We can't go on like this Photo: JOSH EDELSON/AFP via Getty Images

July 21, 2020   7 mins

One of the hardest things, for me at least, about the whole Covid-19 period has been coming to terms with the timescale. In March, back when it was all unfolding, it was frankly weird to think that this — the strange new rules on who you can see and when you can leave your house and what life is like now — was going to be how we lived for months. Even four months later, the idea that it’s going to carry on for many more months or a year is kind of difficult to grasp.

But the big hope has always been the vaccine. At some point, we’ll develop a vaccine that works, we’ll build factories that can churn out millions of doses a day, and then we’ll all be set free. That’s the way out. If you’ve been anything like me, you’ve had that as a sort of promised land. I kind of imagine a great hand emerging from a cloud to deliver it, like God in a Terry Gilliam animation, with a plainchant choir singing in the background.

That hope is important, so you may have been alarmed if you read something in the last few days that seemed to say the hope was false: that herd immunity is impossible, and that vaccines won’t work, because immunity to Covid-19 only lasts a few months. “With coronavirus antibodies fading, vaccine hopes fade too,” said one particularly doom-laden headline. “Immunity to the coronavirus may only last a few months,” says another.

This comes off the back of a preprint from scientists at King’s College London, which looked at 64 patients who had had coronavirus (and 30 uninfected controls), and tracked their level of antibodies in the months after infection. It did, indeed, find that in a lot of patients — the ones who’d had a milder case of the disease — the number of antibodies in the blood decline quite quickly.

What that doesn’t mean, necessarily, is that there is no long-term immunity, or that vaccines are pointless, or that our only hope is for treatments. Let me try to explain why.

First, it absolutely is the case that with many coronaviruses, reinfection is possible. (Although we call Covid-19 “the coronavirus”, there are many others. There are four different seasonal coronaviruses that cause the common cold; SARS and MERS, much more severe diseases, were both coronaviruses too.) 

With the seasonal cold coronaviruses, says Rupert Beale, group leader at the Crick Institute’s Cell Biology of Infection laboratory, about 90% of people have had at least three of them by adulthood, and will, on average, be reinfected with each of them about once every five years.

Babak Javid, a professor of immunology at the University of California San Francisco, agrees. But both of them make an important point: just because you can be reinfected, doesn’t mean that the reinfection is as bad as the original. Javid says that “the only definitive data we have with immunity and coronaviruses” comes from studies from a few decades ago, so-called “human challenge” studies, in which people were deliberately given the common cold and then their immune responses were tracked. 

Crucially, they found that if patients had detectable levels of antibodies before they were given the virus, they were immune. But, as you’d expect, people who didn’t have the antibodies got a cold — and then developed antibodies. The studies found, as with the current coronavirus, that the number of antibodies in the bloodstream then tailed off rapidly. 

A year later, the scientists tried infecting them again. They “were virologically affected”, says Javid – that is, if you swabbed them and tested for a virus, you would find it — but “they had no symptoms whatsoever, even in people with no antibody response”. The period in which they were themselves infectious appears to have been much shorter, as well.

Part of what’s going on here is that antibodies are only part of your body’s immune response to infection. Antibodies are proteins that latch onto infectious agents like viruses and bacteria and destroy them; they are produced by cells in your blood called B-cells. The precise shape of an antibody determines whether it will be able to destroy an invader. Very roughly, if an antibody turns out to be effective at killing some virus in your bloodstream, your B-cells will start producing loads more of that antibody to kill the virus. Then those antibodies will hang around in your blood, and your B-cells will “remember” how to make them, so next time the virus turns up, your body will be ready to fight them off.

But there’s more going on. There are also T-cells. Again, very roughly (and with apologies to Dr Beale if I’ve misunderstood his explanation), there are two kinds of T-cell. One helps B-cells produce antibodies. The other, “killer” T-cells, detect the presence of (usually) virus in infected cells of your body, and then kills those cells. They’re important in this story.

“Broadly speaking,” says Beale, “antibodies prevent infection; killer T-cells help you get over disease.” Just like antibodies, killer T-cells have disease-specific responses – if you have one of the common cold coronaviruses, your body will have a T-cell response to that coronavirus.

That is, probably, part of why — even in people who have no detectable antibodies — reinfections are much less severe than first infections. You may not have enough antibodies to prevent you getting the disease in the first place, but you do have the T-cell response to fight it off quickly and easily. It is very possible, says Beale, that that is what will happen with Covid-19. “It’s behaving exactly as you’d expect for a respiratory coronavirus,” he says. Javid points out that in both SARS and MERS there is a very strong T-cell response “even five or 10 years later”.

There’s another point, which is that the tests used to detect antibodies are (deliberately) not very sensitive. The tests are designed to avoid false positives, because telling someone that they’ve had the disease and are immune when they’re not is much more dangerous than telling them that they’re not immune when they are. So the threshold for a “positive” response is quite high.

The necessary tradeoff is that there are more false negatives. By analogy: say you’re trying to decide whether someone is a Terry Pratchett fan. You set up a very simple test: you ask them “how many Terry Pratchett books have you read?” Then you set a minimum threshold.

If you set the threshold very low, say one or two, you’ll accurately identify all the fans, but you will also let in a lot of non-fans who just happen to have read Good Omens. Whereas if you set it very high, say 15, you’ll accurately weed out all the non-fans, but you might miss out on loads of real fans who are still working their way through the Discworld series for the first time. This is a zero-sum game: setting your threshold higher means more false negatives; setting it lower means more false positives. It’s unavoidable.

The same goes on with antibody tests. If you want to be more sensitive (get fewer false negatives), you can lower your threshold, but that means you’ll have more false positives; if you want to be more specific (fewer false positives), you raise your threshold, but that means you’ll have more false negatives. Unless you develop a new and better test, there is no third option, The tests have usually gone for specificity, rather than sensitivity.

That’s probably wise, but it presents a problem in that even very low levels of antibodies may offer protection. “In some monkey studies where they re-challenged,” says Javid — meaning giving the virus to monkeys who’d already had it — “very very low levels of antibodies were protective.” A lot of the tests use cut-offs that are higher, he says, “than what animal studies say are protective”. 

All that said, it probably is the case that people who’ve had milder cases will have antibody responses that wane quite quickly, in months rather than years, according to both Javid and Beale.

It’s important to say that this doesn’t mean that vaccines won’t work. For one thing, says Javid, vaccine immune responses are “both qualitatively and quantitatively different” from immune responses to natural infection, when it comes to both antibodies and T-cells. That’s not a good thing or a bad thing, necessarily; it could be either, and “there’s no way to tell a priori”. But it does mean that it’s perfectly possible that natural infection could cause this tailed-off antibody response, while a vaccine has a more long-lasting one; or it could be the other way around. Or they could be largely the same. We’ll just have to wait and see, although the early noises from Oxford and elsewhere seem largely optimistic.

What it probably does mean, according to Beale, is that the most likely outcome is that Covid-19 becomes “the fifth seasonal coronavirus”. Assuming that a mild case of the disease only sparks a relatively short-term immunity, but that reinfection with the virus is far less dangerous because of the T-cell response (or whatever else is going on), then once everyone’s immune system is prepared for it, as with the common cold coronaviruses, we could see it going around as an inconvenient but not devastating seasonal illness. “If you’re vaccinated or infected, then it’ll be much less severe,” he predicts. “I’ll be genuinely surprised if that’s not where we end up.”

(He notes, however, that he has been genuinely surprised once already by Covid-19, specifically by the effectiveness of the drug dexamethasone in treating it. It’s still a novel disease and we shouldn’t get complacent that we know everything about it.)

Javid agrees. “This was predictable,” he says. “We kind of knew this was going to be the case months ago, it’s just that now we’ve had long enough to check.”

Even if we don’t ever eradicate the disease, we will probably end up in a place where we can live with it. In the meantime, though, it’s still absolutely crucial that we control the disease. Beale in particular says that “policy-wise, it’d be completely insane to let it get out of control now,” because “by this time next year” it’s likely that we will have an effective and widely available vaccine. “Better to lock down even at considerable economic cost and wait for that,” he says, although he acknowledges the calculus changes if the vaccine will be much longer coming. 

Javid adds that “As with everything Covid, it’s going to come down to how well we protect our vulnerable population,” which in the UK at least isn’t something we’ve conspicuously excelled at.

So there’s no need to panic; it seems to be unfolding as expected. I don’t want to be too harsh on anyone for the perhaps over-alarmist reporting; science writing in a fast-moving pandemic is bloody hard (I’ve discovered). But, so far, at least, I don’t think we need to start worrying that vaccines won’t work or that we’ll all be getting a life-threatening course of Covid-19 every calendar year. The hope that we’ll get out of this in a reasonable timescale, and life will return to something relatively normal before we’re all too old to enjoy it, is still very much there.


Tom Chivers is a science writer. His second book, How to Read Numbers, is out now.