No evidence Vitamin D affects Covid severity in people of European ancestry | Hacker News Reader

No evidence Vitamin D affects Covid severity in people of European ancestry

brunel.ac.uk

35 pointsbdz14 comments

Studies that show no evidence of a particular phenomenon tend not to get the same degree of publicity that studies showing something do – but it’s great that scientists are making the effort to determine if a particular correlation is a result of causation. The conclusion of this study is that Europeans (people with other ancestry were not covered in the study) with naturally high levels of Vitamin D don’t suffer less from Covid 19 than those without. Regardless, it’s still not a bad idea to take Vitamin D supplements for other health benefits.

Too late to edit: I meant to qualify the last sentence that Vitamin D supplements are probably more beneficial in winter, for people not exposing their skin to sun or those with high levels of melanin.

For context, related HN submission on this issue – Lack of Vitamin D is associated with SARS-CoV-2 infection and mortality: https://news.ycombinator.com/item?id=25618273

Not all studies are correlational. There is a small Spanish RCT that showed quite phenomenal results: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7456194/

Here's the paper: https://nutrition.bmj.com/content/bmjnph/early/2021/01/07/bm...

Essentially, they appear to have used stratification on genetic factors to reduce potential confounders and improve their ability to get a less biased estimate.

This seems reasonably solid to me, from a brief overview of the methods section (I'm not an expert though, so there may be flaws I missed).

Definitely some countervailing evidence against Vit-D being directly related. There's almost certainly some third factor between Vit-D and health, but we don't yet have a good sense for what that might be.

Note that the study is only analysing data for White British individuals, so if you are of a different ethnicity there may still be hope for you ;)

>There's almost certainly some third factor between Vit-D and health

A healthy balanced diet and (outdoor) life/exercise?

Regardless it seems that a lack of vitamin D is bad for you, so even if it doesn't directly help with Covid taking suplements seems a nobrainer to me.

Yeah, that's my prior.

Rather difficult to measure and put into an RCT though, unfortunately.

Seems flawed to me. When the individuals with genetic factors that make them Vit-D deficient go to the doctor, they will almost certainly be told to take Vit-D supplements at a higher rate than those without the genetic factor. This confounds the results and makes this type of study useless as it does not measure actual vitamin-D levels.

Actual text. https://nutrition.bmj.com/content/early/2021/01/07/bmjnph-20...

The analysis is a mess. Why can't they put it in a Google Colab notebook to check their calcuations?

There really needs to be a proper big trial conducted on this. There's too much evidence pointing to a possible link between Vitamin D supplements and better Covid outcomes for people with Vitamin D deficiency to dismiss it based on one meta-study.

This wasn't a meta study, but perhaps you are referring to one you know of.

When I heard the original report on vitamin d deficiency and getting seriously ill from COVID, my first thought was people who have sedentary lives tend not to be outdoors been active.

I’ll add an assumption that this may also follow with another correlation that diabetics suffer more from COVID for perhaps the same sedentary reasons.

Might be nonsense but correlation doesn’t have to mean cause - arguing against myself and bad science.

Any opinions from a clinician?

This is one meta-study, it's not exactly proof.

There are no proofs in science.

No, but randomized controlled trials provide much more evidence than correlational studies. I'll admit I don't fully understand the methods these researchers use, so it's possible that they address the common problems, but historically conclusions from trying to establish causal relationships through data mining have not had a great track record. The bits I did understand seem solid enough (vitamin D and obesity levels 10-15 years ago, when the participants were enrolled, and now should be pretty highly correlated, and the sample size is pretty large) but there are a lot of non-obvious pitfalls when trying to establish causality with purely observational methods.

Reply on news.ycombinator.com