I just wanted to say that I'm really grateful that you're still posting Scott :). I check my email multiples times a day to see if there's a new post, and seeing a notification from ACX makes the day that much better
Another great example showing why science is actually pretty hard to do well, even when we are trying to answer questions that appear simple at first glance.
If it's so difficult to answer whether a major vitamin helps with one of the most significant diseases we've seen in decades a full year in, one should consider to what extent they should place strong confidence in their convictions in similar areas with similar scientific processes, for example nutrition and diet (these are also very hard, like most areas, and you should exercise extreme epistemic humility and adaptability, imo)
One more point: a significant proportion of readers will have been staying inside (so away from sunlight) much more than they did before in the past 11 months; just think about not commuting to work everyday.
This will have reduced their (already on average pretty bad, for UK and some US states) level of Vitamin D... in which case a supplement would hurt them even less than the very little theoretical hurt it did in those pre-2020 studies.
I also side with "take Vitamin D supplements". My parents have also been taking one since November because they trust me (and their doctors are fine with them taking it).
Could seasonal, latitude, and minority trends not be linked directly toward indoors exposure and multigenerational households?
Individuals are much more likely to congregate outdoors during the summer than the winter. This corroborates with the known facts that outdoors exposure poses a much lesser risk to corona than indoors exposure. In addition, we know that cold air carries the virus droplets much better than warm air. It seems to me that these facts would almost entirely explain the seasonal and latitude trends.
To explain the "vitamin d" trend in skin color, asian and african american households are much more likely to be intergenerational and/or involved in service jobs.
Regarding vitamin D pharmacokinetics: the actual active form is calcitriol, which has three hydroxy groups. This is produced by a series of enzymatic hydroxylations starting with cholecalciferol (1 hydroxy), and then calcifediol (2 hydroxy groups). So giving people calcifediol is going to act more quickly than cholecalciferol.
It would be interesting to know what were the criteria for ICU admission in the Indian study, as well as the baseline O2 saturation in both groups. 50% of the patients needing ICU care is a little unexpected
Doesn't vitamin D have a build-up time and decay time, in the body. If so, this filter effect would rule out instant-on attempts to use it (e.g. hospital data)
There's at least one thing you get from sunlight that isn't vitamin D-- people have subcutaneous nitric oxide precursors. The really fun part is that you won't find this out by studying mice. Mice are not only covered with fur, they prefer going out at dusk and pre-dawn, so they don't have that body system.
Nitric oxide and sunlight for people was discovered as a result of studying why there were more heart attacks in the UK than in Australia.
I read a bit about this when I had a patient with GFAP encephalomyelitis who had a dramatic recovery after getting sunburned. I was really pissed off with the staff who let a very cognitively impaired man sit in the sunshine and get sunburned. But then he recovered, from literal psychosis and profound memory and executive impairment, to seeming fairly normal, and Neuropsych testing confirmed this. I wasn’t sure if it was vitamin D or not.
So I read about MS, because GFAP encephalomyelitis is incredibly rare, and found that some researchers were arguing that the latitude effect is due to UV rather than Vitamin D itself, and had a mouse model of multiple sclerosis showing mice under UV did better than either controls or vitamin D mice. I can try to find the reference again if you like.
Obviously my sample was n=1 of an extremely rare disease, but it got me thinking.
If it was UV, could that explain the latitude effect but also why vitamin D supplementation studies with good designs had mixed results.
"Do the benefits of taking a Vitamin D supplement at a normal dose equal or outweigh the costs for most people?"
I'd say the percentage is closer to 100%.
There have been numerous studies showing that low vit D levels are correlated with increased risk of depression, diabetes, cancer and influenza (though I haven't looked at all of these studies with the rigor or Mr Alexander so who knows after you disect all of these studies). Anyway, I tell my patients to take it.
"For example, the Spanish study used a version of Vitamin D called calcifediol; the Brazilian one used a slightly different version called cholecalciferol. Calcifediol becomes active more quickly than cholecalciferol, enough so that if I were in the business of defending the Spanish study I might argue that it takes a few weeks for cholecalciferol to work, so giving it to someone who will already be dead or recovered by then is meaningless."
Biology is perversely complicated enough that this could indeed matter. When I was looking up magnesium supplements to take, there are ten different types https://www.mindbodygreen.com/articles/magnesium-supplement-types and depending what you want to achieve (ease sore muscles? help blood pressure? treat constipation?) then the type you should be taking differs.
And looking up about vitamin D types, the differences between the two types seem to be:
"The recommended form of vitamin D is vitamin D3 or cholecalciferol. This is the natural form of vitamin D that your body makes from sunlight. Supplements are made from the fat of lambs' wool.
However, a clinical study reported in 2008 suggested that vitamin D2 works as well as vitamin D3.
Many supplements contain vitamin D as vitamin D2 or calciferol. It's derived from irradiated fungus."
Irradiated fungus? Funky, if we want to turn into Godzilla I suppose?
"Vitamin D2 and D3 are not equal when it comes to raising your vitamin D status.
Both are effectively absorbed into the bloodstream. However, the liver metabolizes them differently.
The liver metabolizes vitamin D2 into 25-hydroxyvitamin D2 and vitamin D3 into 25-hydroxyvitamin D3. These two compounds are collectively known as calcifediol.
Calcifediol is the main circulating form of vitamin D, and its blood levels reflect your body’s stores of this nutrient.
For this reason, your health care provider can estimate your vitamin D status by measuring your levels of calcifediol.
However, vitamin D2 seems to yield less calcifediol than an equal amount of vitamin D3.
Most studies show that vitamin D3 is more effective than vitamin D2 at raising blood levels of calcifediol.
For example, one study in 32 older women found that a single dose of vitamin D3 was nearly twice as effective as vitamin D2 at raising calcifediol levels.
If you are taking vitamin D supplements, consider choosing vitamin D3."
So the Spanish study was using the inferior (?) D2 yet got good results, while the Brazilians were using the superior (?) D3 and found no difference.
Yeah, I'm puzzled too. It *may* be that the Spanish patients, who were already hospitalised when they received the supplementation, were *so* badly off for Vitamin D that anything was beneficial. The only difference I could come up with was that the Spanish study seems to demonstrate that it "drastically reduced" admissions to ICU, while the Brazilian one says it didn't reduce length of stay:
"A small randomized controlled trial in Córdoba, Spain, of calcifediol (25-hydroxyvitamin D3, or 25(OH)D3) for hospitalized COVID-19 patients (henceforth, “the Córdoba study”) found dramatic reduction in the need for ICU admission. This study has been viewed as a small preliminary study, suggesting at most that further study might be warranted. It has gotten relatively little attention, though its strengths and weaknesses were discussed in this article, and a Bayesian cost/benefit analysis found the expected benefits, in terms of lives saved and severe illness avoided, of immediately adopting the treatment protocol were considerably higher than the expected costs. Two other randomized trials of vitamin D formulations for COVID-19 had mixed results: a small trial in India found a high dose of vitamin D3 shortened time to viral clearance in asymptomatic or mildly symptomatic SARS-CoV-2 individuals, and a larger study in Sao Paulo, Brazil did not find a statistically significant benefit of a high dose of vitamin D3 in hospitalized COVID-19 patients."
"Conclusions and Relevance Among hospitalized patients with severe COVID-19, vitamin D3 supplementation was safe and increased 25-hydroxyvitamin D levels, but did not reduce hospital length of stay or any other relevant outcomes vs placebo. This trial does not support the use of vitamin D3 supplementation as an adjuvant treatment of patients with COVID-19."
So *maybe* it's "Vitamin D supplementation if you're hospitalised with Covid-19 means you're less likely to be admitted to ICU, but you won't get out of hospital any sooner than someone with Covid-19 that did not have to be admitted to ICU and wasn't getting Vitamin D supplementation". That is, "if you're run-down/deficient in Vitamin D, that makes it worse if you get Covid-19 and have to be hospitalised, but taking Vitamin D won't stop you getting Covid-19 although it will make it less likely you'll be so badly off you need to go to ICU".
Could there be something like an undershoot effect involved, where people who would ordinarily have lots of vitamin D because they spend a lot of time outside and live close to the equator actually have much less because they're locked down and stay inside instead, which does... something... to their immune systems?
One of the challenges with using serum vitamin D levels as an input variable is that we don't really know whether those levels are a cause or an effect; possibly they are a confounding factor when included with other inputs. Example: although patients with higher vitamin D levels may have fewer M.I.s, we can't produce that result by supplementing vitamin D in the diets of low-vitamin-D patients. See, e.g., https://jamanetwork.com/journals/jamacardiology/fullarticle/2735646.
Well sunlight seems way more promising, so why bother studying Vitamin D?
So here's my experimental design for testing the "something something sunlight"-hypothesis:
Set up a hospital wing with a bunch of UV lights, and another with UV lights and infrared and another with just infrared and treat only people with a specific shade of dark skin in them for Corona. [uhm.... also a control group would be nice]
If there's no difference from any of the light setups, we..... are very confused.
Maybe not enough light.
If we see improvements from all rooms except UV-lights only, we know it's not vitamin D.
If we see.... ugh, lighting is hard, make that guy design the experiment:
"Despite their superficially lighter skin, they have Vitamin D deficiencies almost as bad as blacks." Unless I'm missing something, the linked study included a total of 60 Asians all from the greater Boston area. Obviously not very representative of ~4.7 billion Asians in the world or even the Asian population in the U.S. I would point to larger studies in Japan (e.g https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3755751/) that appear to point to significantly lower vitamin d deficiency than the US population and much lower than the US black population.
My understanding is that hydroxychloroquine is ineffective on its own but helpful when combined with other antiviral medication, thus possibly explaining some significant variation in study results (since many studies actually did combine it with other things):
Interesting review. I only suggest that you consider British epidemiologist, Edgar Hope-Simpson's search for a parsimonious explanation in his study: The transmission of epidemic influenza (1992). There is a clear connection between these seasonal phenomena. Here's a very interesting follow up on EHS's comprehensive work in the BioMed Cenral Virology Journal. Flu is still a surprisingly mysterious phenomenon in many ways. There are probably some read across matters to SARS-CoV-2. The intriguing relationship with VitD and even it's centrality is also well covered. Worth checking out the early C20th methods for transmission research if nothing else. There are some very interesting questions and conclusions that challenge some givens about these phenomena.
Good article but the Brazil and Spain studies were carried out solely among hospitalised COVID-19 patients. The India study is very interesting because it indicates that Vitamin D helps mild cases recover faster.
The next step should be to find out whether Vitamin D can help potentially moderate-to-severe cases remain mild and therefore whether it can prevent hospitalisations in the first place. Perhaps it can be given to a sample group of people of all ages who have never had COVID-19 and who haven't been vaccinated and this group can be compared to a control group.
So anything wrong with just taking L-arginine and L-citrulline as supplements to get nitric-oxide levels up? Like, if I'm going to take a bunch of random stuff sunlight-associated stuff like Vitamin D, I might as well be a completionist about it.
Perhaps at least a portion of the seasonal and latitudinal variation is due to that COVID (and many viruses) are rather vulnerable to ultraviolet.
Thus, the seasonal variation in daily insolation would be the true causal driver, with Vitamin D correlating for the obvious reason.
If this holds water, I would expect the degree of seasonal variation to be weaker closer to the equator, although this would probably be weakened because the directly-caused variation will generate more cases that will spread to the equatorial regions anyway...
Also, I would expect the degree of the effect to vary based on the amount of UV reaching the ground - so if the data exists, plotting the health of ozone layer against the degree of seasonal variation might be informative....
On Zvi's advice, I took 4000 IU of Vitamin D per day for about 2 months.
Then I got kidney stones, and one of the doctors I chatted with said it could likely be from the Vitamin D, and I reasoned that I'd stored up enough Vitamin D (I think it gets stored in fats?) that I should stop taking it. Because kidney stones was no fun.
Laboratory tests should be designed so that around 5% of tests are reported as abnormal. I worked for almost 30 years as a hospital pathologist, and, without doing a formal study, we were reporting way more than 5% of our vitamin D levels as abnormal. Before going off and making any recommendations, I would want to work on the test and make sure that there are no issues with it. Working in a community hospital, I never had the resources to do that sort of large study, but I still have the nagging feeling that something is amiss here, and I am not sure what it is.
As for outcomes in people of East Asian descent, I know through friends in the Chinese community that many people of Chinese descent remembered the original SARS epidemic and changed their behavior way before the government was willing to admit that there was a problem. That confounder will somehow have to be corrected for before drawing any conclusions about outcomes in those patients.
I have been taking the maximum recommend daily dose of 4,000 units Vitamin D3 since Feb 2020 on recommendation from Doctors at the Harvard Medical School. Presumably Vitamin D interacts with ACE2 which was identified as the viral binding site. So presumably not having a Vitamin D deficiency can be beneficial with respect to possibly reducing the risk of becoming infected.
There are a lot of people who definitely do need more Vitamin D because they're deficient, and there are a lot of other people who don't take Vitamin D when they should (or take less than they should) because they're incorrectly worried about taking too much. If you might be in either group, you can ask your doctor for a Vitamin D test, or just buy one for $50 here:
So let’s pretend that I’m a rube who don’t hold with all that high falutin’ college talk about n=stuff and chemistry and randomized study sessions and such.
What is wrong with blurring the conclusion enough for common sense to assert itself? “Vitamin D makes you healthier, and healthy people handle the ‘Rona better.”
If somebody did the exact same studies testing Vitamin C instead of D, do you reckon there would be a similar spread along the effective-ineffective spectrum?
Most online retailers in the UK seem to sell tablets with 4000IU each, and 6000 seems common in the US. Then the Brazilian trial you mentioned used a single megadose 200,000IU (!)
I find it hard to square the NHS's cautious position with what seems like a wide range of dose choices in practice. Any idea what gives?
I would also look at the enthusiasm for Vitamin D in a broader context. Over the past 75 years there has been sequential excitement and enthusiasm for different vitamins/minerals as panacea for many human diseases. Cycles of Vitamin C (the great Pauling first one), various B vitamins, Selenium, Zinc, Vitamin E (for a long time). Medical specialities in diverse fields (immunology, oncology, ID) all simultaneously become excited by the new trendy vitamin/supplement. Initial small studies often show dramatic benefit in whatever metric is being assessed. Sadly, almost universally when large, controlled studies are done, there is great disappointment (and sometimes even negative benefit). Vitamin D is the latest of these cure-all supplements. You could probably look at a graph of medical literature/reports on each of these magic answers, and Vitamin D is certainly trending. But skepticism is in order. But hey, it probably can't hurt...
My general physical health got a lot worse when I first started staying at home all the time (I've literally left my house/yard only twice since last March, and I think it's been a month or so since I've been in sunlight). I had back pain, wrist pain, joint cracking, and fatigue at various points (I'm in my mid 20s). Since I started taking vitamin D or other multivitamins, I have basically not had any such issues. I expect that my COVID avoidance is way more dramatic than most people's, but I think most people have been getting less sunlight than they used to, and if so, Vitamin D seems great for improving their general physical health.
Please correct: >>we can be pretty sure that their COVID is caused by Vitamin D.<< ==> >>we can be pretty sure that their COVID is caused by lack of Vitamin D.<<[and delete this commentary]
My strong, strong prior here is that Vitamin D may be a decent biomarker, but it's a lousy therapy. It's been thrown at nearly every indication under the sun in RCTs, and almost always returns null results. To me, the most plausible explanation is that low Vitamin D levels are correlated with biological states (malnutrition, sedentary living, inflammation) that predispose to poor outcomes, but short-term Vitamin D supplementation does not address the underlying predisposition.
*grumbles about how "vitamin" is not a meaningful category, and, in the case of cholecalciferol derivates, we're actually talking about hormonal replacement instead of vitamin supplementation*
Great summary. A common "bad take" submission I've never run is along the lines of "the race disparity is explained due to vitamin D." Probably not, but maybe a little. It just strikes some people as a completely bonkers suggestion, so I get sent it regularly.
Also gets tarred by association because same people pitching also hock zinc and HCQ. (And now ivermectin, another "bad take" I can't pull the trigger on.)
1) Mendellian randomization and the Biobank studies are both invalid because people supplement in response to genetically low vitamin D, which violates one of the assumptions underlying MR. This is especially true since government guidance in the UK changed pretty dramatically on vitamin D supplementation
2) If you're looking at a racial case for vitamin D, you should really look at immigrant pairs like African immigrants to Sweden, or Indian immigrants to the UK. In all cases you see low Covid rates in poor sunny countries and high covid rates among immigrants from those countries to rich northern countries. This is the Ricketts pattern and strongly indicates vitamin D's involvement. https://shotwell.ca/posts/africa-covid/
3) The problem with the Brazillian study is that vitamin D was given way to late. If it's active against Covid it's because it works basically as an anti-viral, in that Brazillian study they gave vitamin D on average 10 days after symptom onset by which time the virus had cleared. All the supplementation studies where vitamin D has been given in or before the viral replication phase have been positive.
As an aside, I think one of the biggest unanswered questions about COVID-19 is why so few people in equatorial Africa have died. Equatorial South America has done pretty terribly (Manaus = 3.1190°S), and outcomes from the first world indicate that African ancestry does not confer a survival advantage.
Seems for individuals, the first question is whether one might be borderline vitamin-D deficient to begin with (little sunlight, milk-free diet for some reason, etc.). It might help more for Covid in such folks (and only in such folks?). And all sorts of below-the-radar effects of low-D may well get solved besides.
This article seems to miss discussion on the most important claim I've heard around Vitamin D supplementation: that daily Vitamin D supplementation substantially reduces your risk of contracting severe cases of covid.
This is quite different from a) vitamin D reduces your risk of getting covid (why would it) or b) taking vitamin D when you get covid improves outcomes.
My understanding is that *it takes a long time to increase serum vitamin D levels through supplementation.* See this article from the NCBI: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3820059/ . In short: infrequent, low doses (1-2k IU) of vitamin D do not raise serum vitamin D. Dose and frequency (per week) mattered more than duration, but duration was at least a month, not 7 days! And duration effects were increasing - the longer you took vitamin D, the bigger the effect!
Scott dismisses this entire argument by saying "people who can do anything consistently for a year will do better" without actually analyzing it in more depth. No, you actually have to take vitamin D supplements frequently over time (at least one month, with effects increasing up to 6 months), at higher dosage for them to have an effect. Dismissing the underlying mechanism as being a selection effect, means that Scott misses the best argument in favor of long-term vitamin D supplementation.
This hypothesis may be wrong - maybe that research paper is incorrect, maybe there's more evidence behind Scott's handwave - but right now it seems to me that Scott is ignoring the crux of the debate: can long-term vitamin D supplementation reduce your odds of a severe covid case, conditional on you getting covid? Scott unfortunately does not address this question, which is the most important one, IMO.
Given that long-term vitamin D supplementation does increase serum vitamin D levels significantly more than short-term vitamin D supplementation (at least according to the research), I am inclined to believe the research does suggest that vitamin D supplementation can reduce the odds of contracting *severe* covid cases much more than Scott does.
To what degree can it be that people with low vitamin D are more conscientious (indoor workers) and therefore have a lower risk of getting infected?
Seems like the studies shouldn’t control just for race, but also for behavior.
I’d say low vit D white people will be more likely to get infected due to low vit D, but less likely to get infected due to behavior. Which could easily mask the relation between vit D and covid incidence completely.
It would be better to look at severity for this reason, since it’s presumably affected less by conscientiousness than incidence — but it still *is* affected. Conscientious people can do other things that decrease the severity.
I think this qualifies as an enema recommendation, though perhaps there's enough evidence in favor that it doesn't. An old joke:
An actor collapses on stage, and a doctor is found in the audience. As he tends to the actor, an old lady calls from the cheaper seats: "Give him an enema!" She is ignored. She cries again, "Give him an enema!", and again, nobody pays attention. After a third "Give him an enema" is heard, the doctor turns, and says, "Ma'am, I assure you, giving this man an enema would do him no earthly good whatsoever." The lady replies, "Well, it couldn't hurt."
I think you took the wrong side regarding the Spanish Vs Brazilian trials. The Spanish one was peer reviewed. It was better conducted with more activated form of vitamin D. It is totally not wise to give D3 to already hospitalised person. It takes precious time to be converted to the active form. The Spanish trial also was small achieved remarkable p value. This is the unbiased estimator which means that no matter it was small study, the results proves causality. Patients were divided for the true treatment or placebo randomly by a computer. How can one argue against this method ...
The latitude thing might be exterior UV killing COVID on surfaces. In Asia, there's been a buying spree for indoor UV lamps to provide some viral load reduction indoors, complete with timers and motion sensors and shut off whenever a person walks into a room.
It would be useful to check the methodologies utilized for measuring vitamin D levels in these studies, since it seems that the measured value for 25-hydroxyvitamin D (predominant in testing) largely depends on differences in vitamin D binding protein (VDBP) between individuals. Ginsburg & coworkers identified a vitamin D metabolite ratio (VMR) which is dependent on concentration and not on the amount of VDBP that is present in an individual.
> The VMR was independent of VDBP concentration, whereas VDBP was strongly directly associated with the individual vitamin D metabolite concentrations. Prior studies evaluating only 25(OH)D3 may have been confounded by absence of data on VDBP status.
"The Vitamin D Metabolite Ratio Is Independent of Vitamin D Binding Protein Concentration", Clinical Chemistry, 2020, doi: 10.1093/clinchem/hvaa238 , https://pubmed.ncbi.nlm.nih.gov/33188595/
I can help with the seasonal pattern. The smallish bump in the summer was in the hotter, southern states primarily. The much larger bump in the winter was in the cooler, northern states primarily. Common theme? People spending lots of time indoors in climate controlled environments.
It looks like there's another randomized controlled trial in the works for Vitamin D supplementation being planned, from Wang et al. (N = 2700): https://doi.org/10.1016/j.cct.2020.106176
That's a pretty substantial sample size. Can someone take a look at the trial design plan and see if it's reasonable? That isn't my skillset.
A veritable cacophony of Vitamin D studies. I'm frankly surprised you implicitly lend credence to the results. Given the file drawer problem, replication issues and paradoxes within your speculations, there's really no basis for your last 3 paragraphs except intuition/ guessing/ idiosyncratic heuristics. You certainly know that mechanistic thinking just doesn't go very far in medicine; and that case-control and other observational studies are sometimes useful but overall weak. To argue seasonal variation may occur related to sunlight/ vit D while later noting levels are usually stable over long periods is not consistent.
Those criticisms aside, your collection of studies just shows once again how poorly medicine and the lay public have been served by academia in a time of great need. Large well-designed studies capable of definitive answers are sorely lacking among the utter cacophony.
> And maybe some day, after countless false leads and stupid red herrings, one of the claims people make about this substance will actually pan out. Who knows?
One thing that always struck me as weird about vit D is how ubiquitous it is among health-conscious people
I wonder if th lis is a case of memes spreading with no reason, or of the studies missing an important factor.
It's been years since I looked at it but the kidney stone risk should be lowered by simultaneously supplementing vitamin K (which is another one of these quite likely positive supplements).
Regarding summer: sars2 is highly unstable when it comes to light, heathings dries the air which makes your mucous membrane dry and live inside has no wind. Source: christian drosten's podcasts.
My thoughts are that possibly low vitamin d could also simply be a marker of poor diet and lifestyle.
A possible downside of supplementation could be that it makes more postive change less likely.
I.e. more outdoor physical activity, combined with nutrient rich diet will likely* sort vitamin d levels, plus have a host of other benefits that would be missed by supps alone.
Supplements in general have a very poor track record, possibly as many people end up taking them as "substitutes" to a nutrient rich diet rather than supplements? (Purely speculative).
Given the choice between go outside, exercise and eat well, vs take a pill, many will opt for the pill. Which is fair enough providing we know the true efficacy of the pill and can make the trade-off. But I don't think this is the case (for vit d and covid or any other supp and health outcome).
* Possibly for darker skinned individuals, living at northern latitudes, working indoors all day it might be impossible to achieve optimum vitamin d levels through sun exposure and diet?
n all of this you haven't mentioned to what level of Vitamin D you are testing against. The level described as "deficient" for bone health isn't the one described as "deficient" for respiratory health (if that is indeed a thing).
If you want a null hypothesis to aim at, here's one
"The bottom line is that there is no downside to increasing our intake of vitamin D to maintain serum 25(OH)D at at least 30 ng/mL (75 nmol/L), and preferably at 40–60 ng/mL (100–150 nmol/L) to achieve optimal overall health benefits of vitamin D."
Reassess against that much higher than standard level and everybody can avoid talking past each other.
In this instance the obsession with the precautionary principle is to avoid hypercalcemia at all costs - even if that means people succumbing to respiratory illnesses. Something they didn't do, for some reason, with masks.
We know that getting covid is correlated with breathing it in. We can be fairly confident that better ventilation reduces the risk of covid. What do people do in warm sunny weather? Open windows, and go outside. We should expect the latitude and seasonal effects, even if vitamin D was nothing to do with it.
My intuitive take is that vitamin D has some effect on immunity but compared to bone health, it is much less pronounced. The bone mass is comparatively large and we need a lot of calcium therefore vitamin D deficit can quickly cause problems. Whereas the immune system can adapt with lower vitamin D levels.
Very low vitamin D levels in people could make the immunity much worse too but in practice we notice problems with the bones first and treat them accordingly. Supplementing subclinical vitamin D deficit provides only marginal benefits.
In conclusion, it is good to maintain vitamin D levels in the recommended range, and take supplements if you need to, mainly, to help your bones. Maybe it will have a secondary benefit by slightly improving your immunity too but don't rely on that too much.
Given that almost all of us get way less sunlight than our ancestors, I would advocate supplements anyway, to bring us up to the levels that our bodies are adapted to.
Vitamin D has low toxicity, so it's very unlikely that you err on the side of having too much.
The best explanation I've seen for the seasonal and latitude nature of the virus so far is indoor humidity. If the temperature is below 15 degrees C, the air indoors is heated and becomes dry. This causes droplets expelled from the mouth that would settle in more humid conditions to evaporate and become smaller and float around for longer providing a more effective transmission in confined spaces.
A leimilar but lesser effect can occur from airconditioning in summer.
In places with really sealed up indoor spaces the humidity is lowish in winter but not as low as some not as cold places because the less sealed up spaces have more cool outdoor air comes in to be heated. Places with winter daily maximums in the 5 to 15 degree range tend to have the dryest indoor air in winter.
Seasonality/latitude dependence could be completely explained by people gathering indoors vs. outdoors. Very hard to get aerosol transmission in a windy park.
Regarding the Mendelian randomization study: When you buy a 1/4" drill bit (or perhaps a 6.5 mm drill bit if you're into that sort of thing), it's not because you want to own a drill bit, but because you want to own some 1/4" holes. In the same way, we don't want vitamin D in our systems because the vitamin D molecules strut around the body punching corona viruses in their little faces, ripping apart cancer cells, and blocking the depression rays emitted from the dirty laundry pile. We want it in our system because it does something else. Perhaps the people with hereditary vitamin D deficiency have some other source for this, uh, "something else". Maybe they make their 1/4" holes with punches, lasers, or well-disciplined termites.
You should really clarify your definition of "Asian". That term covers a much, much broader range of latitudes than white or black does. Probably difficult to suss out small lineage differences in the entire subpopulation, but you could just focus in on two groups with different historical ranges, like Koreans and Indonesians. Or Native Siberians, if you can find data.
> The loose ends that bother me the most are the seasonal pattern, the latitude data, plus the increased risk of hospitalization and death in Asians. I don't have a great explanation for those.
Couldn't the first two be explained simply by the fact that during the winter in northern countries, people tend to meet indoors rather than outdoors? This facilitates the spread more.
"For example, infection rate from coronavirus seems latitude dependent; in general, the further north an area, the worse it's been hit."
This isn't obvious; the statistics from the southerly countries aren't as reliable as those from the northern ones, and it is known many southerly countries (Mexico, Peru, Bolivia, Ecuador, Indonesia) have been very hard hit. It was true during the first wave in the U.S. and at the beginning of the third, but Mississippi, Louisiana, Alabama, and Arizona have been some of the hardest-hit states. In the E.U., Spain has been hit rather badly. It is true that Russia, the Dakotas, and the U.S. Northeast have been very hard-hit, though, but the U.S. Northeast pandemic seems to have been due to its better connectedness, while the harsh pandemic in the Dakotas and Russia was likely due to lack of social distancing.
Does anyone have a quantitative idea, which proportion of the skin should be exposed to the sun for how long for normal cholecalciferol production, let's say for a maximally light-skinned bloke? Where I live, from october to april, the only exposed body part is face, and we get to see the sun every 3rd day or so. Is there any chance that we're still producing adequate amounts of cholecalciferol if we spend, say, 1 hour a day outdoors?
Lastly, I'm not sure where the idea came from that levels are stable over time. There are studies showing the 10-15yr old UK Biobank levels don't correlate well with recent levels. IIRC one correlation coefficient was around 0.4, which makes more intuitive sense in the first place. The idea of stability over time is reasonable in the face of no measurement or intervention/change, but the UK Biobank participants didn't have their levels hidden from them. It's pretty reasonable to assume those people who discovered they were deficient as a result of the UK biobank participation took steps to correct their deficiency. Their doctors certainly should have encouraged them to do so.
So trust Kaufman et al, Merzon et al, Meltzer et al more.
Dismissing the Indian RCT is inappropriate. This was small n but a well designed trial registered in advance and regardless of the importance of seroconversion clinically in terms of severity to individual patients, this data clearly shows a causal link whereby vitamin D affects SARS-CoV-2 infection.
Favoring the Brazil trial over Cordoba, Spain is also picking the wrong one. The Brazil trial didn't start giving patients D3 until they already needed supplemental oxygen (90% of them at baseline) because they were already 10 days post symptom onset and they were only average 7 days from discharge from hospital in terms of recovery. D3 doesn't ramp up fast enough to make a difference that late in infection is the only conclusion that can be drawn. See the comments of the medRxiv page of that Brazil study for many comments along these lines and other issues: https://www.medrxiv.org/content/10.1101/2020.11.16.20232397v1#disqus_thread
The Cordoba study used calcifediol which skips liver conversion and restores D levels within hours and thus had a much better effect. If they had used it in Brazil maybe the results would have been better. What's more, there's now a new preprint which has statistical issues but nonetheless adds to the positive data. It also used calcifediol and had 10x the patients (though the randomization was by hospital ward not patient, but it still makes the D helps conclusion more likely despite its flaws):
I'm not sure why the Brazil hospital in that study got the patients so late in the infection. Worse healthcare infrastructure in that country?
But the trend is clear: 3 of 3 randomized controlled intervention studies started at hospital admission in countries in which that hospital admission came before patients were so far along that they mostly needed supplemental oxygen all showed positive results. And 2 of 2 calcifediol studies showed extremely positive results.
You said, "in the past I have learned to side with negative RCTs rather than astoundingly massively positive ones when the two conflict." That's a reasonable default for clinical trials of still-on-patent pharmaceutical agents because those tend to be sponsored by the pharma company that has strong financial incentives. For trials of natural agents especially those run by non-industry groups, the opposite should probably be the default. Eg see: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7487184/
You said, "I am going to assume that whichever doctors ran these studies did not completely bungle them and use forms of Vitamin D that could not possibly have worked in the time period involved." Completely bungled in the Brazil study is a reasonable conclusion. It's not completely absurd that it could have worked in the time involved, and the study organizers may not have known how far from recovery the patients would be when they started, so we don't have to fault them for running it. But the late stage of initial administration compared with the time it takes for that version of D to ramp up means its negative result cannot be applied to initiating the same treatment earlier in infection, nor does it cast doubt on D for prophylaxis, nor is it negative for calcifediol which is much faster acting.
Thanks for posting this. Any opinions on N-Acetyl Cysteine (NAC) supplements? I have read/heard claims of reduced covid symptoms associated with it’s regular use.
How many scientists & doctors work on the NICE committee?
Here's a list of hundreds of scientists & doctors who say the evidence is convincing enough that it should be default to use vitamin D for covid as a precaution: https://vitamindforall.org/rollcall.html
It terms of credentials, these groups include many of the top vitamin D researchers in the world (based on citations), a former surgeon general, and many notable names.
Based on the prior 2 corrections I just posted in earlier comments (on Hastie & on the clear pattern from the RCTs) and this appeal to experts, you might want to update your % guesses on D's relevance. My personal guesses are 80% chance that D affects infection risk & 80% that it can help as treatment if given early enough in infection or as calcifediol. And 99% chance that benefits outweigh costs.
Correct reading of the evidence together with understanding the underlying biology makes the situation very clear. It just isn't quite clear at the level of approving a new pharmaceutical drug, and the big problem is that that is the evidentiary bar that NICE and other officials at 3-4 letter acronym bodies are applying. That's not the way to handle public health especially in an acute crisis. It's *astoundingly* clear that the balance of risks is such that increasing vitamin D intakes and reducing deficiency rates is a clear public health win net-net, especially now.
Good piece. Not mainly for the conclusion, which doesn't hinge on the analysis, but for an example of what a careful examination of evidence looks like.
"We assume that people with the genes for low Vitamin D in fact have low Vitamin D. And this isn't confounded by anything; we know their low Vitamin D is genetic. So if these people get COVID more, we can be pretty sure that their COVID is caused by Vitamin D."
Probably the assumption in the first sentence there is at least somewhat false, no? Since people can change their vitamin D levels by taking vitamin D.
Also, it could be that a genetic predisposition to low vitamin D is associated with some *other* genetic thing that has an effect on severity of covid infections (in either direction - so you could get an illusory effect if it's associated with something that makes covid worse, or an illusory absence of effect if it's associated with something that makes covid less bad). (I don't have any specific knowledge that this is the case, but it seems like a gap in the logic as presented here.)
Scott, I suspect it was delayed by the Blog closure but I've definitely been waiting for your thoughts on 'no evidence' and the other scientific messaging in this pandemic. The language used to scientific uncertainty to the general public needs a rethink and I suspect is the source of much public disinformation.
I don't blame Joe Public for thinking 'no evidence' for a *thing* means definitively that the *thing* doesn't work.
Vitamin D for patients admitted to hospital could cut Covid deaths by 60 per cent University of Barcelona preprint for the lancet not peerreviewed yet... I had low Vit D (22ng/ml) when last measured 4 yrs ago (after spending 45min everyday in the midday sun in June...... I started taking 2,000 iu but after reading a few articles associating low vit D levels with the worst outcomes in covid I started taking 10,000 iu (250mcg?) but will be getting tested after my 2nd vaccine shot (1 week) I am aiming for a blood level of 60-80.... Depending on your age and genetics it can take a lot to boost your levels up.... any blood level up to 150 is most likely safe
When the limit of 100 was decided on the actual research showed a level above 300 was toxic....but to err on the safe side an upper limit of 100 was recommended....
The issue was already analyzed probabilistically by Rootclaim. They are also offering a $100,000 bet on it. If you're at 25% it's easy money for you (but don't, you'll lose - your analysis is wrong on many points).
Somebody on facebook argued, that Hydroxychloroquine is still in the game. Because the trial, which said it didn't help was made by Surgisphere, which gave fraudulent data. (I a lazy to look it up on my own).
1) Vitamin D correlates with activity/exercise levels pretty strongly. Consider this as a serious confounder in any observational study.
2) The Mendelian Randomization study I have strong reservations about. To run MR, you need to make strong assumptions about the genetic variants that you are using. So MR makes sense in cases where you know a lot about the variants. But here they are doing a GWAS to get the variants, i.e. taking anything that associates with Vitamin D levels. There's no theoretical reason to assume a priori that these variants have the properties needed to run this MR. Instead, you do some statistical tests and hope you can catch it if it's a problem. But those statistical tests make their own assumptions that I don't find very plausible.
The authors know this and address it and do a reasonable job at that. They try restricting just to variants that lie near genes that are known to directly impact Vitamin D and see the same thing, and they toss out variants that are also known to impact other phenotypes and so could confound. But it's just the reality of the problem: MR is a bit questionable if you haven't study your variants carefully. The fact that they find a null result makes this all less concerning, though it is plausible that confounders could cancel out the true effect.
I always felt, right after reading the original speculation, that Vit. D is a prophylactic rather than a cure. Firstly, you should have been taking this as a supplement already. So, if you have a decent amount of Vit. D taken as supplements cruising in your bloodstream; it is not going to prevent you from getting infected; but do two things: 1) maybe increase the viral load required for infection, 2) when infected the symptoms won't be as severe.
All these studies that you referenced don't seem to contradict this, or am I reading you wrong?
I would note that in general, anyone taking Vitamin D3 regularly in amounts of say 4000 IU a day, would be well served to take Vitamin K2 (probably MK-4) and Magnesium as well. And look into a source of Vitamin A (cod liver oil).
I'm 75 years old, and decrepit enough to be unable to what I used to. Vitamin D pills at Sam's cost pennies per day, hence I have been taking megadoses of it since June, 2020. No downside, possible benefits.
What is the explanation for 80%+ of Covid positive patients in Spain being vitamin D deficient? I would also like to see a regional analysis of Vitamin D deficiency rates in Brazil. Also, what about the 50 other studies? Overall this analysis seems a little weak, although the lack of enthusiasm from authoritative institutions is surprising.
The post-symptom/post-positive RCTs that find a (gigantic!) effect use calcifediol - the form that oral cholecalciferol slowly turns into, over a ~week. The RCTs showing no effect use cholecalciferol, whose conversion into calcifediol won't outrace an infection. I don't blame some researchers for trying plain cholecalciferol - it's a cheap OTC supplement, and maybe megadoses would help! Meanwhile calcifediol is a prescription-drug, & expensive in the US. But assuming "assume that whichever doctors ran these [cholecalciferol] studies did not completely bungle them" is picking one set of researchers arbitrarily to respect, & ignoring the contrast signal between the two choices. Why not assume, conversely, that the researchers who tested calcifediol did not completely bungle their inclination to test something with distinct effects from cholecalciferol? (The effect sizes in the calcifediol studies are gigantic!)
D3 needs to be taken with K2 (prevents kidney stones and other calcium mismanagement effects), magnesium (increases resorbtion) and A (reduces side effects from K2 related enzymes). It also needs to be taken with oil. And in small doses (5-20k IU/day) over a long time, megadoses like 200k IU will have way less resorbtion. I've read dozens of studies, even very recent ones, that showed side effects like kidney stones or artery calcification because K2 was ignored. Never ever assume the people doing these studies are aware of these problems, assume the opposite unless its expressly mentioned in the study. Keep that in mind when facing supposedly simple questions like "did they give an oil-soluble form/gelcap or the much cheaper dry form and if its the latter, did they give it with oil or fatty food". "Surely they couldn't have fucked up this badly"- yes they could lol.
I became a paid subscriber just to post this comment. Here is a hypothesis, which explains the seasonal variance and the latitude data without vitamin D.
Red light.
A very specific wavelength of red light penetrates the skin very deeply and is absorbed by the respiratory enzyme cytochrome c oxidase, which is involved in the electron transport chain in mitochondria, causing an increase in intracellular ATP.
Red light therapy aka low-level laser therapy cures a lot of weird problems. This specific wavelength of red light is also component of sunlight. We get more of it during summer and near the equator.
Vitamin D could be a red herring. Sunlight can do lots of other things too.
Seasonality is an independent variable via temp and humidity, and one study neatly combined them in enthalpy, which you may remember from school! (How do I do hyperlinks?**)
You may find a bigger effect with selenium.
Vitamin D itself has no effect: it's levels in the blood of 25(OH)D3 and its metabolites that count.
The reason it doesn't work if vitamin D given after admission is that it takes 2 weeks for serum levels to rise, so that's way too late to do any good, so the Brazil study was pointless, and was anyway a preprint, and your "use in hospital" figure should maybe be lower than the other one. The Spanish study has the same problem and also didn't check serum levels at all (but did at least use calcifediol/25(OH)D3 ) and still found benefit! An even more metabolised form has been trialed in ICUs and that did indeed work better. Sorry can't find the trial now.
Many prevention trials don't test for D deficiency and have too few participants for significance. Also it takes about 2 years to do a population level prevention RCT, and it's really expensive and hard to get drug companies interested, so that's a tough ask. Prof Adrian Martineau got funding to do one.
I've been going back and forth on these Vitamin D studies. However, unlike HCQ vitamin D has no toxic side effects in moderate doses (<50,000 IU). It's cheap to manufacture. It seems like a no-brainer to use it as a prophylactic. Just don't confuse it with a cure. I'm taking 2,000 IU/day, which is above the generally recommended dose of 600 IU/day, but well within the safe limits.
Vitamin D is needed for lots more than dealing with Covid and like a great many nutrient requirements, more and more past a min. level is a waste. Good ex is Vitamin C where no good data for taking lots of supplemental C, but overwhelming data for what happens if deficient. I did not see the Author of the post say what levels of D3 25 OhD were deemed too low nor did i see what targeted level of D supplementation range was tried for.
D absorption with age decreases and diet for many insufficient to reach appropriate minimal levels needed for overall health and even with supplementation one size not fit all as different people will absorb differently with some needing more D to reach the same level of Blood 25Ohd than others. Add age being an issue to absorbing from the sun and hence it likely that most will need some level of supplementation whether it helps with Covid or not
This is somewhat unrelated, but I've noticed that if my wife takes a vitamin D supplement for an extended period of time, or a larger dose, her anxiety increases. Her vitamin D tends to drop in January/February, but we have to be very careful about trying to increase it. Any ideas why, or if there is a way to raise vitamin D levels without raising anxiety?
I'm absolutely convinced that vitamin D works. It's all about the way how you take it: you have to take it every day over long periods of time, BUT only in the short time span between your cardio exercise and your yoga exercise. Or if you have no time span between these two exercises, then you can take it right before your regular meditation practice. Studies also show that vit. D works better against covid if you unpack a clean face mask anytime you unpack a vitamin D pill. ;-)
If the supplementation studies used accepted recommended doses and levels of vitamin D, these might have been simply too low to produce a therapeutic effect, compared with sunlight exposure. This would account for the discrepancy between those studies and latitude/season effects. To quote a review of vitamin D medicine from 2007 (www.nejm.org/doi/full/10.1056/NEJMra070553):
--
Much evidence suggests that the recommended adequate intakes are actually inadequate and need to be increased to at least 800 IU of vitamin D3 per day. Unless a person eats oily fish frequently, it is very difficult to obtain that much vitamin D3 on a daily basis from dietary sources. Excessive exposure to sunlight, especially sunlight that causes sunburn, will increase the risk of skin cancer.125,126 Thus, sensible sun exposure (or ultraviolet B irradiation) and the use of supplements are needed to fulfill the body's vitamin D requirement.
--
Note also that the relationship between dietary vitamin D intake (including supplements) and serum levels is not linear but logarithmic. This makes it difficult to raise serum concentrations of vitamin D to useful levels with supplements. P.D. Mangan links to more publications here https://roguehealthandfitness.com/how-much-vitamin-d-do-we-need/.
PS: what sort of formatting does Substack support in comments? Not being able to blockquote is very annoying.
Has there been any research on whether sunlight-based vitamin D is good because of the increase of vitamin D vs. the reduction of cholesterol (which is what sunlight converts to vitamin D)? If it's the reduction of cholesterol, then vitamin D pills might not help.
"...if I were in the business of defending the Spanish study I might argue that it takes a few weeks for cholecalciferol to work, so giving it to someone who will already be dead or recovered by then is meaningless.
It would make me more comfortable if someone who did understand Vitamin D biochemistry would confirm this, but I'm not going to hang the whole argument on it."
Not wanting to hang the whole argument on the fact that it takes cholecalciferol 10 days to become available in the patient's body, and that it wasn't even adminstered until 11 days after onset of Covid, is utterly ridiculous. Those are the crucial flaws in the entire study, so it should not only be set aside, but thrown aside with great force. The author is nothing but a flak.
I would just add that there is likely also an age factor - average population age decreasing as you approach the equator because of the old people in the US and Europe.
I am reading this blog for the first time via the Naked Capitalism blog. As a chemist, I tend to focus on perspectives that include a biochemical view. Last fall, there were reports about AI studies that examined the broad array of Covid symptoms, and pointed to a bradykinin storm instead of a cytokine storm as a potential cause of the symptoms. Here is one magazine link that describes some of the biochemical cascades in the bradykinin system:
The article includes a list of potential treatments based on the bradykinin system, including the following quote:
“Interestingly, Jacobson’s team also suggests vitamin D as a potentially useful Covid-19 drug. The vitamin is involved in the RAS system and could prove helpful by reducing levels of another compound, known as REN. Again, this could stop potentially deadly bradykinin storms from forming. The researchers note that vitamin D has already been shown to help those with Covid-19”.
Also, I saw that one commenter here wrote about red light treatment. While I agree that low intensity laser light can be used to treat inflammation and respiratory problems, I think it’s possible that Vitamin D may be helping through a different mechanism. I and my friends and family have had good results with red light therapy for a range of indications. If you are curious, here is Dr. Fred Kahn’s blog post on light therapy and Covid:
He can’t promote the light therapy as a treatment for covid, but offers his opinion. (Spoiler alert: Dr. Kahn is a former vascular surgeon. I have been to him myself and he has very strong opinions about the medical system.) In a nutshell, when properly done using appropriate wavelengths, the light therapy triggers an increase in ATP formation, angiogenesis, and other knock effects including changes in the rate of DNA transcription in some cases.
> "Whenever you see a disease that's better in summer and worse in winter, Vitamin D is one of the possible culprits."
Um, then why haven't I been told to take vitamin D to prevent the flu during winter flu season? What I mean is, isn't there some other explanation you ought to have mentioned, like "people staying longer indoors makes viral spread easier" or "viruses can survive longer outside a host when temperatures are colder and there is less sunlight"?
Later in the article you mention reasons to be skeptical, but in general, I recommend against front-loading all evidence for X (and saving evidence against X for the bottom) unless readers are informed up front that you're structuring the article that way.
You should factor in the absolute huge mess of a country that Brazil is before you put any faith at all in a Brazilian study.
This is not xenophobia. I'm Brazilian. You have to either be here or have lived here for a while to get a grasp of what I'm talking about.
If the study showed no Vitamin D effects, odds are extremely high that the vitamin D used in the trial is simply counterfeit. Brazil has absolutely no QA in its genes. Bribery, on the other hand...
This was good. The CDC and FDA's inability to do simple cost-benefit analysis was shocking and lead to tens if not hundreds of thousands of excess deaths when it came to masks. Sometimes, knowing something won't hurt is enough and the efficacy can be hazy. You need to look at the risk and reward and if something like wearing a mask or taking vitamin d is cheap, easy, and safe you just do it.
I just wanted to say that I'm really grateful that you're still posting Scott :). I check my email multiples times a day to see if there's a new post, and seeing a notification from ACX makes the day that much better
>ask me about schizophrenia rates sometime!
Consider yourself asked! I'd love to hear more about this.
Another great example showing why science is actually pretty hard to do well, even when we are trying to answer questions that appear simple at first glance.
If it's so difficult to answer whether a major vitamin helps with one of the most significant diseases we've seen in decades a full year in, one should consider to what extent they should place strong confidence in their convictions in similar areas with similar scientific processes, for example nutrition and diet (these are also very hard, like most areas, and you should exercise extreme epistemic humility and adaptability, imo)
Seasonality is a whole thing. I love this review, and it does say quite a bit about Vitamin D. There are also a host of other factors that are in general responsibility for coronaviruses doing better during the winter. https://www.annualreviews.org/doi/pdf/10.1146/annurev-virology-012420-022445
One more point: a significant proportion of readers will have been staying inside (so away from sunlight) much more than they did before in the past 11 months; just think about not commuting to work everyday.
This will have reduced their (already on average pretty bad, for UK and some US states) level of Vitamin D... in which case a supplement would hurt them even less than the very little theoretical hurt it did in those pre-2020 studies.
I also side with "take Vitamin D supplements". My parents have also been taking one since November because they trust me (and their doctors are fine with them taking it).
Could seasonal, latitude, and minority trends not be linked directly toward indoors exposure and multigenerational households?
Individuals are much more likely to congregate outdoors during the summer than the winter. This corroborates with the known facts that outdoors exposure poses a much lesser risk to corona than indoors exposure. In addition, we know that cold air carries the virus droplets much better than warm air. It seems to me that these facts would almost entirely explain the seasonal and latitude trends.
To explain the "vitamin d" trend in skin color, asian and african american households are much more likely to be intergenerational and/or involved in service jobs.
Regarding vitamin D pharmacokinetics: the actual active form is calcitriol, which has three hydroxy groups. This is produced by a series of enzymatic hydroxylations starting with cholecalciferol (1 hydroxy), and then calcifediol (2 hydroxy groups). So giving people calcifediol is going to act more quickly than cholecalciferol.
Wikipedia states that, "At a typical daily intake of vitamin D3, its full conversion to calcifediol takes approximately 7 days." And for high doses it's probably slower if the enzymes are saturated. (Source: https://en.wikipedia.org/wiki/Calcifediol which cites https://academic.oup.com/ajcn/article/87/6/1738/4633505)
So I think giving emergency doses of cholecalciferol is a dumb move by doctors.
It would be interesting to know what were the criteria for ICU admission in the Indian study, as well as the baseline O2 saturation in both groups. 50% of the patients needing ICU care is a little unexpected
Doesn't vitamin D have a build-up time and decay time, in the body. If so, this filter effect would rule out instant-on attempts to use it (e.g. hospital data)
There appears to be a more recent calcifediol study from Spain, with some of the same authors: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3771318 Apparently has some significant flaws, which makes me question the earlier study: https://twitter.com/fperrywilson/status/1360944814271979523?s=20
There's at least one thing you get from sunlight that isn't vitamin D-- people have subcutaneous nitric oxide precursors. The really fun part is that you won't find this out by studying mice. Mice are not only covered with fur, they prefer going out at dusk and pre-dawn, so they don't have that body system.
Nitric oxide and sunlight for people was discovered as a result of studying why there were more heart attacks in the UK than in Australia.
Is there a study which includes people's actual skin color? This would be more informative than knowing their race.
I read a bit about this when I had a patient with GFAP encephalomyelitis who had a dramatic recovery after getting sunburned. I was really pissed off with the staff who let a very cognitively impaired man sit in the sunshine and get sunburned. But then he recovered, from literal psychosis and profound memory and executive impairment, to seeming fairly normal, and Neuropsych testing confirmed this. I wasn’t sure if it was vitamin D or not.
So I read about MS, because GFAP encephalomyelitis is incredibly rare, and found that some researchers were arguing that the latitude effect is due to UV rather than Vitamin D itself, and had a mouse model of multiple sclerosis showing mice under UV did better than either controls or vitamin D mice. I can try to find the reference again if you like.
Obviously my sample was n=1 of an extremely rare disease, but it got me thinking.
If it was UV, could that explain the latitude effect but also why vitamin D supplementation studies with good designs had mixed results.
I still take vitamin D 😁
"Do the benefits of taking a Vitamin D supplement at a normal dose equal or outweigh the costs for most people?"
I'd say the percentage is closer to 100%.
There have been numerous studies showing that low vit D levels are correlated with increased risk of depression, diabetes, cancer and influenza (though I haven't looked at all of these studies with the rigor or Mr Alexander so who knows after you disect all of these studies). Anyway, I tell my patients to take it.
Here's a recent one on influenza
https://pubmed.ncbi.nlm.nih.gov/32252338/
The world would be a better place if Vitamin D supplementation, especially at higher latitudes, becomes routine for most people.
So, are you taking a daily dose of vitamin D?
Disclosure: I am taking 25mcg of D3 just because
via negativa
"For example, the Spanish study used a version of Vitamin D called calcifediol; the Brazilian one used a slightly different version called cholecalciferol. Calcifediol becomes active more quickly than cholecalciferol, enough so that if I were in the business of defending the Spanish study I might argue that it takes a few weeks for cholecalciferol to work, so giving it to someone who will already be dead or recovered by then is meaningless."
Biology is perversely complicated enough that this could indeed matter. When I was looking up magnesium supplements to take, there are ten different types https://www.mindbodygreen.com/articles/magnesium-supplement-types and depending what you want to achieve (ease sore muscles? help blood pressure? treat constipation?) then the type you should be taking differs.
And looking up about vitamin D types, the differences between the two types seem to be:
"The recommended form of vitamin D is vitamin D3 or cholecalciferol. This is the natural form of vitamin D that your body makes from sunlight. Supplements are made from the fat of lambs' wool.
However, a clinical study reported in 2008 suggested that vitamin D2 works as well as vitamin D3.
Many supplements contain vitamin D as vitamin D2 or calciferol. It's derived from irradiated fungus."
Irradiated fungus? Funky, if we want to turn into Godzilla I suppose?
D3 being the form that is made when sunlight hits our skin, another site advises that there is a difference in bioavailability: https://www.healthline.com/nutrition/vitamin-d2-vs-d3
"Vitamin D2 and D3 are not equal when it comes to raising your vitamin D status.
Both are effectively absorbed into the bloodstream. However, the liver metabolizes them differently.
The liver metabolizes vitamin D2 into 25-hydroxyvitamin D2 and vitamin D3 into 25-hydroxyvitamin D3. These two compounds are collectively known as calcifediol.
Calcifediol is the main circulating form of vitamin D, and its blood levels reflect your body’s stores of this nutrient.
For this reason, your health care provider can estimate your vitamin D status by measuring your levels of calcifediol.
However, vitamin D2 seems to yield less calcifediol than an equal amount of vitamin D3.
Most studies show that vitamin D3 is more effective than vitamin D2 at raising blood levels of calcifediol.
For example, one study in 32 older women found that a single dose of vitamin D3 was nearly twice as effective as vitamin D2 at raising calcifediol levels.
If you are taking vitamin D supplements, consider choosing vitamin D3."
So the Spanish study was using the inferior (?) D2 yet got good results, while the Brazilians were using the superior (?) D3 and found no difference.
Yeah, I'm puzzled too. It *may* be that the Spanish patients, who were already hospitalised when they received the supplementation, were *so* badly off for Vitamin D that anything was beneficial. The only difference I could come up with was that the Spanish study seems to demonstrate that it "drastically reduced" admissions to ICU, while the Brazilian one says it didn't reduce length of stay:
"A small randomized controlled trial in Córdoba, Spain, of calcifediol (25-hydroxyvitamin D3, or 25(OH)D3) for hospitalized COVID-19 patients (henceforth, “the Córdoba study”) found dramatic reduction in the need for ICU admission. This study has been viewed as a small preliminary study, suggesting at most that further study might be warranted. It has gotten relatively little attention, though its strengths and weaknesses were discussed in this article, and a Bayesian cost/benefit analysis found the expected benefits, in terms of lives saved and severe illness avoided, of immediately adopting the treatment protocol were considerably higher than the expected costs. Two other randomized trials of vitamin D formulations for COVID-19 had mixed results: a small trial in India found a high dose of vitamin D3 shortened time to viral clearance in asymptomatic or mildly symptomatic SARS-CoV-2 individuals, and a larger study in Sao Paulo, Brazil did not find a statistically significant benefit of a high dose of vitamin D3 in hospitalized COVID-19 patients."
"Conclusions and Relevance Among hospitalized patients with severe COVID-19, vitamin D3 supplementation was safe and increased 25-hydroxyvitamin D levels, but did not reduce hospital length of stay or any other relevant outcomes vs placebo. This trial does not support the use of vitamin D3 supplementation as an adjuvant treatment of patients with COVID-19."
So *maybe* it's "Vitamin D supplementation if you're hospitalised with Covid-19 means you're less likely to be admitted to ICU, but you won't get out of hospital any sooner than someone with Covid-19 that did not have to be admitted to ICU and wasn't getting Vitamin D supplementation". That is, "if you're run-down/deficient in Vitamin D, that makes it worse if you get Covid-19 and have to be hospitalised, but taking Vitamin D won't stop you getting Covid-19 although it will make it less likely you'll be so badly off you need to go to ICU".
Can you look at Fluvoxamine? People have been saying "there is a 95% chance that it will reduce hospitalization/death rates by 75% or more" here: https://www.quora.com/What-is-the-current-treatment-for-Covid-19/answer/Steve-Kirsch
Could there be something like an undershoot effect involved, where people who would ordinarily have lots of vitamin D because they spend a lot of time outside and live close to the equator actually have much less because they're locked down and stay inside instead, which does... something... to their immune systems?
One of the challenges with using serum vitamin D levels as an input variable is that we don't really know whether those levels are a cause or an effect; possibly they are a confounding factor when included with other inputs. Example: although patients with higher vitamin D levels may have fewer M.I.s, we can't produce that result by supplementing vitamin D in the diets of low-vitamin-D patients. See, e.g., https://jamanetwork.com/journals/jamacardiology/fullarticle/2735646.
Well sunlight seems way more promising, so why bother studying Vitamin D?
So here's my experimental design for testing the "something something sunlight"-hypothesis:
Set up a hospital wing with a bunch of UV lights, and another with UV lights and infrared and another with just infrared and treat only people with a specific shade of dark skin in them for Corona. [uhm.... also a control group would be nice]
If there's no difference from any of the light setups, we..... are very confused.
Maybe not enough light.
If we see improvements from all rooms except UV-lights only, we know it's not vitamin D.
If we see.... ugh, lighting is hard, make that guy design the experiment:
https://www.lesswrong.com/posts/7izSBpNJSEXSAbaFh/why-indoor-lighting-is-hard-to-get-right-and-how-to-fix-it
And someone who understands Vitamin D and Nitric Oxide biochem to write the interpretation.
Preregister the paper, get a hospital to run this and just fill in the numbers they report.
https://www.lesswrong.com/posts/7izSBpNJSEXSAbaFh/why-indoor-lighting-is-hard-to-get-right-and-how-to-fix-it
And make the Thiel foundation, Kickstarter or whatever fund it. And if you're reading this, you can organize it :)
"Despite their superficially lighter skin, they have Vitamin D deficiencies almost as bad as blacks." Unless I'm missing something, the linked study included a total of 60 Asians all from the greater Boston area. Obviously not very representative of ~4.7 billion Asians in the world or even the Asian population in the U.S. I would point to larger studies in Japan (e.g https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3755751/) that appear to point to significantly lower vitamin d deficiency than the US population and much lower than the US black population.
My understanding is that hydroxychloroquine is ineffective on its own but helpful when combined with other antiviral medication, thus possibly explaining some significant variation in study results (since many studies actually did combine it with other things):
https://twitter.com/__ice9/status/1336164630851428357
Hmm, not sure about this. Arizona, the sunniest state in the US, also has the sixth highest covid infection rate per capita of the 50 states.
Interesting review. I only suggest that you consider British epidemiologist, Edgar Hope-Simpson's search for a parsimonious explanation in his study: The transmission of epidemic influenza (1992). There is a clear connection between these seasonal phenomena. Here's a very interesting follow up on EHS's comprehensive work in the BioMed Cenral Virology Journal. Flu is still a surprisingly mysterious phenomenon in many ways. There are probably some read across matters to SARS-CoV-2. The intriguing relationship with VitD and even it's centrality is also well covered. Worth checking out the early C20th methods for transmission research if nothing else. There are some very interesting questions and conclusions that challenge some givens about these phenomena.
https://virologyj.biomedcentral.com/articles/10.1186/1743-422X-5-29
Good article but the Brazil and Spain studies were carried out solely among hospitalised COVID-19 patients. The India study is very interesting because it indicates that Vitamin D helps mild cases recover faster.
The next step should be to find out whether Vitamin D can help potentially moderate-to-severe cases remain mild and therefore whether it can prevent hospitalisations in the first place. Perhaps it can be given to a sample group of people of all ages who have never had COVID-19 and who haven't been vaccinated and this group can be compared to a control group.
Dr. Atlas is at it again. Pied Piper much?
So anything wrong with just taking L-arginine and L-citrulline as supplements to get nitric-oxide levels up? Like, if I'm going to take a bunch of random stuff sunlight-associated stuff like Vitamin D, I might as well be a completionist about it.
There's all sorts of factors that go into the seasonal pattern. At least when it comes to influenza, high humidity seems to be the most important (see eg https://journals.plos.org/plosbiology/article/file?id=10.1371/journal.pbio.1000316 and https://www.pnas.org/content/106/9/3243 and https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0057485). But there's also temperature, UV radiation, and of course behavioral changes.
Perhaps at least a portion of the seasonal and latitudinal variation is due to that COVID (and many viruses) are rather vulnerable to ultraviolet.
Thus, the seasonal variation in daily insolation would be the true causal driver, with Vitamin D correlating for the obvious reason.
If this holds water, I would expect the degree of seasonal variation to be weaker closer to the equator, although this would probably be weakened because the directly-caused variation will generate more cases that will spread to the equatorial regions anyway...
Also, I would expect the degree of the effect to vary based on the amount of UV reaching the ground - so if the data exists, plotting the health of ozone layer against the degree of seasonal variation might be informative....
On Zvi's advice, I took 4000 IU of Vitamin D per day for about 2 months.
Then I got kidney stones, and one of the doctors I chatted with said it could likely be from the Vitamin D, and I reasoned that I'd stored up enough Vitamin D (I think it gets stored in fats?) that I should stop taking it. Because kidney stones was no fun.
Laboratory tests should be designed so that around 5% of tests are reported as abnormal. I worked for almost 30 years as a hospital pathologist, and, without doing a formal study, we were reporting way more than 5% of our vitamin D levels as abnormal. Before going off and making any recommendations, I would want to work on the test and make sure that there are no issues with it. Working in a community hospital, I never had the resources to do that sort of large study, but I still have the nagging feeling that something is amiss here, and I am not sure what it is.
As for outcomes in people of East Asian descent, I know through friends in the Chinese community that many people of Chinese descent remembered the original SARS epidemic and changed their behavior way before the government was willing to admit that there was a problem. That confounder will somehow have to be corrected for before drawing any conclusions about outcomes in those patients.
I have been taking the maximum recommend daily dose of 4,000 units Vitamin D3 since Feb 2020 on recommendation from Doctors at the Harvard Medical School. Presumably Vitamin D interacts with ACE2 which was identified as the viral binding site. So presumably not having a Vitamin D deficiency can be beneficial with respect to possibly reducing the risk of becoming infected.
https://elemental.medium.com/what-science-says-about-vitamins-and-supplements-for-covid-19-b5feaedbb9f1
There are a lot of people who definitely do need more Vitamin D because they're deficient, and there are a lot of other people who don't take Vitamin D when they should (or take less than they should) because they're incorrectly worried about taking too much. If you might be in either group, you can ask your doctor for a Vitamin D test, or just buy one for $50 here:
https://www.everlywell.com/products/vitamin-d-test/
So let’s pretend that I’m a rube who don’t hold with all that high falutin’ college talk about n=stuff and chemistry and randomized study sessions and such.
What is wrong with blurring the conclusion enough for common sense to assert itself? “Vitamin D makes you healthier, and healthy people handle the ‘Rona better.”
If somebody did the exact same studies testing Vitamin C instead of D, do you reckon there would be a similar spread along the effective-ineffective spectrum?
Thanks Scott - super interesting as ever.
Any thoughts on safe dosing? The NHS recommends 400 IU/day and cautions against taking over 4000 and warns that more can be harmful (https://www.gov.uk/government/publications/vitamin-d-supplements-how-to-take-them-safely/vitamin-d-supplements-how-to-take-them-safely).
Most online retailers in the UK seem to sell tablets with 4000IU each, and 6000 seems common in the US. Then the Brazilian trial you mentioned used a single megadose 200,000IU (!)
I find it hard to square the NHS's cautious position with what seems like a wide range of dose choices in practice. Any idea what gives?
I would also look at the enthusiasm for Vitamin D in a broader context. Over the past 75 years there has been sequential excitement and enthusiasm for different vitamins/minerals as panacea for many human diseases. Cycles of Vitamin C (the great Pauling first one), various B vitamins, Selenium, Zinc, Vitamin E (for a long time). Medical specialities in diverse fields (immunology, oncology, ID) all simultaneously become excited by the new trendy vitamin/supplement. Initial small studies often show dramatic benefit in whatever metric is being assessed. Sadly, almost universally when large, controlled studies are done, there is great disappointment (and sometimes even negative benefit). Vitamin D is the latest of these cure-all supplements. You could probably look at a graph of medical literature/reports on each of these magic answers, and Vitamin D is certainly trending. But skepticism is in order. But hey, it probably can't hurt...
My general physical health got a lot worse when I first started staying at home all the time (I've literally left my house/yard only twice since last March, and I think it's been a month or so since I've been in sunlight). I had back pain, wrist pain, joint cracking, and fatigue at various points (I'm in my mid 20s). Since I started taking vitamin D or other multivitamins, I have basically not had any such issues. I expect that my COVID avoidance is way more dramatic than most people's, but I think most people have been getting less sunlight than they used to, and if so, Vitamin D seems great for improving their general physical health.
Please correct: >>we can be pretty sure that their COVID is caused by Vitamin D.<< ==> >>we can be pretty sure that their COVID is caused by lack of Vitamin D.<<[and delete this commentary]
My strong, strong prior here is that Vitamin D may be a decent biomarker, but it's a lousy therapy. It's been thrown at nearly every indication under the sun in RCTs, and almost always returns null results. To me, the most plausible explanation is that low Vitamin D levels are correlated with biological states (malnutrition, sedentary living, inflammation) that predispose to poor outcomes, but short-term Vitamin D supplementation does not address the underlying predisposition.
*grumbles about how "vitamin" is not a meaningful category, and, in the case of cholecalciferol derivates, we're actually talking about hormonal replacement instead of vitamin supplementation*
Great summary. A common "bad take" submission I've never run is along the lines of "the race disparity is explained due to vitamin D." Probably not, but maybe a little. It just strikes some people as a completely bonkers suggestion, so I get sent it regularly.
Also gets tarred by association because same people pitching also hock zinc and HCQ. (And now ivermectin, another "bad take" I can't pull the trigger on.)
Three things to consider:
1) Mendellian randomization and the Biobank studies are both invalid because people supplement in response to genetically low vitamin D, which violates one of the assumptions underlying MR. This is especially true since government guidance in the UK changed pretty dramatically on vitamin D supplementation
2) If you're looking at a racial case for vitamin D, you should really look at immigrant pairs like African immigrants to Sweden, or Indian immigrants to the UK. In all cases you see low Covid rates in poor sunny countries and high covid rates among immigrants from those countries to rich northern countries. This is the Ricketts pattern and strongly indicates vitamin D's involvement. https://shotwell.ca/posts/africa-covid/
3) The problem with the Brazillian study is that vitamin D was given way to late. If it's active against Covid it's because it works basically as an anti-viral, in that Brazillian study they gave vitamin D on average 10 days after symptom onset by which time the virus had cleared. All the supplementation studies where vitamin D has been given in or before the viral replication phase have been positive.
Re: Asians
Older population?
As an aside, I think one of the biggest unanswered questions about COVID-19 is why so few people in equatorial Africa have died. Equatorial South America has done pretty terribly (Manaus = 3.1190°S), and outcomes from the first world indicate that African ancestry does not confer a survival advantage.
you don't mention this study here. https://nypost.com/2021/02/15/covid-19-study-revives-heavily-disputed-vitamin-d-theory/
Seems for individuals, the first question is whether one might be borderline vitamin-D deficient to begin with (little sunlight, milk-free diet for some reason, etc.). It might help more for Covid in such folks (and only in such folks?). And all sorts of below-the-radar effects of low-D may well get solved besides.
This article seems to miss discussion on the most important claim I've heard around Vitamin D supplementation: that daily Vitamin D supplementation substantially reduces your risk of contracting severe cases of covid.
This is quite different from a) vitamin D reduces your risk of getting covid (why would it) or b) taking vitamin D when you get covid improves outcomes.
My understanding is that *it takes a long time to increase serum vitamin D levels through supplementation.* See this article from the NCBI: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3820059/ . In short: infrequent, low doses (1-2k IU) of vitamin D do not raise serum vitamin D. Dose and frequency (per week) mattered more than duration, but duration was at least a month, not 7 days! And duration effects were increasing - the longer you took vitamin D, the bigger the effect!
Scott dismisses this entire argument by saying "people who can do anything consistently for a year will do better" without actually analyzing it in more depth. No, you actually have to take vitamin D supplements frequently over time (at least one month, with effects increasing up to 6 months), at higher dosage for them to have an effect. Dismissing the underlying mechanism as being a selection effect, means that Scott misses the best argument in favor of long-term vitamin D supplementation.
This hypothesis may be wrong - maybe that research paper is incorrect, maybe there's more evidence behind Scott's handwave - but right now it seems to me that Scott is ignoring the crux of the debate: can long-term vitamin D supplementation reduce your odds of a severe covid case, conditional on you getting covid? Scott unfortunately does not address this question, which is the most important one, IMO.
Given that long-term vitamin D supplementation does increase serum vitamin D levels significantly more than short-term vitamin D supplementation (at least according to the research), I am inclined to believe the research does suggest that vitamin D supplementation can reduce the odds of contracting *severe* covid cases much more than Scott does.
To what degree can it be that people with low vitamin D are more conscientious (indoor workers) and therefore have a lower risk of getting infected?
Seems like the studies shouldn’t control just for race, but also for behavior.
I’d say low vit D white people will be more likely to get infected due to low vit D, but less likely to get infected due to behavior. Which could easily mask the relation between vit D and covid incidence completely.
It would be better to look at severity for this reason, since it’s presumably affected less by conscientiousness than incidence — but it still *is* affected. Conscientious people can do other things that decrease the severity.
Seems like taking the other side of the rootclaim challenge would be very positive EV if those estimates are correct?
https://blog.rootclaim.com/treating-covid-19-with-vitamin-d-100000-challenge/
This article reminded me to take my daily Vitamin D supplement! Thanks Scott for the help.
I think this qualifies as an enema recommendation, though perhaps there's enough evidence in favor that it doesn't. An old joke:
An actor collapses on stage, and a doctor is found in the audience. As he tends to the actor, an old lady calls from the cheaper seats: "Give him an enema!" She is ignored. She cries again, "Give him an enema!", and again, nobody pays attention. After a third "Give him an enema" is heard, the doctor turns, and says, "Ma'am, I assure you, giving this man an enema would do him no earthly good whatsoever." The lady replies, "Well, it couldn't hurt."
I think you took the wrong side regarding the Spanish Vs Brazilian trials. The Spanish one was peer reviewed. It was better conducted with more activated form of vitamin D. It is totally not wise to give D3 to already hospitalised person. It takes precious time to be converted to the active form. The Spanish trial also was small achieved remarkable p value. This is the unbiased estimator which means that no matter it was small study, the results proves causality. Patients were divided for the true treatment or placebo randomly by a computer. How can one argue against this method ...
The latitude thing might be exterior UV killing COVID on surfaces. In Asia, there's been a buying spree for indoor UV lamps to provide some viral load reduction indoors, complete with timers and motion sensors and shut off whenever a person walks into a room.
It would be useful to check the methodologies utilized for measuring vitamin D levels in these studies, since it seems that the measured value for 25-hydroxyvitamin D (predominant in testing) largely depends on differences in vitamin D binding protein (VDBP) between individuals. Ginsburg & coworkers identified a vitamin D metabolite ratio (VMR) which is dependent on concentration and not on the amount of VDBP that is present in an individual.
> The VMR was independent of VDBP concentration, whereas VDBP was strongly directly associated with the individual vitamin D metabolite concentrations. Prior studies evaluating only 25(OH)D3 may have been confounded by absence of data on VDBP status.
"The Vitamin D Metabolite Ratio Is Independent of Vitamin D Binding Protein Concentration", Clinical Chemistry, 2020, doi: 10.1093/clinchem/hvaa238 , https://pubmed.ncbi.nlm.nih.gov/33188595/
The lead author also, Dr. Charles Ginsburg, a nephrologist & prof discussed the research in a 10-minute interview for the Clinical Chemistry Podcast: https://directory.libsyn.com/episode/index/show/clinchem/id/17760140
>There were some early astoundingly massively positive RCTs for hydroxychloroquine
I don't remember any, just a lot of extremely sketchy-looking stuff showing some mild effect. Which RCTs are you thinking of?
I can help with the seasonal pattern. The smallish bump in the summer was in the hotter, southern states primarily. The much larger bump in the winter was in the cooler, northern states primarily. Common theme? People spending lots of time indoors in climate controlled environments.
It looks like there's another randomized controlled trial in the works for Vitamin D supplementation being planned, from Wang et al. (N = 2700): https://doi.org/10.1016/j.cct.2020.106176
That's a pretty substantial sample size. Can someone take a look at the trial design plan and see if it's reasonable? That isn't my skillset.
A veritable cacophony of Vitamin D studies. I'm frankly surprised you implicitly lend credence to the results. Given the file drawer problem, replication issues and paradoxes within your speculations, there's really no basis for your last 3 paragraphs except intuition/ guessing/ idiosyncratic heuristics. You certainly know that mechanistic thinking just doesn't go very far in medicine; and that case-control and other observational studies are sometimes useful but overall weak. To argue seasonal variation may occur related to sunlight/ vit D while later noting levels are usually stable over long periods is not consistent.
Those criticisms aside, your collection of studies just shows once again how poorly medicine and the lay public have been served by academia in a time of great need. Large well-designed studies capable of definitive answers are sorely lacking among the utter cacophony.
> And maybe some day, after countless false leads and stupid red herrings, one of the claims people make about this substance will actually pan out. Who knows?
One thing that always struck me as weird about vit D is how ubiquitous it is among health-conscious people
I wonder if th lis is a case of memes spreading with no reason, or of the studies missing an important factor.
It's been years since I looked at it but the kidney stone risk should be lowered by simultaneously supplementing vitamin K (which is another one of these quite likely positive supplements).
Regarding summer: sars2 is highly unstable when it comes to light, heathings dries the air which makes your mucous membrane dry and live inside has no wind. Source: christian drosten's podcasts.
My thoughts are that possibly low vitamin d could also simply be a marker of poor diet and lifestyle.
A possible downside of supplementation could be that it makes more postive change less likely.
I.e. more outdoor physical activity, combined with nutrient rich diet will likely* sort vitamin d levels, plus have a host of other benefits that would be missed by supps alone.
Supplements in general have a very poor track record, possibly as many people end up taking them as "substitutes" to a nutrient rich diet rather than supplements? (Purely speculative).
Given the choice between go outside, exercise and eat well, vs take a pill, many will opt for the pill. Which is fair enough providing we know the true efficacy of the pill and can make the trade-off. But I don't think this is the case (for vit d and covid or any other supp and health outcome).
* Possibly for darker skinned individuals, living at northern latitudes, working indoors all day it might be impossible to achieve optimum vitamin d levels through sun exposure and diet?
> and die of it 3x more often
Is that number correct?
n all of this you haven't mentioned to what level of Vitamin D you are testing against. The level described as "deficient" for bone health isn't the one described as "deficient" for respiratory health (if that is indeed a thing).
If you want a null hypothesis to aim at, here's one
"The bottom line is that there is no downside to increasing our intake of vitamin D to maintain serum 25(OH)D at at least 30 ng/mL (75 nmol/L), and preferably at 40–60 ng/mL (100–150 nmol/L) to achieve optimal overall health benefits of vitamin D."
https://www.mdpi.com/2072-6643/12/7/2097/htm
Reassess against that much higher than standard level and everybody can avoid talking past each other.
In this instance the obsession with the precautionary principle is to avoid hypercalcemia at all costs - even if that means people succumbing to respiratory illnesses. Something they didn't do, for some reason, with masks.
We know that getting covid is correlated with breathing it in. We can be fairly confident that better ventilation reduces the risk of covid. What do people do in warm sunny weather? Open windows, and go outside. We should expect the latitude and seasonal effects, even if vitamin D was nothing to do with it.
My intuitive take is that vitamin D has some effect on immunity but compared to bone health, it is much less pronounced. The bone mass is comparatively large and we need a lot of calcium therefore vitamin D deficit can quickly cause problems. Whereas the immune system can adapt with lower vitamin D levels.
Very low vitamin D levels in people could make the immunity much worse too but in practice we notice problems with the bones first and treat them accordingly. Supplementing subclinical vitamin D deficit provides only marginal benefits.
In conclusion, it is good to maintain vitamin D levels in the recommended range, and take supplements if you need to, mainly, to help your bones. Maybe it will have a secondary benefit by slightly improving your immunity too but don't rely on that too much.
Given that almost all of us get way less sunlight than our ancestors, I would advocate supplements anyway, to bring us up to the levels that our bodies are adapted to.
Vitamin D has low toxicity, so it's very unlikely that you err on the side of having too much.
The best explanation I've seen for the seasonal and latitude nature of the virus so far is indoor humidity. If the temperature is below 15 degrees C, the air indoors is heated and becomes dry. This causes droplets expelled from the mouth that would settle in more humid conditions to evaporate and become smaller and float around for longer providing a more effective transmission in confined spaces.
A leimilar but lesser effect can occur from airconditioning in summer.
In places with really sealed up indoor spaces the humidity is lowish in winter but not as low as some not as cold places because the less sealed up spaces have more cool outdoor air comes in to be heated. Places with winter daily maximums in the 5 to 15 degree range tend to have the dryest indoor air in winter.
"In the US, more blacks than whites have gotten coronavirus."
In some alternate reality version of the US, perhaps. But not in this US.
Seasonality/latitude dependence could be completely explained by people gathering indoors vs. outdoors. Very hard to get aerosol transmission in a windy park.
Regarding the Mendelian randomization study: When you buy a 1/4" drill bit (or perhaps a 6.5 mm drill bit if you're into that sort of thing), it's not because you want to own a drill bit, but because you want to own some 1/4" holes. In the same way, we don't want vitamin D in our systems because the vitamin D molecules strut around the body punching corona viruses in their little faces, ripping apart cancer cells, and blocking the depression rays emitted from the dirty laundry pile. We want it in our system because it does something else. Perhaps the people with hereditary vitamin D deficiency have some other source for this, uh, "something else". Maybe they make their 1/4" holes with punches, lasers, or well-disciplined termites.
You should really clarify your definition of "Asian". That term covers a much, much broader range of latitudes than white or black does. Probably difficult to suss out small lineage differences in the entire subpopulation, but you could just focus in on two groups with different historical ranges, like Koreans and Indonesians. Or Native Siberians, if you can find data.
> The loose ends that bother me the most are the seasonal pattern, the latitude data, plus the increased risk of hospitalization and death in Asians. I don't have a great explanation for those.
Couldn't the first two be explained simply by the fact that during the winter in northern countries, people tend to meet indoors rather than outdoors? This facilitates the spread more.
"For example, infection rate from coronavirus seems latitude dependent; in general, the further north an area, the worse it's been hit."
This isn't obvious; the statistics from the southerly countries aren't as reliable as those from the northern ones, and it is known many southerly countries (Mexico, Peru, Bolivia, Ecuador, Indonesia) have been very hard hit. It was true during the first wave in the U.S. and at the beginning of the third, but Mississippi, Louisiana, Alabama, and Arizona have been some of the hardest-hit states. In the E.U., Spain has been hit rather badly. It is true that Russia, the Dakotas, and the U.S. Northeast have been very hard-hit, though, but the U.S. Northeast pandemic seems to have been due to its better connectedness, while the harsh pandemic in the Dakotas and Russia was likely due to lack of social distancing.
Have you considered posting these as questions on Metaculus?
Does anyone have a quantitative idea, which proportion of the skin should be exposed to the sun for how long for normal cholecalciferol production, let's say for a maximally light-skinned bloke? Where I live, from october to april, the only exposed body part is face, and we get to see the sun every 3rd day or so. Is there any chance that we're still producing adequate amounts of cholecalciferol if we spend, say, 1 hour a day outdoors?
The UK Biobank studies are flawed, especially Hastie.
See: https://www.sciencedirect.com/science/article/pii/S1871402121000394 (peer reviewed & published)
and https://www.researchgate.net/publication/346922274_Serious_Statistical_Flaws_in_Biobank_Analyses (preprint with more details)
Lastly, I'm not sure where the idea came from that levels are stable over time. There are studies showing the 10-15yr old UK Biobank levels don't correlate well with recent levels. IIRC one correlation coefficient was around 0.4, which makes more intuitive sense in the first place. The idea of stability over time is reasonable in the face of no measurement or intervention/change, but the UK Biobank participants didn't have their levels hidden from them. It's pretty reasonable to assume those people who discovered they were deficient as a result of the UK biobank participation took steps to correct their deficiency. Their doctors certainly should have encouraged them to do so.
So trust Kaufman et al, Merzon et al, Meltzer et al more.
On the RCTs:
Dismissing the Indian RCT is inappropriate. This was small n but a well designed trial registered in advance and regardless of the importance of seroconversion clinically in terms of severity to individual patients, this data clearly shows a causal link whereby vitamin D affects SARS-CoV-2 infection.
Favoring the Brazil trial over Cordoba, Spain is also picking the wrong one. The Brazil trial didn't start giving patients D3 until they already needed supplemental oxygen (90% of them at baseline) because they were already 10 days post symptom onset and they were only average 7 days from discharge from hospital in terms of recovery. D3 doesn't ramp up fast enough to make a difference that late in infection is the only conclusion that can be drawn. See the comments of the medRxiv page of that Brazil study for many comments along these lines and other issues: https://www.medrxiv.org/content/10.1101/2020.11.16.20232397v1#disqus_thread
The Cordoba study used calcifediol which skips liver conversion and restores D levels within hours and thus had a much better effect. If they had used it in Brazil maybe the results would have been better. What's more, there's now a new preprint which has statistical issues but nonetheless adds to the positive data. It also used calcifediol and had 10x the patients (though the randomization was by hospital ward not patient, but it still makes the D helps conclusion more likely despite its flaws):
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3771318
I'm not sure why the Brazil hospital in that study got the patients so late in the infection. Worse healthcare infrastructure in that country?
But the trend is clear: 3 of 3 randomized controlled intervention studies started at hospital admission in countries in which that hospital admission came before patients were so far along that they mostly needed supplemental oxygen all showed positive results. And 2 of 2 calcifediol studies showed extremely positive results.
You said, "in the past I have learned to side with negative RCTs rather than astoundingly massively positive ones when the two conflict." That's a reasonable default for clinical trials of still-on-patent pharmaceutical agents because those tend to be sponsored by the pharma company that has strong financial incentives. For trials of natural agents especially those run by non-industry groups, the opposite should probably be the default. Eg see: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7487184/
You said, "I am going to assume that whichever doctors ran these studies did not completely bungle them and use forms of Vitamin D that could not possibly have worked in the time period involved." Completely bungled in the Brazil study is a reasonable conclusion. It's not completely absurd that it could have worked in the time involved, and the study organizers may not have known how far from recovery the patients would be when they started, so we don't have to fault them for running it. But the late stage of initial administration compared with the time it takes for that version of D to ramp up means its negative result cannot be applied to initiating the same treatment earlier in infection, nor does it cast doubt on D for prophylaxis, nor is it negative for calcifediol which is much faster acting.
Thanks for posting this. Any opinions on N-Acetyl Cysteine (NAC) supplements? I have read/heard claims of reduced covid symptoms associated with it’s regular use.
On appeal to authority:
How many scientists & doctors work on the NICE committee?
Here's a list of hundreds of scientists & doctors who say the evidence is convincing enough that it should be default to use vitamin D for covid as a precaution: https://vitamindforall.org/rollcall.html
It terms of credentials, these groups include many of the top vitamin D researchers in the world (based on citations), a former surgeon general, and many notable names.
Based on the prior 2 corrections I just posted in earlier comments (on Hastie & on the clear pattern from the RCTs) and this appeal to experts, you might want to update your % guesses on D's relevance. My personal guesses are 80% chance that D affects infection risk & 80% that it can help as treatment if given early enough in infection or as calcifediol. And 99% chance that benefits outweigh costs.
Correct reading of the evidence together with understanding the underlying biology makes the situation very clear. It just isn't quite clear at the level of approving a new pharmaceutical drug, and the big problem is that that is the evidentiary bar that NICE and other officials at 3-4 letter acronym bodies are applying. That's not the way to handle public health especially in an acute crisis. It's *astoundingly* clear that the balance of risks is such that increasing vitamin D intakes and reducing deficiency rates is a clear public health win net-net, especially now.
Good piece. Not mainly for the conclusion, which doesn't hinge on the analysis, but for an example of what a careful examination of evidence looks like.
This part seems unconvincing:
"We assume that people with the genes for low Vitamin D in fact have low Vitamin D. And this isn't confounded by anything; we know their low Vitamin D is genetic. So if these people get COVID more, we can be pretty sure that their COVID is caused by Vitamin D."
Probably the assumption in the first sentence there is at least somewhat false, no? Since people can change their vitamin D levels by taking vitamin D.
Also, it could be that a genetic predisposition to low vitamin D is associated with some *other* genetic thing that has an effect on severity of covid infections (in either direction - so you could get an illusory effect if it's associated with something that makes covid worse, or an illusory absence of effect if it's associated with something that makes covid less bad). (I don't have any specific knowledge that this is the case, but it seems like a gap in the logic as presented here.)
Scott, I suspect it was delayed by the Blog closure but I've definitely been waiting for your thoughts on 'no evidence' and the other scientific messaging in this pandemic. The language used to scientific uncertainty to the general public needs a rethink and I suspect is the source of much public disinformation.
I don't blame Joe Public for thinking 'no evidence' for a *thing* means definitively that the *thing* doesn't work.
Vitamin D for patients admitted to hospital could cut Covid deaths by 60 per cent University of Barcelona preprint for the lancet not peerreviewed yet... I had low Vit D (22ng/ml) when last measured 4 yrs ago (after spending 45min everyday in the midday sun in June...... I started taking 2,000 iu but after reading a few articles associating low vit D levels with the worst outcomes in covid I started taking 10,000 iu (250mcg?) but will be getting tested after my 2nd vaccine shot (1 week) I am aiming for a blood level of 60-80.... Depending on your age and genetics it can take a lot to boost your levels up.... any blood level up to 150 is most likely safe
When the limit of 100 was decided on the actual research showed a level above 300 was toxic....but to err on the safe side an upper limit of 100 was recommended....
The issue was already analyzed probabilistically by Rootclaim. They are also offering a $100,000 bet on it. If you're at 25% it's easy money for you (but don't, you'll lose - your analysis is wrong on many points).
https://blog.rootclaim.com/vitamin-d-can-likely-end-the-covid-19-pandemic/
https://blog.rootclaim.com/treating-covid-19-with-vitamin-d-100000-challenge/
Somebody on facebook argued, that Hydroxychloroquine is still in the game. Because the trial, which said it didn't help was made by Surgisphere, which gave fraudulent data. (I a lazy to look it up on my own).
1) Vitamin D correlates with activity/exercise levels pretty strongly. Consider this as a serious confounder in any observational study.
2) The Mendelian Randomization study I have strong reservations about. To run MR, you need to make strong assumptions about the genetic variants that you are using. So MR makes sense in cases where you know a lot about the variants. But here they are doing a GWAS to get the variants, i.e. taking anything that associates with Vitamin D levels. There's no theoretical reason to assume a priori that these variants have the properties needed to run this MR. Instead, you do some statistical tests and hope you can catch it if it's a problem. But those statistical tests make their own assumptions that I don't find very plausible.
The authors know this and address it and do a reasonable job at that. They try restricting just to variants that lie near genes that are known to directly impact Vitamin D and see the same thing, and they toss out variants that are also known to impact other phenotypes and so could confound. But it's just the reality of the problem: MR is a bit questionable if you haven't study your variants carefully. The fact that they find a null result makes this all less concerning, though it is plausible that confounders could cancel out the true effect.
I always felt, right after reading the original speculation, that Vit. D is a prophylactic rather than a cure. Firstly, you should have been taking this as a supplement already. So, if you have a decent amount of Vit. D taken as supplements cruising in your bloodstream; it is not going to prevent you from getting infected; but do two things: 1) maybe increase the viral load required for infection, 2) when infected the symptoms won't be as severe.
All these studies that you referenced don't seem to contradict this, or am I reading you wrong?
I would note that in general, anyone taking Vitamin D3 regularly in amounts of say 4000 IU a day, would be well served to take Vitamin K2 (probably MK-4) and Magnesium as well. And look into a source of Vitamin A (cod liver oil).
I'm 75 years old, and decrepit enough to be unable to what I used to. Vitamin D pills at Sam's cost pennies per day, hence I have been taking megadoses of it since June, 2020. No downside, possible benefits.
The most obvious difference between Spain and Brazil is latitude, which we already know has an effect (for whatever reason).
What percentage of people in Spain vs. Brazil do you think have a vitamin D deficiency?
It is easy to hypothesize that vitamin D supplementation in Spain alleviated deficiencies in many patients, whereas in Brazil it was superfluous.
All your other critiques apply. But this seems relevant too.
What is the explanation for 80%+ of Covid positive patients in Spain being vitamin D deficient? I would also like to see a regional analysis of Vitamin D deficiency rates in Brazil. Also, what about the 50 other studies? Overall this analysis seems a little weak, although the lack of enthusiasm from authoritative institutions is surprising.
The post-symptom/post-positive RCTs that find a (gigantic!) effect use calcifediol - the form that oral cholecalciferol slowly turns into, over a ~week. The RCTs showing no effect use cholecalciferol, whose conversion into calcifediol won't outrace an infection. I don't blame some researchers for trying plain cholecalciferol - it's a cheap OTC supplement, and maybe megadoses would help! Meanwhile calcifediol is a prescription-drug, & expensive in the US. But assuming "assume that whichever doctors ran these [cholecalciferol] studies did not completely bungle them" is picking one set of researchers arbitrarily to respect, & ignoring the contrast signal between the two choices. Why not assume, conversely, that the researchers who tested calcifediol did not completely bungle their inclination to test something with distinct effects from cholecalciferol? (The effect sizes in the calcifediol studies are gigantic!)
D3 needs to be taken with K2 (prevents kidney stones and other calcium mismanagement effects), magnesium (increases resorbtion) and A (reduces side effects from K2 related enzymes). It also needs to be taken with oil. And in small doses (5-20k IU/day) over a long time, megadoses like 200k IU will have way less resorbtion. I've read dozens of studies, even very recent ones, that showed side effects like kidney stones or artery calcification because K2 was ignored. Never ever assume the people doing these studies are aware of these problems, assume the opposite unless its expressly mentioned in the study. Keep that in mind when facing supposedly simple questions like "did they give an oil-soluble form/gelcap or the much cheaper dry form and if its the latter, did they give it with oil or fatty food". "Surely they couldn't have fucked up this badly"- yes they could lol.
I became a paid subscriber just to post this comment. Here is a hypothesis, which explains the seasonal variance and the latitude data without vitamin D.
Red light.
A very specific wavelength of red light penetrates the skin very deeply and is absorbed by the respiratory enzyme cytochrome c oxidase, which is involved in the electron transport chain in mitochondria, causing an increase in intracellular ATP.
Red light therapy aka low-level laser therapy cures a lot of weird problems. This specific wavelength of red light is also component of sunlight. We get more of it during summer and near the equator.
Vitamin D could be a red herring. Sunlight can do lots of other things too.
Seasonality is an independent variable via temp and humidity, and one study neatly combined them in enthalpy, which you may remember from school! (How do I do hyperlinks?**)
You may find a bigger effect with selenium.
Vitamin D itself has no effect: it's levels in the blood of 25(OH)D3 and its metabolites that count.
The reason it doesn't work if vitamin D given after admission is that it takes 2 weeks for serum levels to rise, so that's way too late to do any good, so the Brazil study was pointless, and was anyway a preprint, and your "use in hospital" figure should maybe be lower than the other one. The Spanish study has the same problem and also didn't check serum levels at all (but did at least use calcifediol/25(OH)D3 ) and still found benefit! An even more metabolised form has been trialed in ICUs and that did indeed work better. Sorry can't find the trial now.
Many prevention trials don't test for D deficiency and have too few participants for significance. Also it takes about 2 years to do a population level prevention RCT, and it's really expensive and hard to get drug companies interested, so that's a tough ask. Prof Adrian Martineau got funding to do one.
Studies on vitamin D, selenium etc:
https://docs.google.com/document/d/1q5IH2hGjjdPi-vcs4zOBlArgFJ9iSDdZVoceevUPI9c/edit?fbclid=IwAR1nEL2wsIWIi6XXYGs-QWHX_-pgjgg88cU7zQRrfxtCtdXChTqXMn78fyA
I've been going back and forth on these Vitamin D studies. However, unlike HCQ vitamin D has no toxic side effects in moderate doses (<50,000 IU). It's cheap to manufacture. It seems like a no-brainer to use it as a prophylactic. Just don't confuse it with a cure. I'm taking 2,000 IU/day, which is above the generally recommended dose of 600 IU/day, but well within the safe limits.
Vitamin D is needed for lots more than dealing with Covid and like a great many nutrient requirements, more and more past a min. level is a waste. Good ex is Vitamin C where no good data for taking lots of supplemental C, but overwhelming data for what happens if deficient. I did not see the Author of the post say what levels of D3 25 OhD were deemed too low nor did i see what targeted level of D supplementation range was tried for.
D absorption with age decreases and diet for many insufficient to reach appropriate minimal levels needed for overall health and even with supplementation one size not fit all as different people will absorb differently with some needing more D to reach the same level of Blood 25Ohd than others. Add age being an issue to absorbing from the sun and hence it likely that most will need some level of supplementation whether it helps with Covid or not
This is somewhat unrelated, but I've noticed that if my wife takes a vitamin D supplement for an extended period of time, or a larger dose, her anxiety increases. Her vitamin D tends to drop in January/February, but we have to be very careful about trying to increase it. Any ideas why, or if there is a way to raise vitamin D levels without raising anxiety?
Didn't see this in the comments, so FWIW https://jamanetwork.com/journals/jama/fullarticle/2776738
I'm absolutely convinced that vitamin D works. It's all about the way how you take it: you have to take it every day over long periods of time, BUT only in the short time span between your cardio exercise and your yoga exercise. Or if you have no time span between these two exercises, then you can take it right before your regular meditation practice. Studies also show that vit. D works better against covid if you unpack a clean face mask anytime you unpack a vitamin D pill. ;-)
Noob question from somebody who knows nothing about this:
If you think that sunlight modulates immune response to Coronavirus, why would you test vitamin D, rather than, um, sunlight?
Did you notice that the spanish and brazilian rcts used different forms of vitamin d?
If the supplementation studies used accepted recommended doses and levels of vitamin D, these might have been simply too low to produce a therapeutic effect, compared with sunlight exposure. This would account for the discrepancy between those studies and latitude/season effects. To quote a review of vitamin D medicine from 2007 (www.nejm.org/doi/full/10.1056/NEJMra070553):
--
Much evidence suggests that the recommended adequate intakes are actually inadequate and need to be increased to at least 800 IU of vitamin D3 per day. Unless a person eats oily fish frequently, it is very difficult to obtain that much vitamin D3 on a daily basis from dietary sources. Excessive exposure to sunlight, especially sunlight that causes sunburn, will increase the risk of skin cancer.125,126 Thus, sensible sun exposure (or ultraviolet B irradiation) and the use of supplements are needed to fulfill the body's vitamin D requirement.
--
Note also that the relationship between dietary vitamin D intake (including supplements) and serum levels is not linear but logarithmic. This makes it difficult to raise serum concentrations of vitamin D to useful levels with supplements. P.D. Mangan links to more publications here https://roguehealthandfitness.com/how-much-vitamin-d-do-we-need/.
PS: what sort of formatting does Substack support in comments? Not being able to blockquote is very annoying.
Has there been any research on whether sunlight-based vitamin D is good because of the increase of vitamin D vs. the reduction of cholesterol (which is what sunlight converts to vitamin D)? If it's the reduction of cholesterol, then vitamin D pills might not help.
You should watch the medcram deep dive into Vitamin D and covid - https://www.youtube.com/watch?v=ha2mLz-Xdpg
The author says,
"...if I were in the business of defending the Spanish study I might argue that it takes a few weeks for cholecalciferol to work, so giving it to someone who will already be dead or recovered by then is meaningless.
It would make me more comfortable if someone who did understand Vitamin D biochemistry would confirm this, but I'm not going to hang the whole argument on it."
Not wanting to hang the whole argument on the fact that it takes cholecalciferol 10 days to become available in the patient's body, and that it wasn't even adminstered until 11 days after onset of Covid, is utterly ridiculous. Those are the crucial flaws in the entire study, so it should not only be set aside, but thrown aside with great force. The author is nothing but a flak.
Excellent article.
I would just add that there is likely also an age factor - average population age decreasing as you approach the equator because of the old people in the US and Europe.
Hi,
I am reading this blog for the first time via the Naked Capitalism blog. As a chemist, I tend to focus on perspectives that include a biochemical view. Last fall, there were reports about AI studies that examined the broad array of Covid symptoms, and pointed to a bradykinin storm instead of a cytokine storm as a potential cause of the symptoms. Here is one magazine link that describes some of the biochemical cascades in the bradykinin system:
https://elemental.medium.com/a-supercomputer-analyzed-covid-19-and-an-interesting-new-theory-has-emerged-31cb8eba9d63
The article includes a list of potential treatments based on the bradykinin system, including the following quote:
“Interestingly, Jacobson’s team also suggests vitamin D as a potentially useful Covid-19 drug. The vitamin is involved in the RAS system and could prove helpful by reducing levels of another compound, known as REN. Again, this could stop potentially deadly bradykinin storms from forming. The researchers note that vitamin D has already been shown to help those with Covid-19”.
Also, I saw that one commenter here wrote about red light treatment. While I agree that low intensity laser light can be used to treat inflammation and respiratory problems, I think it’s possible that Vitamin D may be helping through a different mechanism. I and my friends and family have had good results with red light therapy for a range of indications. If you are curious, here is Dr. Fred Kahn’s blog post on light therapy and Covid:
https://fredkahnmd.com/2020/04/28/bioflex-laser-therapy-in-the-treatment-of-covid-19-infection/
He can’t promote the light therapy as a treatment for covid, but offers his opinion. (Spoiler alert: Dr. Kahn is a former vascular surgeon. I have been to him myself and he has very strong opinions about the medical system.) In a nutshell, when properly done using appropriate wavelengths, the light therapy triggers an increase in ATP formation, angiogenesis, and other knock effects including changes in the rate of DNA transcription in some cases.
> "Whenever you see a disease that's better in summer and worse in winter, Vitamin D is one of the possible culprits."
Um, then why haven't I been told to take vitamin D to prevent the flu during winter flu season? What I mean is, isn't there some other explanation you ought to have mentioned, like "people staying longer indoors makes viral spread easier" or "viruses can survive longer outside a host when temperatures are colder and there is less sunlight"?
Later in the article you mention reasons to be skeptical, but in general, I recommend against front-loading all evidence for X (and saving evidence against X for the bottom) unless readers are informed up front that you're structuring the article that way.
You should factor in the absolute huge mess of a country that Brazil is before you put any faith at all in a Brazilian study.
This is not xenophobia. I'm Brazilian. You have to either be here or have lived here for a while to get a grasp of what I'm talking about.
If the study showed no Vitamin D effects, odds are extremely high that the vitamin D used in the trial is simply counterfeit. Brazil has absolutely no QA in its genes. Bribery, on the other hand...
To get a feel of what I'm talking about: an entire major city, Rio de Janeiro, has been drinking brownish water that tastes like dirt for close to two years, and nothing is being done: https://www.thetimes.co.uk/article/smelly-brown-water-is-safe-rio-residents-told-k9hbhsdvp
Don't trust the Brazilian study. Don't trust any Brazilian study. Brazil is not a normal place by any measure.
This was good. The CDC and FDA's inability to do simple cost-benefit analysis was shocking and lead to tens if not hundreds of thousands of excess deaths when it came to masks. Sometimes, knowing something won't hurt is enough and the efficacy can be hazy. You need to look at the risk and reward and if something like wearing a mask or taking vitamin d is cheap, easy, and safe you just do it.