Thread by @michaelzlin

Curious why Pfizer's 2-dose vax for <5yo failed and if Moderna will do better? I was too, so I looked into it.

It was a bit of a brainteaser, but basically Pfizer played it safe and got blindsided by Omicron but Moderna should do better if side effects are tolerable

a thread

Basic facts: Pfizer's 2-dose vax for 2 to <5yo is 3mcg, compared to 10mcg for 5 to <12yo and 30mcg for ≥12yo. It's also 3mcg for 0.5 to <2yo, by the way, which is an interesting clue in the failure analysis. Ph3 enrollment started in summer 2021.
pfizer.com/news/press-rel…

WSJ reported that the 2 to <5yo vax (going to write 2–4yo now) was not working well against Omicron. VE was ~50% for Delta and then dropped for Omicron. It was also reported that the vax induced lower Abs in 2-4yo than the approved 30mcg for adolescents.
wsj.com/articles/lower…

As everyone knows, Pfizer is now waiting for the third dose to see if it provides adequate Omicron protection. As a 3rd dose broadens immunity (via activation of hypermutated memory B cells) this has a chance of getting over the 50% threshold.
biorxiv.org/content/10.110…

But a question is if Moderna can get VE >50% with just 2 doses, and I think the answer is likely yes, for two reasons. First, because Moderna's dose is much higher than Pfizer's: 25mcg for 2-4yo if the below article is correct;
nytimes.com/2022/03/14/us/…

and second, because a failure analysis suggests Pfizer's 3mcg for 2-4yo was indeed underdosed. This conclusion comes from a careful look at something that hasn't been discussed so much: VE of the 10mcg dose for 5-11yo against Omicron.

The Pfizer vaccine for 5-11yo (10mcg x2) showed 90% protection against disease from 6/2021 to 9/2021. This is very good, especially since some cases were likely Delta. The dosing looks reasonable as NAb levels are similar to 16-25yo getting 30mcg x2.
nejm.org/doi/full/10.10…

However against Omicron, 2x 10mcg Pfizer for 5-11yo only had VE = 31%, compared to VE = 59% for 2x 30mcg Pfizer for 12-15yo. So there was a larger drop in VE for 5-11yo than for 12-15yo which couldn't really have been expected ahead of time. cdc.gov/mmwr/volumes/7…

Now the 2x 3mcg dose for the 2-4yo groups actually gave lower antibody levels than the 2x 10mcg dosing for 5-11yo (1302 vs 3409).

So if 3409 gave VE = 31% against Omicron for 5-11yo, it seems unlikely 1302 would do better than that for 2-4yo.

From Pfizer investor presentation:

Here we have to digress into some speculation. Why is Omicron protection worse for the 2x 10mcg vax in 5-11yo than the 2x 30mcg vax in 12-15yo, if it elicits the same Ab levels (actually higher per the Pfizer slide)? And will 2-4yo act like 5-11yo or magically more like 12-15yo?

Since the earliest days of SARSCoV2, kids <12 seemed to get infected less than those >12, and recover more quickly if they did. I took a position on this early (7/2020) when most studies on kids grouped 1-18yo together:

https://twitter.com/michaelzlin/status/1278742738628775937

The outcome difference has reduced as strains have become more virulent, but it suggested a mechanistic difference in immune function. I proposed it was better innate immunity in preadolescents, and multiple papers in late 2021 suggested that was true.

https://twitter.com/michaelzlin/status/1435868817595330561

Anyway the digression is to propose one reason why similar Abs in preadolescents and adolescents might confer different VEs for Omicron. Pre-Omicron, innate immunity and vax-induced Abs may have worked together to clear viruses before symptoms appeared most of the time.

But Omicron replicates faster in nasal epithelial cells; maybe that overcomes innate immunity in some young kids. And then some other difference in younger kids comes into play so that they are more likely to present with symptoms than older kids.

https://twitter.com/PeacockFlu/status/1477227929557999617

If Omicron overcoming innate immunity is the explanation for 5-11yo showing lower VE at the same Ab levels as ≥12yo, then that would likely apply to 2-4yo as well.

The next question is, would it be helpful to increase the dose? It seems yes. 11yo's who got 2x 10mcg did far worse vs Omicron than 12yo who got 2x 30mcg (11% vs 67%).
cnbc.com/2022/02/28/pfi…

The physiological difference between 11 and 12yo is far smaller than the vax difference between 10mcg and 30mcg, so it's likely 2x 30mcg would help the 11yo's.

And the 11% VE for 11yo's vs 31% for the larger group of 5-11yo's suggests the dose/weight ratio matters, as we expect.

So why wasn't the dose for 5-11yo's higher than 10mcg to begin with? Well with kids who have a lower chance of hospitalization and death than adolescents and adults, it was understandable that the emphasis was on safety.

The Phase 3 trial for 5-11yo experienced 1 fever >40ºC out of ~1000 participants, and the excellent efficacy against infection prior to Omicron seemed to prove the 10mcg was the right choice. Omicron upset the apple cart.
nejm.org/doi/full/10.10…

But that side effect profile is quite mild (0.1% rate of high fever which resolved after tylenol), so there may have been room to go a little higher for 5-11yo. Pfizer tested 20mcg but I don't know what they found in terms of side effects.

Now getting back to what to do for 2-4yo: In hindsight, would it have been better to use something higher than 3mcg? In terms of efficacy, most certainly yes, by analogy from the 5-11yo experience above.

In terms of safety, Pfizer had its reasons to choose 3mcg for 2-4yo: they had seen the next dose up of 10mcg to experience more fevers: 2 out of 32 (6%) vs 0 at 3mcg. (Would be useful to know if those fevers were in the smallest kids.)

If the 2 high fevers in the 32 2-4yo's were in the smallest kids, then that would suggest Pfizer would have benefited from a different age segmentation. And the fact that 0.5-1yo's were also getting 3mcg suggests that it might have been fine to give 3-4yo's something like 6mcg.

Age-specific differences in efficacy within the 2-4yo group could explain the reported lower immunogenicity in Phase 3 compared to 0.5-1yo. Note this lower immunogenicity was not seen in the Phase 1 dose-ranging study.

But you can imagine Phase 3 enrolled a larger % of 4yo's than Phase 1. Seems very likely that parents would be more worried about side effects in 2yo's, and more motivated to vaccinate 4yo's to allow participation in larger social events. Another argument for finer age splits.

It's also possible that Pfizer just got unlucky with the 2 kids aged 2-4yo who got high fevers after 3mcg, because Moderna is giving a whopping 25mcg in this same age group in their Phase 3! So they must have seen an acceptable AE profile in their dose-ranging up to 25mcg.

All indications are Moderna’s 100mcg adult dose is more efficacious (it makes 2x the antibodies) and has more side effects than Pfizers’s 30mcg, so 25mcg would be more efficacious and have more side effects than Pfizer’s 7.5mcg.

In fact you’d expect Moderna’s 25mcg for 2-4yo would be equivalent to >10mcg for Pfizer, where Pfizer saw high fever. That Moderna went through suggested they didn’t see the same. Hopefully that holds up in their Phase 3. It hasn't been stopped for safety, so it seems okay so far

Thus if we could change the past, Pfizer could have done instead:
0.5-2yo: 3mcg
3-4yo: 7mcg
5-8yo: 11mcg
9-11yo: 15mcg
≥12yo: 30mcg

Even better would have been to test an Omicron vax starting in Jan; then we'd likely have a proven good vax this month.