IT SEEMS that every time we think we are turning the tide in the coronavirus pandemic, another new variant emerges. The latest threat is the B.1.617.2 variant that is playing a large role in the terrible outbreak in India and is spreading in many other nations. The big question is, will existing vaccines work well enough to prevent major new outbreaks?

We already know that several vaccines are somewhat less effective at preventing symptomatic infections by new variants. For B.1.617.2, the drop in efficacy appears to be small, but even a small drop matters when most people are only partially vaccinated or unvaccinated, says Deepti Gurdasani at Queen Mary University of London. “Any degree of escape at this point in time is concerning,” she says.

A drop in efficacy not only means vaccinated people have a higher risk of being infected, it also makes it harder to reach the herd immunity threshold – beyond which the virus cannot spread widely – via vaccination. What’s more, variants that are more transmissible raise this threshold, making it even harder to reach. There is growing evidence that B.1.617.2 is more transmissible than the B.1.1.7 variant first identified in the UK.

“88% Effectiveness of Pfizer/BioNTech vaccine against Indian variant”

On the plus side, existing vaccines still appear to provide substantial protection against serious illness or death for all variants. “All these vaccines tend to be able to limit severe infection and hospitalisation against those different variants,” says Jamie Triccas at the University of Sydney.

But there is still a risk. Ravi Gupta at the University of Cambridge says he has heard many reports from India of people dying despite being vaccinated, though mainly after having had just one dose. Control measures must therefore be maintained to keep infections down, says Gupta. “To open up with a partially vaccinated population is worrisome.”

In fact, says Gurdasani, modelling studies suggest that a more transmissible variant with some ability to evade vaccines could cause a bigger wave of hospitalisations and deaths in the UK than the one in January.

Establishing how well vaccines work against particular variants can be hard. Where new variants became dominant in countries as trials were carried out, we do have good data (see table, below).

For instance, a small trial of the Oxford/AstraZeneca vaccine was under way in South Africa as the B.1.351 variant evolved and spread in that country. In February, South Africa halted the vaccine’s roll-out after the trial suggested it doesn’t prevent most mild or moderate illnesses caused by B.1.351.

“60% Effectiveness of Oxford/AstraZeneca vaccine against Indian variant”

Similarly, trials showed that the Novavax vaccine is about 96 per cent effective at preventing symptomatic infections caused by older variants in the UK, about 86 per effective against the B.1.1.7 variant and about 51 per cent effective in South Africa, where B.1.351 was causing almost all cases.

Once a vaccine has been rolled out, its efficacy can be estimated in the “real world” by monitoring the vaccination status of people who test positive for a certain variant, or by looking at the proportion of cases with the variant relative to the main circulating virus and vaccination status.

On 22 May, Public Health England (PHE) published this information for B.1.617.2. It found small, “non-significant” drops in efficacy against symptomatic infections for people who were fully vaccinated.

For the Pfizer/BioNTech vaccine, the study found 93 per cent efficacy against B.1.1.7 and 88 per cent against B.1.617.2 after both doses had been given. For the AstraZeneca vaccine, it was 66 per cent against B.1.1.7 and 60 per cent against B.1.617.2, after both doses.

But there was a bigger fall after just one dose. For both vaccines, one-dose efficacy was just 34 per cent against B.1.617.2, compared with 51 per cent against B.1.1.7.

A better way

These post roll-out studies can only be done when a variant is already widespread. Ideally, we would want to know sooner if new variants can escape vaccines.

One way to do this is to carry out neutralisation assays. These involve taking antibodies from vaccinated people, mixing increasing amounts with the virus and pouring it on cells to see what antibody level prevents infection, or “neutralises” the virus.

Nathaniel Landau’s team at New York University recently showed that antibodies from people who had received the Pfizer or Moderna vaccines are two to threefold less effective at neutralising the B.1.617.2 variant. That is a relatively small decrease, says Landau. “Our prediction is that [these] vaccines are going to remain protective, certainly for the vast majority of people,” he told New Scientist last week, before the PHE study came out.

“100% Effectiveness of most vaccines against death from older variants”

But neutralisation studies don’t tell us exactly how protective a vaccine will be. This is what Triccas and his colleagues are trying to address. In a paper published last week, they analysed data from several studies and identified a strong correlation between the level of neutralising antibodies that vaccines elicit and the amount of protection those vaccines provide against symptomatic infections.

This could give us a way to work out the efficacy of booster shots and new vaccines, and of existing vaccines against new variants, without carrying out expensive and time-consuming trials.

However, the model that Triccas and his colleagues have developed based on their findings has some worrying implications. For starters, we know that the level of neutralising antibodies wanes over time, which suggests that the efficacy of vaccines will wane too.

According to the model, the less effective a vaccine is, the faster its efficacy will wane. For example, a vaccine with an efficacy of 95 per cent would fall to 77 per cent after 250 days, but one with an initial efficacy of 70 per cent would drop to 33 per cent over the same time.

A similar effect would be seen with antibody-evading variants, the model predicts. So if a vaccine were, say, 95 per cent effective, a fivefold drop in the effectiveness of neutralising antibodies would reduce that to 77 per cent. A 70 per cent effective vaccine, however, would decline to 32 per cent.

This could be why there was such a big fall in the AstraZeneca vaccine’s effectiveness against B.1.351 in the South African trial. The PHE study shows signs of this too: “The reduction in vaccine effectiveness appeared to be greater with [AstraZeneca],” it says.

However, Landau isn’t convinced Triccas’s model is correct. The level of neutralising antibodies reflects how many B-cells you have churning out antibodies, says Landau. These factories stop producing antibodies over time, but they don’t necessarily go away, meaning antibodies might be ramped up again very quickly.

Even if Triccas’s team is right, decreasing efficacy against symptomatic infections doesn’t necessarily mean people will get severely ill. Even a low level of neutralising antibodies can still provide protection, says Landau. “I think people will maintain antibodies for some time that will stop them getting very sick.”

And in addition to antibodies, so-called T-cells also help protect us against severe disease. It is harder for viruses to escape the T-cell response than it is for them to evade antibodies.

“Our prediction is that these vaccines will remain protective against variants for the vast majority”

But if the efficacy of certain vaccines rapidly wanes or is much lower against variants, it is going to be even harder than we thought to halt the virus’s spread. We don’t know what proportion of people must be immune to reach the herd immunity threshold, yet estimates range from 70 to 90 per cent.

Achieving this threshold requires very effective vaccines and very high vaccine uptake, especially if children aren’t eligible, says Gurdasani. “It may not even be possible to achieve herd immunity with these new variants,” she says.