Is there any reason to believe that higher infectivity should correlate with lower lethality? i.e. the virus which keeps its host alive the longest has greatest opportunity to replicate and spread?
Yes, there's kind of a triad of attributes that affects a virus and it's ability to cause a pandemic. Virulence (how sick people become), transmissibility (how easy does it spread), lethality (can it kill at high rate).
-higher virulence causes people to stay at home more, and less spread in community
-higher transmissibility means more spread person to person
-high lethality means people die faster and not spread as fast
SARS and MERS before SARS-COV2(COVID) all had high lethality and therefore those that were infected didn't get a chance to transmit to others that well, and likely helped slow the spread. COVID seems to have a good balance between the three.
The above is a big simplification, but the general idea sticks.
Your model is missing one big variable and that is latency, the time difference between when people are most contagious and when they are symptomatic.
SARS-1 had low latency, and so people tended to become contagious and symptomatic at around the same time, reducing transmission. Whereas SARS-2 has a high latency which helps it to spread while people are unaware they even have it. This also insulates SARS-2 from selective pressure to reduce its virulence.
It should be noted that for Sars-Cov-2 specifically (i.e. given a virus that has the properties of Sars-Cov-2 already) lethality will likely be a neutral factor for the evolution going forward. The reason being that viral titres peak around day 2 in a window of 1-5 days, whereas it takes at lot longer to die from Covid. In other words there is same sort of cap where lethality prevents transmission, but Sars-Cov-2 is nowhere near it.
In both the natural and leaky-vaccine scenarios, the virus evolves towards an ecological equilibrium, but in the leaky-vaccine scenario, that equilibrium is shifted in the direction of greater lethality. See the linked article: as a direct consequence of mass vaccination of chickens, Marek's diseased evolved from a minor illness into a highly-virulent, highly-transmissible disease that causes 100% lethal cancer in chickens unless the chicks are vaccinated before they hatch from their eggs.
What do you think about this article, by the author of the Marek’s disease study, that criticizes extrapolating from the chicken scenario to the covid scenario?:
The core argument is that Marek kills in 10 days or less, which is too soon for viral spread to happen. MDV vaccines keep chickens from dying, thus spread of hot variants can occur. Since covid doesn't kill that much, Marek phenomenon can't happen. The obvious issue with the argument is that the hot lethal variants are observed post vaccine availability. Earlier strains were "mild" or "virulent", but not "very virulent plus". "Death was relatively rare".
From an 1998 article on Marek, there appear to have been at least 3 jumps in Marek virulence, one in the late '50s pre vaccines (mild => virulent) and two post vaccine availability (virulent => very virulent, very virulent => very virulent plus), see chart on page S50.
"MDV became increasingly virulent over the second half of the 20th century [19,21–24]. Until the 1950s, strains of MDV circulating on poultry farms caused a mildly paralytic disease, with lesions largely restricted to peripheral nervous tissue. Death was relatively rare. Today, hyperpathogenic strains are present worldwide. These strains induce lymphomas in a wide range of organs and mortality rates of up to 100% in unvaccinated birds."
The conclusion of the rebuttal article you linked to is rather terrifying. If there ever was a "the road to [chicken] Hell is paved with good intentions", this is it.
"In the history of human and animal vaccines, there have not been many cases of vaccine-driven evolution. But in every one of them, individuals and populations have always been better off when vaccinated. At every point in the 50-year history of vaccination against Marek’s disease, an individual chicken exposed to the virus was healthier if it was vaccinated."
(1) from m MDV to v MDV strains in the late
1950s;
(2) from v MDV to vv MDV in the late 1970s;
(3) the appearance of the putative vv + pathotype
in the early 1990s (Figure 4).
Each occurrence has been associated with greater disease losses which have persisted until introduction of the next generation of vaccine. The reason for the shifts remain speculative. Shifts in the 1970s and 1990s may have been in response to certain MD vaccines. Kreager (1996) has suggested vaccines may have only a 10-year useful life before new strains evolve. The shift in the 1950s occurred prior to vaccine use, but may have been related to changes in poultry husbandry, especially the move towards larger houses and the increasing density of poultry operations in certain geographic regions"
m = mild, v = virulent, vv = very virulent, vv+ = very virulent plus, as explained on page S46, Table 1.
To be very clear, Marek is not covid. I am drawing no equivalence. I just point out that Andrew Read's 2021 articlet is built around a weak argument.
There is selection pressure on the virus to avoid severe symptoms, because hosts that are up and being sociable are better vectors for the virus than hosts that are in bed or hooked up to tubes in hospital.
However, there is also selection pressure for the virus to infect cells more efficiently and reproduce in greater quantities. These traits tend to correlate with the virus being more dangerous to the host.
The way the virus transmits before the host becomes symptomatic is quite an effective strategy for the virus, and means the selection pressure for the virus to not cause symptoms is less than it otherwise would be.
