Ricochet is the best place on the internet to discuss the issues of the day, either through commenting on posts or writing your own for our active and dynamic community in a fully moderated environment. In addition, the Ricochet Audio Network offers over 50 original podcasts with new episodes released every day.
Coronavirus Speculations on Old Men and Bad Bugs
I have a couple of speculations relating to the COVID-19 pandemic.
I. No Country for Old Men?
It’s been widely reported that COVID-19 fatalities are heavily concentrated among the old, and that the fatality rate for men is significantly higher than for women. There is quite a bit of evidence for this, but it appears is that the risk of death may be extraordinarily concentrated among old men. This could explain some of the differences in death rates between countries. Perhaps it is countries with a high proportion of very old men (70+ and even 80+) that would be expected to have unusually high death rates.
I found the following graph from the WHO’s regional office in Europe (here), showing proportions of death by age and sex. They do not give a date, but it says that the source is data aggregated from Italy and Spain from reports of 18,590 deaths.
I found two further data sources for this and prepared my own graphs. The first is for Italy, based on a report of over 12,500 deaths through April 2 (the paper is here; population data by age and sex from here). Here’s the graph:
The green bars are the deaths per 100,000 overall, with the blue bars for men and the pink bars for women. You can see how the risk is extraordinarily concentrated among men over 70, and especially among men over 80. Overall, the deaths in Italy were 69% male and 31% female.
I found good information about this from New York state, which reports data by age and by sex but, unfortunately, not by age and sex (the data is here; population information from here). New York does not separately report the 90+ age category, but I was able to put together this graph comparing New York state and Italy:
The data follow the same pattern, though the rates are substantially higher for New York. Remember that this is data from Italy about a week ago, and Italy’s total deaths have increased by about 50% since then.
Overall, in New York state, COVID-19 deaths have been 61% male, 39% female. It would be interesting to have this information by age category.
There could be an interaction between population age and the degree of social contact among the old, which could cause a differential effect between countries. This is just a hypothesis, without supporting data (yet). The idea is that old men are particularly vulnerable, and death rates will be higher in areas: (1) with a large number of old men (obviously), and (2) where the old men have more social contact with others.
II. The Good Bug and the Bad Bug?
In a recently posted paper (here), Dr. Knut Wittkowski stated that there was evidence of at least two strains of the virus causing COVID-19, with different levels of fatality associated with the different strains. He suggested that a deadlier strain (I’ll call it the “Bad Bug”) emerged in Wuhan and then spread to South Korea, Iran, and Italy. One or more less deadly strains (I’ll call them the “Good Bug”) spread elsewhere, particularly to other parts of Asia and beyond.
There are now new stories reporting that the new coronavirus was present in New York earlier than previously thought, based on an analysis of mutations in the virus. Researchers found that “early city cases were not linked to later ones, which they found came from Europe,” according to the NY Post (here). An article in the NY Times (here) states that “the earliest cases identified in New York were not linked to later ones” and that the later viruses “were practically identical to viruses found around Europe.” Other teams cited in the NYT article found some infections traced to Washington State.
In contrast, a paper by some Yale researchers (here) analyzed cases in Connecticut, and found that most resulted from interstate travel, tracking back to Washington state, with few from Europe or Asia. Connecticut has a much lower death rate than New York, at present.
The Good Bug/Bad Bug hypothesis might explain the strange progression of COVID-19 in many places. I am frankly skeptical of efforts to contain the virus. Yet most of China, and South Korea and Japan, and many Asian countries have low death figures. In addition, though there is evidence that the disease arrived earlier on the West Coast, the death rates in Washington and California remain quite low (particularly compared to New York). There are also European countries with very high death rates (notably Italy and Spain, and increasing in France and the UK), while Germany has a very low rate.
What if the Good Bug infected much of China, much of Asia, and then the West Coast of the US and perhaps some fortunate European countries like Germany? And the Bad Bug infected Iran, then Italy and Spain, then other parts of Western Europe, and finally New York and New Orleans?
The Good Bug may have passed with little or no effect, and given immunity. Areas hit hard by the Bad Bug are the ones suffering higher numbers of deaths.
This is just a hypothesis, for discussion.
ChiCom delenda est.
Published in Healthcare
I don’t have numbers to verify it, but I would wager that the proportions in your Point I would look pretty much the same with or without the current virus.
The idea of two strains with distinctly different virulence is appealing because it resolves what I think is the most nagging problem about this epidemic: it seems almost inconceivable that the virus didn’t travel widely on the west coast, given the high volume of passenger travel to/from China; yet if that’s the case then we would expect to see a far greater impact on hospital resources, certainly too much to be rolled into even a peculiar influenza season and overlooked.
The report today that the New York virus is genetically similar to the European version of the virus makes it all very plausible.
It would be interesting to know what virus is prevalent in Germany. Is it more like the virus in California, or like the one in New York? Hopefully we can get the various strains cataloged and tracked soon.
Good post.
This is in line with a conspiracy theory based on “L” and “S” strains that I heard a few weeks ago. That theory goes, that somebody invented two strains of a virus: one was deadly (the weapon) and the other was mild (the vaccination). The mild strain was set free in the originating country and shortly thereafter, the weapon was unleashed. As long as you are speculating about one engineered virus, why not speculate about two!
California got the virus earlier and is close to herd immunity
We don’t know how widespread the virus is in California. I do think there’s probably a lot of it there, but we’re still speculating.
And that explanation alone wouldn’t account, I don’t think, for the fact that we didn’t have an ICU crisis in California, comparable to the one in New York. I can accept that a lot of the illness attributed to influenza in California over the last few months might have been the Wuhan virus, but I still find it hard to believe that the peculiar virulence of the disease wasn’t noticed.
Unless, that is, there are in fact two (or more) strains, very different in their lethality, as has been suggested.
The quality of statistics in different countries and different regions of different countries makes me wonder about the good vs bad virus. Have the different viruses been isolated ans shown to be different?
I’d like to know the smoking rates and co-morbidity differentials of old men vs old women. We know life expectancy is lower for men. Did more men die “with” the virus than “ of it” as compared to women?
The demographic story has merit. One interesting phenomenon about coronavirus (from what I’ve read in medical journals, I’m not an MD) is that it not only affects people with comorbidities, but specially with cardiovascular-related comorbidities. This apparently sets it apart from other, more classical respiratory viruses.
The fact that men tend to have higher rates of cardiovascular issues (although not necessarily hypertension, which is a major comborbidity) may explain part of the sex discrepancy. The fact that smoking is much more pervasive in Italy (and among men over women within Italy) may also explain some of its lethality there.
I wouldn’t put too much stock in the good bug vs. bad bug idea. It’s important to remember that most Western countries are sequencing virus isolates continuously. If there were any difference in the apparent lethality, this would be fairly clear and would immediately be reported since mutations leading to divergent outcomes are an obvious question in any viral epidemic.
The good bug vs. bad bug discussion seems to be an attempt to convince ourselves that there is an easy way out of the epidemic, especially since it supports the unproven hypothesis that the mild strain of the virus has already infected the entire West Coast so that most people there can go back to work. If the difference in strains were that clear, we would almost certainly have heard about it by now.
Also, keep in mind that most of the strains circulating in Germany are nearly identical to the strains in either northern Italy or eastern France, both areas that are getting pummeled by the virus. In one striking example, the province of Haut Rhin has one of the worst outbreaks in the world and is apparently routinely denying ventilators to patients over 80, while a few miles across the border in Germany (i.e. in people exposed to the exact same virus strain), hospitals are empty and are even having French patients medivacced in.
The less emotionally satisfying but more plausible explanation for major differences in spread/lethality is a combination of a) the occurrence of “super-spreader” events (like sporting events or religious gatherings that led to the major outbreaks in Europe), which is fairly random, and b) the fact that most states on the West Coast imposed lockdowns before the virus could take hold. That doesn’t mean lockdowns are the ONLY way to achieve low infection/death rates, but the effectiveness of lockdowns can’t be disproven based on the data available, and appears much more plausible at the moment than some type of demographic/virus genetics hypothesis.
Also, if anyone is interested in the demographics question and other factors that may have caused the outbreak in northern Italy to be so bad (and why it might not be so bad in the US), I refer you to this excellent article in JAMA:
https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2764369
It’s written by two front-line Italian doctors and John Ionnidis, the noted Stanford contrarian who has been critical of the conventional wisdom on coronavirus. Not everything in the article is fully justified by the actual data, but it’s still a much more serious and realistic article than many others floating around.
Mendel, this isn’t convincing to me.
The “super-spreader” event idea isn’t plausible. We have these all the time in the US. At this time of year, they were called NBA, NHL, and NCAA basketball games.
I’m going to analyze your hypothesis about the West Coast lockdowns. They were only a few days before the East Coast, and I find it implausible that this could have had much of an effect. Before looking, I expect the case rates to have been similar at the time that lockdowns were imposed, so this explanation would not work.
My motivation is to explain the strange, high fatality rate in the NYC area, which I would have expected on the West Coast as well. I don’t see any way that the lockdowns could have had an effect on deaths for around 3 weeks (say 18-24 days), which is not what I see in the data. I see no effect, generally.
I will be interested in whatever you learn. I am attracted to the two-strains hypothesis because it seems to explain things better than anything else I’ve heard, but it has its own challenges as well.
Religious gatherings in Europe? Whatever happened to the Europe we thought we knew?
Also keep in mind that this was present on the west coast long before it appeared in New York, so the timing difference in shutdowns is minimal. I am very skeptical of the idea of “super-spreader” events. Certainly, at a sporting event, you are in proximity to the 5 or 10 people immediately around you… it’s not like a literal bug that might jump from one person to the next, it would need to infect before spreading beyond that distance. If it is truly possible for a single individual at a megachurch or other event to infect hundreds (and I find it hard to believe that it actually is), then this would happen at every grocery store, market, workplace, school, etc… and the disease would be everywhere. The whole idea of “super-spreader events” is implausible when applied to any setting other than a nursing home or a hospital.
One more note: I attended a concert in Seattle in late February. People shoulder to shoulder. This was well after the virus was detected in WA and just before the Kirkland outbreak happened. On any given night in Seattle (this was a Tuesday!), there are dozens upon dozens of these events, and they were not stopped until the middle of March.
It should be easy.
It takes one day to DNA sequence a virus.
You are immune?
The west coast got the virus in November, earlier than NY and Italy.
regarding super spreaders, what caused covid 19 outbreak in New Orleans? Mardi Gras
I’d love to find out! If there aren’t enormous swaths of us who are immune, then all that stuff about the virus being airborne and “super-spreader events” is likely nonsense.
I hear that they’ve been having some pagan Mother Goddess rituals in Italy, some time back. Some place called Vatican City, I think.
I’ve been there. The place was full of tourists and pickpockets. Hardly any room for religion.
https://www.arcpointlabs.com/blog/covid-19-antibody-testing/
There are numerous examples already in which one or a very small number of infected people spread the disease to several hundred within 1-2 days’ time. One of the best characterized events was at the ski area of Ischgl in Austria, where one employee at a very popular bar returned to work after a few days in northern Italy (where the virus was rapidly spreading). Within two days several hundred people (other employees but mostly tourists) had been infected.
Based on the sequences of the viruses detected in all of those tourists, we know that they all got the virus from the same source, and that the original source was almost certainly northern Italy. Based on detailed investigations of individual travel histories and movements within the ski area, it’s fairly clear that the majority of infected people got infected at one of two bars (the second bar being the site of a good friend of the index patient).
It’s not a question of whether super-spreading events are possible with coronavirus: we have very solid evidence of numerous examples. The question is: why haven’t there been more observable events so far, i.e. in the Pacific NW? There’s no clear answer here, but the most obvious explanation is stochastics: they didn’t occur because the virus is not evenly disseminated across the world. It may well have gained a critical mass of local transmission in northern Italy much sooner than in the US or elsewhere. Or there may be some other confounding factor we haven’t identified yet.
My point isn’t to say that the discrepancies are definitely due to super-spreading, but that the hypothesis is plausible enough that we can’t rule it out when deciding which restrictions to roll back.
I’m not sure what you mean by “strange, high fatality rate” in the NYC area. The crude fatality rate (i.e. total deaths/total cases) is fairly consistent across several key states in the northeast vs. west coast:
NY: 4.5%
CT: 4.3%
MA: 2.8%
WA: 4.7%
CA: 2.9%
I’m assuming you’re referring to the fact that there has been so much less transmission on the west coast? If that’s the case, I refer you back to my previous comment.
Your expectation is pure speculation based on no solid data.
First off, remember that the virus doesn’t immediately spread homogeneously throughout a country as large as the US, which is what your comment implies. Also remember that just because one infected person travels from state A to state B while another travels from state A to state C at the same time doesn’t mean the dynamics of transmission in state B and state C will at all resemble each other for the first few weeks – a lot depends on random factors, like how many people each index patient comes into contact with.
Also, we all know that testing in the US has been spotty and sparse at the best of times; back when the total case numbers were still in the low triple digits, testing was even more sparse. So any assumption of case loads early in the epidemic based on actual figures is questionable at best.
Indeed it is. And that’s why there’s a public database with almost 6,000 SARS-CoV-2 sequences: https://www.gisaid.org/
One reason I keep saying the “two bugs” hypothesis is less likely is because we have more information on the viral sequence than almost anything else in this pandemic. And on top of that, because the data is publicly accessible and fairly easily analyzed, we can be fairly confident that thousands, if not tens of thousands of people, have looked at this data and tried to discern any meaningful patterns.
If there was an obvious correlation between viral sequence and transmission dynamics, we would be seeing numerous reports backed by copious data, not just a quick hypothesis by a former epidemiologist.
Yes, it was speculation at the time. I checked and posted on this after my comment (here). It seems pretty accurate, though I was using state-level data and not more local data.
Sorry, I used bad terminology. In my sentence, I used “fatality rate” to mean fatalities per 100,000, not fatalities per infection.
I would say my point still stands: we will likely never have a good sense of how many people were actually infected in mid- to late-February in any state in the US, because our testing regime flat-out sucked at the time and we don’t have enough tools to determine those rates retrospectively.
I didn’t mean to dispute that such transmission can occur at specific events and locations. My comment was directed toward the hypothesis that such events in Italy caused the wider spread, which I found implausible because we have similar events in the US.
If your hypothesis is that sporting events caused a big spread in Italy, but not in the US, because the virus was already more widespread in Italy, this begs the question of why it was so much more widespread in Italy in the first place. I realize that it might be earlier introduction. It might be necessary to analyze this at a more local level to get meaningful answers.
Understood. But here’s the issue: if the west coast was indeed being predominantly infected by a strain with such different transmission properties than the east coast virus, we would also expect to see some difference in the fatality rates between the two coasts. Maybe there is also some huge testing gap between the west coast and the northeast, but for now I still find this hypothesis less likely (although certainly not implausible) than the simple stochastic explanation. Both are obviously speculative.