COVID-19

Discussions about serious topics, for serious people
Locked
User avatar
lpm
Junior Mod
Posts: 5944
Joined: Mon Nov 11, 2019 1:05 pm

Re: COVID-19

Post by lpm » Mon Sep 13, 2021 10:49 am

Thread on who is still dying from Covid.
https://twitter.com/john_actuary/status ... 6521977860

The splits for deaths of the fully vaccinated are interesting - a return to the male/female difference we speculated about way back in Feb 2020. To a very very strong degree, Covid can now only kill the old, particularly the vulnerable old, plus immunocompromised - and those who opt in to Covid mortality by refusing the vaccine.

Image
⭐ Awarded gold star 4 November 2021

User avatar
Woodchopper
Princess POW
Posts: 7057
Joined: Sat Oct 12, 2019 9:05 am

Re: COVID-19

Post by Woodchopper » Fri Sep 24, 2021 7:24 am

Interesting from the NYT,

Since the pandemic began, Covid has often followed a regular — if mysterious — cycle. In one country after another, the number of new cases has often surged for roughly two months before starting to fall. The Delta variant, despite its intense contagiousness, has followed this pattern.

[…]

We have asked experts about these two-month cycles, and they acknowledged that they could not explain it. “We still are really in the cave ages in terms of understanding how viruses emerge, how they spread, how they start and stop, why they do what they do,” Michael Osterholm, an epidemiologist at the University of Minnesota, said.

But two broad categories of explanation seem plausible, the experts say.

One involves the virus itself. Rather than spreading until it has reached every last person, perhaps it spreads in waves that happen to follow a similar timeline. How so? Some people may be especially susceptible to a variant like Delta, and once many of them have been exposed to it, the virus starts to recede — until a new variant causes the cycle to begin again (or until a population approaches herd immunity).

The second plausible explanation involves human behavior. People don’t circulate randomly through the world. They live in social clusters, Jennifer Nuzzo, a Johns Hopkins epidemiologist, points out. Perhaps the virus needs about two months to circulate through a typically sized cluster, infecting the most susceptible — and a new wave starts when people break out of their clusters, such as during a holiday. Alternately, people may follow cycles of taking more and then fewer Covid precautions, depending on their level of concern.

[…]

There have been plenty of exceptions to the two-month cycle around the world. In Brazil, caseloads have followed no evident pattern. In Britain, cases did decline about two months after the Delta peak — but only for a couple of weeks.
https://www.nytimes.com/2021/09/01/brie ... eload.html

raven
Catbabel
Posts: 645
Joined: Mon Nov 11, 2019 8:58 pm

Re: COVID-19

Post by raven » Fri Sep 24, 2021 9:15 pm

The second plausible explanation involves human behavior. People don’t circulate randomly through the world. They live in social clusters, Jennifer Nuzzo, a Johns Hopkins epidemiologist, points out. Perhaps the virus needs about two months to circulate through a typically sized cluster, infecting the most susceptible — and a new wave starts when people break out of their clusters, such as during a holiday. Alternately, people may follow cycles of taking more and then fewer Covid precautions, depending on their level of concern.
Ding, ding, ding. We have a winner.

Social networks are granular, behaviour changes.

Herainestold
After Pie
Posts: 2029
Joined: Mon Nov 25, 2019 1:23 pm

Re: COVID-19

Post by Herainestold » Fri Sep 24, 2021 11:43 pm

This "two month surge" rule reminds me of the stuff that technical stock market analysts do. Take some kind of complex cyclical phenomenon that no one understand and look for some kind of cyclical function that you can superimpose on it,and then claim that explains everything.

We don't know what drives covid epidemics well enough to predict it, we can only make educated guesses.

Just about every optimistic prediction has been wrong. The countries that have done best have imposed robust lock downs.Vaccines have worked very well to reduce high death rates, but have been less successful, in stopping transmission. Especially against Delta.

Immunity wanes, people get complacent, new variants are more virulent.

Don't throw your masks out yet. You will need them.
Masking forever
Putin is a monster.
Russian socialism will rise again

User avatar
bob sterman
Dorkwood
Posts: 1123
Joined: Mon Nov 11, 2019 10:25 pm
Location: Location Location

Re: COVID-19

Post by bob sterman » Sat Sep 25, 2021 9:22 am


Since the pandemic began, Covid has often followed a regular — if mysterious — cycle. In one country after another, the number of new cases has often surged for roughly two months before starting to fall. The Delta variant, despite its intense contagiousness, has followed this pattern.

[…]

There have been plenty of exceptions to the two-month cycle around the world. In Brazil, caseloads have followed no evident pattern. In Britain, cases did decline about two months after the Delta peak — but only for a couple of weeks.
https://www.nytimes.com/2021/09/01/brie ... eload.html

So it follows a regular 2 month cycle - except in the plenty of countries where it doesn't. Hmmmm.

User avatar
sTeamTraen
After Pie
Posts: 2551
Joined: Mon Nov 11, 2019 4:24 pm
Location: Palma de Mallorca, Spain

Re: COVID-19

Post by sTeamTraen » Sat Sep 25, 2021 9:29 am

bob sterman wrote:
Sat Sep 25, 2021 9:22 am
https://www.nytimes.com/2021/09/01/brie ... eload.html

So it follows a regular 2 month cycle - except in the plenty of countries where it doesn't. Hmmmm.
Now turn to page 7 for your horoscope!
Something something hammer something something nail

User avatar
sTeamTraen
After Pie
Posts: 2551
Joined: Mon Nov 11, 2019 4:24 pm
Location: Palma de Mallorca, Spain

Re: COVID-19

Post by sTeamTraen » Sat Sep 25, 2021 9:34 am

lpm wrote:
Mon Sep 13, 2021 10:49 am
Thread on who is still dying from Covid.
https://twitter.com/john_actuary/status ... 6521977860

The splits for deaths of the fully vaccinated are interesting - a return to the male/female difference we speculated about way back in Feb 2020.
I suspect that the excess mortality for males there may be being driven by the fact that 84yo men are just less healthy than 84yo women, as evidenced by the difference in life expectancy between the two. I would expect about twice as many 84yo men vs women to die in any given month even without Covid.
Something something hammer something something nail

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: COVID-19

Post by shpalman » Sat Sep 25, 2021 9:54 am

bob sterman wrote:
Sat Sep 25, 2021 9:22 am

Since the pandemic began, Covid has often followed a regular — if mysterious — cycle. In one country after another, the number of new cases has often surged for roughly two months before starting to fall. The Delta variant, despite its intense contagiousness, has followed this pattern.

[…]

There have been plenty of exceptions to the two-month cycle around the world. In Brazil, caseloads have followed no evident pattern. In Britain, cases did decline about two months after the Delta peak — but only for a couple of weeks.
https://www.nytimes.com/2021/09/01/brie ... eload.html

So it follows a regular 2 month cycle - except in the plenty of countries where it doesn't. Hmmmm.
https://twitter.com/nathanwpyle/status/ ... 7475208193
future.jpeg
future.jpeg (139.89 KiB) Viewed 2203 times
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: COVID-19

Post by shpalman » Sat Sep 25, 2021 10:00 am

sTeamTraen wrote:
Sat Sep 25, 2021 9:34 am
lpm wrote:
Mon Sep 13, 2021 10:49 am
Thread on who is still dying from Covid.
https://twitter.com/john_actuary/status ... 6521977860

The splits for deaths of the fully vaccinated are interesting - a return to the male/female difference we speculated about way back in Feb 2020.
I suspect that the excess mortality for males there may be being driven by the fact that 84yo men are just less healthy than 84yo women, as evidenced by the difference in life expectancy between the two. I would expect about twice as many 84yo men vs women to die in any given month even without Covid.
I did once find some Europe-wide life expectancy at 65 data.
life-expectancy-at-65.png
life-expectancy-at-65.png (109.77 KiB) Viewed 2201 times
Mainly because I didn't really like this "higher death rate because more old people and old people are less healthy" argument. Or in other words what matters for your covid fatality probability isn't the number of years since you were born but the number of years until you would have died anyway.

But life expectancy of course only makes sense at the population level.
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

User avatar
sTeamTraen
After Pie
Posts: 2551
Joined: Mon Nov 11, 2019 4:24 pm
Location: Palma de Mallorca, Spain

Re: COVID-19

Post by sTeamTraen » Sat Sep 25, 2021 10:08 am

At the age of 10 I read a factoid with activities carrying a 1 in 1 million risk of death. I vividly remember that "being a 60 year old man for 15 minutes" was on the list. Now that I am a 60 year old man, I'm aware that I'm dodging several bullets, albeit mostly fired from a considerable distance, every day.

I was looking up the mortality figures for the US by age a few weeks ago, for some other purpose, and I checked. It seems to hold: 1 in a million per 15 minutes is about 1 in 10,000 per day, and indeed about 2.8% (10,000/365) of US 60 year old males die per year, as do 2.8% of 65 year old females. (One should of course use one of those formulas with a couple of "1 divided by" clauses to calculate this, but for small chances it's close enough to add rather than multiply.)
Something something hammer something something nail

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: COVID-19

Post by shpalman » Sun Sep 26, 2021 10:57 am

Norway to end coronavirus-related restrictions.

Norway's full vaccination coverage has only recently caught up with the rest of us but it seems like first dose rates are picking up again there. They recently had a bit of a peak in case rates which is coming down, but still higher than Italy for example. However, their death rate is extremely low.

At this point I'm frustrated that Italy's death rate is still so high compared to the case rate. The CFR here is still about 1% as compared to 2% in the wave last winter before vaccination became available; it's about 0.4% in the UK. Cases and hospitalizations are still coming down though, although of course if things are getting better in Sicily it's not because they went from White to Yellow but because everyone went back to work in the north.

Well, we'll see how the next couple of weeks plays out.
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

User avatar
Woodchopper
Princess POW
Posts: 7057
Joined: Sat Oct 12, 2019 9:05 am

Re: COVID-19

Post by Woodchopper » Mon Sep 27, 2021 2:35 am


Quantifying impacts of the COVID-19 pandemic through life-expectancy losses: a population-level study of 29 countries

Abstract
Background
Variations in the age patterns and magnitudes of excess deaths, as well as differences in population sizes and age structures, make cross-national comparisons of the cumulative mortality impacts of the COVID-19 pandemic challenging. Life expectancy is a widely used indicator that provides a clear and cross-nationally comparable picture of the population-level impacts of the pandemic on mortality.

Methods
Life tables by sex were calculated for 29 countries, including most European countries, Chile and the USA, for 2015–2020. Life expectancy at birth and at age 60 years for 2020 were contextualized against recent trends between 2015 and 2019. Using decomposition techniques, we examined which specific age groups contributed to reductions in life expectancy in 2020 and to what extent reductions were attributable to official COVID-19 deaths.

Results
Life expectancy at birth declined from 2019 to 2020 in 27 out of 29 countries. Males in the USA and Lithuania experienced the largest losses in life expectancy at birth during 2020 (2.2 and 1.7 years, respectively), but reductions of more than an entire year were documented in 11 countries for males and 8 among females. Reductions were mostly attributable to increased mortality above age 60 years and to official COVID-19 deaths.

Conclusions
The COVID-19 pandemic triggered significant mortality increases in 2020 of a magnitude not witnessed since World War II in Western Europe or the breakup of the Soviet Union in Eastern Europe. Females from 15 countries and males from 10 ended up with lower life expectancy at birth in 2020 than in 2015.

Key Messages

This is the first study to assemble a high-quality data set of harmonized mortality estimates, life tables and age by cause decomposition for 29 countries representing most of Europe, Chile and the USA to provide novel evidence of the cumulative, comparative impacts of the pandemic on population health.

Out of 29 countries analysed, the COVID-19 pandemic led to losses in life expectancy in 27, with large losses of life expectancy of >1 year in 11 countries for males and 8 among females.

Losses in life expectancy observed in Central and Eastern European countries in 2020 exceeded those observed around the dissolution of the Eastern Bloc (with the exception of Lithuania and Hungary), whereas similar magnitudes of losses in Western Europe were last seen around World War II.

Compared against recent trends, females from 15 countries and males from 10 ended up with lower life expectancy at birth in 2020 than in 2015—a year when life expectancy was adversely impacted already due to an especially bad flu season.

Losses in life expectancy were largely attributable to increased mortality above age 60 years and linked to official COVID-19 deaths.
This should end the arguments that the deaths were of people who died with Covid and would have died anyway.

Probably won’t though.

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: COVID-19

Post by shpalman » Mon Sep 27, 2021 8:28 am

Woodchopper wrote:
Mon Sep 27, 2021 2:35 am

Quantifying impacts of the COVID-19 pandemic through life-expectancy losses: a population-level study of 29 countries...

... Males in the USA and Lithuania experienced the largest losses in life expectancy at birth during 2020 (2.2 and 1.7 years, respectively), but reductions of more than an entire year were documented in 11 countries for males and 8 among females. Reductions were mostly attributable to increased mortality above age 60 years and to official COVID-19 deaths...
This should end the arguments that the deaths were of people who died with Covid and would have died anyway.

Probably won’t though.
If the main result is a reduction of life expectancy of 1-2 years then people who died of covid would have died anyway the following year, or in the worst case, the year after that.

Yes I'm sure that's not the correct way to interpret these results, but then, everyone dies eventually, and your risk of death from covid was comparable to your chance of otherwise dying in the next n years anyway where n was a not necessarily very large number.
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

User avatar
Woodchopper
Princess POW
Posts: 7057
Joined: Sat Oct 12, 2019 9:05 am

Re: COVID-19

Post by Woodchopper » Mon Sep 27, 2021 8:47 am

shpalman wrote:
Mon Sep 27, 2021 8:28 am
Woodchopper wrote:
Mon Sep 27, 2021 2:35 am

Quantifying impacts of the COVID-19 pandemic through life-expectancy losses: a population-level study of 29 countries...

... Males in the USA and Lithuania experienced the largest losses in life expectancy at birth during 2020 (2.2 and 1.7 years, respectively), but reductions of more than an entire year were documented in 11 countries for males and 8 among females. Reductions were mostly attributable to increased mortality above age 60 years and to official COVID-19 deaths...
This should end the arguments that the deaths were of people who died with Covid and would have died anyway.

Probably won’t though.
If the main result is a reduction of life expectancy of 1-2 years then people who died of covid would have died anyway the following year, or in the worst case, the year after that.
Surely that would only occur if Covid was the only cause of death. As people are dying of everything else as well, a reduction of 1-2 years in total would mean that Covid caused people to lose many more years.
shpalman wrote:
Mon Sep 27, 2021 8:28 am
Yes I'm sure that's not the correct way to interpret these results, but then, everyone dies eventually, and your risk of death from covid was comparable to your chance of otherwise dying in the next n years anyway where n was a not necessarily very large number.
Probably not Elizabeth 1 wrote:All my possessions for a moment of time.
Saying that people are going to die anyway sounds nice from the perspective of someone who probably has many decades left. An infection cutting someone's life expectancy from, say, five years to one year would probably seem very harsh to them.

User avatar
bob sterman
Dorkwood
Posts: 1123
Joined: Mon Nov 11, 2019 10:25 pm
Location: Location Location

Re: COVID-19

Post by bob sterman » Mon Sep 27, 2021 8:57 am

Woodchopper wrote:
Mon Sep 27, 2021 8:47 am
shpalman wrote:
Mon Sep 27, 2021 8:28 am
If the main result is a reduction of life expectancy of 1-2 years then people who died of covid would have died anyway the following year, or in the worst case, the year after that.
Surely that would only occur if Covid was the only cause of death. As people are dying of everything else as well, a reduction of 1-2 years in total would mean that Covid caused people to lose many more years.
Indeed - estimates of average years of life lost for each COVID death have been around 10-15 years.

E.g.

https://www.nature.com/articles/s41598-021-83040-3

User avatar
Sciolus
Dorkwood
Posts: 1313
Joined: Mon Nov 11, 2019 6:42 pm

Re: COVID-19

Post by Sciolus » Mon Sep 27, 2021 10:10 am

Around 1 in 10 deaths are from covid*, each covid death is 10 years early, therefore overall loss of life expectancy is 1 year. There might be some second-order effects but I think that's a sensible calculation (all numbers approximate for simplicity).

*Sounds about right given roughly 1 million deaths per year and 100,000 covid deaths per year in the UK.

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: COVID-19

Post by shpalman » Mon Sep 27, 2021 10:13 am

bob sterman wrote:
Mon Sep 27, 2021 8:57 am
Woodchopper wrote:
Mon Sep 27, 2021 8:47 am
shpalman wrote:
Mon Sep 27, 2021 8:28 am
If the main result is a reduction of life expectancy of 1-2 years then people who died of covid would have died anyway the following year, or in the worst case, the year after that.
Surely that would only occur if Covid was the only cause of death. As people are dying of everything else as well, a reduction of 1-2 years in total would mean that Covid caused people to lose many more years.
Indeed - estimates of average years of life lost for each COVID death have been around 10-15 years.

E.g.

https://www.nature.com/articles/s41598-021-83040-3
So if you catch COVID your risk of death over the next month or so is similar to the risk that you would have (non-accidentally) died over the next 10-15 years.

This is more effective than saying that COVID existing in your country somehow shortens your life by 1-2 years, which is more a statement about how many people in your country suffered from serious COVID and is only valid at the population level.
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

User avatar
Sciolus
Dorkwood
Posts: 1313
Joined: Mon Nov 11, 2019 6:42 pm

Re: COVID-19

Post by Sciolus » Mon Sep 27, 2021 10:52 am

shpalman wrote:
Mon Sep 27, 2021 10:13 am
So if you catch COVID your risk of death over the next month or so is similar to the risk that you would have (non-accidentally) died over the next 10-15 years.
No, that's not correct. If you *catch* covid, your risk of death is similar the risk that you would have died in the next 1 year (that it is 1 year, about the same as the population LLE, is coincidence). If you *die* of covid, you will on average have lost 10 years of life expectancy, because that's the LE of an 80-ish-year-old which is the age at which most covid deaths occur.
This is more effective than saying that COVID existing in your country somehow shortens your life by 1-2 years, which is more a statement about how many people in your country suffered from serious COVID and is only valid at the population level.
Both are useful. One is the conditional probability at an individual level, which tells you about the nastiness of an infection. The population LLE tells you about both spread and national response to the pandemic.

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: COVID-19

Post by shpalman » Mon Sep 27, 2021 11:10 am

Sciolus wrote:
Mon Sep 27, 2021 10:52 am
shpalman wrote:
Mon Sep 27, 2021 10:13 am
So if you catch COVID your risk of death over the next month or so is similar to the risk that you would have (non-accidentally) died over the next 10-15 years.
No, that's not correct. If you *catch* covid, your risk of death is similar the risk that you would have died in the next 1 year (that it is 1 year, about the same as the population LLE, is coincidence). If you *die* of covid, you will on average have lost 10 years of life expectancy, because that's the LE of an 80-ish-year-old which is the age at which most covid deaths occur.
Covid kills lots of older people because their LE at the moment right before they caught covid was shorter than the years covid takes off a person's life.* And it kills fewer younger people because they aren't actually that likely to otherwise die over the next 10-15 years.

Alternatively the "your risk of dying of covid is equivalent to your otherwise risk of death over the next n years" has a value of n which depends on age and so the statement isn't actually that useful.

* - this is probably a poor way to put it, because we don't expect that covid shortens a person's life expectancy even if they "fully" recover from it, and we won't know that until the younger cohort with infection in their medical history start to reach ages similar to the normal life expectancy. Or, alternatively, fail to reach those ages.
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

User avatar
Sciolus
Dorkwood
Posts: 1313
Joined: Mon Nov 11, 2019 6:42 pm

Re: COVID-19

Post by Sciolus » Mon Sep 27, 2021 11:31 am

I'm afraid I don't understand what you're trying to say. My statements were a reflection of the unusually strong age-related severity of covid, and would not apply to other causes of death with a very different age-risk profile. The severity of covid is nothing to do with LE as such, it just so happens that the age-related change in covid severity is approximately proportional to the age-related change in all-causes risk of death (which is why, for covid, "If you *catch* covid, your risk of death is similar the risk that you would have died in the next 1 year" works); that is not the case for most other infections AFAIK and there is AFAIK no a priori reason why you expect it to be the case. Replace "covid" in my post with, I dunno, "scarlet fever" and it would be completely wrong.

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: COVID-19

Post by shpalman » Mon Sep 27, 2021 12:04 pm

Ok let me put it this way. If I, in my mid-40's, were to catch covid and die, then we could say that it took 30-40 years off my life because I would have been expected to live until whatever the life expectancy is around here.

(Let's say we're doing the calculation a year ago before vaccines became available; vaccination seems to reduce a person's risk of serious outcomes in a manner equivalent to being 30 years younger.)

However, my chance of dying of covid if I'd caught it would actually probably have been very low, since I don't have any obvious risk factors. Not having any risk factors is another way of saying (for covid at least) that I'm extremely unlikely to have died of natural causes in the next 10-15 years anyway. In other words, if I had such bad risk factors so as to be at non-trivial risk of dying of covid, my life expectancy probably wasn't 30-40 years. You have to compare me with other people of my age of a similar health profile to me.
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

User avatar
Woodchopper
Princess POW
Posts: 7057
Joined: Sat Oct 12, 2019 9:05 am

Re: COVID-19

Post by Woodchopper » Mon Sep 27, 2021 12:18 pm

shpalman wrote:
Mon Sep 27, 2021 12:04 pm
In other words, if I had such bad risk factors so as to be at non-trivial risk of dying of covid, my life expectancy probably wasn't 30-40 years. You have to compare me with other people of my age of a similar health profile to me.
Maybe not as much as you assume. People are often able to carry on living for decades with chronic medical conditions that would still reduce their chances of surviving pre-vaccination Covid. Obviously that doesn't apply to things like aggressive cancers.

Herainestold
After Pie
Posts: 2029
Joined: Mon Nov 25, 2019 1:23 pm

Re: COVID-19

Post by Herainestold » Tue Sep 28, 2021 3:12 am

I am curious how Long Covid will affect life expectancy.
Masking forever
Putin is a monster.
Russian socialism will rise again

Chris Preston
Snowbonk
Posts: 529
Joined: Tue Nov 12, 2019 8:05 am

Re: COVID-19

Post by Chris Preston » Tue Sep 28, 2021 10:22 pm

Results from UK infection survey

Some interesting outcomes. Looks like protection from the AZ vaccine is not as good as the Pfizer, but fades less so that after 90 days there is no significant difference. Having previously had COVID-19 offers no more protection than vaccines.

Refusing to wear a mask modestly increases risk, but that may be tied to other risky behaviours.

Interacting with a lot of, particularly young, people increases risks.
Here grows much rhubarb.

User avatar
Woodchopper
Princess POW
Posts: 7057
Joined: Sat Oct 12, 2019 9:05 am

Re: COVID-19

Post by Woodchopper » Wed Sep 29, 2021 2:40 am


Smoking and COVID-19 outcomes: an observational and Mendelian randomisation study using the UK Biobank cohort

Abstract

Background Conflicting evidence has emerged regarding the relevance of smoking on risk of COVID-19 and its severity.

Methods We undertook large-scale observational and Mendelian randomisation (MR) analyses using UK Biobank. Most recent smoking status was determined from primary care records (70.8%) and UK Biobank questionnaire data (29.2%). COVID-19 outcomes were derived from Public Health England SARS-CoV-2 testing data, hospital admissions data, and death certificates (until 18 August 2020). Logistic regression was used to estimate associations between smoking status and confirmed SARS-CoV-2 infection, COVID-19-related hospitalisation, and COVID-19-related death. Inverse variance-weighted MR analyses using established genetic instruments for smoking initiation and smoking heaviness were undertaken (reported per SD increase).

Results There were 421 469 eligible participants, 1649 confirmed infections, 968 COVID-19-related hospitalisations and 444 COVID-19-related deaths. Compared with never-smokers, current smokers had higher risks of hospitalisation (OR 1.80, 95% CI 1.26 to 2.29) and mortality (smoking 1–9/day: OR 2.14, 95% CI 0.87 to 5.24; 10–19/day: OR 5.91, 95% CI 3.66 to 9.54; 20+/day: OR 6.11, 95% CI 3.59 to 10.42). In MR analyses of 281 105 White British participants, genetically predicted propensity to initiate smoking was associated with higher risks of infection (OR 1.45, 95% CI 1.10 to 1.91) and hospitalisation (OR 1.60, 95% CI 1.13 to 2.27). Genetically predicted higher number of cigarettes smoked per day was associated with higher risks of all outcomes (infection OR 2.51, 95% CI 1.20 to 5.24; hospitalisation OR 5.08, 95% CI 2.04 to 12.66; and death OR 10.02, 95% CI 2.53 to 39.72).

Interpretation Congruent results from two analytical approaches support a causal effect of smoking on risk of severe COVID-19.

https://thorax.bmj.com/content/early/20 ... 021-217080

Locked