15 Apr, 25 tweets, 6 min read
Let's talk about the background risk of CVST (cerebral venous sinus thrombosis) versus in those who got J&J vaccine.

We are going to focus in on women ages 20-50.

We are going to compare the same time period and the same disease (CVST).

DEEP DIVE🧵 KEY NUMBERS!
Spoiler: After an engaging convo with @ashishkjha tonight, I now believe we already have enough info to END the pause on J&J for everyone over 50, probably enough for all men (assuming we don't get a rush of bad news in the next couple of days), and women 20-50 in most situations
We have 6 CVST out of 1,402,712 doses in women 20-50.

That’s the known quantity we can work from.

That’s around 1 out of 233,785 in the 21-day post vaccination period (or 0.43 per 100,000 women 20-50 in the 21-day post vaccination period).
Meanwhile, the background rate of CVST is 0.5 to 2.0 per 100,000 per year for women 20-50.

Let’s say it is 1.0 per 100,000 to make it easy for the rest of the numbers.
But that's annual background rate of CVST in women 20-50

We need to get the 21-day background rate in order to compare it to the J&J risk (because that's how long the risk appears to last after jab).

So we multiply 1 per 100,000 by 0.056 (because 21 days is 5.6% of a year).
1.0 per 100k x 0.056 = 0.056 CVST per 100,000 women 20-50 in any given 21-day period.

THAT, my friends, is our background rate.

0.056 per 100k women ages 20-50, over 21-day risk period.

(Note: comparing the same time periods matters greatly. You can't do yearly vs 21 days!)
Recall the we found that the 21-day rate of post-J&J CVST is 0.43 per 100,000 women ages 20 to 50.

So, we divide to get the relative rate…
0.43/0.056= 7.6.

That means that in the 21-day post-vaccine period, women 20-50 are 7.6 more likely to have CVST than background.

But....
The real range might be 3.8-15.2x.

That just depends on whether the true background rate of CVST is 1 per 100k versus lower (say 0.5 per 100k) vs higher (2.0 per 100k).

This sounds alarming until you realize the risk of COVID-19, *even in this age group*.

Let's look at that.
In 2019, there were 67 million women in the US ages 20-50 (Census data).

If we don't get rid of COVID, eventually most would get it.

Let's say 80% eventually would get it (because of herd immunity).

That means 54 million women age 20-50 would get COVID.

How many would die?
Just for the sake of 'reductio ad absurdum,' allow us to imagine that just 1 in 10,000 in this demographic would die of COVID.

That's a low estimate ON PURPOSE to make a point.

5,400 deaths would occur from COVID, if all US women 20-50 were infected.

Many more in the ICU, etc.
Now let's imagine that all 67 million US women ages 20-50 women get J&J instead of COVID.

The fatality rate so far from J&J-related CVST is 1 in 1.4 million.

That means J&J induced CVST would kill 0.00000071 x 67 million = 48 people.

hmm....
Let's look at the scorecard.

If COVID infects all US women ages 20-50: kills 5,400 (and actually it's far higher; recall I used the ridiculously low infection fatality rate of 1 out of 10,000, just to make the numbers super conservative)

J&J-induced CVS: kills 48.

⚖️🤔
That means that J&J would save 5,352 lives of women ages 20-50 in the US, if it were our only option, compared to COVID.

So far, the COVID death rates for US women ages 20-50 is ~670 PER MONTH (>8000 total).

That means that every day 22 women ages 20-50 are dying of COVID.😰
CVST associated with the vaccine has killed one US woman TOTAL so far.

That one death is tragic, but way outnumbered by the daily COVID deaths.

(Recall I said 5400 COVID deaths if all women 20-50 got the disease. But in fact the real rate would be 5-10x worse, depending on IFR)
What does it all mean?

I had a really interesting convo tonight with @ashishkjha.

He asked point blank if I would give J&J to a 55 year old man who won't get access to another option for weeks or month.

Depends. On what? ....
1. How prevalent is COVID in the region (the higher, the more urgent vaccination is)
2. How long until they can get an alternative.
3. Do they have risk factors for bad COVID.

In most cases the answer would still be YES, but not in all cases.
But could the rates of vaccine-related CVST actually be worse?

I am starting to doubt it.

Why?

Because if so, the rate of reported SEVERE CVST would go up VERY quickly after the "pause" was announced.

Every clinician in the US is looking for this, even in retrospect.
If many severe cases have not been reported en masse in the last couple days, they are unlikely to be many.

Caveat: There is likely to be an increase in the number of MILD CVST-like events, other clots.

(If this happens, I hope media don't confuse mild and severe disease)
But those won't really change the risk calculation--a mild problem is, well, mild.

What we need to watch is not just clots, but clots that cause dangerous problems.

Similarly, we are not comparing this to mild COVID, but to SEVERE COVID. I've been careful to do that here, yes?
So in my view, it is likely that within 72 hours of the pause, we will know 90% of what we need to know about whether this condition is an ORDER OF MAGNITUDE worse than we thought with respect to SEVERE life threatening or organ-harming effects.
Unless things are FAR worse than what we already seem to know, the J&J pause should end for:
1. Everyone over 50
2. Likely for men of all ages.

Ok so what about women 20-50?
And it's even likely a good idea that we should end the pause for women ages 20-50 soon, though I think in a couple of weeks we'll better know how often mild CVST is occurring (again, the number of mild cases is likely to increase, and the number of severe cases by very little)
So: why give J&J to women 20-50?

Basically if there's no alternative available soon in that particular area.

If there's an alternative, sure why not save that 1 in a million life.

But if there's not, why surrender MANY more to COVID while awaiting a late-arriving alternative.
In sum, we are waiting more data.

But I suspect that we already have enough to have resumed J&J.

I think ACIP is in a tough spot here.

If they ended the pause tonight, they might not gather enough information to reassure people.

But sadly, doing so delayed life-saving doses.
Last thing: I am thinking aloud in this public forum.

If I were advising FDA or CDC, I'd INVITE dissent.

I am curious what people think.

Thanks to @ashishkjha and @Atul_Gawande for pressing me to think aloud on this.

I hope it's worthwhile, illuminates, and spurs debate.
🙌

• • •

Missing some Tweet in this thread? You can try to force a refresh

This Thread may be Removed Anytime!

Twitter may remove this content at anytime! Save it as PDF for later use!

# More from @jeremyfaust

9 Apr
Two highly concerning papers now live in the New England Journal of Medicine regarding the adenovirus-based AstraZeneca/Oxford vaccine.

A small number of patients appear to have had serious clotting events with implications.
🧵
nejm.org/doi/full/10.10…
nejm.org/doi/full/10.10…
First, the top-line findings are concerning.

In Germany and Austria, 11 patients developed clotting problems and of these 6 died. The median age was 36.

In Norway, 5 had the condition, 3 died.

Some if not many of these patients were previously healthy.
Keep in mind that in these nations, several million AZ doses have been given.

So this is rare.

But, it looks real.

There's a well described bio-molecular explanation described in these papers that is plausible and consistently measured here.

Impressive and important work.
3 Mar
📰Major update on “external causes” of death (CW suicide, homicide, accidents).

Our team carefully analyzed new CDC datasets for US deaths from January thru July of 2020.

Drug overdoses: up.
Homicide: up.
Suicide: down (!)
Vehicle: down, then up.

medrxiv.org/content/10.110…
🧵/
We took a close look at these causes, modeling the changes over time, and generating expected ranges for 2020.

Drug overdoses: Way up. Even though they were up already before the pandemic, it's clear that things got worse.

Homicide: Very much went up during but prominently after stay-at-home period in US.
15 Jan
Update: most of my colleagues had no symptoms or just arm pain after coronavirus vaccination.

Naturally, *I had to be one of the 10-15% with symptoms (chills, fatigue) bothersome enough to disrupt my usual activities for 24-36 hours.

I affectionately call it “manaphylaxis” 😂
But after a couple doses of Tylenol or Advil I’m mostly ready to do my usual stuff. Similar to my first dose (less arm pain this time).

I mention this because it’s a known thing. This happens to some people.

It’s a lot better than getting actual COVID-19, lemme tell you that!👇
But we need to support people who may need to miss work.

Paid sick leave for the 36-48 hours after vaccination for the small but significant subset of people like me will make a huge difference.

We need to lower all barriers!
30 Dec 20
No time to write threads lately, but I must now.

A 41-year old Congressman-elect from Louisiana has died of COVID-19.

But this kind of thing has been happening all along.

This is the crux of my recent work with @RWalensky @hmkyale @Cleavon_MD etc

jamanetwork.com/journals/jama/…

1/🧵
We found that thousands of young adults have died of COVID-19 and that thousands more have died of other causes during the pandemic.

Things have gotten much WORSE since re-opening.

Do not attempt to blame lockdowns.

Data do not support that.

Below is ALL CAUSE mortality.
It appears to be a combo of COVID, unintentional overdoses, homicide to some extent.

Data on that emerging. cdc.gov/media/releases…

We can eliminate ALL of these deaths with better public health policies. @DrSarahWakeman @MaxJordan_N @ml_barnett @AlisterFMartin @meganranney
17 Dec 20
This Spring, I noticed my colleague @Cleavon_MD tweeting about young people dying of COVID.

Many people of color.

I wondered if these sad cases were statistical anomalies.

Must be that these were "rare" cases, I thought.

I was wrong.

It was real. ja.ma/38b4ayh
🧵
I gathered together a team of people to help look at this.

In various ways, we assessed the data.

It led to the work above and this @nytimes op-ed as well.

nytimes.com/2020/12/16/opi…
This has been an important example, for me, of testing assumptions.

I assumed one thing but wanted to check if I was right or wrong.

When I realized the opposite of my initial assumption was true, I knew we had to get this message out.