And the existing vaccines were designed to mimic the original strain of Covid. We’ve been broadly lucky so far – the vaccines all still work against the variants we’re aware of. But they work less well against the new Delta variant; a Lancet paper last month found that people who had had the Pfizer vaccine had a reduced immune response to the new variant, and a recent Public Health England report says that a single dose of either Pfizer or AstraZeneca only reduces the likelihood of infection by about 33%. 

I don’t want to be too downhearted. It is worth remembering that even though we don’t have data from this vaccine, on the whole the immune system has an excellent memory, and vaccine protection usually lasts for years. And a double dose of either vaccine is much more effective, even against Delta: PHE says 60% in the AZ case, 88% with Pfizer. And they are much more protective than those headline figures against severe disease and death: another PHE paper finds that people who have had both jabs, of either vaccine, are more than 90% less likely to end up in hospital.

But Delta is a warning shot. We might well see a new variant – I was going to say Epsilon, but apparently we’re already up to Kappa – which escapes the vaccines even more effectively.

That’s why we need to start talking about third jabs.

If we want to maintain the freedoms we currently have, and certainly if we ever want to get rid of the disease, we are going to have to have a third innoculation. “The virus is very likely to remain endemic, for the next year at least,” says Dr Rupert Beale of the Crick Institute. Even with everyone double-jabbed, Delta is transmissible enough and vaccine-resistant enough to keep circulating. 

The latest SPI-M-O modelling suggests that it has an R0 of seven. (R0 is the number of people the average infected person spreads the disease to, in a population that has no immunity and no countermeasures against the disease.) If that’s true, then somewhere around 85% of people would need to be fully immune to reach herd immunity. With two jabs in every arm, we still wouldn’t reach that. In that situation it may not kill all that many people, but it will not go away, either. 

More from this author

Who isn't getting vaccinated?

By Tom Chivers

And it may still kill quite a few. Vaccinated people are protected from severe disease, but not perfectly: another Lancet paper suggests that the risk of hospitalisation, if you get the disease, is about a quarter what it would be if you weren’t. SPI-M-O suggests that an uncontrolled third wave could cause hundreds of deaths a day by autumn, even though almost everyone would be double-jabbed, hence Monday’s announcement of the delay to the roadmap.

So we need to talk about what the strategy should be.

So for starters, should we wait for variant-specific vaccine updates? The pharma companies aren’t idiots: they’ve already started creating vaccines for the Alpha to Delta variants. The mRNA vaccines, such as Pfizer and Moderna, are especially quick to turn around, since you can code a new one almost like computer software. 

Amin Khan, head of vaccines at the biotech firm GreenLight[1. Full disclosure: I’ve done some paid work with GreenLight during the pandemic], says that you can get a new variant-specific mRNA vaccine ready to go in a few weeks. And if the new version simply targets a slightly modified version of the spike protein, as the existing vaccines do, it won’t need much in the way of testing and regulatory approval. Changing your manufacturing system is more complicated, “but within two or three months, you can get a new variant to the market”.

Playing whack-a-mole with new variants isn’t a long-term solution, though. The hope is that “third-generation” vaccines will be capable of covering all the existing variants and most foreseeable future ones. But, says Khan, that’s a bit more complicated. A more complete version might target other parts of the virus than the spike protein; that would mean a much more rigorous testing and approval regime, and it may take months longer to get such a vaccine to market. 

We needn’t necessarily wait for a third-generation vaccine to start giving boosters, though. There is good reason to think that a third jab of an existing vaccine will provide a boost to protection, and there’s an ongoing trial at the University of Southampton to see whether that’s true — but it is true for other vaccines, so it’s probably true for Covid.

Assuming that it does offer greater protection, the next question is whether the booster jab should be the same vaccine you’ve already had, or a new one. Again, there are theoretical reasons to think that a different (“heterologous”) booster from the one you’ve already had might provide greater protection. But “there’s no substitute for large amounts of data”, says Beale. 

Suggested reading

Why we should welcome the lab leak hypothesis

By Bret Weinstein

At the moment there isn’t much data. A Lancet paper came out recently which found that a heterologous boost had greater “reactogenicity” – that is, it caused more of the mild, expected side effects (feverishness, headache, joint pains) that tell you a vaccine is probably working – than a same-vaccine jab, while also finding no severe adverse reactions. So that’s promising, as are a few other studies. 

And the Southampton trial will be assigning people one of seven different vaccines at random, so it should naturally show if there’s a significant difference between using the same jab and using a different one. That said, it has a relatively small sample size (about 3,000 people, divided between seven Covid vaccines and a control), so it may not be powerful enough to detect small differences. 

One thing that might be worth trying is letting people say that they’re happy to take a heterologous third dose at their own risk; I (for instance) would be quite happy to take a Pfizer or Moderna jab on top of my two AZ doses, and accept the risk of any negative consequences, because I’m extremely confident that there wouldn’t be any. And if a few hundred thousand people did that, then we would soon get some good data on the effectiveness of the strategy.

Suggested reading

France is infected with anti-vaxxers

By John Lichfield

Most likely, it will turn out that it’s best to do heterologous boosting. Not only might it be more effective, but it reduces the logistical difficulties of finding the right dose for the right person. A really crucial thing should be that we drop the restrictions on using the AstraZeneca vaccine in the under-40s; if we’re going to accelerate second jabs, let alone consider doing thirds, we’re going to need it.

The third question is who we should be giving it to. There is, as I understand it, good evidence that, insofar as vaccine-induced immunity gets weaker over time, it gets weaker fastest in the old; and since the old are the most vulnerable, it makes sense to boost from the eldest to the youngest again. I might add one detail, though, which is that since there is a clear difference in protection against Delta — at least according to the PHE data — between the Pfizer and the AstraZeneca vaccines, it may be worth preferentially giving boosters to people who had Ox/AZ. 

And the last question is: when? 

This is complicated; we need data. But assuming there is a real boost in protection from a third jab in people whose first jab was six months ago or more, then we probably need to start thinking about giving those elderly and vulnerable people the boost in the next couple of months. Professor Saul Faust, who heads the Southampton trial, says that the JCVI will make decisions in August or September, by which time the first vaccines will have been in people’s arms for nine or 10 months. 

If the UK hits its target of giving every adult a first dose by the end of July, then every adult should have had a second dose by the end of September or early October. Whether we ought to start giving third doses to the elderly before we give second doses to younger groups – or even whether we vaccinate children beforehand – is a question that can only be answered with a cost-benefit analysis, looking at how many lives (or rather years of life) we can expect to save with each. Until some actual data is in from the Southampton trial and others, we can’t reasonably do that analysis.

There’s another question, of course. Can we reasonably justify giving third doses to British citizens when billions of people around the world haven’t had a first dose? From a global perspective, the answer is obvious: a life in Malawi or Sri Lanka is worth no less than a life in the UK and should be given equal value. I instinctively want to say that we should throw huge amounts of money and all our spare vaccines into GAVI and Covax.

But from a UK government perspective, it’s not so clear. “The first duty of the Government is to afford protection to its citizens,” goes the axiom. If you hire a babysitter, you are paying that babysitter to look after your child, rather than every child in the world.

This doesn’t extend to a complete disregard for lives in other nations. If a babysitter saw a child drowning across the street, I think most of us would think it was justified for them to rush out and save it even if it meant leaving your child unattended for a few minutes; at some point, a large chance to save a life in the developing world outweighs a small risk to a life in your own country. But it is at least not obvious that, from a national government’s point of view, the right thing to do is share vaccines equitably. 

Most likely, I’m still well protected against Covid, including the Delta variant. But whether that will still be true by the time the Lambda, Upsilon, Pi and Omega variants arrive is not clear; each new variant that arrives is a new chance for it to escape the vaccines more effectively. 

This is still uncharted water. We need to chart them, as quickly as possible, and to get third jabs into people who need them, so that we can keep on opening up and not slide back into the dark days. I’ve put my name down for the Southampton trial.