Study published in JAMA Network Open: Vitamin D signficantly cuts ICU admission rate

In a small double blind randomized control trial (pdf at link) at a Spanish hospital early administration of calcifediol (a Vitamin D metabolite) significantly cut ICU admission rate and was associated with a less severe clinical course.

Calcifediol (25-hydroxyvitamin D) is what the liver turns vitamin D3 into as a normal part of vitamin metabolism in a process that normally takes about 7 days. Calcifediol is what is actually tested for when testing a patient’s vitamin D3 levels. Calcifediol is further processed in the kidneys to make calcitriol which is the active form of vitamin D used by the body.

In the study patients who had a prior medical record of a vitamin D deficient test result within the one year prior to testing positive for COVID-19 were randomly assigned to receive the best standard of care (control group) or best standard of care and early administration of calcifediol (treatment group).

The treatment group had a far lower rate of rate of ICU admission and a less severe course of the disease.

The trial, run out of the Reina Sofia University Hospital in Córdoba Spain, only had 76 participants.

“Of 50 patients treated with calcifediol, one required admission to the ICU (2%), while of 26 untreated patients, 13 required admission (50 %) p value X2 Fischer test p < 0.001.”

The authors suggest a larger trial.

I’d be quite skeptical of this study until confirmed at other centres and larger volumes.

OTOH lots of people are deficient in Vitamin D, it’s cheap and OTC, and it’s unlikely to hurt anything.

Although vitamin D is cheap and some are deficient, and it is unlikely to hurt. But it is very hard to credit the vast improvement in Covid morbidity claimed.

There are numerous clinical trials underway including sizable numbers of participants, to see if vitamin D supplementation cuts risk of getting Covid-19 or limits severity. Some but not all preliminary studies have been positive.

One quibble about all of this (actually a common one to these types of studies). This study was of people who were Vitamin D deficient, and to the point that they were put on drugs for it prior to the pandemic. It’s frequently the case that alleviating an extreme deficiencies (or the opposite) can have an impact that would not be present in cases where the deficiency was not as severe (or non-existent).

IOW, even if this study is completely valid, it doesn’t necessarily follow that any random person is accomplishing anything at all by upping their Vitamin D levels.

[Another area where this is a hot issue is the question of whether an average person would benefit from lowering their salt intake, even though people with high blood pressure have been shown to benefit. And so on.]

It’s important to remember that when they say things like “p < 0.001” it doesn’t mean that there’s that level of confidence that the specific measured difference is accurate. It just means that it’s that (un)likely that there is zero difference between the two groups and that the entire difference in outcome is random chance. But there’s still a very good chance that the true difference between the two groups is of a very different magnitude than what was measured in this study.

It looks like they didn’t actually have the participants pre-study vitamin D levels, so they weren’t selected based on vitamin D deficiency.

Serum 25OHD concentrations at baseline or during treatment are not available [39,40]. Overall, adults living in the Córdoba area are relatively vitamin D deficient (16 ng/mL on average) in late winter and early spring [17].

I’m sure I saw in one of the media summaries of the study that it said they used vitamin D deficient patients, but I’m not seeing that in the actual study. Maybe it’s there someplace else, and I’m just missing it. If the subjects’ vitamin D deficiency is representative of the general population, then that suggests the results apply to the general population (or at least those of the general population who get severe enough COVID to end up in the hospital).

It is possible the effect is not as large as was found in this study. The study is small, and found a large effect. A small study will only find a large effect, because they don’t have enough power to find small effects. Was the study just “lucky” and the real effect size is much smaller or non-existent? Absolutely, and if this is the one study out of a 1000 done on vitamin D and COVID that is getting published because it found something, then it is certainly possible that the finding is due to random chance.

This is just a pilot study, so it doesn’t have to be perfect. The point is to do something that is just big enough to make a decision about doing something much larger. Unless it turns out there are horrible problems that aren’t being reported, the results are interesting enough to justify doing much larger trials.

I was going by the third paragraph of the OP, which stated "In the study patients who had a prior medical record of a vitamin D deficient test result within the one year prior to testing positive for COVID-19 …

I’m not sure you understood the context of my quote. I was responding to Dr_Paprika, who seemed to be casting doubt on the credibility of the study based on the magnitude of the difference, which he did not find credible. My point was that it’s possible that the magnitude was inflated by random fluctuation but that the study was still validly finding a genuine difference.

The magnitude of the difference is very surprising despite the possible role of vitamin D in activating T cells. It means the results should be viewed skeptically until a larger study is done with further quantification of initial vitamin D levels and better understanding of the mechanism. But it does not mean the study is unimportant, wrong or should be ignored. Particularly due to the low cost and high safety of the intervention, as pointed out. My guess is there is a much smaller effect - by all means more studies should be done.