PLOS One has retracted a paper linking vitamin D levels and COVID-19 morbidity three years after a critic flagged the data in the study as “deeply bizarre.” The authors objected to the retraction, with one calling it “outrageous” and pointing to flaws in the published notice.
The article, which appeared in February 2022, claimed people with low levels of vitamin D were at increased risk for severe COVID-19 and were more likely to die of the disease than other patients. It has been cited 65 times, according to Clarivate’s Web of Science.
The paper had a “huge, immediate impact,” said Gideon Meyerowitz-Katz, a senior research fellow from the University of Wollongong in Australia, citing the fact that the paper had been viewed over 1 million times within six weeks of being published. The article joins others, many also flagged by Meyerowitz-Katz, purporting to find links between vitamin D intake and COVID-19 severity that have been retracted or removed.
In an April 2022 thread on X, Meyerowitz-Katz noted the study’s “MASSIVE effect, whereby virtually everyone who died” had vitamin D levels below the normal range, he wrote at the time.
Meyerowitz-Katz also wrote that, while the paper had “no obvious signs of fakery,” the data were “deeply bizarre.” He noted all of the patients with chronic obstructive pulmonary disease in the study also had hypertension, a trend not in itself unusual besides the complete overlap.
The study considered low vitamin D levels to be less than 20 ng/mL, and 15 people had a vitamin D level of exactly 10 ng/mL. Ten of those 15 died, a larger percentage than any other bracket. “This makes having a vitamin D of exactly 10ng/mL the most dangerous range seen in the study,” Meyerowitz-Katz wrote.
The September 8 notice cites a “fatal flaw” with the study design “which prevents testing of the hypothesis.” David Knutson, the head of communications at PLOS, told us the flaw refers to the fact the authors included only patients who had already been tested for vitamin D. The notice states the fact the patients had been previously assessed for vitamin D levels and that that level may have changed prior to COVID-19 infection may have introduced “confounding factors” or bias.
Eyal Sela, director of the Otolaryngology Head & Neck Surgery Department at Galilee University in Israel, and the senior author of the article, called the journal’s decision “outrageous.”
“We did the research during COVID, which put all of us in a position to examine any patient who entered the hospital without any selection,” he told Retraction Watch, maintaining patient selection was unbiased. The research was conducted in Sela’s lab at the university.
After responding to Meyerowitz-Katz’s posts on X, the journal asked a member of the editorial board and an independent statistical expert to reassess the work.
Meyerowitz-Katz had followed up with the journal in March 2023 and was told by Maddy Ghose, an editor at PLOS One, the journal was still investigating. He said he hasn’t heard anything since.
As for why the retraction ultimately took three years, “the case was delayed at different points for various reasons,” Knutson told us.
All but three of the study’s 18 authors disagreed with the retraction. The rest did not respond, the retraction notice states.
An email to Retraction Watch signed by three of the authors — Amiel Dror, Michael Edelstein and Orly Yakir, writing “on behalf of the authors” — says the possibility of “measurement and selection bias is there and acknowledged in the paper.” They also noted the patients’ vitamin D values the researchers used weren’t recorded as part of their study; instead, they included patients who both tested positive for COVID-19 and had existing prior levels available in the hospital database.
Both Dror — the study’s corresponding author — and Yakir are researchers at Galilee University, and Edelstein is a professor of public health and epidemiology at Bar Ilan University in Ramat Gan, Israel.
“Because of these limitations we were very cautious regarding claims of causation, and we also explicitly state that our findings should not be interpreted as evidence for using vitamin D therapeutically for COVID patients,” the authors told us.
The retraction notice also mentions a delay in measurement of vitamin D and COVID-19 infection — anywhere from 14 to 730 days — which the authors said is “relatively short compared to other papers using a similar approach” but was mentioned in the study limitations.
The authors also pointed out what they considered errors in the retraction notice, including the statement “the analyses performed were inadequate to test the association between 25-hydroxyvitamin D3 levels at the time of infection and severity of COVID-19 illness.” The authors told us in their email the “association tested is NOT about levels at the time of infection (since the acute infective process affects vit D levels) but PRE-infection- this is even in the title.”
The researchers continued their “offer to conduct additional analyses to check for the extent of were rejected as coming too late” and the retraction came “without a warning.”
“We think our results have been a useful addition to the literature, especially in the context of the acute phase of a pandemic,” the authors wrote.
Meyerowitz-Katz told us the reasons cited in the retraction notice “make little sense” to him as well and “are very straightforward limitations to the methodology.”
“And yet, the journal has now retracted the paper for reasons that were immediately obvious on a quick perusal of the manuscript, three years too late,” he continued.
Meyerowitz-Katz has flagged several papers from the COVID-19 literature, including attempting to debunk the controversial drug ivermectin as a COVID-19 treatment.
By our count, nearly 600 COVID-19 studies have been retracted so far, bringing into question how robust the publishing process was in the midst of the pandemic.
Correcting COVID-19 literature has been largely slow and laborious. A study published in April in the Cambridge University Press, coauthored by Ferric Fang, who is on the board of The Center for Scientific Integrity, of our parent nonprofit organization, noted more than half of COVID-19 retractions “resulted from problems that could have been detected prior to publication.”
Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on X or Bluesky, like us on Facebook, follow us on LinkedIn, 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].