looking at the South Africa wave, I still feel like the attack rate, the proportion of the population that got sick, is still very low. Just the people who had prior immunity — that was expected to be 70 to 80 percent based on seroprevalence studies.
But the other thing that could explain it is a shorter generation time, which many of us are speculating about now.
That would mean that what appeared at first to be a genuine transmission advantage could just reflect how fast the reproduction time was. In other words, we thought each Omicron case was infecting more new people than each Delta case had. But it may be that it was just cycling through that transmission cycle much more quickly, so in a period of a week, say, we had twice as many links in the chain of transmission — not that each link was linking out to more chains.
It would mean that you could kind of go up quickly, but also come down quickly. But these are things that are just being speculated right now.
And it could mean that some of those astronomical figures for R0 or Rt — estimates for how many people each case would infect — were too high?
Yeah, it could certainly mean that it could certainly have implications for R0 being lower.
That makes some sense to me — it’s been puzzling how we might have gone from an Rt of six to an Rt of about one in the space of a week or two without really dramatic policy or social-behavior change. There has been some change, but that would be a massive, massive shift.
These are the questions many modelers have been asking, since it doesn’t match with what we know about transmission dynamics. But I’m not entirely sure what’s happening. I think it’s either a change in behavior or a shorter generation time. I think both of those things or a combination of them could explain it.
And getting back to the simultaneous-wave hypothesis … If Delta is basically stable and Omicron is growing in a distinct immunological niche, what does that mean going forward? Will those people now getting Omicron have an immune advantage with Delta, or not? Is it possible this whole wave is going to wash over the population and confer no additional immunity at all on those getting these Omicron infections? For a while, it seemed likely that a breakthrough infection offered a real advantage going forward, but that requires robust cross-protection from one strain to the next.
This takes us into the realm of conjecture, but I’ll tell you what we know so far. We know that previous infection from previous variants offers only about 20 percent protection from Omicron infection. That is down from 80 percent protection before, with the earlier variants, to 20 percent, with this one. The cross-protection is not zero, but it appears massively reduced. And given that the antigen is so different and people who were infected with previous variants have a much lower level of protection against Omicron, I would think that the opposite would be true as well, although I haven’t seen any reports about it. I think that’s an open question, but I think it is entirely plausible that Omicron immunity will not protect to the same extent against Delta as infections from earlier variants have.
Which could mean that the whole wave may not mean all that much for overall immune protection in the population.
If that is the case, then the real worry is that you might get infected with one variant, not develop immunity to the other, and then still be susceptible to another variant that is circulating at high levels in the population. It’s kind of like two different pandemics in a sense. They have a little bit of cross-protection, but not a lot. And this idea of herd immunity, which people have talked about a lot, it becomes very hard to attain — herd immunity against which variants, you know?
To some degree, it seems like the more we’ve learned about Omicron, the more it’s approached what might’ve seemed at the outset a best-case outcome — it does appear to be causing milder disease, for instance, and these waves do appear to be crashing much sooner than we might’ve feared. But if it doesn’t materially protect you against infection with other variants, including future ones, to be infected with Omicron, that’s not such an optimistic case.
I don’t know about severity. I know about those other reports in South Africa, but I don’t know why it appears so mild there. I think partly it has to do with previous infection, which protects against severe disease, but not against infection. And for South Africa, their fourth wave was massively different in terms of the prior immunity in the population. For the U.K., that wasn’t true; our last wave and our current wave are not hugely different in terms of the level of vaccination.
Which means you’d expect to see a much larger drop-off in severity in South Africa, because the share of the population with immune protection grew so much since the beginning of the last wave, when the country as a whole was much more vulnerable. That much-talked-about analysis suggesting the hospitalization rate had fallen by 90 percent — the hospitalization rate in the previous wave was 19 percent. That is astronomically high.
And here in the U.K., the recently published Imperial College report showed that the hospitalization rate appears to be not significantly different from Delta, which is really worrying for us.
Although that was based on some very preliminary data, right?
Yeah. It was based on preliminary data. That’s true. But, I mean, that’s the data we are seeing, and Denmark’s data as well. If you look at it and account for lags in hospitalization, I think the last analysis I saw was 0.6 percent hospitalization ratio versus Delta’s one percent or something.
That’s a meaningful decline. But it’s not 90 percent.
So I’m not convinced that in western Europe, at least, that what happens in South Africa is transferable. And, in fact, it already doesn’t seem to be unfolding. For us, actually, it’s been quite shocking. Most of us were thinking we would start seeing a rise in hospitalizations in about two weeks, but London hospitalizations are already rising very rapidly.
So even though this is likely to move quite quickly, you’re still expecting a quite grim next month or six weeks?
I think so. I mean, the thing is, even if it is mild, the rate of spread is going to affect so many more people that it’s going to overwhelm that advantage. Imagine that it’s half as severe as Delta, but the cases are doubling every two days.
You’re basically in the same place two days later.
Exactly. Yeah. Two days and you have the same number of overall hospital admissions.
That’s a good baseline for expectations, I think, given how many people get excited about drops in virulence. But that set of facts seemed scarier to me a week or two ago, when I could imagine Omicron just growing exponentially for quite a while. Now the fact that we’re seeing the waves slowing and peaking pretty quickly feels like at least a small comfort, since it implies the total number of cases we’re likely to see is smaller than we might’ve feared — and therefore, that the relative severity may matter a little more, rather than being just overwhelmed into irrelevance by the rate of spread.
But that’s really South Africa. We don’t know what the course is gonna be here. I don’t know if you can extrapolate the South African experience of that, either. I hope we can.
I definitely agree that the number of doubling times was fewer than I expected in Gauteng. And it’s not just Gauteng. We are seeing it across South Africa. And I hope that’s what happens here, but I think we don’t know yet. In the U.K., the cases are so high that we’re not able to track them anymore. And that means we just don’t have any idea what the doubling time is right now, or whether the rate of growth is reducing or not.
Is that what you’re most focused on right now, trying to determine whether the U.K. wave is really cresting?
Well, I’m definitely watching the hospitalizations, though that’s a really lagging indicator. But yes, the main thing that I’m looking at right now is positivity because I think that’s the only indicator that’s telling us about growth. It’s not a perfect indicator, but then cases are not reliable anymore.
And what is the positivity data that you’re seeing now in the U.K. telling you? That, in fact, cases aren’t slowing down as much as the raw caseload might suggest?
Yeah. I mean, certainly the positive rate in London shows that positivity is sort of vertical. It’s around 20 percent, despite cases appearing to sort of perhaps not show the same rate of growth.
Is that just fewer people getting tested?
Well, there are huge lags with reporting, too, so you don’t have a real picture of what kids’ numbers are on a given day until much later. But, yeah, there is definitely a testing-availability problem as well now.
Even more so in the U.S., I think.
Yeah, I heard about that.
This interview was edited and condensed for clarity.
More on Omicron
- Where to Get Home COVID Tests and Why They’re So Hard to Find
- Beijing’s Olympics Will Probably Be As Depressing As Tokyo’s
- How Wastewater Became a COVID Crystal Ball