A paper that purported to find vitamin D could reduce the severity of COVID-19 symptoms has been retracted from PLOS ONE, four years after the journal issued an expression of concern about the research.
The article, “Vitamin D sufficiency, a serum 25-hydroxyvitamin D at least 30 ng/mL reduced risk for adverse clinical outcomes in patients with COVID-19 infection,” appeared online September 25, 2020. Michael F. Holick, a professor of medicine at Boston University and a proponent of the use of vitamin D, was the last author among a group of other researchers from the Tehran University of Medical Sciences in Iran.
Soon after publication, the paper gained traction on platforms like Twitter (now X) as evidence vitamin D could treat COVID-19 symptoms.
Amidst this discussion, Nick Brown, a science integrity researcher at Linnaeus University in Växjö, Sweden, and Gideon Meyerowitz-Katz, a research fellow at the University of Wollongong, Australia, pointed out potential flaws with the work, including the small sample size of the study and a lack of patient information such as how the patients died.
PLOS ONE posted an expression of concern for the article on October 14, 2020. According to the notice, “concerns were raised about the validity of results and conclusions reported in the article and about undisclosed competing interests.” The expression of concern also noted “statements in the article, including in its title and conclusions, that suggest a causal relationship between vitamin D levels and the clinical outcome of COVID-19 infections which is not supported by the data.”
The competing interests refer to Holick’s “non-financial interests based on his vitamin D research and other activities focused on vitamin D; contributions to an app that tracks vitamin D; and interests that include consultancies, funding support, and authorship of books related to vitamin D usage.” In 2018, the New York Times reported on Holick’s financial ties to the vitamin D industry.
“This project did not receive any grants. Prof. Holick did not have any funding support for this project,” Mohammad Ali Sahraian, the paper’s corresponding author, told Retraction Watch.
Regarding the competing interests, Holick said:
The app dminder.info is free. I do not derive any income from it. I am no longer a consultant for Quest Diagnostics and therefore it was not listed. I do not receive any book royalties related to any information that is in this publication. I do not have any conflict of interest regarding any aspect of the study design, the results, or conclusions.
The paper has been cited 189 times, according to Clarivate’s Web of Science, with the bulk of citations coming after the expression of concern.
On June 6, nearly four years later, PLOS ONE retracted the study. A statistical reviewer and members of the PLOS ONE Editorial Board found that the study design was inadequate to address the research question and the methods used were not detailed enough to reproduce the study, according to the retraction notice. “As such, we have concluded that the article’s conclusions are not supported by the reported data,” the notice said.
Sahraian said he and his colleagues were “very surprised” and “upset” by the retraction, which should “not be decided without any ethical problem.”
“It would have been preferable for these issues to have been identified and addressed during the pre-publication review process, rather than after the work had already been made public,” Sahraian told Retraction Watch. “However, we did not claim a causal role of vitamin D in the clinical outcome of COVID-19 infections. The current study was cross-sectional and only considered the association between COVID-19 and circulating vitamin D levels.”
Holick, Saharian, and the paper’s first author do not agree with the retraction, while others could not be reached, the notice states. Sahraian said they don’t agree because they provided detailed responses to the editorial board’s concerns on multiple occasions, but did not receive any response from the journal.
“I believe it is unfair, unethical, and inhumane to retract an article due solely to the faults of the editorial team rather than addressing the flaws inherent to the study itself,” he said.
David Knutson, head of communications at PLOS, said, “the authors’ comment that the communications in this case were one-sided, and the implication that our editorial decision was biased, are not true.”
PLOS completed an objective evaluation and communicated with the authors on multiple occasions, Knutston said. He acknowledged a “long gap in our communications with the authors between March 2021 and December 2023” and apologized for one instance in which PLOS did not update an author following their query about the retraction decision.
Meyerowitz-Katz said he has mixed feelings about this retraction. “It is to the credit of the editorial team that they even bothered to investigate, and retracting on the basis of low quality – rather than misconduct or similar – shows a commitment to integrity that few journals display,” he said.
But retracting it “at this point will have no impact whatsoever on the pollution of the literature – the damage has well and truly been done. I think that the PLOS ONE editorial team should explain exactly why it took an entire pandemic’s worth of time to come to a decision, given that the issues they cite as reasons to retract are the same ones I pointed out on Twitter in September 2020,” Meyerowitz-Katz said.
Brown said it would have been better if the retraction “didn’t take quite so long,” but given the expression of concern was issued rapidly, “I think the journal handled it overall fairly well.”
PLOS ONE cited an ethics case backlog for their slowness in retracting this paper, among other papers.
Knutson told us the case was “complex,” and the four year time period “reflects the time we required to complete a rigorous assessment and investigation,” as well as “some internal delays due to competing priorities.” He said:
Given the complexity of this case, the nature of the concerns, and potential clinical implications of the article and the concerns raised, we published an interim Expression of Concern soon after the concerns were raised to our attention.
Knutson said the competing interests mentioned in the expression of concern did not play a part in the retraction of the paper. “The retraction decision was based on concerns about the study design, methodological reporting, and the reliability of the article’s conclusions,” he said.
Like Retraction Watch? You can make a tax-deductible contribution to support our work, subscribe to our free daily digest or paid weekly update, follow us on Twitter, like us on Facebook, or add us to your RSS reader. If you find a retraction that’s not in The Retraction Watch Database, you can let us know here. For comments or feedback, email us at [email protected].
I always wondered how some of the colleagues publish 20-30 papers a year. I always tell my PhD students if they can write a major paper or two based on the work that would be excellent. But the academia is now full of people who just count the numbers and H-index. Your work helps me to understand this proliferation of publications, and this is not good for science long termly.
From within our genuine clinical practice with enthusiastic colleagues, the maximal number of papers we can publish on annual bases is 6-8. Every single paper must encompasses full clinical documentation, not fake numbers of imaginary patients. I reviewed more than 500 papers from different international medical journals. 90 percent are fake and have nothing to do with authentic clinical practice. The falacies in published papers are endless and the level and the quality of the reviewers are unacceptable .
The paper not only gained traction on Twitter but was endorsed by Anthony Fauci.
Did he? I’ve only seen Fauci comment on vitamin D supplementation in case you are deficient, not specifically referring to this paper.
The retraction of the article is a fatal mistake on the part of the PLOS ONE editorial team, as several other observational studies have confirmed the results of the study.
In the meantime, the results of observational studies, such as the result of thestudy by the University of Heidelberg are no longer are no longer doubted.
“Vitamin D Deficiency and Outcome of COVID-19 Patients ”
https://www.mdpi.com/2072-6643/12/9/2757
„In our patients, when adjusted for age, gender, and comorbidities, VitD deficiency was associated with a 6-fold higher hazard of severe course of disease and a ~15-fold higher risk of death.“
Unfortunately, like many other scientists, the PLOS ONE editorial team doubts the results of the observational studies only because they could not be replicated in intervention studies.
It is more likely that the results of the many intervention studies can be doubted because they did not use the form of vitamin D that was already present in the observational studies.
In observational studies, the storage form of vitamin D calcidiol is already present, which acts quickly. In most intervention studies, however, vitamin D3 in the form cholecalciferol was administered, which must first be converted into calcitriol before it becomes effective through further conversion into calcitriol.
In particular, the initial conversion of D3 to calcidiol can take several days. Since patients are typically only admitted to intervention studies on the day of hospitalisation and sepsis is already present on this day, which has to be treated within a few studies, D3 supplementation can not help much.
If calcidiol levels were measured longitudinally in intervention studies, this could be recognised. However, as this is not common practice, it is not recognised that a severe 25(OH)D deficiency can occur despite D3 supplementation, causing the immune system to fail.
At least calcidiol must therefore be administered, which can be converted into calcitriol within a few minutes
ChatGpt summarised this very aptly after a discussion:
>> The form of vitamin D administered is critical to its effectiveness in acutely ill patients. Cholecalciferol (vitamin D3) may not have the rapid effect required in acute infections and incipient sepsis. Calcidiol and calcitriol, which are more rapidly bioavailable and act directly, have tended to show better results in studies.
For future studies, it is important to consider the choice of vitamin D form and possibly favour fast-acting forms such as calcidiol or calcitriol, especially in patients with acute illness or sepsis.<<
Unfortunately, there are about 140 intervention studies in which vitamin D3 was administered, but only 3 studies in which the fast-acting forms calcidiol or calcitriol were administered. The typical result of these 3 studies was that no patient died and very few had to be ventilated.
The best known of these is this one:
Entrenas Castillo M, Entrenas Costa LM, Vaquero Barrios JM, et al.
"Effect of calcifediol treatment and best available therapy versus best available therapy on intensive care unit admission and mortality among patients hospitalized for COVID-19: A pilot randomized clinical study."
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7456194/
However, it will probably take some time before it is recognised that the 140 studies are flawed because the wrong form of vitamin D was used in them, and only the 3 studies with calcidiol or calcitriol delivered correct results.
The main reason why it will still take some time is that it must first be generally recognised that vitamin D is a "negative acute phase reactant" and that the 25(OH)D value can fall by up to 2.5ng/ml per day during an infection because this is used to activate T-cells, which are then used to fight viruses.
More on vitamin D use during an infection
Vitamin D the life insurance against Long Covid and autoimmune diseases
https://www.facebook.com/share/p/LzLmk3ycoXZ568vT/
This sharp drop in 25(OH)D levels cannot be stopped quickly by D3 supplementation, but can be stopped by calcidiol supplementation.
Only when it is recognised that the time factor plays a role due to the high daily vitamin D use will many people realise why the results of intervention studies are so strongly dependent on the form of vitamin D used.
Then hopefully all patients who are hospitalised with signs of sepsis will be given one of the fast-acting forms of vitamin D immediately.
I remember engaging in a battle with the Iranian authors on ResearchGate in 2020, when they first pitched their absurd ideas. One went as far as suggesting that Africans of Sub-Saharan origin had a particular propensity for COVID-19 due to their higher melanin makeup. At the time, I had expected more support (being from that group myself), but the silence was deafening. There’s some consolation (and personal pleasure) in reading of the retraction is such hogwash studies.