Viruses Don’t Always Get Less Deadly
What prior pandemics can tell us about the future of Covid-19.
I see the claim all the time on Twitter:
“Viruses get weaker over time. Covid will turn into a cold.”
Here’s one widely circulated tweet making this claim:
The argument is simple. Wouldn’t a virus spread better if it didn’t kill its host? It sounds plausible at first glance.
So far, this isn’t happening with Covid. It looks like Delta is actually more lethal than the original strain of Covid. Delta is maybe twice as likely to hospitalize or kill an infected person, according to data in Scotland, Canada, and Singapore.
There are other viruses that don’t get weaker over time. HIV is 100% fatal without treatment. It’s not getting any less deadly. The disease takes years to kill you, there’s plenty of time for it to infect other people before symptoms set in. There’s no pressure for it to become less fatal.
HIV is notable for its ability to hide from the immune system. Your body can fight off many viruses within days or weeks. For those, there’s no sense in staying asymptomatic. Their goal is to multiply quickly, produce some symptoms, and spread to another host before they get defeated.
Rabies is 100% fatal and it’s not getting any weaker. After you get bitten by a rabid animal, the virus works its way slowly through the nerves to your spinal cord and brain, over the course of 3–12 weeks. The virus multiplies in the brain and also spreads to the salivary glands. Symptoms start with fever, headache, and anxiety. They move on to aggressive behavior, frothing at the mouth, hallucinations, and paralysis.
The virus kills brain cells by overstimulating nerve receptors, a process called excitotoxicity. There’s no good treatment for rabies. You can take an effective vaccine after you get bitten but before the virus has reached the brain. But if you wait too late, you’re probably dead. The body can mount a successful immune response, but nerves are destroyed before that happens. Doctors have tried putting patients in a coma to prevent excitotoxicity. That’s worked 3 times it’s been tried, it’s failed 50 times, and every other rabies patient has died without treatment.
Destroying the brain is a feature, not a bug. The agitation gets an infected animal to go out and bite others. As long as each animal bites one other, the virus keeps spreading. There’s no pressure on the virus to become less lethal. If it didn’t damage the brain, it wouldn’t spread as easily. If it stayed dormant without symptoms, the immune system might clear it.
Another case of sustained viral lethality can be found in Australia. Australia is troubled with invasive species, where many things brought over from Europe take over, with no natural competitors.
European rabbits came to Australia in 1788. They quickly became feral and grew to large numbers. In 1950, scientists introduced a virus called Myxomatosis to try to kill the rabbits. The virus was 99.8% lethal, so this seemed like a chance to eradicate the species.
The experiment didn’t work. Over time, the surviving rabbits became more resistant to the virus. The virus also became less deadly. Various strains were found in the wild that were anywhere from 50% to 100% fatal. Over time, the virus settled on about 70%. Too lethal and it killed rabbits too quickly to spread. Not lethal enough and it didn’t produce enough symptoms to spread. The virus didn’t wipe out the rabbits, it just made life much worse for them. Now any given rabbit can fall ill with only a 30% chance of survival.
There is no rule saying that all viruses evolve to become mild.
There is only one rule for evolution: anything that reproduces, survives.
None of the examples above were respiratory viruses, so maybe it’s not a fair comparison. But they can help us reason about the problem a little.
Covid can infect people before any symptoms start. The deadly effects of Covid happen 2–3 weeks after the patient is most contagious. Where is the selective pressure for it to become milder?
Maybe we can predict its course by looking at other respiratory viruses.
There are 4 coronaviruses other than covid circulating around the human species. All of them cause mild colds. About 25% of all colds are caused by these 4 viruses. Most others are caused by rhinoviruses, of which there are over 100. Making a vaccine for the common cold would be difficult because of this diversity.
We have a guess as to when one of these coronaviruses, called OC43, jumped from animals into humans.
We can make a history of a virus from its mutations. This is easiest for Covid because we’ve been watching it since the beginning. Covid started with one case and mutated to become millions of cases. They all fit into one family tree:
For other viruses, we can work backward to look at the current diversity, the current mutation rate, and guess when the tree started.
For OC43, this calculation points to sometime around 1890. In 1889, there was a disease outbreak called the Russian Flu that killed a million people around the world.
So, there is precedent for one coronavirus going from a deadly pandemic to a mild cold over 130 years. It’s not obvious how many seasons it took to become a milder virus. Researchers write, “It probably lasted a fair number of years, like bad flu seasons, until it slowly lost pathogenesis.”
It’s also not obvious how much of the change is from mutation and how much from immunity. Most people today encounter OC43 as children, the body learns to deal with it while they’re young. OC43 occasionally causes outbreaks in nursing homes; in one example, it killed 8% of the elderly people infected. If you released the same virus on an unexposed population, would it kill millions of older people? We’re not entirely sure if the virus changed at all.
We have better records from 1918, when the king of modern pandemics hit. The Spanish flu was contagious and deadly, it killed about 2% of infected patients, a number 3–4 times worse than Covid-19. The age distribution of deaths was different, as well. Spanish flu killed many healthy people in their 20’s and 30’s, while Covid deaths are concentrated among older people.
The Spanish flu faded out after 3 waves of infections:
The first wave of the Spanish flu wasn’t the worst. The virus mutated to become deadlier before the second wave.
The virus did eventually get milder, but that wasn’t why the main reason that deaths stopped. Deaths stopped after enough people got immunity. Some milder versions of this H1N1 flu still come back from time to time, most recently in 2009.
The common flu tends to be a bit worse than a cold. The flu doesn’t become a bit milder every year, it’s more like the severity comes and goes as new strains arise.
Other than the famously awful year in 1918, we had moderately bad flu years like 1957 and 1968.
We don’t develop permanent immunity because the virus keeps changing. Part of this is just random genetic drift. Part of the problem is that flu infects other animals, like birds and pigs, and recombines into different forms that are more or less deadly.
Flu vaccines are only 50–60% effective. We don’t know which strain to anticipate, when we start making next year’s vaccine. Your body may also put up an immune response tailored to an older version, not the current one, in a process called the original antigenic sin.
We’ve found drugs that fight the flu, which make each case milder. Vaccines need to be changed out every year, but those drugs can be used for longer. The virus still evolves drug resistance in the long run, however. Amantadine was approved by the FDA in 1968. By 1980, we found some flu strains with resistance to the drug. By 2005, 96% of flu strains in China were resistant to this treatment.
We have different names for each flu variant. The Spanish flu was called H1N1. The 1968 Hong Kong flu was H3N2.
H and N are the surface proteins for flu. The H stands for Hemagglutinin and N for Neuraminidase.
There are types from H1 through H18, but only H1, H2, and H3 can infect human cells.
But there are other flu strains at risk of jumping to humans. H5N1 is a type of bird flu that occasionally infects people. It’s 50% fatal when that happens, but it doesn’t transmit from one person to another, only from birds to people.
It’s possible that at some point in the future, the virus could gain this ability and become a 50% lethal flu pandemic.
A 2012 experiment tried to prove that bird flu could make this transition by genetically modifying the H5N1 virus to have airborne transmission between ferrets. The experiment shows that such a pandemic could happen naturally. It also shows that a lab could create such a pandemic, if the genetically modified virus got out.
It’s also possible that laboratory work could cause another Spanish flu pandemic. A 2007 experiment reconstructed the Spanish flu virus using “reverse genetics, relying on tissue from victims of the early-day flu pandemic.”
In 2014, the Obama administration took notice and banned gain of function research. Dr Fauci, a proponent of the research since at least 2007, quietly restored funding in 2017. Much of the research continued in Chinese labs, and some of the funding came from US agencies like the NIH.
A few years later, Covid-19 showed up in Wuhan. Okay, let’s be clear: there’s no proof it came from a lab, just some circumstantial evidence. There’s also no proof it came from nature, since we haven’t found the original host animal or the wild source virus. We don’t yet know where it came from.
I digress. Let’s get back to the real questions. Will Covid evolve to be more mild? How long will the pandemic last? Are we close to herd immunity? Will that be the end or will Covid mutate to reinfect everyone?
People say that Covid will turn into the common cold because they’re looking for hope. They’re sick of lockdowns and masks and not enjoying the life we had before 2020. They want a mental model that says things will turn out alright. For the same reason, some people want to think that Covid is a hoax or the news is lying about the death toll. Others want to think that Hydroxychloroquine and Ivermectin are miracle drugs that cure Covid, if only the government would let us use them.
There is no rule that says that everything will turn out okay. We need to move past the simple mental models, look at data, and do experiments, to figure out what will happen.
Right now, we can see that Delta is taking over the world. It’s displacing every other known strain:
Delta outcompetes alpha, beta, and gamma. We don’t have new data from Peru yet, but we presume it outcompetes Lambda. Data from within the US shows the same situation:
Delta is taking over simply because it’s more contagious than any other strain. That’s the one and only factor. People on Twitter argue whether vaccines or anti-vaxxers are pushing covid’s evolution. Right now, neither one is. The strain that grows fastest is winning.
It’s easy to explain why delta is taking over — it spreads more easily, each patient infects more people, it grows exponentially to become the dominant strain.
It’s less obvious why delta is evolving to become deadlier.
As I said earlier, death and disability from Covid come after transmission, so there’s no strong evolutionary pressure for it to become milder.
Let’s review the lifecycle of covid, as depicted in this article in Nature:
A small region of the spike protein, called a furin site, interacts with the furin enzyme in each cell. This is complicated, better that I just quote the Nature article directly:
By snipping the bond between the S1 and S2 subunits, the furin cut loosens up virion spike proteins so that during cell entry they respond to a second cut by TMPRSS2, which exposes the hydrophobic area that rapidly buries itself in a host-cell membrane, says Gallagher. If spikes are not pre-clipped by furin — and they aren’t always — they bypass TMPRSS2, and enter through the slower endosomal pathway, if at all.
Two coronavirus variants, Alpha and Delta, have altered furin cleavage sites. In the Alpha variant, the initial proline amino acid is changed to a histidine (P681H) ; in the Delta variant, it is changed to an arginine (P681R). Both changes make the sequence less acidic, and the more basic the string of amino acids, the more effectively furin recognizes and cuts it, says Barclay. “We would hypothesize that this is the virus getting even better at transmitting.”
More furin cuts mean more spike proteins primed to enter human cells. In SARS-CoV, less than 10% of spike proteins are primed, says Menachery, whose lab group has been quantifying the primed spike proteins but is yet to publish this work. In SARS-CoV-2, that percentage rises to 50%. In the Alpha variant, it’s more than 50%. In the highly transmissible Delta variant, the group has found, greater than 75% of spikes are primed to infect a human cell.
Okay, the virus enters a cell, it replicates, a bunch of viruses come back out. With the original variant, half of the viruses that came out were ready to reinfect other cells. With Delta, 75% are.
With Delta, viral load will get higher, faster. For one generation, it’s only a 50% increase in the number of virus particles. But, it’s an exponential process; after a few cycles, the gains add up. Researchers in China found that viral load was 1,000 times higher with Delta, after a few days of infection.
The viral load with Delta becomes detectably high after 4 days, instead of 6. So, you get sick a bit faster. You can infect others sooner. And you get a bit sicker, overall. The disease is maybe twice as bad. The mutation that made it spread faster seems to also make you sicker.
Perhaps there’s another furin site variant that could spread even faster or get people a bit sicker. I would like to hope that the virus has already maxed out its abilities here. But the current furin site only activates 75% of the viruses. Maybe there’s room for that number to go a bit higher.
A team of researchers in Washington state also found that every covid variant besides Delta has also been deadlier than the original strain:
Delta has a more efficient furin site than Alpha, and it also makes you get sicker. That fits in with the theory I sketched out above. On the other hand, Beta and Gamma are also worse than Delta, and they don’t have those efficient furin sites. They have gotten deadlier for other some other reasons. It’s possible that some combination of Delta and Gamma’s mutations could make a strain that’s more lethal than either. There’s no guarantee that we’ve seen the worst strain yet.
At the beginning of the pandemic, we hoped that herd immunity would stop the virus. We tossed around numbers like R0. The original strain had an R0 of 2-3, which means that the average infected person spreads the virus to 2-3 others. Delta has an R0 between 5 and 10.
With an R0 of 3, we’d get herd immunity once 75% of people are immune.
With an R0 of 5, 80% need antibodies.
With an R0 of 10, we’d need 90%.
If R0 is 10, and vaccination is only 80% effective, then herd immunity is impossible.
R0 is not an exact number. There is no standard person or standard environment. One person might spread the virus to 100 others in a crowded room. Another might stay home and spread it to zero. One country might have more crowded rooms than another.
The best we can say is that, all else being equal, a patient with Delta will spread the virus to more people than the original strain. Herd immunity will be harder to achieve than we thought.
It’s possible to suppress the virus with some combination of immunity and behavioral restrictions. It might be impossible to suppress it with immunity alone.
Anyways, how close are we to those numbers?
The UK government produces regular reports on herd immunity:
More than 90% of adults in the UK have antibodies. That might be high enough for herd immunity. But rates aren’t that high among young people. Adults aged 16–24 are only at 70% immune, and kids under 12 aren’t eligible for vaccination. Immunity among kids could still be as low as 20%. Kids could continue to spread the virus to each other and then occasionally give breakthrough infections to vaccinated adults. Since the vaccines aren’t perfect, maybe 10–20% of those kids will infect their vaccinated parents.
We won’t really know if herd immunity works until some country tries vaccinating children. Or, until we send kids back to school without masks and just let them all get infected.
We’ll know if it works from the UK sooner than we will in the US.
There isn’t as much data available in the US, but we can still estimate where the country is at. The CDC thinks that 22% of the US has been infected with Covid. Some states are worse than others:
22% seems implausibly low to me. This would imply the virus is 0.86% fatal for the average case.
Data scientist Youyang Gu thinks that about 30% of Americans had been infected, as of March. That puts the virus at around 0.6% fatal.
50% of the US is fully vaccinated. Maybe 30% have had natural infection. Those two groups overlap, but I’m not sure how much. Overall, maybe 65–70% of the US has antibodies.
During last winter, 20% of the US caught Covid. To catch up to the UK, the US would need another wave of Covid as big as last winter. Or the country would need to convince another 20% of people to get vaccinated. And, the US will also need to vaccinate or infect children to have any hope of reaching herd immunity.
Suppose we get to a point where 90% of people have antibodies and we have herd immunity. Is the pandemic over?
No. Herd immunity would stop the current wave of the virus, but antibody levels decline over time. You can get reinfected by the 4 other human coronaviruses.
Antibody levels fade within 12 months for most people. Some people can get reinfected in as little as 6 months. We don’t all get re-exposed to the viruses each year, so we don’t all catch them every year. On average, it takes 2–3 years to catch them again.
That sounds depressing but, on the plus side, previous infection should make covid more mild.
In the short term, immunity comes from antibodies. Those antibodies should prevent immediate reinfection. Over time, there’s also a T and B cell response to the virus. Those cells give your immune system a memory of the infection. You might still get sick the next time, but you’ll make antibodies more quickly.
A study in the UK found 200 people that had been reinfected with covid. The patients had much lower viral loads than their first case, suggesting that each case is milder.
People who’ve been vaccinated sometimes get sick. But they’re half as likely to transmit the virus, also suggesting that the viral load is lower.
Two things should give us hope about the future of the pandemic. Repeat infections should be weaker. And it should be possible to revaccinate people yearly to prevent infections, just as we do with flu shots.
Can covid evolve to escape the immune response and become just as dangerous as the original?
Possibly. The best evidence here comes from an experiment that tried to create an immune-resistant strain of covid. The paper is entitled “SARS-CoV-2 escape in vitro from a highly neutralizing COVID-19 convalescent plasma”.
Scientists grew a strain of Covid in a solution with antibodies. They cultured the viruses that survived and tracked which mutations arose. Over time, 3 mutations popped up that made the virus completely evade the antibodies.
One of those mutations, called E484K, has already been seen in some covid variants. It’s in the Gamma variant but not the Delta variant.
It seems that the mutations work in two ways. The spike protein is covered with sugar molecules called Glycans which hide some of its proteins from the immune system:
The mutations work by either moving the Glycans around or changing the proteins that are exposed.
It isn’t 100% certain that the natural evolution will be the same as the accelerated laboratory evolution or how long it will take. But the experiment shows it’s highly likely we’ll have future waves of covid from immune resistant variants.
People will probably get reinfected, with or without these variants. But reinfection might happen sooner with variants and the illness from variants should be worse than reinfection by the same strain.
We can update covid vaccines for the latest strains.
That experiment shows the benefits and harms of gain of function research. Such research could predict which variants are coming, and let us prepare vaccines for them. It could also create a variant that escapes immunity and reinfects people, if that new virus somehow got out of the lab.
Conclusions:
Covid will stick around forever, like the other 4 endemic human coronaviruses. It will probably go through waves that evade the current immune response and make reinfections more likely.
One prediction market guesses that we’ll see 70,000 covid deaths every year going forward. That’s nowhere near as bad as 2020, but it’s still about twice as bad as the average flu season. Like the rabbits in Australia, life might just have gotten permanently worse for us.
Prior infection or vaccination won’t give permanent immunity. But having a prior immune response to it should hopefully make reinfections milder. And repeated vaccination should be available for those most concerned or at risk.
For an individual, the takeaway from all this is simple. You can’t just wait for covid to evolve, it’s not going to become more mild anytime soon. To be more resistant, you need antibodies. It’s your choice how you get them: you can get vaccinated or you can get sick.
