I think it’s a timeframe thing. I believe (but am not sure) that CDC is predicting what variants are around right now, while daily case reports reflect infections that happened a few days ago. In any case, UW is one of a handful of places tracking variants from PCR test results, and they are basically consistent with PNW being 90+% Omicron today:
https://twitter.com/pavitrarc/status/1473050978026803200?s=21
…but the CDC is stating that it’s over 96% in that region, and Washington is the only state in the region whose overall numbers aren’t declining (implying it probably is the most likely to have a significant Omicron outbreak). And their confidence intervals are insanely wide—like 14-95% in some regions.
It’s probably all academic right now anyways. Omicron will take us soon one way or the other.
(ETA Just saw that it’s a projection from the 16th, which helps explain it.)
Yeah, the CDC is projecting, because the sequencing data lag. But bobman’s figure there is a great illustration of how omicron can be rapidly taking over while overall cases decline.
I think we can expect boosted to have very low hospitalization rate, regular two dose low but the unvaxxed likely fucked. Even if the rate per case is lower the total number of hospitalizations will be up due to the rapid spread.
That sharp spike is going to create havoc. Mostly for the unvaxxed but there will be collateral damage as the system overloads.
You have to be careful though, because that chart is by sample collection date, not testing date. So the later numbers are smaller, at least in part, because not all samples for those dates have been tested yet.
Question for the science bros: to what extent, if any, do mutations impact the rate of false positives/false negatives on Covid tests? My naive, definitely-not-a-science-bro, assumption would be that if a thing changes a test that was designed to detect an earlier version of that thing would become less accurate. Otoh, depending on what the test is actually detecting I could see how it could be possible for the test to identify multiple versions equally well…
That happens but it’s easy to figure out when that’s happening and not a big deal and can be caught with another test that looks at a different part that hasn’t changed. And it’s actually good in some ways because you can detect different strains without doing sequencing.
Weird wrinkle in the data out of Denmark:
Of cases in 12+ year olds, only 8.6% of Omicron cases are in the unvaccinated, vs 23.8% for other variants. Denmark has something like 77% of 12+ year olds vaccinated. So on the surface this would suggest that the unvaccinated are LESS likely to get Omicron than the vaccinated.
Of course, I don’t believe this. Some sort of ADE effect seems more or less impossible given that vaccines still protect against severe disease. The main explanation I can think of is that unvaccinated people are less likely to get tested. This would also explain why the unvaccinated % for other variants looks low (i.e. you’d expect it to be higher than, not equal to, the population share of unvaccinated people). But still, it’s a hell of an effect size to chalk up to that. Anyone have any other ideas?
The PCR tests typically test for 3 specific genes, so even if there’s a mutation in one, the other genes will still be detected. Omicron actually does have a mutation in one of the tested genes used by certain tests, which allows likely Omicron cases to be detected without sequencing the full genome.
Minor quibble, but PCR tests with DNA or RNA probes of specific sequences, not entire genes. Wookie would know better than I though.
If you see people talking about “S gene dropout” this is what they mean, the S gene codes for the spike protein, which in Omicron is highly mutated. The test still gives a positive result because of the other two sequences it checks but the spike “drops out” of detection. This is specifically due to a deletion mutation. The tests can cope with slightly different sequences, but a deletion causes an out-by-one error, like this:
A few letters are wrong in the second sentence and that’s fine, but in the third sentence the deletion means all the subsequent letters are wrong.
So far the antigen folks say Omicron detection is good.
But the idea that we will get a different detection profile in some future variant is not unlikely.
The vaccinated may be doing higher risk activities and thus exposed more often. I don’t know Denmarks public health policies but if things like eating indoors is limited to vaccinated individuals their risk will be a lot higher.
Seems unlikely, but maybe. The unvaxxed husband of my mother was worried about having Christmas. I was quite glad as I didn’t want to go but hadn’t in five years. Make it six!
Possibly age effects too. Older are more vaccinated and more likely to show symptoms.
Lots of possible causes really but the vaccine being responsible is pretty low likelihood.
https://twitter.com/laurenthehough/status/1472847303518171136
https://twitter.com/laurenthehough/status/1472847355875672067
To nitpick, insertion and deletion mutations outside of the primer binding site may be lethal for the protein in question, but they don’t affect detection by qRT-PCR. The substitution in the primer binding site in your example is much more detrimental.
A newspaper reporting on this mentioned none of them were boosted.
I don’t know if that’s real or not, but I’m sure there are plenty of docs who feel like that for pretty much those reasons.
