Finally Confirmed! Vitamin D Nearly Abolishes ICU Risk in COVID-19

Vitamin D and COVID-19

September 3, 2020

The first randomized controlled trial (RCT) of vitamin D in COVID-19 has just been published. The results are astounding: vitamin D nearly abolished the odds of requiring treatment in ICU. Although the number of deaths was too small to say for sure, vitamin D may actually abolish the risk of death from COVID-19.

The First Randomized Controlled Trial on Vitamin D and COVID-19

The trial was conducted at the Reina Sofía University Hospital in Córdoba, Spain. The trial included 76 patients with COVID-19 pneumonia. Although this is no longer the standard of care, all patients were treated with hydroxychloroquine and azithromycin and, when needed, a broad-spectrum antibiotic. Admission to the ICU was determined by a multidisciplinary committee consisting of intensive care specialists, pulmonologists, internal medicine specialists, and members of the ethics committee.

The patients were randomly allocated to receive or not receive vitamin D in a 2:1 ratio. This resulted in 50 patients in the vitamin D group and 26 patients in the control group.

The Vitamin D Treatment Protocol

The vitamin D was provided as oral calcifediol, also known as calcidiol, 25(OH)D, and 25-hydroxyvitamin D. This is a metabolite of vitamin D that our livers make. It is is the principle form of vitamin D that circulates in the blood, and we use it as a measure of vitamin D status.

Traces of 25(OH)D occur in food, and it is five times as potent as vitamin D. As described on page 255 of the 1997 DRI Report, 25(OH)D is given an international unit (IU) value that equates it to vitamin D. Whereas one microgram (mcg) of vitamin D is 40 IU, 1 mcg of 25(OH)D is 200 IU.

The treatment in this RCT was soft capsules of 532 mcg 25(OH)D on day 1 of admission to the hospital, followed by 266 mcg on days 3 and 7, and then 266 mcg once a week until discharge, ICU admission, or death.

This is equivalent to 106,400 IU vitamin D on day 1, 53,200 IU on days 3 and 7, and 53,200 IU weekly thereafter. If this were given as daily doses, it would be the equivalent of 30,400 per day for the first week, followed by a maintenance dose of 7,600 IU per day.

The vitamin D status of the patients was not measured. However, the average vitamin D status in this region of Spain during the time of year in which the study was conducted is 16 ng/mL. A single dose of 100,000 IU vitamin D tends to raise a 25(OH)D of 10 ng/mL into the 20-30 ng/mL range. My suspicion is that the bolus dosing in the first week brought the patients' vitamin D status into the 30-40 ng/mL range by the end of the week, and that most of the healing took place in the circa 40 ng/mL range.

The Results: Near Abolition of ICU Risk

The results are absolutely stunning. 50% of the control group (13 people) required admission to the ICU. Only 2% of those in the vitamin D group (one person) required admission to the ICU.

Expressed as relative risk, vitamin D reduced the risk of ICU admission 25-fold. Put another way, it eliminated 96% of the risk of ICU admission. Expressed as an odds ratio, which is a less intuitive concept but is often used in statistics because it gives an estimate of the effect of the treatment that would be constant across scenarios with different levels of risk, vitamin D reduced the odds of ICU admission by 98%. Either way, vitamin D practically abolished the need for ICU admission.

This was statistically significant at p<0.001, and the 95% confidence interval was 0.002-0.17. This means that the probability of observing differences this large or greater if there is no true effect of vitamin D is less than one in a thousand, and that the probability is 95% that the true effect lies somewhere between an 83% and a 99.8% reduction in the odds of ICU admission.

No matter how you slice it, the effect of vitamin D is extremely compelling.

Because the study is small, the potential confounding variables were not perfectly evenly distributed between the two groups. There was more high blood pressure in the control group, and there was a borderline greater number of patients with diabetes in the control group. Though not statistically significant, the vitamin D group had more people over the age of 60 and five times as many people with organ transplants or who were otherwise on immunosuppressive drugs. To account for all of these differences, they adjusted for them statistically. In the adjusted model, vitamin D still reduced the odds of ICU admission by 97%, with the 95% confidence interval ranging from a 75% to a 99.7% reduction in the odds.

Did Vitamin D Also Abolish the Risk of Death?

All of the vitamin D patients were discharged without complications. Half the control group was discharged without ICU admission. Among the other half, 11 were eventually discharged from ICU, and 2 died.

Although there were, thankfully, too few deaths to run statistics on, since patients generally would be admitted to the ICU before dying, and since vitamin D nearly abolished the risk of ICU admission, we can infer that in a larger study with more deaths, vitamin D would probably abolish or nearly abolish the risk of death.

Comparison With Observational Studies

These results are consistent with the first observational study on vitamin D, which found that 96% of severe and critical cases occurred at 25(OH)D under 30 ng/mL, whereas 97.5% of mild cases had 25(OH)D above 30 ng/mL.

This study couldn't measure the effect on mortality, but all two deaths were in the control group and the near abolition of ICU risk suggests that death would also be nearly abolished. This is consistent with the second observational study finding that only 4% of those with vitamin D status above 30 ng/mL died, while 88% of those with vitamin D status at 20-30 ng/mL died, and 99% of those with vitamin D status lower than that died.

It is less consistent with the weaker observational studies that came out later. For example, the fourth observational study found that prevalence of 25(OH)D below 20 ng/mL differed across three categories of severity judged by CT scan in males but not females. It is also less consistent with more recent studies in Iran and England. In Iran, 25(OH)D below 30 ng/mL only increased risk of severe infection by 21%, although mortality was roughly cut in half at that level. In England, ICU patients were half as likely to have 25(OH)D above 20 ng/mL as non-ICU patients, and mortality wasn't associated with vitamin D status.

In other words, the first RCT shows the effect of vitamin D is closer to the more extreme estimates of the first observational studies than it is to the more moderate estimates of the more recent studies.

The observational studies on infection risk are weaker than those on severity and mortality (see here, here, here, and here), but this RCT didn't look at infection risk.

The RCT contradicts the findings of the two Mendelian randomization studies. These found either no evidence for causality in the vitamin D/COVID-19 connection, or only weak evidence. However, these studies infer causality by looking at the effect of genetics. Genetics only explain 4.2% of the variation in 25(OH)D and many of the genes involved have non-specific, indirect relationships to vitamin D status. This RCT overwhelmingly takes precedence over the findings of the Mendelian randomization studies.

This Study Is the Single Most Important Vitamin D and COVID-19 Study

Since the first vitamin D study was released as a preprint* on April 23, we have been waiting for data that could settle whether the association between vitamin D and COVID-19 incidence, severity, and mortality is a causal one.

This study settles the question: yes, it is causal.

It is important for scientists to replicate each other's findings. Future studies with more diverse populations may reveal differences in the response between different populations. Future larger studies will more precisely refine the exact effect of vitamin D. Future studies with different dosing protocols, in different contexts (for example, without the use of hydroxychloroquine and azithromycin, or with the administration of other drugs or nutrients) will add nuances to our understanding of the effect of vitamin D.

However, given the degree to which it is nearly entirely harmless, and almost without exception beneficial, to maintain 25(OH)D above 30 ng/mL, it would be irresponsible not to interpret this study as definitive evidence of causality. At an absolute minimum, maintaining vitamin D status in this range should be part of the public health message to reduce COVID-19 risk, and 25(OH)D should be universally screened in all hospitals to be treated in anyone with COVID-19, and should be universally screened in all testing centers when antibodies and PCR testing is done, so that everyone knows not only their COVID-19 exposure but also their vitamin D status. If it's low, they should be given advice on how to bring it back up to normal.

How I've Changed My Position

On March 17, when I released Version 1 of The Food and Supplement Guide for the Coronavirus, I expressed concern that vitamin D might increase infection risk or severity by increasing ACE2, the entryway of SARS-CoV-2, the coronavirus that causes COVID-19, into our cells. In response to the first observational study of vitamin D released on April 23, which I covered in this newsletter on April 24, I released version 3 of the guide on April 28, in which I revised this stance and took the position that 25(OH)D should be maintained at 30-34 ng/mL.

As observational studies accumulated, they converged on the common point that 30-40 ng/mL is the sweet spot where infection risk, severity, and mortality are all lowest, without any risk of getting too much vitamin D.

I remained concerned that there might be some risk of a U-shaped curve, where risk increases at vitamin D levels higher than 40 ng/mL. Evidence for or against this was scant and contradictory. Data from Switzerland provided hints of an increased infection risk above 40 in those under 70 years old, but not in those older than 70. Data from Chicago suggested no difference in infection risk above 40 ng/mL, but it didn't separate the people by age. Data from Israel suggested that 25(OH)D above 53 ng/mL completely abolishes infection risk, but everyone with levels that high was under the age of 50. In the Israeli study, the relevant data were only in the young, and risk seemed abolished. In the Swiss study, the relevant data existed for all ages, and it suggested risk increased in the young. All of these studies concern infection risk, not severity or mortality, they all have too little data at high 25(OH)D levels, and they contradict each other.

While I remain agnostic whether there is some level of vitamin D above which infection risk is increased, the new RCT concerns severity and mortalityAlthough 25(OH)D levels weren't measured, they must have hit at least 40 ng/mL by the time the patients were released, and they possibly exceeded 50 ng/mL. That they gave these patients such massive doses of vitamin D without measuring their vitamin D levels, and that this nearly abolished the need for ICU admission, suggests that there should be no concern about a U-shaped curve with severity or mortality for short-term dosing of up to 8,000 IU per day over several weeks or for temporarily pushing 25(OH)D above 50 ng/mL during the course of treatment.

Given that observational studies around infection risk can consistently show that it is increased at low vitamin D status but cannot show any consistent picture at high vitamin D status, I am, for now, relieving myself of concern about this U-shaped curve. If high vitamin D status can almost eliminate severity and mortality, the off-chance that in some people at some high level it could increase infection risk should be tolerated.

You can see the evolution of my stance on vitamin D through these past posts:

Taking Action on These Findings

I think the best way to take action on these findings is to maintain 25(OH)D in 30-40 ng/mL. This is the sweet spot according to the observational literature, and if we assume the patients in the Spanish RCT came in with 16 ng/mL, 30-40 ng/mL probably approximates their average blood level over the course of their recovery.

This is also consistent with other metrics of optimal vitamin D status. All-cause mortality is lowest at about 28 ng/mL according to one meta-analysis of eight European studies, seven conducted in the general population, pooling data from just under 27,000 people. Another widely circulated meta-analysis showing a bottoming-out in the 40-60 ng/mL range derived its data in that range entirely from a conference abstract. If that one outlier is removed, as can clearly be seen in Figure 4 of the paper, the bottoming out of all-cause mortality is below 40 ng/mL and probably close to 30 ng/mL. This is consistent with the amount needed to maximally suppress parathyroid hormone (PTH) in most people, which is the body's own sign of inadequacy in the vitamin D and calcium economy. This is reflected in the widespread use of 30 ng/mL as the boundary of adequacy.

If one is maintaining 25(OH)D this high, one could mimic the maintenance dose of vitamin D used in this study by taking 7,600 IU per day upon the onset of symptoms through recovery.

If one goes into an illness with 25(OH)D below 20 ng/mL, one could quickly bring this level up using a bolus dosing approach like in the RCT. Their approach would equate to using 100,000 IU on the first day, followed by 18,000 IU per day for the next six days, then switching to the maintenance dose; or by using 30,000 IU each day for the first seven days, then switching to the maintenance dose; or by using 100,000 IU on the first day, 50,000 IU on the third day, 50,000 IU on the seventh day, and then switching to the maintenance dose.

Version 6 of the Food and Supplement Guide for the Coronavirus

I have now released Version 6 of The Food and Supplement Guide for the Coronavirus to reflect the new study on vitamin D. Purchases of the guide are greatly appreciated, as they help sustain my work on this newsletter and will help me start finishing my Vitamins and Minerals 101 book.

The Bottom Line

The data are now in. The effect of vitamin D on COVID-19 severity, and likely mortality, is causal. Maintaining 25(OH)D 30-40 ng/mL is likely to be strongly protective against having a severe or fatal case. Use of bolus dosing as described in the “taking action” section upon the first sign of symptoms, if one has levels much lower than this, and otherwise supplementing with a maintenance dose of 7-8,000 IU per day during illness, is likely to be strongly protective against severe and fatal cases.

Stay safe and healthy,

Share This on Social Media

Join the Next Live Q&A

Have a question for me? Ask it at the next Q&A! Learn more here.


Subscribe or upgrade your subscription here.

Masterpass members get access to premium content (preview the premium posts here), all my ebook guides for free (see the collection of ebook guides here), monthly live Q&A sessions (see when the next session is here), all my courses for free (see the collection here), and exclusive access to massive discounts (see the specific discounts available by clicking here). Upgrade your subscription to include Masterpass membership with this link.

Learn more about the Masterpass here.

Take a Look at the Store


*  The term “preprint” is often used in these updates. Preprints are studies destined for peer-reviewed journals that have yet to be peer-reviewed. Because COVID-19 is such a rapidly evolving disease and peer-review takes so long, most of the information circulating about the disease comes from preprints.