I mean, sure, Delta clearly has this fitness advantage that allows it to reproduce itself more rapidly. It makes higher viral loads more quickly, so it loads aerosols with more virus particles, and therefore it bangs on the door of every prospective susceptible individual more vigorously than Alpha had done before it. It therefore acquired higher transmissibility than Alpha, just as Alpha managed the same trick with respect to EU1 before that. But so what? Why are Alpha and EU1 not still around, in some less prominent form?
One might argue that the older variants do persist, but since only a sample of the infected population is genomically sequenced, they become less and less likely to be observed. Fair enough, but here’s a sharper form of the same question: Why do we so rarely see co-infections by different strains of the virus? Co-infections do occur, but they are certainly rare enough to merit individual journal articles when discovered, and rare enough to never be recorded (or at least brought to public attention) in routine genomic surveillance of SARS-CoV-2 specimens, even in the UK’s genomic surveillance program which is actually based on Whole Genome Sequencing (WGS), and ought to be capable of noticing such co-infections. Why is that? When Alpha ruled the roost, but Delta was intent on knocking it off its perch, there was a time when they had about equal prevalence (in the UK that occurred between 3 May and 17 May). At that time, infections had about equal chance of being Alpha or Delta. Why were there not at least a few “both” infections detected last summer? Why was it always either one or the other?
What I’m trying to suggest here is that these two observations—essentially complete displacement of one rapid-transmission variant by another, and rarity of co-infections by those variants—are in fact connected, and the connection is related to the discussion of breakthrough infections in this series of posts.
Cross-Immunity
The connection that I am suggesting is the following: I believe that so far, we have not really seen a variant capable of effectively hiding from the short-term immunity created by infection due to other variants. As a consequence, every new rapid-transmission variant that has shown up to-date—notably Alpha and Delta—has announced itself to any immune systems that it infected as an exemplar of the SARS-CoV-2 genome. This has the effect of poisoning the susceptible host landscape for any slower-transmitting variant, because by the time the slower variant gets around to presenting itself to a potential host, that host’s immune system already knows that it is under attack from SARS-CoV-2, and presents a hostile environment to the attacker. Obviously, this would be all the more true if the potential infectee’s immune system was primed by vaccination against SARS-CoV-2.
If this is correct (and again, I could be wrong about this) then this is the reason we have so seldom seen co-infections by multiple strains: there is only a narrow time window for the second strain to infect the host before the enraged immune system mounts its defense. The contact rate for new infections would have to be faster than the timescale for the rise of immune response, and even if it were, by the time the new infection got a foothold in the organism the first infection would likely be many exponential growth times ahead of it anyway, making the new virus that much lower in viral load and hence less detectable in an assay. So co-infections ought to happen, statistically, but they also ought to be rare, and hard to observe.
And under this hypothesis the answer to the question of why new SARS-CoV-2 variants have essentially annihilated their predecessors is that in effect, they present the same genomic signal to human immune systems, creating cross-immunity, so that the more rapidly-transmitting variant in effect starves the less rapidly-transmitting variant of susceptible hosts to infect, unless the slower variant can find some local isolated tidepool of susceptible hosts to hide out in.
There Have Been No True Vaccine-Escape Variants Yet!
If one accepts this view, one consequence is that to date we have not yet seen a single genuine vaccine-escape SARS-CoV-2 variant—one capable of producing Type 1 breakthrough infections! If we had, such a variant could not have been driven out of the genome by a later variant, because the later variant would not be capable of creating cross-immunity against it. The escape variant would have learned to hide from the immune response adapted to SARS-CoV-2, so such responses would not affect it. Nor would such an escape variant drive out a previous variant, for the same reason! The two variants would circulate concurrently, like flu and rhinovirus, and one would certainly expect to see frequent co-infections.
This is an odd revelation for me. Until recently I was sure that Beta (B.1.351) was in fact a Type 1 vaccine-escaper, based on the number that it did on vaccines in phase 3 clinical trials in South Africa (where Beta was prevalent), which I wrote about at the time. But looking back at some of those analyses, and in particular at the J&J clinical trials, one can clearly see the signature of Type 3 breakthrough in the South Africa trials by comparing the protective efficacy against infection (poor) with the efficacy against severe disease (pretty decent). This is what one expects in Type 3 breakthroughs, with neutralizing antibodies struggling to keep up with the rapidly-growing infection, but the killer T-cells doing precisely the job they were designed to do, and preventing severe disease. Also, look at the Covariants Per-Variant page, at the panel for Beta: that variant is definitely gone (except for an ostensible trace in the Philippines which is actually a data corruption artifact, since there is no data from the Philippines since 6 September) [Note added 10 January 2022: In updated data, as of 2021 November 29, Beta is officially gone everywhere. RIP.]. That would definitely not have happened, if Beta were a true escape variant, untroubled by the T-cells that kill Delta.
If this view of the matter is correct, there’s some very good news, some interesting news, and some somewhat worrying news.
The Very Good News: Delta Is A Goner
As of this writing, the reporting on genomic sequencing lags by far too long (weeks) to keep up with the progress of Omicron in Europe and the US. In South Africa, as of 29 November, Omicron was 96%, and everything else was anecdote-level numbers. More up-to-date numbers are available from local health systems and hospitals, and they show that Omicron is, as of late December 2021, already prevalent in the USA and in Europe. The trend suggests that Delta will be entirely gone from community transmission by some time in January. Given that the symptoms associated with Omicron are (clinical consensus now has it) milder, this is as if Omicron were tearing around madly, alarming human immune systems at a much faster rate than Delta can burgle them, but is itself not much of a burglar.
The new variant appears to be in the process of turning the COVID-19 disease into something like the flu, in a classic demonstration of genetic virulence drift, played out at worldwide scale and at high speed. It’s not that I particularly like being in the Petri dish, but the experiment is really fascinating, and the results are welcome.
The Interesting News: A New Epidemic Surveillance Signal For Vaccine-Escape Variants
Here’s an interesting implication of this discussion: if a real vaccine-escape variant were to emerge—that is, a variant capable of producing Type 1 breakthrough infections, as described in the typology that I gave in this post, due to its ability to conceal itself from vaccine-primed antibodies and T-cells—then we would expect to suddenly see a new phenomenon that we have not observed in the past: frequent co-infections by the contending variants should become common. The reason is that the new variant will not, by definition, present itself as an exemplar of the known SARS-CoV-2 genome, so vaccinated hosts’ immune systems will not immediately set out to slaughter it, even if they are presently engaged in a campaign against an infection by the older variant.
A sudden surge of double-variant infections would therefore signal the advent of a true vaccine-escape variant, capable of producing the Type 1 breakthrough infections, the development we should be most fearful of. This is, in a sense, “good” news, because it’s a signal we ought to be able see immediately, should it arise, given the right type of genomic surveillance, and a real-time signal produces more time for policy to adapt, by tuning mRNA vaccines, say, or preparing a new round of lockdown measures.
A co-infection surge would, in due course, be followed by a second characteristic signal: the two variants would co-exist in circulation, without one driving the other to extinction. However this would be a time-series signal, and time series take, well, time to develop and be observed. The co-infection signal, by contrast, would be immediately obvious, and a clear advance warning that the threat had just changed for the worse.
It might not be easy for some of the next-generation sequencing (NGS) techniques to catch a co-infection signal, since some of them are optimized for high-throughput at the expense of high genomic resolution. However there appear to be some NGS techniques that are sensitive to co-infection (here I must own up to having merely Google-search level of expertise, which is to say not much…). And as I pointed out above, the UK at least has an impressive high-throughput (about 50,000 specimens per week) genomic surveillance program that uses WGS, and should be able to detect a co-infection signal, if they’re looking for it.
The Worrying News: Omicron May Not Be The Exit
I sincerely hope they are looking, and that other genomic surveillance programs start looking too, if they aren’t already.
I wrote in the last post that perhaps the advent of Omicron is the pandemic exit ramp. It would be nice to think so. But honestly, remember the “mutation casino” analogy from that post? The casino is going to still be open after Omicron takes it over, and viruses will still be pulling one-armed bandit levers all over the world, looking for fitness payoffs. In the “Immunosuppressed Patient” theory of Omicron’s origin, the variant learned not to kill its host, which is nice. But it never learned any tricks to dodge a real human immune system. Learning such a trick, it seems to me, would constitute a new fitness advantage, because such a variant could then more easily re-infect “immune” hosts, potentially at about the same rate that Omicron is currently capable of (global dominance in 2-3 months).
I have no idea why no variant appears to have learned this trick yet. Presumably it’s not so easy to spoof the human cellular immune system. But the more virus particles in circulation, the more the one-armed bandit lever-pulls, and the greater the chances that one day it may happen. Then we could have two SARS-CoV-2 variants circulating concurrently, and capable of concurrent infection. And the new one wouldn’t even have to contend with a vaccinated host population, at least until new vaccines were developed and rolled out on a mass scale. Not to put too fine a point on it, that would really suck.
This is a huge, and utterly foreseeable risk. It’s past time we shut down the casino, by manufacturing and shipping billions of COVID-19 vaccine doses to all the parts of the world where there are plenty of unvaccinated arms not belonging to militantly ignorant, scientifically illiterate, sociopathic anti-vaxxers. There is plenty of willingness to take a jab in regions of the world with little or no vaccine availability, and vaccinating those regions needs to become a top priority for wealthy nations. It’s not just a moral duty (although it is also that): it’s a strategy to abate the risk of prolonging and extending the pandemic, or at least to minimize its impacts and costs—compared to which the cost of a global vaccination campaign is mere accounting noise.
I hear the whirring of the slots. It’s an ominous sound, which should be giving everyone the heebie-jeebies. Let’s get it done.