Factors associated with Covid-19 deaths in records of 17M adult NHS patients
twitter.com
twitter.com
This is interesting data. I want to see Vitamin D status included in a large population study like this because I've been following two smaller studies covering about a thousand cases total that shows Vitamin D deficiency has a risk ratio of 10 to 20 (more even than being age 80+ in the above study). The studies:
[1] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3585561
[2] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3571484
Table 1 in each of those papers show Vitamin D status vs Outcomes. The correlation between Vitamin D status, where Normal is >30 ng/ml (or 75 nmol/L), and death rates is stark.
From [1] (which had n=780 cases) here is the punchline: "98.9% of Vitamin D deficient cases died while only 1.1% of them were active cases. 87.8% of Vitamin D insufficient cases died while only 12.2% of them were active cases. Only 4.1% of cases with normal Vitamin D levels died while 95.9% of them were active cases."
From [2] (which had n=212 cases) here is the punchline: "Of the 212 (100.0%) cases of Covid-2019, 49 (23.1%) were identified mild, 59 (27.8%) were ordinary, 56 (26.4%) were severe, and 48 (22.6%) were critical (Table 1). Mean serum 25(OH)D level was 23.8 ng/ml. Serum 25(OH)D level of cases with mild outcome was 31.2 ng/ml, 27.4 ng/ml for ordinary, 21.2 ng/ml for severe, and 17.1 ng/ml for critical."
Note: the classification for outcomes was "(1) mild – mild clinical features without pneumonia diagnosis, (2) ordinary – confirmed pneumonia in chest computer tomography with fever and other respiratory symptoms, (3) severe – hypoxia (at most 93% oxygen saturation) and respiratory distress or abnormal blood gas analysis results (PaCO2 >50 mm Hg or PaO2 < 0 mm Hg), and (4) critical – respiratory failure requiring intensive case monitoring."
I want to see a dozen more studies like [1] and [2] to see if this holds up to replication with larger populations.
Wait, only 13.3% (1.1 + 12.2) of all the active cases in total were vitamin D insufficient or worse? I thought the majority of people had insufficient vitamin D.
This data doesn't fully make sense to me. How come that so few people with a vitamin D insufficiency were active cases? I get that most of them died. But I don't get that so few, less than 15 percent, got to become an active case.
Edit: ah, it's an Indonesian study. Nevermind, I wouldn't know the numbers from there.
Without sun exposure, you'd have to be eating several pounds of fatty fish every day to get to "healthy" levels from diet alone. You need a supplement.
Keep in mind, non-severe Vitamin D deficiency doesn't cause any significant evolutionary pressure, so migrating humans just rolled with it - to this day.
Not really. We should be blasting "get your Vitamin D supplements" from the rooftops, because maybe that's a significant factor in COVID-19 and it's safe and people should get them anyway. There's no downside.
If in a few years it then turns out that Vitamin D had nothing to do with it, it doesn't matter. If it turns out that Vitamin D really was a big factor, we'll be sorry for being hesitant with a safe and cheap intervention.
one cause of vitamin D deficiency is the overuse of sunscreen
sunscreen should be used as protection against sunburn, but not put on all the time in every instance
be outside, stop lathering yourself with sunscreen and get natural, low continuous levels of vitamin D.
supplements may be better than nothing but they far from providing you with the same dose and availability of vitamin D
The generally safe amount is 4000 IU daily, which should get almost everybody to healthy levels. However, if your doctor finds that you have low levels, he may prescribe far higher amounts for a while.
Vitamin D is an easily manufactured supplement, it's not medicine that some people absolutely have to take every day.
Also, we're heading into summer and the sun is still rising every day.
> However the main issue that I think we want to prevent is luring people into a false sense of security that causes them to decrease the proven preventive measures.
We've had the same discussion about masks, so I think that one's settled.
> We don't need large groups of people gathering on beaches because they think the sun will be a miracle cure and they don't have to worry about getting sick.
If you don't want people gathering on the beaches, prohibit it. Having said that, the risk of infection outdoors is generally considered to be quite low.
> So if you want to personally start taking it without changing your behavior, be my guest. But any advice to take it needs to be clear that the evidence it helps is still very suspect at this point.
The evidence that Vitamin D supplementation works is there and it has already been recommended for everyone with low serum levels, well before COVID-19.
The problem is that people often aren't even aware that their serum levels are low, so that is highly prevalent - especially in the elderly and people with darker skin.
The difference between masks and vitamin D is that one is proven to work and one might work. If vitamin D doesn't help as much as people here are claiming, there is a real downside if people are behaving as if it works. That applies to my arbitrary example of beach gatherings as well as other behavior the government has no hope of preventing.
Once again, this doesn't necessarily mean that vitamin D doesn't help. We just need better data. The early indication was that hydroxychloroquine helped too and we all know how that turned out.
Hand washing and social distancing appear to be the most important. That means soap and 6+ feet for at least the next 2 years. That is definitely doable. I don't see masks as a long-term strategy.
When the primary vector of spread is through water droplets expelled while talking, coughing, or sneezing, I would think the importance of reducing the travel distance of those droplets would be self evident.
It's likely to be causal.
also I was out shopping recently and they seemed to sell many varieties of D3, but I didn't see any other D vitamins offered for sale. Is D deficiency really D3 deficiency? or have the other D vitamins sold out?
https://ods.od.nih.gov/factsheets/VitaminD-HealthProfessiona... (excellent series for all vitamins)
Different manufacturers = varieties? It's the same stuff or should be.
The last thing in the world a cancer patient needs is well nourished cancer cells, most grow and reproduce quickly.
There are a number of other studies on this. From what little I've read, if the effect is real,it seems the it is relatively weak, but still worth doing.
It may explain why heavily indoors NYC had so many lethal cases, even as most infected were already keeping themselves indoors.
Also, SARS 1 ended at the beginning of June 2002. There is clearly a correlation that needs to be looked into.[2]
[1] https://www.medscape.com/answers/302460-86798/what-are-the-s...
[2] https://en.wikipedia.org/wiki/Severe_acute_respiratory_syndr...
TBH I am not super solid on every one of these details.
“ Coronavirus update Consider taking 10 micrograms of vitamin D a day to keep your bones and muscles healthy.
This is because you may not be getting enough vitamin D from sunlight if you’re indoors most of the day.
There have been some news reports about vitamin D reducing the risk of coronavirus. However, there is no evidence that this is the case.”
[1] https://www.nhs.uk/conditions/vitamins-and-minerals/vitamin-...
"No evidence" makes it sound like there is actually no evidence. There is evidence - it just got linked in the two comments that make up the root of this thread.
> COVID-19 appears to disproportionately affects black and minority ethnic individuals. The underlying mechanism is unknown.
> One potential mediator could be their higher prevalence of apparent vitamin D deficiency.
> We explored whether blood 25 hydroxyvitamin D (25(OH)D) concentration was associated with COVID-19 risk.
> We found no evidence that (25(OH)D) explains susceptibility to COVID-19 infection, either overall or between ethnic groups.
[2] is pretty scary. As a 30 year old I know that my chance of death is low, but I’d like to avoid a “severe” or “critical” case. There are a large number of unpleasant states between “fine” and “dead” that I’d like to not experience. A 50/50 chance of it being no big deal are worse odds than I’d prefer.
[1] is sending me back out onto the patio this weekend for tanning.
Get some cod liver oil, drink a tablespoon every day. Does wonders.
https://www.cebm.net/covid-19/vitamin-d-a-rapid-review-of-th...
Unsurprisingly there's no direct clinical evidence.
"We searched for trials and didn't find any"
Unsurprisingly, there haven't been any (completed) trials on this particular combination of novel disease and possible prevention/treatment yet.
I'm concerned that people conflate "no evidence for efficacy" with "it doesn't work" instead of "we don't know if it works". You always start out with "no evidence".
There are studies that put serum Vitamin D levels against COVID-19 outcomes and they show a highly significant correlation. That's not "causal evidence", but it should put you on alert, you shouldn't be waiting for 2021 for possible Vitamin-D trials to complete.
As the authors of that review are saying, you should be supplementing anyway, whether there is a causal relationship or not.
Until they correct for socio-economic factors it's a potentially misleading bit of data.
Absence of evidence because you haven't tried much/ there's no data yet is not the meaningful kind.
‘We found no trials of vitamin D in COVID-19 that have reported results.’
‘As our searches returned no relevant results, [...]’
But keep on posting the paper though.
"There is some evidence that daily vitamin D3 supplementation over weeks to months may prevent other acute respiratory infections, particularly in people with low or very low vitamin D status. This evidence has limitations, including heterogeneity in study populations, interventions, and definitions of respiratory infections that include upper and lower respiratory tract involvement."
"The current advice is that the whole population of the UK should take vitamin D supplements to prevent vitamin D deficiency. This advice applies irrespective of any possible link with respiratory infection."
Although vitamin D is not mentioned in this study, it does show that black people have a hazard ratio of about 2 compared to whites. In the UK black people would be more vitamin D deficient than white people.
Now that I think about it, isn't the potential for vitamin D deficiency the reason there are white people in the first place?
“Some people will not get enough vitamin D from sunlight because they have very little or no sunshine exposure.” .. “ If you have dark skin – for example you have an African, African-Caribbean or south Asian background – you may also not get enough vitamin D from sunlight.”
Yes. Pale skin has better vitamin D production in low light, while black skin avoids skin damage even in extreme sunlight. Skin color is just evolution finding the right trade-off for the local climate.
Might not be something I'd do every day forever, but I could see doing daily samples for a week to get a feel for aboroption, then occasionally to check how my supplements and sun exposure are working.
If you could build a thing that can measure several different parameters depending on which test strips you load into it, I'd buy a pack of vitamin D strips, a pack of vitamin B, a pack of testosterone, and pack that looks for markers of inflammation. Heck yeah. My doctor would eat that data right up.
I haven't found and answer to this, if Vitamin D is such big factor then why aren't vitamin D deficient populations in the subcontinent ( india/pakistan ..) being hit hard.
Aging reduces vitamin D production in skin. There is a decrease in the concentration of 7-dehydrocholesterol in the epidermis in old compared with young individuals and a reduced response to UV light, resulting in a 50% decrease in the formation of previtamin D3
Here's a fantastic explanation of why you should not let your brain tell you that you're learning about things that cause Covid complications and death:
It might even purify your second-hand smoke.
but yes, you have a very real point. I remember reading about some drug that seemed to give people heart attacks, but it actually cured what it was supposed to and heart attacks were the next statistically likely thing to get you.
Imagine people at a bar. The smokers would stand outside in the cold, smoking, while the non-smokers inside are warmer but at higher risk of catching an infection.
Seems more likely that nicotine itself might have a direct interaction, or that there is no benefit at all, and it's all just noise in the data.
You're much more likely to have lung cancer if you smoke than if you don't -- but once you DO have lung cancer, a history of smoking makes keytruda more likely to work.
https://duckduckgo.com/?q=Ben+Goldacre+of+Bad+Science&t=fpas...
However, we are seeing differences, for example with diabetes.
17M just provides the baseline of "normal". Tons more data isn't needed to get that right. 170M wouldn't make a difference. I bet 1.7M wouldn't make a big difference either.
What we really need to know right now is how the Covid deaths behave. All that this offers is the 5,683 sample, but the title inflates it to emphasize the 17M. We need the 5,683 sample to be merged with other samples to reach 56,830, etc. That is the significant thing.
Sadly there will be more data as there are more deaths, and this team will be able to pull more analytics out. However 5,683 is a big enough sample size to find some interesting information on hazard rates.
For the team doing the analysis, plugging data analytics into 17 million primary care records is definitely a significant achievement. Medical records are intended for use by doctors and letting researchers into them has confidentiality issues. The process they followed was covered in the paper and seemed thorough.
> People from Asian and black groups are at markedly increased risk of in-hospital death from COVID-19, and contrary to some prior speculation is only partially attributable to pre-existing clinical risk factors or deprivation
Early in the outbreak I remember seeing some research that indicated people of Asian descent (East Asian) are genetically predisposed to more severe cases of this coronavirus, something to do with higher expression of the ACE-2 receptor.
It's unclear to me (only skimmed the figures) if the various ethnic/racial hazard factors correct for poverty (deprivation) or not.
> People from Asian and black groups are at markedly increased risk of in-hospital death from COVID-19, and contrary to some prior speculation this is only partially attributable to pre-existing clinical risk factors or deprivation;
Asian in Britain will primarily be Indian/Pakistani descent.
I wonder if this is vitamin d related? Darker complexion of black and Asian could cause higher risks of vitamin d deficiency in UK.
> Asian/Asian British
> 9. Indian
> 10. Pakistani
> 11. Bangladeshi
> 12. Chinese
> 13. Any other Asian background, please describe
I'd expect more it is attributable to less access to premium healthcare (remember, in the UK everyone gets healthcare, but not everyone gets privately-sponsored add-on healthcare)
"As a public health emergency, COVID-19 is being treated through the public health system."
https://www.bupa.co.uk/coronavirus/health-insurance
Not that private healthcare providers are actually providing resources to the NHS so its possible that you could end up in a private hospital even though you don't have private health insurance.
But the endocrine system is ferociously complex - maybe endocrine abnormalities are a proxy for obesity, or an underlying thyroid issue etc.
According to [0], vitamin D deficiency is associated with higher free testosterone in men and lower estradiol in women. Higher testosterone is associated with cardiovascular problems, which are a big COVID risk factor.
But that doesn't jib well with the study you find, so I dunno what to make of it.
0. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5218632/#__ffn_...
https://www.iihs.org/topics/fatality-statistics/detail/gende...
> The number of driver fatal crash involvements per 100 million miles driven in 2017 was 62 percent higher for males (2.1 per 100 million miles traveled) than for females (1.3 per 100 million miles traveled). Rates were substantially higher for males than for females ages 16-29, but were only slightly higher for ages 30 and older. The gender difference was largest among drivers ages 20-29.
So in some sense the real figure is 1.62x for men as women per mile driven, although (1) men driving more miles is still a real thing and as a result more men die, and (2) I don't know to what extent super-distance drivers (like commercial truck drivers, who skew heavily male) are putting their finger on the scale.
If you ignore teenagers, that difference comes out to 1.59x; and leaving out persons <30, 1.46x. Still much higher than 1.0.
This distinction might be important if there is some sort of vitamin-D/skin-melanin correlation to outcomes in northern latitudes.
Surely the geographical relationship between dense cities and their inhabitants demographics is more than enough to explain it.
> The overall cumulative incidence of COVID-19 hospital death at 80 days from the study start date was <0.01% in those aged 18-39 years, rising to 0.35% and 0.17% in men and women respectively aged ≥80 years, with a trend by age.
How do those rates compare to the seasonal flu in the UK?
If you could preselect a group of people that you know had covid and measure their death rate, the numbers are likely to be much higher. This study did not do this. This is probably because this would require that all patients are tested for covid as they visit their doctor and this was obviously not happening.
You can be skeptical, but it does seem to be the case that flu deaths is actually a made up number.
https://blogs.scientificamerican.com/observations/comparing-...
It would also be fantastic if you could provide a credible source for this surprising claim that US flu deaths are greatly overstated.
What data is this based on?
https://www.gov.uk/government/statistics/annual-flu-reports
The last bad hit on the hospitals was in 2017-2018
"Through the USISS mandatory scheme, a total of 3,454 ICU/HDU admissions of confirmed influenza were reported across the UK from week 40 2017 to week 15 2018, including 372 deaths, based on combined data from England, Scotland and Northern Ireland."
The death rate from anybody admitted to ICU from flu seems to be about 10%.
Admission rates for flu during that bad season was 0.22 per 100k
From what little data there is on Covid-19 it looks like that admission rates to ICU are about 0.30 per 100k, with a wide range (0.20 to 0.40). We won't know the average until it is over.
About 30% die (and again the range is anything up to 50%).
The UK data for Covid-19 in any depth is incomplete.
https://www.medrxiv.org/content/10.1101/2020.05.06.20092999v...
Obesity, of course, is a major factor. It's hard to pull apart Type II Diabetes and breathing disorders from it, too, so a realistic assessment of disorders that would be mitigated if the patient lost weight might show it as the absolute major contributor.
Even Vitamin D deficiency is linked to obesity: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3705328/
Part of what makes me start again is the desire to avoid spooking others with those flu-like symptoms in the middle of an epidemic, since I live in a densely populated area.
I don’t have a clear idea of what mechanism is at work here, my hypothesis is that the withdrawal somehow suppresses my immune system.
Unfortunately, as far as I can tell, nicotine patches aren’t as good (or useful at all) at reducing the respiratory symptoms in short term. I suppose it’s not purely nicotine withdrawal that’s causing this.
The good news is that these symptoms are temporary.
Public health communications on the benefits of cessation don’t usually mention this. I wonder if that’s a mistake. I would have thought it puts a lot of people off quitting, and maybe if they knew what was happening it would be easier to deal with.
When I have a week or so apart from the crowds, I’ll see how long it takes to push through those symptoms.
(I thought if withdrawal causes immune system suppression, this effect could explain the apparent COVID statistic correlation mentioned by 0-_-0 in parent comment, but from your information it’s unlikely to.)
(This would explain why former smokers are worse off, since they have residual lung damage but no protective nicotine.)
Of course this is all quite hypothetical, and may turn out to be wrong, but I gather it’s biochemically plausible.
In the paper they used this index which measures deprivation by living location in the UK: https://en.wikipedia.org/wiki/Multiple_deprivation_index
Don't mean to sound like a pedant here even though it totally reads that way. I figured with enough epidemiology in the news these days, people may see a lot of "Hazard ratios" or "Cox regression" or "survival analysis" and be at risk of some confusion. I work with these concepts and I get tripped up myself sometimes.
What does this mean? Sleep deprivation?
https://en.wikipedia.org/wiki/Multiple_deprivation_index
For example suppose you are very poor but you live where I grew up in Metroland. Lack of money sucks pretty badly, but in Metroland other deprivation factors are mostly low.
There are jobs in Metroland. The schools are pretty good. There's a hospital, and a nearby doctor with a friendly manner (remember Britain has universal healthcare, but universal does not mean universally excellent). You can't afford private rent here, but local government can find you somewhere to live, and not in six months but maybe six days.
The local crime isn't too bad. Vandals, maybe the odd burglary you suspect a neighbour is on the game (prostitution), but you don't fear for your life. The surrounding space is nice, there's a play park, rolling hills, fresh air if you like that sort of thing.
So in Metroland even the poor aren't very deprived.
In contrast, in some deprived Northern cities in particular, even the middle class suffer pretty badly, and will strive to get out.
It’s an interesting counter factual to imagine how differently the media would be covering the situation if the genders were reversed and women had about twice the chance of dying from COVID-19..
- Bullet point
- After bullet point
- In the form of a tweet per bullet point
- Trying to skip the bullshit
- And find the conclusion
(I'll stop now) ...and having the anxiety to be disappointed that I'll read so much bullshit and end up with nothing. Yes, in the end he links to the abstract which has the (unfortunately harder to parse) meat, but also yes, the Twitter thread disappoints, because he ends up tooting his own horn that they were the first, "this [study is] extremely important" and how he needs a sandwich... and seemingly that's where the thread ends? (TBH, thank god for that).
a) the thread, in readable form https://threadreaderapp.com/thread/1258372975004389379.html
b) the full publication https://www.medrxiv.org/content/10.1101/2020.05.06.20092999v...
So:
!thread https://twitter.com/bengoldacre/status/1258372975004389379
TL:DR; civilisation's patterns (cities, trade, travel) co-evolve diseases. Populations tend to evolve resistance to local endemic disease, but are susceptible to novel agents, including foreign infections. Rome infected China and vice versa.
Cystic fibrosis is a predominantly Caucasian genetic disorder because one copy of the gene helps protect against certain historical plagues in Europe without being too debilitating. But two copies kill you gruesomely at a young age. And it kills you primarily by making you more prone to infection.
Kind of like how Sickle Cell works. It helps protect against malaria, but two copies of the gene are bad news.
In western countries at least, an obvious confounded for those of asian or black ethnic descent is _poverty_ and thus access to healthcare, malnutrition, sleep deprivation, poor education and everything else that goes with poverty.