Widely shared vitamin D-COVID-19 preprint removed from Lancet server

A preprint promoted by a member of the UK Parliament for claiming to show that vitamin D led to an “80% reduction in need for ICU and a 60% reduction in deaths” has been removed from a server used by The Lancet family of journals.

The preprint, “Calcifediol Treatment and COVID-19-Related Outcomes,” was posted to Preprints with The Lancet on January 22. On February 13, David Davis, a Conservative member of UK’s Parliament, tweeted:

The tweet has been retweeted and liked tens of thousands of times. Almost immediately, the paper began attracting criticism, particularly about its claims to have been a randomized study. Here’s one comment on the preprint site:

Although the paper claims it is a randomised study, it also says that all patients treated in 5 wards received calcifediol treatment, while all three of the other wards received no calcifediol. How this study can be considered randomised is therefore questionable (maybe the wards were randomised but that is a very poor level of randomisation). It is also presumably open label, meaning that the attending physicians and decision makers would have been well aware whether the patients were receiving calcifediol or not. Its concerning to me that that in the calcifediol group more patients apparently died than were referred to the ICU. In the control group approx. 50% of the patients referred to the ICU died (assuming all those who died were ICU patients). This raises some troubling questions about the decision making process in the calcifediol group, were patients not referred to ICU who should have been?

Others, including Aurora Baluja, an anesthesiologist and critical care doctor in Spain, raised concerns on Twitter. And on February 15, The Lancet Infectious Diseases asked Baluja for an “urgent” review of the paper. 

In that review, which Baluja has posted to Github, she raises questions about the alleged randomization, and also whether the trial was properly registered — a requirement for publication by most medical journals. Baluja also asked for the preprint to be taken down from The Lancet’s servers and posted instead to medRxiv, “to allow for continuous improvements and extensive prepublication review.”

Today, The Lancet Infectious Diseases removed the preprint, replacing it with:

We have removed this preprint due to concerns about the description of the research in this paper. This has led us to initiate an investigation into this study.

The comments that have been posted on this preprint will remain available on this page. Please note that this comment thread is now closed to further posts.

This is the 85th paper about COVID-19 that has been retracted or withdrawn, according to our records.

Nick Brown, who brought the preprint to Baluja’s attention, tweeted:

Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on Twitter, like us on Facebook, add us to your RSS reader, or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].

8 thoughts on “Widely shared vitamin D-COVID-19 preprint removed from Lancet server”

  1. This article unfortunately has a connotation to the untrained eye, like mine, that vitamin D has no role to play in the fight against Covid-19. It also seems to contradict the theory that vitamin D deficiency also does not play a role….. and that there is no co-relation to the disproportionate death of people of colour in the northern hemisphere.

  2. Oddly if you read the paper it does not claim to be an RCT and the only place it says it is right at the end where it says under Outcomes ” The effect of calcifediol administration was studied in a prospective open randomized controlled trial. ” In the rest of the paper it says patients randomised to wards.

  3. Also of interest is that headlines like “Vitamin D Flops in Moderate-Severe COVID — Randomized trial in Brazil disappoints” in Medicalpagetoday and others raises no concerns or Twitter storm, because it matches people beliefs. Yet the original design of the random trial is fatally flawed. You cannot use un-activated vitamin d in seriously ill people because of the time it takes for the liver to convert it to 25(OH)D. This delay means that the people either die or recover before it has any effect, as pointed out by the Spanish team.

    1. 25(OH)D is not really the active compound either – and the 25(OH)D levels were measured and shown to be substantially and rapidly increased by supplementation. The authors even note that “These analyses indicate that a single oral dose of 200 000 IU of vitamin D3 can rapidly increase 25-hydroxyvitamin levels, so the present null findings cannot be attributed to the failure of increasing serum 25-hydroxyvitamin D levels.”

    2. “unactivated” vitD is such idiotic take from biochemistry perspective: neither ‘unactivated’ vitD, nor ‘activated’ vitD – calcifediol (25(OH)D) are biochemically active; 25(OH)D STILL NEEDS to be additionally hydroxylated why no one uses 1,25(OH)2D (the real ‘activated’ vitD(3))? because hypercalcemia is deadly in people with cardiac complications -_-

  4. Yes, I wholly agree with Pete above. The paper couldn’t prove effectiveness of Vitamin D when it had no chance of being metabolized in the time frame.

    It’s one of those things about medical literature itself. Even when there is the direct ability to measure things like serum level, studies simply don’t do it. I don’t care how many pills one pops, it all depends upon what gets into the system and whether people were deficient in the first place.

  5. It’s a bit sad you just parrot Pete and not read my response. The authors in that study from Brazil DID measure serum levels of 25(OH)D, showed a substantial increase, meaning there WAS enough time to metabolize the vitamin D.

    I quote once again: “These analyses indicate that a single oral dose of 200 000 IU of vitamin D3 can rapidly increase 25-hydroxyvitamin levels, so the present null findings cannot be attributed to the failure of increasing serum 25-hydroxyvitamin D levels.”

  6. marco, the 25(OH)D was measured at discharge from hospital, so we do not know how much they had inside them when they were struggling with covid. You do not therefore know.

    Secondly the charts shown in the final paper are different from those in the preprint oops why?

    But what is most idiosyncratic about the Brazil study it was measured on time spent in hospital. so an early death contributed the same as an early discharge to the result

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.