Can vaccinating for COVID-19 cause a selection pressure for more virulent strains?

Question

The steelman of an argument I've seen floating among some critical members of society is as follows:

Since COVID-19 vaccines are applied as a prophylactic (reducing symptoms although people may still be infected), or in any other way leaky, we are creating an environment where the virus is more likely to mutate to a more lethal variant.

This argument is usually made in reference to Marek's diease, or paralleled with some other form of immune escape, as in antibiotic-resistent bacteria or malaria.

I guess Marek's diseases is most salient, because mass vaccination of chicken has selected a strain that is so lethal that chicklets are sure to die unless vaccinated.

Decades after the first vaccine was introduced, current strains of Marek Virus cause lymphoma formation on throughout the chicken's body and mortality rates have reached 100% in unvaccinated chickens. The Marek's disease vaccine is a leaky vaccine, which means that only the symptoms of the disease are prevented.

Infection of the host and the transmission of the virus are not inhibited by the vaccine. This contrasts with most other vaccines, where infection of the host is prevented. Under normal conditions, highly virulent strains of the virus are not selected. A highly virulent strain would kill the host before the virus would have an opportunity to transmit to other potential hosts and replicate.

The vaccine's inability to prevent infection and transmission allows the spread of highly virulent strains among vaccinated chickens. The fitness of the more virulent strains is increased by the vaccine.

Perhaps one of the highest-profile people to make this claim (in my part of the world) is Geert Vanden Bossche.

I feel like, if the argument were true, medical agencies would not advocate for mass vaccination including people less at risk of serious disease, like young people.

So apparently the claim should be false and the COVID-19 vaccines do not cause these unintended fitness pressures. What is the hole in the argument? What is the meaningful difference between the epidemiological effects of the MDV vaccine and COVID-19 vaccines?

Answer 1

Well, the question is theoretical, so will be the answer. In the form of a 2015 commentary (in Science) on the paper that seemingly first proposed this concept:

Vaccines have saved millions of human lives, but according to evolutionary biologist Andrew Read they sometimes may also cause pathogens to turn deadlier. Read first put forward the theory 15 years ago. Now, in a new paper, he presents evidence that that is what happened with the virus causing Marek's disease, an infectious disease in chickens. Read acknowledges that the effect has never been seen with human vaccines, but he argues that future vaccines that prevent disease rather than infection could have the same effect. Other researchers say that no general conclusions should be drawn from the paper. Even if he turns out to be right, the study offers no support whatsoever for those who oppose vaccination, Read stresses. If "leaky" vaccines are proven safe and effective, they should be used, he says, but perhaps with closer monitoring and additional measures to reduce transmission, such as bed nets for malaria. [...]

But other researchers say the study has little relevance for public health. Read “should stop scaremongering,” says vaccine researcher Adrian Hill of the University of Oxford in the United Kingdom. [...]

It’s a convincing study, says Michael Lässig, who studies influenza evolution at the University of Cologne in Germany, “But it’s a very special set of circumstances … I would be careful about drawing general conclusions.” Hill also thinks that Marek’s disease may be a special case; nothing suggests that human vaccines have ever made a disease more virulent, he says. What’s more, natural immunity is “leaky,” too, Hill argues, allowing infected people to survive and transmit a disease that is deadly to others. “For malaria, whatever today’s vaccine does is a drop in the ocean of all the immunity that is happening in Africa from all the infections,” he says.

There's some more expert commentary on that 2015 paper.

“So, what have they shown? Put simply, they have shown that chickens infected with a lethal strain of the virus (Marek’s disease virus, MDV) can only transmit it until they die (whereas those infected with a non-lethal strain can transmit it over a prolonged period). If chickens were immunised with a “perfect vaccine”, it would both protect them and prevent them transmitting MDV to other chickens. Current vaccines against MDV, however, are “imperfect”, in that vaccinated chickens will be protected and will survive but will still be able to transmit the virus to other chickens (which will, unless they too have been vaccinated, die). Vaccinated chickens therefore present a mechanism for prolonged shedding and transmission of the lethal virus.

“What have they not shown? As the authors acknowledge, they have not demonstrated evolution of MDV from non-lethal to lethal forms in vaccinated chickens. Their work does, however indicate that, if such a mutant virus arose, it could be maintained in chickens vaccinated with the “imperfect vaccine” but not in unvaccinated chickens.

So, in other words, the putative evolution of the virus hasn't really been demonstrated in that paper. Not sure how you'd expect it to be proved or disproved for a human-hosted virus, if doing that for one that infects animals is so difficult...

And by the way, the "leaky" term is not really that widespread... because a number of researchers disapprove of it:

“A word of warning on terminology – the authors use various terms to describe the vaccines: “sterilising” versus “non-sterilising”, “perfect” versus “imperfect” and “leaky” (versus presumably “non-leaky”). Apart from the matter that technically it is the immunity induced by vaccines that is “sterilising “ or “non-sterilising”, the term “leaky” is to me the least desirable as it could be seen to imply an implicit defect of the vaccine, which could be translated by some as meaning that the vaccine itself poses some threat – this is not the case.”

On the other hand, there are some recent reviews (e.g. Milller & Metcalf 2019) that are more favorable to Read's hypothesis, and find some corroborating evidence from how (imperfect) natural immunity affects virulence) although as I mentioned the "leaky" terminology is not favored.

Read et al. [16] showed convincingly that when vaccines blocked the mortality effects of Marek’s disease virus in commercial poultry the virus evolved higher virulence to the point where vaccinated hosts lost all benefits of vaccine-induced immunity and non-vaccinated individuals experienced increased mortality. Recently, Fleming-Davies et al. [17] showed that incomplete acquired immunity (analogous to vaccination) to the bacterial pathogen Mycoplasma gallisepticum in finches selects for increased virulence, as low-virulence strains are unable to infect previously infected hosts while high-virulence strains retain this ability.

[citing 17]: Fleming-Davies AE, Williams PD, Dhondt AA, Dobson AP, Hochachka WM, Leon AE, Ley DH, Osnas EE, Hawley DM. 2018 Incomplete host immunity favors the evolution of virulence in an emergent pathogen. Science 359, 1030–1033. (doi:10.1126/ SCIENCE.AAO2140)

What I would call false however is any attempt to imply that only vaccines can cause this effect. As Fleming-Davies et al. put in the conclusion of their paper...

The effects of incomplete immunity described here are arguably a specific case of a broader phenomenon whereby increased virulence is favored by quantitative host variation in susceptibility, whether due to host genetic variation, imperfect vaccines, or innate immune priming.

A bit more searching does find one preprint that discusses via modelling/simulation how it might apply to SARS-CoV-2... apparently it doesn't based on the title "No current evidence for risk of vaccine-driven virulence evolution in SARS-CoV-2". I'm not up-to-speed with the research on how the Delta variant might have arisen, but for the "Kent" one it's been suggested/suspected that it came about in an immunocompromised individual (no vaccine was suggested to be involved) due to the number of accumulated mutations.

And since you name-dropped Bossche, he's selling [at least as an idea] some kind of "universal vaccine" holly grail that doesn't currently exist, and probably can't really exist either. He should not be really confused with Read though; as far as I know the latter is more reputable researcher...

Answer 2

No, it's false. Plus, it's speculation - so you can't even directly disprove it. That is not how evolution works though, so it's not something to worry about.

https://science.sciencemag.org/content/336/6085/1157

in many cases, organisms need to perform multiple tasks that all contribute to fitness; thus, fitness is an increasing function of the performance at all tasks F(Pi(v),…,Pk(v)), where Pi(v) is the performance at task i. The best phenotype for one task is usually not the best for other tasks—resulting in a trade-off situation. Maximizing fitness is thus a multi-objective optimization problem.

Fitness is how good an organism is evolutionarily - how good it is at producing offspring which have the same genetic makeup. However, this fitness is the combination of a lot of traits. For a virus, one trait might be immune evasion, one might be viral load, one might be resistance to elements, etc (If you've played Plague, Inc. this will sound familiar). Improving one trait in the real world usually means that some other trait will get worse.

https://www.nature.com/articles/s41564-020-0690-4

Mutations can also make a virus either more or less virulent. A common idea is that virulence will only change — either upwards or downwards — if it increases the transmission rate of the virus, which effectively means an increase in the number of virus ‘offspring’. However, high virulence may (although by no means always) reduce transmissibility if the host is too sick to expose others. Without information on the precise evolutionary forces and selection pressures in operation, predicting how virulence might evolve is an extremely difficult and perhaps futile task.

The claim is that vaccinations will cause more lethal variants to be selected. That's not what vaccinations select for, though - they select for variants which have a trait that improves vaccine escape (by definition). The only way for this claim to be true is if there is some causal linkage between escape and lethality. It's not just that there's no evidence for a link; we'd expect the opposite to happen most of the time. If a mutation could occur that would improve both immune evasion and transmissibility, we expect that it would have already occurred. It's more likely that mutants will trade off one trait for another. This isn't a certainty, it's just what evolutionary theory says the odds are.

What isn't speculative is that vaccines are far and away preferred over spread without vaccines. In that situation you have even more risk of viral mutation, as the overall viral load (e.g. the number of virus replications in the entire population of people) is substantially higher.