How Bad Could the Delta Variant Get?

By David Wallace-Wells, New York Magazine

02 August 21

 

t present, there are two big anchors to conventional-wisdom thinking on the Delta variant: that those already vaccinated remain exceedingly well protected against the new, more transmissible strain, and that those who aren’t remain exceedingly vulnerable.

But a third fact seems to me to be just as significant in assessing the COVID risks the country faces going forward: that the age skew of the disease and the age skew of vaccine penetration, taken together, mean that the country as a whole has probably had at least 90 percent of its collective mortality risk eliminated through vaccines. Death isn’t the only outcome worth worrying about, of course — being hospitalized or even ventilated is no happy outcome for anyone, and the possibility of long COVID looms over each case (though precisely how prevalent that phenomenon is remains up for debate). But for most of the course of the pandemic, cases and deaths have guided our sense of the trajectory of the disease and proceeded together almost in lockstep. Almost certainly that relationship has been severed by mass vaccination, since the overwhelming majority of the most at-risk are now very well protected.

Indeed, both in the United States and in those similarly well-vaccinated countries whose Delta waves precede ours, that is what we’ve seen. In the U.K., which has had, after India, perhaps the world’s most striking Delta wave, the infection fatality rate may have fallen as much as 20-fold from previous waves. And it appears possible, at least, that for all the public alarm about British reopening in the midst of the Delta wave, that new case numbers may already have peaked and begun to decline. If this holds up, it is likely that the country’s seven-day rolling death total during the Delta wave will peak, a few weeks from now, below 100. In January, it was over a thousand.

In Israel, the story is much the same. According to the invaluable tracker at Our World in Data, the number of new cases has climbed from a peak below ten per day in early June to more than a thousand now, toward the end of July — a tenfold increase in about six weeks. And while, again, the death totals will likely continue to grow in the weeks ahead, over the same period of time, the country’s rolling seven-day death total has climbed from about 1.5 to 1.7.

Americans fretting about our vaccination rates often point to those two countries as models, but, as with pandemic performance before the vaccines, when it comes to vaccine rollout, we may be less far behind our peers than we tell ourselves. Israel has fully vaccinated 58.4 percent of its population and about 90 percent of those over the age of 60. In England, 53 percent of its total population is fully vaccinated, and 94 percent of its over 50s. The U.S. is at 49 percent of its total population and 90 percent of its seniors.

There is surely some dampening culture-war and disinformation effect on vaccination rates here, and the gap between 49 percent and 58 percent is not insignificant — in the U.S., getting to 58 percent would mean 30 million additional vaccinations. But among the truly vulnerable elderly, the rates across this cohort of countries are quite similar. Canada looks poised to become the world leader in vaccinations, after a slow start, with 52.6 percent of the country fully protected and no signs of slowing down (though it hasn’t yet reached 90 percent of even it’s over-80 population). It may be the case that there is something of a natural social equilibrium, at least in this early stage of rollout, at about 60 percent of total population and 90 to 95 percent of the vulnerable elderly. You get almost all of the old, many of the middle-aged, and frustratingly few of the young, who face the least risk from the disease itself.

While American caseloads have been growing at alarming rates in certain parts of the country, the effects of vaccination are just as striking here. In recent weeks, amid all the alarm over Delta, the seven-day rolling average of American deaths per million has, according to Our World in Data, ticked up from a nadir of 0.66 on July 8 to 0.74 today. It was over two as recently as late May, and over ten in January. In Los Angeles, where new mask mandates are being issued and in-door dining has been temporarily suspended, new cases have grown tenfold since mid-June. The seven-day average of deaths has in fact fallen from 6.3 to 5.9, though some amount of future rise is likely, since death peaks invariably trail case peaks. An NYU modeling team recently predicted that New York City might see 900 deaths in the Delta wave; in its horrific first wave, last spring, the city recorded more than 20,000.

A different team of modelers, called the COVID-19 Scenario Modeling Hub, working with the CDC, unveiled a starker set of projections Wednesday. In their most likely scenario, NPR reports, the modelers suggested a mid-October Delta peak of 60,000 new daily cases across the country and 850 daily deaths — about one-quarter as bad as the worst previous wave last winter. In their worst-case scenario, the Delta wave would peak at 250,000 new daily cases and 4,000 deaths. Both projections for deaths seem quite implausible to me, given that each would represent an even higher case-fatality rate than we saw in January, when the country’s new daily case rate peaked at 251,000 and the rolling seven-day death total reached just 3,400. Hardly anybody in the entire country was vaccinated then; today, we’re just below 50 percent of the total population and 90 percent of the senior population, where mortality risk is so concentrated.

In the U.K. today, there are about 47,000 new cases (about one-fifth the group’s projection for an American Delta peak), and while it is the case that death totals in the coming weeks will likely grow, at the moment the country’s seven-day rolling death total is, according to Our World in Data, just 55. Even extrapolating that trajectory out for a few additional weeks, it would imply, in the American context, that a peak of 60,000 new daily cases would produce something on the order of 100 daily deaths, not 850, and a peak of 250,000 new daily cases would imply something like 500, not 4,000. The group also assumes that the Delta wave would not just continue but continuing growing until October, several months from now; the indication of caseload declines from the U.K. and elsewhere suggest that may not be the case — that the wave may indeed peak and begin to subside well before. (On top of which, mid-October is a strange time for a projected peak, given that, last year, it was around then that seasonal fall effects began to kick in, accelerating spread.)

Of course, 500 deaths is still 500 deaths — and it is worth keeping in mind that these are figures that would’ve horrified us 18 months ago, as comparatively small as that number of tragic and preventable deaths now seems, given the brutality of past waves. On Thursday, I spoke about all this with Eric Topol, the head of Scipps and a sort of one-man clearinghouse for new coronavirus data the world over. He was somewhat less sanguine about the next few months.

I wondered if it might be useful to start by laying out how I’m processing what seem to be two contrasting trends. The first is that, according to the White House, something like 90 percent of American seniors are fully vaccinated, and since we know an awful lot of middle-aged people are too, and know the vaccines are effective, we can conclude that we have dramatically reduced the country’s overall mortality risk — probably by 90 percent, maybe 95 percent. There are lots of places with plenty of unvaccinated people, and overall the rates aren’t what we’d like them to be, but even in those places, vaccinations have skewed toward the most vulnerable — at least speaking relatively.

Nevertheless, speaking in absolute terms, there are still a lot of vulnerable people, and given a worst-case scenario of pervasive spread through the vulnerable population, you could see a lot of serious illness and death. It would be much, much less than would’ve been the case if that worst-case scenario of total spread had taken place a year ago, but the total figures could still be pretty grim. And, unfortunately, Delta may just represent that worst-case scenario. What am I missing?
I think overall you’ve got it pegged, but I would modulate some of that. Right now, about 85 percent infections are with Delta. We’re not at 99, but we will be probably in a week. And then we’ll be there for several weeks. So we’re going to have a Delta wave that comes to go through just like it did India. But eventually, as the Delta wave proceeds, it will start to come down — it’s starting to come down, perhaps, in the U.K. and Netherlands, though it hasn’t started to come down yet in Israel and other places. But it doesn’t get to everybody. It didn’t get to everybody in India — it was horrific, but there’s plenty of people still that didn’t get touched there by any of the 607 lineages, which include Delta.

Why is that? Is it because of changes in social behavior in response to rapid spread? Something particular about this variant? A reflection of the dispersion of unvaccinated through the country? Or some other factor, some combination, some dynamic we don’t truly understand?
There are many reasons why Delta will die out before getting to everyone vulnerable — you have listed some like change in behavior. It’s a combination of factors. But the best evidence is from India and now Russia, without vaccines at any appreciable percent — even there, however efficient the virus is, it’s not capable of reaching everyone. Just as the 1918 flu pandemic didn’t get to everyone. These pandemic pathogens burn through a population, but they invariably leave many behind who are vulnerable, not because they had prior COVID or some genomic host insulation. I believe the U.K. is clearly heading down now, which is a quite important prognosticator for the U.S. pattern in the weeks ahead. How many weeks and what will be the peak cases (and other outcomes) is the only unknown in my mind.

So the proclamation that some have made, saying you’re going to either get a vaccine or you’re going to get COVID, the Delta version — that’s not exactly accurate. Because even though it’s really efficient, this variant, it doesn’t find everybody. It just can’t get to everybody, but it gets to a lot of people.

How many?
We’re tracking right with the U.K., if you want to look at the log charts. They got to 50,000-plus cases. And if you multiply that by five, for the population difference, we’d get to 250,000 — that’s easy extrapolation. That could be where we’re heading.

That’s nationally, you mean — 250,000 new cases per day, right?
Some states look like they’re in really bad shape — worse, if you look at the arc of increase, including hospitalizations, than at any prior point in the pandemic. That’s not great. It doesn’t look pretty. But, as you aptly pointed out, we’re blunting the deaths, and to a lesser extent blunting the hospitalizations, because the younger people, they do get to the hospital, they just don’t die, fortunately.

The age skew for hospitalizations, while dramatic, isn’t as dramatic as the age skew for mortality. I think, according to the CDC, mortality risk is 600 times as high for someone in their 80s than someone in their 20s; for hospitalization, it’s just a 15-fold increase.
Right. And in this Delta wave, the hospitalized are mostly unvaccinated younger people. The other thing I’d say is a lot of people discount long COVID, but that’s a big deal. If we do get to 200,000 cases a day, that’s a lot of long COVID.

From what I can tell, estimates of that prevalence are really all over the place — some studies suggest rates as high as 30 percent or even 50 percent of all cases, but those don’t seem to me to be very good surveys and would suggest something like 50 million Americans are dealing with a debilitating chronic condition already. Some other estimates are very, very low — considerably under one percent, even. How do you ballpark?
Ten percent. Probably it’s either high single digits or low double digits is the real deal. When you get north of that, with those surveys showing higher figures, those people are not necessarily dealing with serious symptoms for, say, a year plus — they’re getting better, or their symptoms aren’t as worrisome. They’re not as debilitated. But for the real-deal cases — the ones that can’t work, the real, significant brain fog, the ones that really are suffering — it’s probably one out of ten.

Wow.
It’s a nasty, horrible condition for some people. It doesn’t get enough respect still.

The one other point is this breakthrough-infection thing is probably worse than we have estimated. I don’t mean for death and hospitalizations, but in terms of the ability of the disease to burn through the vaccinated to make them get them infected — it’s not looking as good as we’d hoped, I think it is fair to say.

Right, though, of course, none of the vaccines were tested in clinical trials to measure their efficacy in preventing transmission. This is a bit of a side note, but I have wondered over the last few weeks what the public response would have been, and if it would’ve been different, if the initial trials had included those measurements and the first announcements about the vaccines hadn’t been 90 percent-plus efficacy against symptomatic disease but 90 percent-plus efficacy against symptomatic disease and only, say, 60 percent against transmission.
I’m really glad you brought this up, because this was a real miscue. All the trials, the end point wasn’t death or hospitalization but symptomatic disease.

Right, though we could also look at the deaths in the data.
And there were very relatively few deaths and hospitalizations in those trials, 75,000 people in the two trials, Moderna and Pfizer — there were like nine deaths in the Pfizer trial and 30 in the Moderna.

But what you’re onto is a really big deal because the vaccines are basically not living up to the trials in terms of symptomatic disease, though they are offering great protection against what the trials didn’t test — death and hospitalization. We’re getting the same 90-some percent reduction of those end points. But with Delta we’re seeing a drop-down of protection in symptomatic disease and transmission. It is being transmitted. It is getting into people who’ve been fully vaccinated.

Is it fair, given that and given the age skew of the disease and of vaccinations, that at this point, the worries of long COVID are sort of at a social level, a bigger worry for you, then, than from death itself?
Well, insofar as we’re going to be seeing a lot more long COVID, yes. But, you know, even one death …

Right, every death is terrible.
And now we’re talking about younger people dying, predominantly. These are all catastrophes, even those who don’t die — prolonged hospitalizations and all that goes along with that. So, no, I wouldn’t want to say that I’m not concerned about these other hospitalizations and deaths — we will see them. They just won’t be like the monster third wave.

How much less bad, do you think?
Hopefully, like you alluded to, one-tenth as bad, 90 percent less mortality. That would probably be the best-case scenario. But, still, those are tragedies, and Delta is already flooding the health systems in Arkansas, Missouri, Louisiana, just like in the beginning. And there’s study after study showing that when a local health system gets overloaded, more deaths occur. So in those unvaccinated pockets, the deaths and hospitalizations will be appreciable. That’s part of the replay — this is the movie we all never wanted to see again.

Speaking of those overloading ICUs and ERs, there’s been a lot of reporting recently about those places being now full of younger and sicker patients in the first wave. How should we think about that? Is it the simple result of a more transmissible variant — that something that moves through the population faster will yield a more compressed wave that strains hospitals just because it is moving so fast? Or is there something about Delta that has changed the virulence or age skew?
There’s two confounders here that make these stories very difficult to interpret. One is, to have something that’s hypertransmissible, you can’t look at the data in the same way — if you look at Russia or Indonesia, say, you might think, Oh my gosh, the deaths are higher than ever before in the pandemic, it must be more deadly. But the deaths could just be explained by transmissibility.

And then the other confounder is the vaccines. In the U.K., the case fatality rate was previously over 2 percent and now it’s 0.12 percent. A lot of that magnitude of reduction is because it’s younger people. So it’s very hard to say. But the way I put it is, you know, we’ve got enough problems with Delta, with its being so remarkably transmissible, and with some immune evasiveness. We don’t need to assign it another feature of being more deadly — we just don’t need to do that. We don’t have proof that it is, for starters. There’s just some dangling, minimal data out there one way or the other. But at this point it’s really not the principal issue. The principal issue is it just spreads so darn easily.

So even with the new research about the high viral loads, suggesting Delta producing 10,000 times more virus than the previous strains, your intuition is that the virulence is in the same neighborhood as the other variants, is that right? Not dramatically different?
That’s my sense. You know, it’s possible there may be, in the weeks ahead, much better data to sort through this, but it’s a very challenging thing to assert, now, because there are just too many things to try and control for. It’s not like we have a randomized trial here.

Right. And what about the effect of these outbreaks on vaccine rates? Are you hopeful that, in a perverse way, they may make people more eager or willing to get vaccinated locally?
Yeah. And finally we’ve seen some Republican leaders wake up, though, you know, months too late.

It’s like there was a memo sent out.
It sure seems like it. It really all came out in the last 48 hours. Unfortunately, by the time they did come out and say that stuff, we already had Delta penetration throughout the country. It takes six weeks minimum to get some immunity. So for them to be calling for people to go out and get vaccinated now — I mean, it’s like the FDA giving us a full approval this coming September. What good is that going to be when the wave is over?

I think we are seeing small evidence of the stage that people are getting scared in Missouri and Arkansas and are lining up to get vaccinated. But the numbers aren’t big. We’re probably at 500,000 doses a day total, and we were at 4 and a half million a day in March and April.

Other countries get a lot of credit for doing better than us with vaccines. But it doesn’t feel to me like it’s categorically better. Israel is at 60 percent of the population. The U.K, is at 53 or 54, and we’re at 48 or 49 or something like that.
Even so, the gap in fully vaccinated is profound. So if you look at Israel at 58.2 percent of the total population now, and the U.S. is 48.7 percent — that’s 10 percent of the total population. And in Israel they have a much younger population, so they can’t vaccinate their kids.

And we’re just not positioned as well. Back in June, our vaccination program just collapsed. It’s been horrible. Canada is the ultimate comparison. I mean, Canada is just chugging along. It’s going to be the top in the world pretty soon. So, no, I am disappointed.

How do you think it all plays out heading into the fall?
Looking ahead to the fall, I’m optimistic. Delta will have passed through by then — it’ll pass through by late August, or September, if it looks like India or the U.K. or Netherlands. We’ll have a rapid descent, and it’ll burn through. We’ll still have lots of COVID in this country, but it’ll be back to where it was before Delta came. It will be at a lower level. The only question is, is there something lurking that’s worse than Delta? There’s no sign of it yet, but there’s too much of this virus circulating to be confident — too many people in Indonesia and sub-Saharan Africa who are getting sick. But I hope not. I’m hoping that this is as bad as it gets. But if you talk to evolutionary biologists, they’ll tell you the variants are going to get worse.

Right, though, just to return to the beginning of the conversation, it’s still a very different-looking pandemic because of the vaccines, right? This variant is bad, future ones will be, too. There will be outbreaks and new cases and some amount of cases, including serious illnesses and death. Certain hospitals may be overwhelmed. None of this is pretty, none of this is happy. But, still, in the big picture, we’ve made so much progress from where we were six or nine months ago.
I don’t think it’s either-or. It’s both. If you emphasize that we made some progress, that’s true, but you could have made a lot more progress with more vaccinates. It’s fantastic that we’re going to see a whole lot less death, but, you know, haven’t we seen enough death already now? Haven’t we seen enough people suffering in the hospital? This nightmare we’ve had — you know, enough of this! To discount potentially hundreds of more deaths a day — it could be a thousand, it could get that high at peak — that’s a lot of people dying. We’ve never gotten it down much below 300.

I think we’ve gotten numb here. We’ve gotten numb to the point that if we had done a much better job vaccinating, like any other vaccine in our history — like polio or many others where everybody got vaccinated — we wouldn’t be dealing with nearly as many deaths, hospitalizations, or the big burden of cases and on and on. I think if you just pick the upbeat side of this, it ignores that. There is a real downside here that we can’t ignore. When you have 35-year-old people — healthy people, perfectly healthy — who wind up in the hospital and are teetering on death, when you have that, you say, God, what are we doing here? We could have prevented this.