Read on Twitter
Will vaccination reduce transmission or just disease?
Do 3 vaccines w interim or final phase 3 results (Pfizer, Astrazeneca, Moderna) reduce asymptomatic infections & does reduced symptomatic infection imply reduced infections?
Thread
Do 3 vaccines w interim or final phase 3 results (Pfizer, Astrazeneca, Moderna) reduce asymptomatic infections & does reduced symptomatic infection imply reduced infections?
Thread
Background
Developing a vaccine for COVID-19 has been a goal since the virus was 1st identified in Jan 2020.
But what is the purpose of vaccines? Many, it turns out!
They can reduce disease, reduce infection, or reduce infectiousness, or some combination. https://nymag.com/intelligencer/2020/12/moderna-covid-19-vaccine-design.html
Developing a vaccine for COVID-19 has been a goal since the virus was 1st identified in Jan 2020.
But what is the purpose of vaccines? Many, it turns out!
They can reduce disease, reduce infection, or reduce infectiousness, or some combination. https://nymag.com/intelligencer/2020/12/moderna-covid-19-vaccine-design.html
Why does it matter whether vaccine reduces disease, infection or infectiousness?
Because it changes who we vaccinate first & whether vaccination protects friends & family of vaccinated person (herd immunity!).
Because it changes who we vaccinate first & whether vaccination protects friends & family of vaccinated person (herd immunity!).
If vaccine only reduces disease but not infectiousness, there are big possible downsides (thread is about health care workers now being vaccinated, but applies to everyone vaccinated!) https://twitter.com/DiseaseEcology/status/1334215150333673472
The 3 phase 3 trials (2 final, 1 interim) that have been released so far (Pfizer, Astrazeneca, Moderna) have all had as their primary endpoint a reduction in "symptomatic COVID-19 cases".
The main goal, based on those options laid out by FDA, was to reduce the probability of getting sick by at least 50% (w/ 95% CI>30%). https://www.fda.gov/media/139638/download
All 3 trials defined "symptomatic COVID-19 cases" with definitions that included very mild symptoms, b/c this would result in a much faster assessment of vaccine efficacy & many argued that vaccine that reduced mild cases would also do so for severe cases.
Some trials also have secondary endpoints that try to assess whether vaccine reduces severe disease, or infection, or both.
https://clinicaltrials.gov/ct2/show/NCT04368728
https://clinicaltrials.gov/ct2/show/NCT04400838 https://clinicaltrials.gov/ct2/show/NCT04470427
https://clinicaltrials.gov/ct2/show/NCT04368728
https://clinicaltrials.gov/ct2/show/NCT04400838 https://clinicaltrials.gov/ct2/show/NCT04470427
We all know that all 3 vaccines meet FDA requirements for being "effective" at preventing (mildly) symptomatic COVID-19 cases:
Pfizer: 94.8% (89.8,96.7)
Moderna: 94.1 (89.3,96.8)
Astrazeneca (SD/SD): 62.1% (41.0,75.7)
Pfizer: 94.8% (89.8,96.7)
Moderna: 94.1 (89.3,96.8)
Astrazeneca (SD/SD): 62.1% (41.0,75.7)
Refs for results:
Pfizer/BioNTech: https://www.fda.gov/media/144245/download
Moderna:
https://www.fda.gov/media/144434/download
Astrazeneca https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)32661-1/fulltext
Pfizer/BioNTech: https://www.fda.gov/media/144245/download
Moderna:
https://www.fda.gov/media/144434/download
Astrazeneca https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)32661-1/fulltext
2 vaccines have (somewhat) convincing data that they prevent cases of severe disease:
Moderna: 30 vs 0
Astrazeneca (interim): 10 (mixed doses) vs 0
Pfizer data aren't convincing yet, b/c too few severe cases after primary endpoint: 3 vs 1 (NOT 9 vs 1: https://twitter.com/DiseaseEcology/status/1336446195284070400).
Moderna: 30 vs 0
Astrazeneca (interim): 10 (mixed doses) vs 0
Pfizer data aren't convincing yet, b/c too few severe cases after primary endpoint: 3 vs 1 (NOT 9 vs 1: https://twitter.com/DiseaseEcology/status/1336446195284070400).
So we know all 3 prevent symptomatic infection. Do any prevent asymptomatic infection or reduce infectiousness & can we use prevention of symptomatic infection to infer effects on asymptomatic infection or infectiousness?
Only 1 of 3 collected periodic data to rigorously assess prevention of asymptomatic infection: Astrazeneca.
Others will try to assess w/ antibody screening in subset; data have not been shown so far. One more (Moderna) used swab from date of 2nd dose as crude assessment.
Others will try to assess w/ antibody screening in subset; data have not been shown so far. One more (Moderna) used swab from date of 2nd dose as crude assessment.
What do these data from Astrazeneca & Moderna suggest? Do vaccines prevent asymptomatic infection & thereby reduce transmission?
A: It's suggestive, but messier than we'd like, & clearly shows that preventing symptomatic infection does NOT imply reduced asymptomatic infection.
A: It's suggestive, but messier than we'd like, & clearly shows that preventing symptomatic infection does NOT imply reduced asymptomatic infection.
Moderna (in supplemental doc: https://www.fda.gov/media/144453/download):
Tests on swabs from date of 2nd dose showed ~2/3 lower PCR+ in asymptomatic people. No stats done; but pretty strong effect (simple binomial prop: P=0.0012).
Tests on swabs from date of 2nd dose showed ~2/3 lower PCR+ in asymptomatic people. No stats done; but pretty strong effect (simple binomial prop: P=0.0012).
Astrazeneca:
No (yes?). As most know, this trial was greatly complicated by giving a subset of people a half-dose:full-dose w/ long delay b/w doses in 18-55 yr in UK & full-dose:full dose w/ ~4 wks in 56+ & elsewhere. In planned trial no reduction at all (3.8%; 95%CI -72.4,46.3)
No (yes?). As most know, this trial was greatly complicated by giving a subset of people a half-dose:full-dose w/ long delay b/w doses in 18-55 yr in UK & full-dose:full dose w/ ~4 wks in 56+ & elsewhere. In planned trial no reduction at all (3.8%; 95%CI -72.4,46.3)
In accidental half-dose:full dose w/ long time gap, reduction by ~half but on edge of including 0 in CI (58.9% 95%CI: 1.0, 82.9).
Most worrying (to me) is that the full-dose:full-dose vaccination reduced symptomatic cases by 62.1% but had NO EFFECT on asymptomatic infections. Thus, one CANNOT argue that a large reduction in symptomatic infections guarantees a significant reduction in infection.
Thus, while it's *possible* that the high efficacy of Pfizer/BioNTech vaccine also reduces asymptomatic infection (& results from very similar Moderna vaccine are encouraging), I wouldn't count on it without seeing data.
What about infectiousness? Do any of the vaccines reduce either how long people are infectious (duration) or how infectious they are (magnitude) when infected?
No data. None at all. Not a single viral load that I've seen. If I missed it, please provide link.
No data. None at all. Not a single viral load that I've seen. If I missed it, please provide link.
Summary
-All 3 vaccines reduce (mild) symptomatic infection
-1(2?) reduce severe disease
-1 (likely) reduces asymptomatic infection; a 2nd does so but only w/ accidental weird dosing regime
-Reducing symptomatic disease does NOT always reduce infection
-no data on infectiousness
-All 3 vaccines reduce (mild) symptomatic infection
-1(2?) reduce severe disease
-1 (likely) reduces asymptomatic infection; a 2nd does so but only w/ accidental weird dosing regime
-Reducing symptomatic disease does NOT always reduce infection
-no data on infectiousness
Given this lack of evidence, vaccinating those at higher risk (elderly) is much more prudent strategy than vaccinating younger people & is likely better even if vaccination does prevent infection: https://twitter.com/DiseaseEcology/status/1338015008152698880.
I hope all 3 trials will share available data on viral loads (which all have, at least for symptomatic cases & 2nd dose swabs), & serological results ASAP. Even better, why not send packet of swabs & tubes to all participants now to get more data on infection/infectiousness?!
Lastly, I hope studies are underway on whether vaccination protects household members of vaccinated to provide clear evidence of reduced transmission. Viral load data would be useful, but correlations b/w viral loads on swabs & exhaled virus can be v low ( https://twitter.com/DiseaseEcology/status/1312537985145950208)
Also, since we don't have data on protection against infectiousness & since vaccination isn't 100% effective, vaccinated should continue to take preventative measures to protect themselves & those around them (masks, distance, etc.) until most are vaccinated.
Clarification: since all 3 vaccines reduce symptomatic cases & none, so far, increase asymptomatic infections, all likely reduce transmission substantially. But insufficient evidence that they block it completely (more data needed).
H/t @JoannaMasel
H/t @JoannaMasel
Addition: @mlipsitch wrote a very helpful blogpost today w details of how efficacy of vaccination in terms of transmission, symptoms, etc. contribute to herd immunity. I strongly recommend it. https://ccdd.hsph.harvard.edu/2020/12/17/covid-19-vaccines-and-herd-immunity/
