Background

Question
info icon

Does Vitamin D reduce the severity of COVID-19 outcomes?

Answer
info icon

Vitamin D reduces the odds for severe COVID-19 by a factor of around 5.

(
49%
probability)

Background
info icon

There is a heated debate regarding the effectiveness of vitamin D in improving COVID-19 outcomes. While numerous studies show a correlation, many claim it is due to confounding factors. A few controlled trials showed a strong effect, but the most notable one, conducted in Córdoba, Spain, had significant flaws that prevented it from being widely accepted.

More studies are underway, and within a year or so we will likely get a more certain answer, but a probabilistic analysis of currently available information can reach a useful conclusion faster, potentially saving hundreds of thousands of lives.

Additionally, the conclusion below is included in the Rootclaim $100,000 challenge.

Note: this is a simplified analysis; the full version was published in September 2020 on the Rootclaim blog.

Hypotheses Considered
info icon

Calculated Results
info icon

Calculated Results
info icon

1

49%
5-fold:

Vitamin D reduces the odds for severe COVID-19 by a factor of around 5.

49%

2

29%
20-fold:

Vitamin D reduces the odds for severe COVID-19 by a factor of around 20.

29%

3

15%
No Effect:

Vitamin D has no effect on COVID-19 outcomes.

15%

4

5%
2-fold:

Vitamin D reduces the odds for severe COVID-19 by a factor of around 2.

5%

5

2.2%
Damage:

Vitamin D worsens COVID-19 outcomes.

2.2%

Starting Point
info icon

Initial Probabilities

Name
Initial Likelihoods
info icon
Damage

10%

No Effect

67%

2-fold

15%

5-fold

5%

20-fold

3%

As a baseline, few drugs are effective for any specific disease. However, there are multiple studies showing a correlation between vitamin D and COVID-19, and several proposed mechanisms of actions. On the flip side, there is the risk that vitamin D could actually exacerbate COVID-19 by increasing the expression of ACE2.

We will represent these facts with the following prior probabilities of using a high-dosage, short term, vitamin D treatment for Covid 19:

  • Damage – 10%

  • Nothing – 67%

  • 2-fold improvement – 15%

  • 5-fold improvement – 5%

  • 20-fold improvement – 3%

Name
Initial Likelihoods
info icon
Damage

10%

No Effect

67%

2-fold

15%

5-fold

5%

20-fold

3%

Evidence
info icon

Effect
info icon

Reduced admittance to ICU

Name

Effect

info icon

Updated Likelihoods

Damage
÷
2000000

2.2%

No Effect
÷
6667

15%

2-fold
÷
4

5%

5-fold
×
125

49%

20-fold
×
83

29%

A randomized controlled study of 76 patients found significant improvements in patients treated with vitamin D relative to patients not treated with vitamin D. Specifically, 13 of 26 patients in the control group were admitted to the ICU, while only 1 of the 50 patients treated with vitamin D were admitted to the ICU.

Name

Effect

info icon

Updated Likelihoods

Damage
÷
2000000

2.2%

No Effect
÷
6667

15%

2-fold
÷
4

5%

5-fold
×
125

49%

20-fold
×
83

29%

Discussion
info icon

userIcon
user icon

Simon

Apr 14, 2022 at 8:45 PM

How does this new study affect the probabilites? https://www.medrxiv.org/content/10.1101/2022.03.22.22271707v1

user avatar

Dan Warren

Jul 28, 2021 at 2:40 PM

It seems ridiculous that the 76 participant preliminary study with obvious methodological problems has such a massive impact on the probabilities. Mauro posted a paper contradicting this claim to this thread more than a month ago, an analysis using a much larger data set than the one that flipped these probabilities. I'm inclined to suggest that rootclaim owes Butler-Laporte , Nakanishi et al $100K for debunking this.

logo icon

Rootclaim

Oct 7, 2021 at 11:05 AM

As we responded to Mauro, the Mendelian randomization studies are not too useful in this context, since genetics account for only a small portion of vitamin D blood level, which is almost exclusively dictated by environmental factors. As you say, the study we input into our analysis has certain limiting factors. However, to reduce the uncertainty that these factors introduce, we built a probabilistic model which took into account these limitations. This means we are using numbers that are more conservative than the results of the trial and still arriving at a high probability that vitamin D significantly improves COVID-19 outcomes.

user avatar

Mauro Orsini

Jun 23, 2021 at 3:06 PM

logo icon

Rootclaim

Oct 7, 2021 at 11:01 AM

The Mendelian randomization studies are not too useful in this context, since genetics account for only a small portion of vitamin D blood level, which is almost exclusively dictated by environmental factors.

user icon

claim_guest_25

Jun 15, 2021 at 7:25 AM

Is any research about N-COV-1 (SARS) and Vitamin D relevant? I heard that in Canada, SARS front-line workers were advised to increase intake of vitamin D; I don't know what that was based on. Apologies if this is old news. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3624188/ 72. Shaheen SO, Jameson KA, Robinson SM, Boucher BJ, Syddall HE, Sayer AA, Cooper C, Holloway JW, Dennison EM. 2011. Relationship of vitamin D status to adult lung function and COPD. Thorax 66:692–698 [PubMed] [Google Scholar] 73. Sundar IK, Hwang JW, Wu S, Sun J, Rahman I. 2011. Deletion of vitamin D receptor leads to premature emphysema/COPD by increased matrix metalloproteinases and lymphoid aggregates formation. Biochem. Biophys. Res. Commun. 406:127–133 [PMC free article] [PubMed] [Google Scholar]

logo icon

Rootclaim

Oct 7, 2021 at 11:57 AM

Thanks for sharing, definitely interesting. However, there is already enough evidence about SARS-2 and Vitamin D that we don’t feel that SARS-1 adds enough information.