Read on Twitter
1/n
Exciting findings on the AZN vaccine:
1. big trial (17k)
2. tested 1 dose vs 2 doses (assessing 1-dose efficacy correctly, unlike Pfizer), and short vs long delay to dose 2
3. first insights on prevention of active infection not just symptoms
https://ssrn.com/abstract=3777268
Exciting findings on the AZN vaccine:
1. big trial (17k)
2. tested 1 dose vs 2 doses (assessing 1-dose efficacy correctly, unlike Pfizer), and short vs long delay to dose 2
3. first insights on prevention of active infection not just symptoms
https://ssrn.com/abstract=3777268
2/n
Dose 1 efficacy=76% against symptomatic COVID19 (correctly assessed 3-8wk afterwards, the 3wks is to build immunity)
Dose 2 efficacy=82% if given >12wk, 55% if <6wk (note longer followup for those receiving 2nd dose)
Dose 1 efficacy=76% against symptomatic COVID19 (correctly assessed 3-8wk afterwards, the 3wks is to build immunity)
Dose 2 efficacy=82% if given >12wk, 55% if <6wk (note longer followup for those receiving 2nd dose)
3/n
The longer followup to measure dose 2 efficacy might mean more time for immunity to decay. So we can't compare dose 2 to dose 1 efficacy directly. However it does seem giving dose 2 >12wk later is better than <6wk later.
The >12wk dose 2 produces higher antibody titers too.
The longer followup to measure dose 2 efficacy might mean more time for immunity to decay. So we can't compare dose 2 to dose 1 efficacy directly. However it does seem giving dose 2 >12wk later is better than <6wk later.
The >12wk dose 2 produces higher antibody titers too.
4/n
This is most interesting: infections were tested in a subset by weekly home swab and RT-PCR, really critical to learning how well vaccines block transmission.
Turns out no difference in asymptomat infections detected. Here's the 1-dose results. Also true for 2-dose group.
This is most interesting: infections were tested in a subset by weekly home swab and RT-PCR, really critical to learning how well vaccines block transmission.
Turns out no difference in asymptomat infections detected. Here's the 1-dose results. Also true for 2-dose group.
5/n
Now it would be easy to say that the vaccine only helps symptomatic disease, doesn't prevent asymptomatic disease. But this would be wrong, because what must be happening is some infections that would have become symptomatic are being suppressed to an asymptomatic level.
Now it would be easy to say that the vaccine only helps symptomatic disease, doesn't prevent asymptomatic disease. But this would be wrong, because what must be happening is some infections that would have become symptomatic are being suppressed to an asymptomatic level.
6/n
As symptomatic dz is suppressed by 76% vs unvaccinated, some of the 76% must become asymptomatic. If the vaccine really had no effect against asymptomatic infection, then you would expect to observe an *increase* in asymptomatic cases in the vaccinated due to this conversion
As symptomatic dz is suppressed by 76% vs unvaccinated, some of the 76% must become asymptomatic. If the vaccine really had no effect against asymptomatic infection, then you would expect to observe an *increase* in asymptomatic cases in the vaccinated due to this conversion
7/n
But we see asymptomatic infections are flat between vaccinated and control. It means virus replication is suppressed enough to reduce symptom rate by 76%, and also suppressed enough to clear out some asymptomatic infections to undetectable levels in the weekly testing regimen
But we see asymptomatic infections are flat between vaccinated and control. It means virus replication is suppressed enough to reduce symptom rate by 76%, and also suppressed enough to clear out some asymptomatic infections to undetectable levels in the weekly testing regimen
8/n
Note we dont' know how good the testing was. There were relatively few asymptomatic cases detected (e.g. 13 asymptomatic vs 71 symptomatic = 15% in unvaccinated), but poor swabbing/compliance cannot be ruled out.
Note we dont' know how good the testing was. There were relatively few asymptomatic cases detected (e.g. 13 asymptomatic vs 71 symptomatic = 15% in unvaccinated), but poor swabbing/compliance cannot be ruled out.
9/n
The actual # of asympt cases is important as it gives us an idea of how protective the vaccine is against transmission. For example assuming no isolation of the symptomatic, we can estimate block of transmission as follows:
The actual # of asympt cases is important as it gives us an idea of how protective the vaccine is against transmission. For example assuming no isolation of the symptomatic, we can estimate block of transmission as follows:
10/n
30% asympt rate x 75% relative transmission + 70% sympt rate x 100% relative transmission x 24% relative risk = 39 units transmission risk in the vaccinated, out of 30% asympt rate x 75% relative transmission + 70% sympt rate x 100% relative transmission = 93 units in nonvax
30% asympt rate x 75% relative transmission + 70% sympt rate x 100% relative transmission x 24% relative risk = 39 units transmission risk in the vaccinated, out of 30% asympt rate x 75% relative transmission + 70% sympt rate x 100% relative transmission = 93 units in nonvax
11/n
So that gives 39/93 = 42% relative risk of transmission in the vaccinated, a 58% drop. Note the assumed 30% asympt rate and 75% relative transmission were those used in a recent CDC model. https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2774707
So that gives 39/93 = 42% relative risk of transmission in the vaccinated, a 58% drop. Note the assumed 30% asympt rate and 75% relative transmission were those used in a recent CDC model. https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2774707
n/n
Ironically if all symptomatic people isolated promptly then the effectiveness of the vaccine on blocking transmission would be close to zero because only asymptomatic people would spread disease, and the vaccine doesn't reduce their numbers overall.
Ironically if all symptomatic people isolated promptly then the effectiveness of the vaccine on blocking transmission would be close to zero because only asymptomatic people would spread disease, and the vaccine doesn't reduce their numbers overall.
Epilogue: In theory a vax could enhance transmission if it converted symptomatics to asymptomatic spreaders without clearing virus, but this is unlikely as the immune system doesn't stop at a certain virus load. IAC such a vax is still beneficial if vulnerable are vaxxed first.
