Thread by @peripatetical

A few comments on endpoints for COVID vaccine trials. For regulators, the main question when considering effectiveness is whether the vaccine reduces the risk of symptomatic COVID.

https://twitter.com/ProfMattFox/status/1340662431115309060

But there are two other relevant questions - whether a vaccine prevents severe disease, and whether it prevents transmission.

For the question of symptomatic COVID, the endpoint of interest is symptomatic confirmed COVID. (there are actually two definitions of this - the European ECDC and US CDC, with slightly different symptom lists - both are collected).

Symptomatic confirmed COVID is the primary endpoint, meaning that the study is designed to answer this question specifically. This is the "headline" result we have seen reported from phase 3 studies.

The question of whether the vaccine prevents severe COVID is important, as there had been some concern that some immune responses could make disease worse.

Although the numbers of patients who are hospitalised are much smaller, the numbers from the published trials are falling in the right direction - fewer, or no hospitalised patients in the vaccine arm.

Whether the vaccine prevents transmission isn't a simple question to answer. There are three potential ways to look at this - the best would be to look at infection rates in the close contacts of those in the study.

But this would turn a 40,000 patient study into a 150,000-plus person study. This would be difficult, and obviously household members could still get infection from someone other than the study participant.

A second way to answer this is to look at asymptomatic infection - for example, in the AstraZeneca trial, they did weekly swabs on participants in some trials.

They found that there were slightly lower rates of asymptomatic infection in the vaccine group (29/3288) vs the control group (40/3350).
thelancet.com/journals/lance…

Moderna also did a pre-dose 2 swab and found prevalence lower in the vaccinated group (14/14134) than the placebo group (38/14073).
fda.gov/media/144453/d…

Note that even if there is little protection against asymptomatic infection, it doesn't mean that the vaccine mightn't still reduce transmission.

If vaccination only turned symptomatic infections into asymptomatic infections, that would be an issue, but a reduction in overall infection (symptomatic + asymptomatic) is probably a good thing.

It is also thought that those who are truly asymptomatic (that is, those who never get symptoms) are probably less infectious, but this needs to be confirmed in the context of vaccination.

A third way to answer this is to do serology to look for infections in those who didn't have PCR positive COVID.

Obviously the usual serology based on the spike protein would be affected by the vaccine, but it is possible to do serology using non-vaccine components (eg N-binding). This is planned in the Pfizer study, but I haven't seen results yet.
pfe-pfizercom-d8-prod.s3.amazonaws.com/2020-09/C45910…

In summary, we primarily want a vaccine to prevent people getting sick and going to hospital. Whether the vaccine prevents transmission is an important question - early data from Moderna and AstraZeneca hints that overall infection may be reduced, but more work is required.