As of this morning, we have a first look at the Oxford/AstraZeneca vaccine’s efficacy in clinical trials via press releases from both organizations. The number in the headlines says about 70% efficacy, but there’s more to the story.

Here’s the landscape so far: we have results from Pfizer and from Moderna, both of them developing mRNA-based coronavirus vaccines, and both showing efficacy in the 90 to 95% range. The Oxford effort is a different platform, though, with key similarities and key differences. It relies on another virus (a chimpanzee-derived adenovirus) that has had its original DNA genetic payload removed and substituted with the appropriate DNA to produce the full-length Spike protein of the coronavirus. In this construct, the original viral “leader sequence” at the beginning of this DNA has been replaced with another, the leader sequence found for the human tissue plasminogen activase (TPA) protein, because this gave better expression and a better immune response. These adenovirus particles can’t replicate – they don’t have the DNA to express the proteins needed to do that. But they do have all the viral machinery needed to infect a human patient’s cells and force them to express the coronavirus Spike protein, which will set off an immune response that should provide protection against later exposure to the real coronavirus.

So both the adenovirus vector and the mRNA vaccines hijack the protein expression capabilities of a vaccinated person’s own cells, making them produce SARS-Cov-2 Spike protein constructs and thus setting off the immune system. The Pfizer and Moderna mRNA vaccines we’ve seen so far actually express a form of the Spike protein that has a couple of proline residues mutated to make it more stable, whereas the Oxford/AZ vaccine is using the straight wild-type Spike sequence – there’s one difference. Another big one is of course that the Oxford/AZ vaccine is using a completely difference virus to deliver a DNA sequence, whereas the mRNA vaccines are skipping up to a later stage in protein production and slipping messenger RNA directly into the cells.

What was announced today is that they have quite different results for two different dosing regimens. This interim analysis was run when 131 cases had been accrued across trials in the UK, Brazil, and South Africa across about 24,000 trial participants (treatment and control groups). In the treatment group, 8,895 participants received two full doses of the vaccine, spaced one month apart, and 2,741 patients got a half dose at first, followed by a full dose a month later. And the efficacy rates for these two dosing regimes were very different: 62% for the two-full-dose group and 90% for the half/full group. I do not see a breakdown of how those 131 cases partitioned across the two groups, but the overall N has to be higher for the first, doesn’t it? I’d like to know what the statistics are for the 90% efficacy number, for sure.

Update: many people had been wondering earlier today why there was a lower-dose initial group at all. Reuters reports that it was a dosing mistake in the UK, that the investigators were then stuck with. Immunology strikes again!

Why might there be such a significant split in efficacy? My own wild guess is that perhaps the two-full-dose protocol raised too many antibodies to the adenovirus vector itself, and made the second dose less effective. This has always been a concern with the viral-vector idea. It is, in fact, why this effort is using a chimpanzee adenovirus – because humans haven’t been exposed to it yet. Earlier work in this field kicked off with more common human-infective adenoviruses (particularly Ad5), but there are significant numbers of people in most global populations who have already had that viral infection and have immune memory for it. Dosing people with an Ad5 vector would then run into patients whose immune systems slap down the vaccine before it has a chance to work. That’s not the case for a chimpanzee-infecting form, naturally (few if any people have ever been exposed to that one!) but the two-dose regime may have run into just that problem. Immunology being what it is, though, there are surely other explanations, but that’s the one that occurs to me. Update: there’s always the outside nasty chance that the smaller N in the 90% group is giving a number that won’t hold up. I would hope this isn’t the case, but without a better look at the statistics, it’s not possible to rule that out.

Now, I’ve seen people speculating this morning that these numbers may be better than they look, because they believe that these trials monitored patients by PCR tests rather than by symptoms. If that were the case, then yes, that’s a finer net than the Pfizer and Moderna trials used and it would certainly affect the efficacy readouts. But I don’t think it is: looking at the published trial protocol for the US trial, the cases are defined as “SARS-CoV-2 RT-PCR-positive symptomatic illness”, and the patients have to show symptoms of the disease (see Table 13). Update: I have been unable to find published protocols for the UK/Brazil/South Africa trials that went into today’s numbers, but I have no reason to think that they differ on this point. So I don’t think we can explain the lower efficacy by saying that they were finding asymptomatic people as well: the trial excludes asymptomatic people from its endpoint definition. The rate of asymptomatic cases in the treatments and controls will be determined in these trials (see section of the protocol) but those aren’t the numbers we’re seeing today.

So from an efficacy standpoint, the choice is clear: if this vaccine is going to be deployed, the half-dose/full-dose regime is the obvious choice, since otherwise you can do the same amount of work dosing your population, use up more vaccine. doing it, and get notably worse results. How about from the safety side of things? The Oxford release says just that “No serious safety events related to the vaccine have been identified”, and the AZ one says “No serious safety events related to the vaccine have been confirmed”. I would have preferred to hear more about local and systemic reactions, as we did with the Pfizer and Moderna releases, but that seems to be it. Readers will recall that a participant in the UK trial developed transverse myelitis, and that the trial was stopped in the US for about a month. (Note: the US trial is the two-full-dose version). Update: they say now that they’re going to ask to switch to the apparently more effective dosing regime in the US trial.

Overall, I would have to think that Oxford and AZ are disappointed with the results from the two-full-dose regime and will be actively trying to track down the reason for the better performance in the the half/full dosing, which one would expect to be the way the vaccine is eventually used. How many of the other trials that are being run are using that protocol, one wonders? This could still be an effective weapon in the pandemic, but the stories are starting to differentiate. Pfizer (very effective, tough distribution and storage), Moderna (very effective, easier distribution/storage than Pfizer, but perhaps stronger safety reactions), and now Oxford/AZ (widely varying efficacy depending on dosing, easier distribution/storage, safety details TBD). The next vaccine effort to report efficacy will be J&J, another adenovirus vector, and this time with a one-shot dose. The landscape is starting to fill in a bit!

Source link