Thread by @hildabast

Preprint of the South African phase 1b/2 trial for the Oxford/AstraZeneca vaccine is online. Primary efficacy analysis is for 2,026 people who weren't infected with HIV (there was also a very small arm of people who were HIV+) medrxiv.org/content/10.110… ...1/n

...This shows you how the new variant spread in the 2 provinces the trial was conducted in...2/n

..As we already knew, the overall efficacy was 22%, but only 10% against the new variant, with a wide range of uncertainty: but enough for the trial's protocol for excluding efficacy. There were no severe illnesses in the vaccinated or placebo groups: a young, low risk group..3/n

...Out of the events (people with symptomatic Covid-19), there were:

Vaccinated group: 15 mild, 4 moderate, 0 severe
Placebo group: 17 mild, 6 moderate, 0 severe ....4/n

...They did an unplanned analysis of the time up to October 31 - before the variant arrived. Efficacy was 75.4% (CI: 8.9 to 95.5). (That's also an extreme amount of uncertainty)... 5/n

...Antibodies weren't able to protect people from mild disease; T-cells could however provide protection against severe illness. But they couldn't answer that in this trial. However, they include data from 17 people in the UK trial...6/n

...Sequencing data are included in a supplementary file. That data suggests that T-cells generated after vaccination are not as affected by the mutation in the variant. But clinical data are needed to show if this = protection against severe disease medrxiv.org/content/10.110… ..7/n

...The authors point to the efficacy of the J&J vaccine at preventing severe Covid-19 in South Africa since the arrival of the variant as another reason it's possible there's still some protection. Only a trial can test that now & that's complicated at this stage...8/n

...Because severe Covid-19 is fortunately so much less common than mild disease, you need a larger trial, with people at risk of severe disease, than for more common outcomes. Very tough situation. /9