Annals of Internal Medicine has retracted a paper on treatments for Long COVID because the authors did not report dosing information for metformin, one of the drugs tested in the study. The editor-in-chief of the journal called the exclusion “sloppy” and “inaccurate.”
The same day the journal retracted the paper, it published a new version addressing the exclusion. But one researcher told Retraction Watch the replacement still leaves unanswered questions.
Published in March, the clinical trial found the antidepressant fluvoxamine – marketed as Luvox – reduced fatigue and metformin showed no significant benefit for patients with long-term fatigue and other chronic symptoms following COVID illness. Other studies have found fluvoxamine helped reduce depression in patients with Long COVID, and researchers have found starting metformin within three days of being diagnosed with COVID-19 can help prevent long-term effects of the disease.
During the trial, doctors reduced the twice-a-day metformin dose from 750 mg to 500 mg because patients were experiencing side effects including gastrointestinal problems, and 9% of patients in that group had dropped out. In their analysis, the authors pooled the two groups because they considered the doses to be “equivalent,” senior author Edward Mills told Retraction Watch.
“Someone didn’t like that and wrote to the editor demanding we keep them separate, so we did the new analysis that has them as two different groups,” Mills said. “It makes no difference to the results or interpretation.”
Carolyn Bramante, a metabolic health and COVID-19 researcher at the University of Minnesota, told us both the retracted and replacement papers had different inclusion and exclusion criteria for the fluvoxamine, placebo and metformin arms. “Having the same inclusion/exclusion criteria is a fundamental, basic requirement of randomization,” she said.
Bramante has studied the use of metformin in people with acute COVID-19 and has advocated for patients with the condition to take the drug to protect against Long COVID.
Bramante said she also was concerned with the high number of adverse events reported in the paper. She pointed to another article published in March in Annals of Internal Medicine led by So Yun Lim, a researcher at the University of Ulsan College of Medicine in Seoul, which found few adverse events when testing the efficacy of 14 days of metformin in Long COVID patients. Conversely, nearly half of patients in Mills’ study who took metformin reported adverse events.
Lim’s study had “values you would expect” for adverse events, she said. “You typically don’t expect trials to have such different findings.”
Christine Laine, the journal’s editor-in-chief, told us Annals of Internal Medicine decided to retract the paper “because the authors did not report the changes in metformin dosing that occurred during the conduct of the trial.”
In the new version of the paper, the authors included the reduction in the dose of metformin and an explanation of why that arm of the study was stopped for futility, both of which were recommended by the trial’s data and safety monitoring committee.
The retracted article concluded metformin showed no benefit to patients with Long COVID, but the new version states the metformin findings were inconclusive.
Several of the same authors, including Mills, a part-time professor of health research methods at McMaster University in Hamilton, Ontario, lost a similar paper on acute COVID-19 from The Lancet for concerns with the metformin arm of an earlier version of the REVIVE-TOGETHER trial. Both the retracted paper and replacement in Annals cite the retracted Lancet paper.
“This isn’t really a retraction but rather a republication as I think formatting the manuscript would have been difficult or impossible,” Mills told us in an email.
Laine told us, “The editors do not know whether the reason for the inaccurate reporting was intentional,” she said. “We do know that it was sloppy, inaccurate, and the scientific record needed to be corrected.”
When we asked why Annals chose to republish the paper, Laine said: “Not making these results available in a republished article would have been disrespectful to the 399 patients who altruistically participated in the study.” She also said the article underwent further peer review prior to publication in October.
Corresponding author Jamie Forrest, a health informatics researcher at Northeastern University’s Vancouver campus who was also on the Lancet article, told us Mills’ comments “represent those of the authorship.”
Before its retraction, the paper was cited four times, according to Clarivate’s Web of Science. The paper joins our list of almost 800 retracted COVID-19 articles, which includes a 2020 retraction from the journal for a paper claiming masks were ineffective in preventing the spread of COVID-19.
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].