The Presidents’ Day holiday interrupted COVID-19 data reporting, so there was no usable update yesterday. Today, as of 2/22/2022, the U.S. stands at roughly 26 new cases per 100K population per day, down 40% in the past seven days, and down 90% from the peak of the Omicron wave five weeks ago. If there is any reduction in the rate of decline, it’s pretty subtle so far.
There are a couple of other things worth noting.
First, for the U.S., son-of-Omicron (BA.2) appears to be a dud. In theory, it’s about half-again as infectious as the original Omicron variant (B.1.1.529), and should be displacing the original variant. In practice, the CDC keeps revising its estimates of spread of BA.2 downward. As of the week ending 2/19/2022, BA.2 was still estimated to account for less than 4 percent of all new U.S. cases.
At the current rate of decline of new cases, I’d guess that the Omicron wave will be well and truly over before BA.2 can even get going.
Second, hospitalizations and deaths have not fallen anywhere near as fast as new cases. That’s completely different from the end of the Delta wave, where those fell in sync. The upshot is that the case hospitalization rate for Omicron is now back up to where it was under the far-more-virulent Delta. That’s puzzling, and I keep thinking that there has to be some significant explanation for it.
I’ve fumbled around trying to explain it. Maybe it’s cases missing out of the official counts, due to the rise of home testing. Maybe more virulent strains of Omicron are gaining ground. But neither of those really explained the magnitude of the effect.
Michael Andreas, a long-time reader of this blog, provided what I think is the most plausible explanation of this: Maybe immunity is waning among the elderly. They were first in line to get vaccinated and boostered, they should be first to see that vaccine-provided immunity decline. And because they are so much more likely to be hospitalized or die from COVID, compared to others, a shift in cases toward the elderly would markedly increase the number of hospitalizations and deaths per case.
Tracking down the numbers by age, the end of the Omicron wave has, in fact, been associated with a shift toward the elderly in terms of new cases, hospitalizations, and deaths. Plausibly, that’s the result of waning immunity in that population.
Case trend
Data source for this and other graphs of new case counts: Calculated from The New York Times. (2021). Coronavirus (Covid-19) Data in the United States. Retrieved 2/23/2022, from https://github.com/nytimes/covid-19-data.” The NY Times U.S. tracking page may be found at https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html
From the second graph, you can see that there is still no clear slowdown in the rate of decline in new cases for the U.S. as a whole. We’re approaching our one-month anniversary of cases declining at a steady 40+% per week.
Maybe there’s the start of a slowdown showing for the states that led the decline.
Son-of-Omicron remains a dud in the U.S.
Source: CDC COVID data tracker.
Recall how COVID strains are named. The two original strains of COVID-19 were termed A and B. Each generation of variants is separated by a dot. Each new variant is numbered in the order it is discovered. And then, after four generations, you give that line its own letter designation, if there are any further sub-variants.
The original strain of Omicron is B.1.1.519. That was given the alias BA. Then BA.2 is the second known sub-variant, dubbed son-of-Omicron. And then there’s BA.1.1, which is the first grandson of Omicron, so to speak.
BA.2 is thought to be half-again more contagious than the original Omicron strain. It has (e.g.) become the dominant strain in Denmark.
But in the U.S., it’s spreading far less rapidly than was originally projected. The CDC’s current estimate (for the week ending 2/19/2022) is that it accounts for 3.8 percent of all new cases. That’s actually down a bit from where the CDC thought it was last week. So this is spreading far less rapidly than the CDC had projected. There now seems to be little danger of it prolonging the U.S. Omicron wave, given how fast cases in general have been falling.
The only thing happening at the moment is that the original Omicron strain is being pushed out by BA.1.1. Near as I can tell, nobody seems to think this has any clinical or epidemiological significance whatsoever. There is no research to suggest that grandson-of-Omicron BA.1.1 is materially different from the original Omicron B.1.1.529.
Are rising case hospitalization and mortality rates due to waning immunity in the elderly?
Source: Calculated from CDC COVID data tracker data, data accessed 2/22/2022.
On the downside of the Omicron wave, we’ve seen a far faster rate of decline in new cases than in hospitalizations or (two-week-lagged) deaths. The result is that the case hospitalization rate for Omicron is back to where it was under the far-more-virulent Delta variant.
I seem to be the only person in the U.S. who thinks that needs some sort of explanation. Taken at face value, Omicron appears to have grown far more virulent as the cases have declined. But surely if that were true, somebody would have noticed it.
I’ve thrown a couple of possible explanations at this, but nothing really stuck. Maybe a rapid growth in home testing has significantly reduced the trend in officially-counted cases. But the reduction in cases needed to account for a tripling of the case hospitalization rate is just too large to be plausible. Maybe this is due to the newer strains of COVID-19 (BA.2 and BA.1.1) being more virulent that the original (B.1.1.529). But nobody says that about BA.1.1., and BA.2 accounts for far too few cases to matter yet.
Michael Andreas, a long-time reader of this blog, suggested what I now think is by far the most plausible explanation: Waning immunity. In particular, waning immunity among the elderly who a) were first in line to get vaccinated and boosted, and b) are far more likely to be hospitalized or die, if infected, compared to the remainder of the population.
First, it’s well-established that immunity from COVID-19 infection declines over time, whether from vaccination or from prior infection. Prior to Omicron, the debate over booster shots centered on this observable decline in immunity, particularly among the elderly. Some of the estimates of decline in immunity were remarkably large, as in this this study of persons using Veterans Administration facilities, a largely elderly population. Other studies significantly slower rates of decline.
Source: Figure 1, SARS-CoV-2 vaccine protection and deaths among US veterans during 2021, Barbara A Cohn et al, Science, Vol 357 no. 6578,
Just to underscore how concentrated those severe outcomes are, here’s a repeat of a graph showing the case mortality data for COVID-19 and flu. For either disease, almost all the deaths occur among the elderly.
(There’s nothing unique about flu or COVID-19 in this regard. If you graphed all-causes mortality, or the most common causes of death, you’d see something very much like this. Other than deaths from accidental causes, the sole exception that comes to mind, among major causes of death, is breast cancer, where roughly 40 percent of deaths are in the under-65 population).
And so, putting the argument together, the elderly:
- are at highest risk for loss of immunity over time,
- were vaccinated and boostered first, and so have had the greatest time elapse since immunization, and
- have vastly higher case hospitalization and case mortality rates compared to the rest of the population.
And, sure enough, as of mid-February 2022, COVID-19 deaths have become ever-more-concentrated in the elderly.
Source: Courtesy, Michael Andreas, source data are Provisional COVID-19 deaths by week, sex, and age, from NCHS.
So it’s not as if the virulence of Omicron has increased across-the-board. It’s that the elderly — and only the elderly — appear to have lost some of their protection against dying from COVID-19.
We see a similar but more muted change in hospitalizations. The elderly’s share of new COVID-19 hospitalizations has been rising for roughly the past month and a half.
Source: Calculated from the CDC unified hospital dataset, accessed 2/23/2022.
As far as I can tell, there’s no way to get reliable national information on cases by age in a timely fashion. Instead, I’m using data from Virginia, and using only the period of time following their recent clean-up of cases with unknown ages.
The results from Virginia reinforce what you’re seeing above. The elderly’s share of all new cases has been rising fairly rapidly during the decline of the Omicron wave.
Source: Calculated from Virginia Department of Health file of total COVID-19 cases by age group.
This is a little bit piecemeal, but taken as a whole, the evidence suggests that the elderly’s share of new cases, new hospitalizations, and deaths is rising.
That is particularly notable given the return to school that occurred under Omicron. If anything, we’d have expected cases in the pediatric and young-parent age groups to increase as a fraction of the total. To the contrary, what we’re actually seeing is a risking fraction of all cases and severe cases in the elderly.
There is no way directly to prove it, but the most plausible explanation is that immunity is waning first among the elderly. And that’s why we’re seeing what looks like an increasing (non-age-adjusted) case hospitalization and case mortality rate at the end of the Omicron wave. Because hospitalizations and (particularly) deaths are so concentrated in that age group, even a modest shift in toward older persons can plausibly result in a marked increase in the rates per case.
I haven’t take this to the final conclusion step yet, to see if I can reconcile all the numbers based on the higher case rates of the elderly. But, back-of-the-envelope, the shift in age mix of new cases from Virginia, combined with the extreme concentration of deaths in the elderly, does seem to be able to account for the much slower decline in deaths compared to cases. As the share of cases in the elderly roughly doubled, all other things equal, we should expect to see a near-doubling of the case mortality rate. And that’s exactly what we’ve seen nationally, from 0.3% of cases to 0.6% of cases.
The upshot is that this seemingly modest demographic shift is probably adequate to explain the rising case mortality rate, and so, probably the rising case hospitalization rate as well.
As a result, the most likely explanation I can find for the apparent uptick in case hospitalization and mortality rates is that immunity is declining more rapidly among the elderly, right now, than it is in the rest of the population.
Obviously, this has some fairly serious implications if you happen to be elderly. Hardly a day goes by when I don’t read some comment, on some newspaper article, that boils down to “I’m vaccinated and boosted, so I no longer need to wear a mask.” That attitude could be a costly mistake for an elderly person if immunity is now rapidly declining.
Given that it’s the CDC’s job to make recommendations on mask use and vaccination, it would be helpful if someone in a position of authority there would directly address this. What I see, as I peruse the CDC website, is reticence to provide detailed information on the loss of vaccination-based immunity over time. But if we’ve got a bunch of older Americans who think they’re good to go because they got the booster, the CDC should be burdened to clarify just how much protection the average older American can expect to have, some months after that last booster shot.
