Suddenly we have a lot more vaccine efficacy data to discuss! Yesterday came an announcement from Novavax about trials they’ve been conducting in the UK and in South Africa with their recombinant protein vaccine, and today comes equally anticipated data from J&J (Janssen) on their adenovirus vector candidate. We’ll have a look at the numbers in that order.
The full Novavax package should be coming out in a few days. But what we know now is that in the UK trial (roughly 15,000 people), the vaccine showed 89% efficacy. And the company says that around half of these participants were infected with the B.1.1.7 variant, so these are excellent results against that one and against “coronavirus classic”. The South African trial was smaller (around 4,000 patients), and the data have correspondingly more scatter in them. But the readout is clear: the vaccine is less effective against the B.1.351 variant, perhaps around 50%. Almost all the cases in this trial were from that one. Novavax says that they have already begun working on another protein vaccine to be used as a booster for those who’ve received another one, and that’s good news. Not least because Novavax also noted that they were seeing people infected with B.1.351 who had already been infected with the earlier strains. There appear to have been no significant safety signals during the trials (an initial report of volunteers being hospitalized appears to have been mistaken). Keep in mind that there’s another trial of the Novavax candidate still underway here in the US as well – the question is now how many countries will authorize the vaccine based on the data we have in hand and when this vaccine can be rolled ou
Now to J&J – then we’ll discuss some implications. Their data for the one-shot dosing protocol show 72% efficacy in the North American volunteers, 66% in Latin America, and 57% in South Africa. That last one has to be another B.1.351 effect. There were also no safety signals – adverse events were actually more frequent in the placebo group. There were also no real differences across different age groups, which is very good to see.
Now, these numbers from J&J are definitely lower than the Moderna and Pfizer/BioNTech vaccines. I wish that they were higher. But this is the landscape we have, and given the development timelines, I think we’d better get used to this it, because now the biggest and most advanced players have reported in. Broadly speaking, it looks like the mRNA vaccines have definitively higher efficacy, but we have to remember that their trials were run before the new variants got much of a foothold. We have in vitro data about how they might respond to B.1.1.7 and B.1.351, but so far we have no real-world efficacy reads. For all we know, the Moderna and Pfizer/BioNTech vaccines also come down to 50-60% efficacy against B.1.351.
But at any rate, with a 95% value (or the 96% seen with Novavax against “coronavirus classic”), you can have a good deal of confidence that you’re going to have solid real-world effects. When we see numbers like J&J’s, or (especially) AstraZeneca’s (the EMA says this morning that overall efficacy for that one is 60%), or probably everyone’s against B.1.351, we have to look more closely at the clinical outcomes. You may not be able to prevent so many people from going PCR-positive, but can you keep them out of the hospital? Out of the morgue?
J&J is saying that overall, their vaccine was overall at 66% efficacy at preventing
all “moderate to severe”coronavirus infection (updated this), but 85% effective at preventing “severe disease” requiring hospitalization. They had a conference call this morning, and my impression is that this statement wasn’t clarified very much (we’ll hear more next week, I believe). But they also say that there were no coronavirus deaths at all in the vaccinated group. So the real-world effects are nothing to turn up one’s nose at – if this had been the first vaccine to report, we’d have been setting off fireworks. Similarly, it looks like the AstraZeneca/Oxford vaccine, for all the problems in figuring out its dosing and its efficacy numbers, also does a good job of preventing severe disease. We don’t have data on severity of disease from the Novavax team yet; I hope that we see it very soon.
Given the advent of the new variants – and the threat of seeing more of the same – all of these vaccines are going to have a part in beating back the pandemic. The AZ/Oxford, J&J, and Novavax vaccines do not have the ultra-cold-chain distribution problems that the mRNA vaccines have, and thus can have greater penetration into areas with less infrastructure. We also have to remember that these vaccines all have different manufacturing processes at many of their stages, which means that ramping them up doesn’t turn into a zero-sum cage match.
There are two more factors to consider. I keep coming back to the idea of antibody maturation via clonal selection of B cells, and I note that J&J says that the protection afforded by their vaccine did increase over time. That’s working in our favor with all the vaccines, but we have no idea (yet) if one of these might do a better job with this process than others. As far as I know, the only way to know is to keep collecting data to see if such differences emerge. The other thing to keep in mind is (as mentioned above) the possibility of retooling these agents to better deal with the new variants. My impression is that this would be easiest for the mRNA and recombinant protein vaccines (that is, for Moderna, Pfizer/BioNTech, and Pfizer) and harder for the viral-vector ones (J&J and AZ/Oxford). But we’ll need to clear up the regulatory framework on this – you’d think it should be much more like the clearance of (say) the yearly flu vaccine and not like coming in with a completely new vaccine.
It looks like we’re going to need these updates. The B.1.351 variant is clearly pushing the efficacy numbers down for the first crop of vaccines, and there is no reason whatsoever to think that the process will end there. We’ve got to stay ahead of the damn viruses however we can.