1829

I did some Twitter reading (from legit scientists only, so should be OK for this thread!) on the biological part of Omicron, to try to understand what the initial reporting around it looking “scary” actually meant.

Here’s a good thread on what is known about the Omicron mutations in isolation:

And here’s one specifically talking about how to reason about the combination of changes, and how the impact on antibody efficacy.

Tweets 2 and 4 are maybe the most intersting ones, since they show mutations happening basically across all most important sites targeted by a wide variety of antibodies. The big takeaways seem to be that a) there will almost certainly be some immune evasion, but just a reduction, b) the changes are largely concentrated on immune evasion rather than improved transmissibility, which would be good news.

And then the for the seemingly maximally informed pessimistic take, from one of the researchers who had previously made an experiment to generate a variant of the spike that was not neutralized by the antibodies of either the people who had Covid or who were vaccinated. They think that the changes in the variant are probably more significant than those in the experiment.

1830

What will actually take to have reliable numbers of transmissibility / severity / vaccine efficacy of Omicron:

And that’s in the UK, who seem to have the some of the world’s most extensive monitoring and best systems for actually linking the data together. In South Africa? Seems like they’d require rather more data, and thus even more time.

This was another good read on the uncertainty:

This is a cool visualization (the second one that is; the first visualization is an unreadable mess, but I can’t figure out how to make Discourse only show one of the pictures):

Then a modeling exercise (you’ll need to read the whole thread) on the data we actually do have so far:

I.e. let’s assume that the kind 500% growth rate advantage over Delta that people have previously computed really is true, and not just due to a founder effect. What level of immune evasion would Omicron need to have to actually achieve that, given reasonable assumptions about the epidemiological status of South Africa pre-variant?

The answer is that fairly small changes in the assumptions make a massive difference: if 75% of the South African population has pre-existing immunity, it’d imply Omicron evading the immunity for 93% of the people With 90% pre-existing immunity, it’d “only” be 31% immune evasion.

1831

Pre-existing immunity is probably lower in South Africa, due to lower vaccine uptake, and lower infections- South Africa’s been stricter than most places, and Africa in general has had fewer COVID cases I believe. 3 Million cases known out of 60 mil Pop. so 5%

South Africa has 25% fully vaxxed.

I’d say given those numbers, pre-existing immunity is at least 30% probably somewhere between 40-50.

What does 50% give you with those calculations?

1832

That’s addressed in the thread. The amount of pre-existing immunity in ZA should be expected to by very high. The excess mortality in South Africa during the pandemic has been 0.39%, despite the population being very young so the fatality rates should be low as well.

But maybe the excess mortality is coming from somewhere else than Covid? To cross-check, there’s a couple of seroprevalancy surveys done in Gauteng province. One is from the start of the year, the other one is from October last year. Unfortunately nothing newer than that.

The results are 20% and 27%. Now add 9 months and two waves worth of Covid infections on top of that, a 25% vaccination rate, and the 75%-90% range of pre-existing immunity looks pretty plausible.

It’d mean that the data could not be explained just by immunity evasion. It’d either have to also be substantially more transmissive than Delta, or people with immunity would need to be more likely to be infected (280% immune evasion). The latter seems quite unlikely…

1833

The shape of the Omicron wave in ZA compared to the last three ones. The full thread is worth reading for the interpretation; hospital admissions “only” growing as fast as in previus waves might be good news since the absolute number of cases is growing way faster.

A thread with a more sophisticated modeling exercise the level of immune escape in Omicron:

1834

The proxy signal for Alpha/Omicron from PCR tests, in England:

There’s a clear increase from the background noise level of 0.1% to 0.3%, starting a week ago. That suggests Omicron cases have been getting missed by the sequencing. (The UK has 50k total cases per day, so 0.3% prevalence would be about 100 new Omicron cases per day). It is a really quick increase compared to the similar computations with Alpha and Delta back in the day, but I guess these could still be mostly imports.

But the signal being this clear already now does suggest that we’ll get some reliable Delta vs. Omicron growth numbers from outside of ZA within a week.

1835

I would have thought that the S gene dropout samples would be prioritized for sequencing. Is the UK still doing completely random ascertainment? Perhaps that’s a change that has only been instituted in the last few days so perhaps the sequencing data is lagging a bit?

1836

My understanding is that they will sequence all SGTFs, but won’t have them skip the queue. So it’ll take about a week for those to get sequenced. But they do prioritize getting the sequencing done ASAP for anyone with a travel history to Africa. (“Missed” was a bad choice of words.)

Also, only about half their PCR tests are with an assay that’s affected by the SGTF.

1837

Preprint: Increased risk of SARS-CoV-2 reinfection associated with
emergence of the Omicron variant in South Africa. A substantial level of immune evasion against past infections. No analysis on vaccines yet.

This graph visualizes it well. A week into the wave that was still barely visible on the charts, they were already seeing as many reinfections as at the peak of the previous wave.

1838

Waiting on information on severity of these reinfections and breakthrough cases before I panic.

1839

Just in time for Christmas, despite the supply issues. Can’t say I enjoyed last year’s beta launch. (ba dum tss)
Sorry. Ah, well, at least I’m coping.

1840

I’m not sure that chart is a good one, since of course there are going to be more reinfections as more population has been infected previously. You would expect this wave to have much more reinfections/primary infections that any previous.

The scary chart for me is this one:

image
image975×638 68.3 KB

Where you see current reinfections exceed their model for no change in reinfections risk (although still close to the confidence intervals).

Also I’m not sure how they have accounted for different R0s. We might be seeing a higher reinfection rate, a faster growth rate, or a combination of both.

1841

I’m not sure that chart is a good one, since of course there are going to be more reinfections as more population has been infected previously. You would expect this wave to have much more reinfections/primary infections that any previous.

The scary chart is this one:

image
image975×638 68.3 KB

Where you see current reinfections exceed their model for no change in reinfections risk (although still close to the confidence intervals).

Also I’m not sure how they have accounted for different R0s. We might be seeing a higher reinfection rate, a much faster growth rate, or a combination of both (most likely).

1842

I saw South Africa is claiming they believe their population was 80-90% COVID immunity just from infections, with 25% or so vaccination rate.

So most of their COVID immunity is through infections (and this is one reason the vaxx rate is so low there)

1843

Sure, and that aspect is shown on the middle graph (B, population at risk for reinfection). I like this graph over the one with a model results, since to understand that graph you need to understand the model. This one is just the raw data. The population at risk for reinfection is 2x larger than during the delta wave, which clearly can’t expect the detected reinfections shooting up 10x faster.

Right, we had that discussion a few days ago. Note that this study is just about reinfections, not about vaccine breakthrough exceptions. So it is still possible that the level of vaccine evasion is much lower, and more heavily vaccinated countries are OK.

1844

A post was merged into an existing topic: Coronavirus 2019

1845

Oops, my bad, wrong thread. Thanks Tom.

1846

The weekly-ish UK variant technical brief is basically all on Omicron this time:

Not a ton of new data there, except an answer to espressojim’s question on sequencing. There’s a graph on page 29 showing their sequencing backlog vs. the SGTF results from PCR. They’re just reaching the knee in the curve, so expect a lot more cases to be reported in the UK starting now. Once that happens, we should also start getting a better idea on whether that sharp increase is due to community transmission or linked to travel.

One more model on immune escape, yet again more sophisticated and apparently with better data on the levels of pre-existing immunity:

Their conclusion is interesting (page 13): even 100% immune escape won’t explain the results from Gauteng. Omicron would also need to be more transmissible.

1847

I continue to be amazed by how damn good the UK healthcare and data gathering systems are. They’ve had some of the best data collection and well designed clinical trials for pharmaceutical interventions out there - really puts the American system to shame. We’d know so much less about how to deal with the pandemic if the UK wasn’t there.

1848

Woof.

Read the thread:

Seems pretty clear:

  1. If you’ve had COVID but weren’t vaccinated, get vaccinated.
  2. If you got vaccinated in the spring or summer, get boosted.
  3. Wear a mask, especially in any indoor public space.
  4. Be smart about gatherings

We still don’t have anything like a full picture yet, but it’s slowly coming into focus.