AstraZeneca’s very strange data release around their COVID-19 vaccine.

AstraZeneca and its partner Oxford University released the final, phase III clinical data on their coronavirus vaccine this morning. At first, I was confused since some of the press releases state that it was 90% effective, others stated it was 60% effective, and yet other press releases came up with an average of 70% effective. These were supposed to be the results from a single clinical trial! Remarkably, this was only the start of the confusion on this vaccine’s data. 

The AstraZeneca vaccine is much more complicated than either the Pfizer or Moderna vaccines. The AZ/Oxford vaccine uses a totally different whole virus (chimpanzee adenovirus). They make the virus in a way that does not produce its own genetic material, so it cannot reproduce. Instead, they substituted the DNA for the SARS-Cov-2 (COVID-19 virus) spike protein. These chimpanzee virus particles still can infect human cells and then force them to make the COVID-19 spike protein. This should induce your body to make an immune response to the spike protein that they believe would provide immune protection to the real coronavirus.

This is much more complex than the Pfizer and Moderna vaccines, which are purified messenger RNA. When injected the vaccine enters the individual directly and produces only the coronavirus spike protein in the recipient’s own cells.

When the AstraZeneca vaccine was given in the intended two doses, it was 60% effective in protecting against COVID-19 infection. This means it reduced the incidence of coronavirus infection by 60% in people who received the vaccine. So where do we get the 90% effective number? It turns out that some of the participants mistakenly got a first dose (of two) of vaccine that was half of what it should have been! In these subjects, the vaccine was remarkably more effective (90%) in preventing COVID-19. The 70% number was an average of the two groups.

Aside from the fact that accidents will happen (kudos to Elvis Costello), judging the outcome of a vaccine on a mistake in the clinical trial does not seem to be a good idea. So, let’s toss out the 90% effectiveness claim in the small number (approximately 2,000) of subjects who got the accidental dose. Sixty percent effective overall clearly beats the 50% effective threshold set by the FDA. However, it is nowhere near the efficacy of the Pfizer and Moderna vaccines. I presume AstraZeneca realized this as they played up the low cost and ease of transport of the vaccine in their press release.

Why would a lower dose of the vaccine be more effective; it seems counter-intuitive? It may be that people make an immune response to the chimp adenovirus that prevents the virus from making the coronavirus spike protein and protecting against COVID-19. This seems to be the only explanation anyone can come up with (including me), but it is pure conjecture.

According to Derek Lowe’s terrific Science Magazine blog, which I consulted for his take on the results (he was equally confused, but gives a more detailed scientific analysis), AZ will switch to the lower, more effective dose in the U.S. They will probably have to prove the effectiveness of this dose in the U.S. before the FDA will approve the vaccine.

While AZ says this is all good news because they can “immunize more people with the same amount of vaccine,” that idea seems to require more data. In fact, the press releases today had no information about the actual infections in the trial subjects, or safety outcomes.

I have said before, I have an unjustified prejudice against injecting myself with chimpanzee adenovirus to prevent bat coronavirus. It does not really help convince me if it is the “cheaper, less effective” vaccine!