We have some more data on the vaccine front that’s worth looking at. J&J has published a bit more on their trials of their adenovirus-vector candidate, with data on the immunogenicity of the vaccine in patients. They have several cohorts evaluated: a single shot of low dose, single shot of high dose, two shots of low dose and two shots of high dose. This article at STAT sums it up: it’s clear that the vaccine generates neutralizing antibodies at both low and high doses, that the higher doses do indeed tend to give higher antibody levels, that the two-dose regime leads to higher antibody titers overall, and that these titer rise over time (present at day 29 after injection and actually higher at day 57).
So far, so good. But what we still don’t know are the correlates of protection: which of these antibody levels are enough to keep a person from getting the coronavirus and to what extent? The same goes, of course, for the T-cell data. The paper shows levels of both CD4+ and CD8+ T cells after vaccination, but how these relate to protection is an open question. And for all of these, such numbers might be different in different patient groups (particularly older versus younger patients). Fortunately, we should soon (within two weeks, I’d say) be getting actual efficacy data from the single-dose trial, so I’m not going to spend much energy speculating.
There’s another thing to watch: the New York Times reported yesterday that J&J has told the US government that they might have fallen as much as two months behind on their production schedule for the vaccine. No word as to what the problem might be. Adenovirus vectors have vulnerabilities of their own, with a whole set of manufacturing steps that are distinct from the other types. What you gain in not having everything reliant on a single pathway, you can lose when unique problems crop up. But overall it’s still far better to have a diverse set of things under development than not.
Now for another vaccine candidate, the inactivated virus one from SinoVac. I discussed that one here after an earlier publication, and when last heard from there was an odd report from Turkey of 90% efficacy in a small trial with no data to back it up. Then just a few days ago there was a figure of 78% efficacy from the trial in Brazil, but that has now been revised to 50%. Meanwhile, Indonesia says that a preliminary analysis shows about 65% efficacy – the government there has approved the vaccine, but the rollout is apparently not going well. To be honest, these are about the efficacy numbers one could expect from an inactivated-virus vaccine. It’s an old technology, and it doesn’t always work that well. The chaotic release of these data (and the lack of comment from SinoVac themselves about any of these numbers) does not build confidence, with the consequences that you can see in that last link.
But that takes us back around to what the true efficacy of another vaccine is: the Oxford/AZ one. The statistics on that are a mess too, and their public presentation has been a mess, and we can only hope that things come into better focus when their US trial reports (other data are coming as well). But at the moment, I wouldn’t be prepared to put money on how much better it is than the SinoVac one. To be sure, a 50% or 60% effective vaccine (with a reasonable safety profile) is surely better than no vaccine at all. But it means that you have to vaccinate a lot more people before you start to make a dent in a pandemic like this one, which is a significant real-world limitation. Of course, here in the US we’re not doing that great a job in rolling out the 95% effective ones, either, if we want to talk real world effects!
Update: let’s compare these numbers with the immunity that a person can get from just being infected with the coronavirus itself. That’s been a matter for speculation, but we now have some numbers from the NHS in the UK. In their SIREN study they’ve been looking at a large cohort of health care workers, monitoring them for infections and antibodies. Out of over six thousand who have had the virus, there have been about 42 re-infections. Comparing that to the cohort of people who were never infected, that comes out to 83% efficacy. So there’s your comparison number – which means that being vaccinated with either of the mRNA agents provides better protection than being infected with the real coronavirus can.