Thread by @profshanecrotty

A very large, well designed study of vaccine efficacy in UK in during Delta or Alpha. 🧵
Take home messages:
🔵 ~86% Pfizer protection from Delta infections Ct < 30
🔵  Peak Delta viral RNA in infected vax and unvaxxed were similar  
🔵  Some evidence of declining immunity

The preprint:
ndm.ox.ac.uk/files/coronavi…
There is a ton in this.

🔵 ~86% Pfizer protection from Delta infections Ct < 30

I very much agree with biostatistics Prof Shiela Bird that focusing on cases with Ct<30 makes sense in these types of surveillance studies. Pfizer Delta protection was 86%, AZ was 69% (Table 2). bit.ly/3sAeqd8

Those numbers are very similar to the NEJM study for symptomatic cases: 88% Pfizer, 67% AZ. With two different study designs.

That suggests modest to moderate reduction in vaccine efficacy against Delta compared to Alpha, in terms of total infections (Pfizer ~6% and ~10%. AZ ~8% and ~16%. Two studies above.) Really quite encouraging for the RNA vaccine.

I also agree that it makes sense to focus on the age 18-64 data, because almost everyone > 65 in 6th UK was vaccinated by the time of Delta (hooray for them!), leaving too few unvaccinated older subjects for controls. Which the authors state.

🔵  Peak Delta viral RNA levels in infected vax and unvaxxed were similar

Unlikely Alpha, peak viral RNA similar in Delta infections. This is large scale, independent data, consistent with several other reports now. Authors note: this doesn't necessarily mean the same thing:

Also, AFAIK we still don’t know that ratios of viral RNA to infectious virus for Delta compared to Alpha.

And, the total amount of Delta virus over time in a vaxxed versus unvaxxed person is still very different, based on this Singapore Study:

medrxiv.org/content/10.110…

So, vaxxed people can surely transmit Delta. How ofter do vaxxed people transmit compared to unvaxxed?

At least 7X less, because vaxxed people get that many fewer infections (< 30 Ct).
14X less seems a conservative estimate (i.e., 93% fewer transmissions),

given the earlier estimates on transmission in the Alpha era and the narrower period of high RNA load for vaxxed persons in the Singapore Delta study. This estimate applies to the window of time covered in this vaccine study, which was ~4 months.

So, the vast majority of transmissions are unvaxxed.

🔵  Some evidence of declining immunity against infection

There is definitely a decline over ~4 months after vaccination. ~92% to ~72% for Pfizer (cases <30 Ct, ages 18-64).
That 72% is pretty similar to the Israeli unpublished % for Pfizer Delta protection at ~3-4 months.

The Pfizer clinical trial (against original and Alpha) had efficacy against symptomatic cases 95% early and 84% at 6 months. With a 10% drop/shift against Delta that would be 85% early and 74% at 6 months.

The authors predict comparable protection by Pfizer and AZ will occur at 4.5 months. That is a projection (the study hasn't gone that far). Immunologically, that would be surprising. I don't know of any immune response data supporting an equivalence between the vaccines.

That being said, it is worth further investigation. If forced to bet, I would be forced to bet on behavioral differences between the groups, as opposed to immunological protective equivalence.

The authors also tried to deal with the timing of vaccinations, but I don't think the analyses provide insights here, since the UK has shifted between 12 week and 8 week schedules, as again noted by Prof Bird.
bit.ly/3sAeqd8

Regarding behavior, as noted by the authors, it can change based on perceived efficacy.