When failure to correct a flawed paper could put patients’ lives at risk

Robert Speth

On April 15, 2021, as COVID-19 was waning several months prior to the surge in deaths associated with arrival of the Delta variant, the journal Cell published an eye-catching paper. 

Titled “Soluble ACE2-mediated cell entry of SARS-CoV-2 via interaction with proteins related to the renin-angiotensin system,” the article stood in stark contrast to the contemporary understanding of the mechanism of SARS-CoV-2 infection, which until then held that ACE2 on the membranes of susceptible cells served as the “receptor” for the virus.

The paper was notable because it claimed that vasopressin, also known as antidiuretic hormone, worsened COVID-19 infections. Vasopressin is known for its ability to promote water retention in the kidneys as well as to constrict blood vessels, but had not previously been associated with COVID-19 infections. 

Upon reading the paper, one of us (MB) noted a large number of inaccuracies. The authors had used the wrong reagent: a high molecular weight precursor of vasopressin rather than vasopressin itself. They also incorrectly portrayed ACE2, the V1B vasopressin receptor, and the AT1 angiotensin II receptor – the primary mediators of their hypothetical mechanism of COVID-19 infection. (PubPeer commenters also pointed out problems in the paper, including a failure of the authors to post their original data.)

Vasopressin offers the advantage of minimal pulmonary artery constriction and respiratory stimulation compared to other vasopressors used to maintain blood pressure in patients with COVID-19 who are on ventilators. But it would be a bad idea in COVID-19 patients if it increased the infectivity of SARS CoV-2. Fortunately, a study showed that wasn’t the case.

Michael Bader

So, if doctors based their treatment decisions on the Cell paper, they’d be depriving patients of an important regimen. MB sent a list of the inaccuracies in the Cell paper to the editor of the journal. Six weeks later, having received no response, he sent a copy of his correspondence to a group of approximately 30 experts in the field of renin-angiotensin system research,  including one of us (RS). Unsurprisingly, his concerns received near unanimous affirmation from the group.

Seven weeks after MB attempted to contact Cell, the editor responded. He suggested that MB contact the authors of the paper to apprise them of their errors. Subsequently the editor forwarded the list of challenges to the authors of the paper. The authors responded to the letter with a total rebuttal, denying any mistakes. 

Accepting the rebuttal at face value, the editor wrote that the only challenge he would accept to the paper was a replication of the study that did not support the findings of the original article. Taking up the challenge, two researchers MB had contacted previously about the article agreed to work with him to repeat some of the experiments. Not unexpectedly, their results, which were published in Cell, did not support the original findings. 

However, following that paper’s publication, the authors of the original paper published yet another article in Cell defending their study. That  despite two other studies that have failed to replicate the Cell study’s claims. 

This past May, MB once again requested that the Cell editors attach a correction to the original paper, but they stopped responding. We have now published a letter to the editor in the Medical Research Archives describing just a few of the major concerns we have about the Cell paper, emphasizing its potential to adversely impact the treatment of COVID-19 patients who need ventilatory support.

Some might say that all of this back and forth is how science should work, rather than a correction or retraction. But we would suggest it is better to follow the “If you see something, say something” dictum. Few do, because of fear of retaliation and putting their careers in jeopardy. Failing to correct blatant errors in the literature is, in our opinion, editorial misconduct. This is not a debatable matter. Science cannot maintain society’s trust if it doesn’t police itself.

Given the refusal of the editor of Cell to even implement a correction of the manuscript to notify readers of the erroneous characterization of vasopressin, let alone the other errors, we can only hope that by publicizing these errors in as many venues as possible, we can minimize the harm that this paper could potentially cause to ventilated COVID-19 patients.

Robert Speth is affiliated with the Department of Pharmaceutical Sciences, College of Pharmacy at Nova Southeastern University, in Fort Lauderdale, Fla., and the Department of Pharmacology and Physiology at the School of Medicine, Georgetown University, in  Washington, D.C. Michael Bader is with the Max-Delbrück-Center for Molecular Medicine in Berlin, Germany.

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 team@retractionwatch.com.

11 thoughts on “When failure to correct a flawed paper could put patients’ lives at risk”

  1. The doi for the Letter to the Editor is incorrect. It should be “https://doi.org/10.18103/mra.v10i9.3079”

      1. MRA was originally published by “Knowledge Enterprises Inc”, and was covered by various scholars of predatory publishing. For convenience I will link to Kaye’s blog rather than to Beall.


        Then the dude in Minnesota who’s behind MRA reinvented himself as European, and changed his company to “European Society of Medicine”, to focus on running conferences.


        Please forgive the following link-farm:

  2. I see (at a cursory glance) that papers marked “pdf final” at MRA have an accompanying doi that functions correctly, while others marked only “pdf” have no doi shown on the web page, but there is one included in the attached pdf. The latter produces an error at this point in the process.

    So I suppose that will settle down shortly.

  3. The URL in the manuscript should have a comma between doi and org rather than a period. If you simply type in doi:10.18103/mra.v10i9.3079 it will take you to a place where you can download the full text of the manuscript

  4. The new link will now take you to the abstract of the paper which has a tab diathermy the publication date that opens the full text of our article

  5. >They also incorrectly portrayed ACE2, the V1B vasopressin receptor, and the AT1 angiotensin II receptor – the primary mediators of their hypothetical mechanism of COVID-19 infection.

    What does this even mean?

  6. To answer your question, it is mind-boggling that they did not even take the trouble to determine that the AT1R angiotensin II receptor and the V1B vasopressin receptor are G protein-coupled receptors. They were also equally unconcerned with accurately representing ACE2, which has only one transmembrane-spanning domain. To me it indicates a degree of carelessness that calls into question the validity of this entire work.

    1. The first citation on the paper is about GPCRs. What is there for them to determine? The world is already aware that V1B and AT1R are GPCRs. Whose lives are at risk here over scientific publication semantics? Are more people going to get Covid because a paper in Cell suggests we don’t fully understand the receptor kinetics involved in entry of Covid into cells?

      You can pick apart 9 out of 10 scientific pubs and find small errors, typos, etc unless it’s coming out of a so called “mega-lab”

      Maybe we’re reading different papers altogether. The one I read states they transfected cells with a plasmid containing the sequence for NM_000490 which is recombinant human vasopressin. If you define that as a vasopressin precursor, well yes the gene that encodes the protein is technically a precursor I guess. What’s next, should we measure the kerning on their figure captions and see if it violate’s Cell’s publication guidelines?

      I enjoy the content on Retraction Watch but lately it seems to be “Covid papers with small mistakes that we think should be retracted but weren’t and are therefore dangerous”. I would argue that the vast body of manipulated/fabricated cancer research is far more dangerous to the world than a flawed in vitro study attempting to elucidate the mechanism of cellular entry for coronavirus

      Also I don’t know why you bothered to share the clinical study suggesting there’s not a link between vasopressin and Covid infection in a small group of patients with hypoxaemic respiratory failure that are in critical care. The paper you’re dissecting is in vitro work.

  7. This is not an issue of semantics. SARS-CoV-2 infects cells by binding to membrane bound ACE2, whereupon it becomes internalized. Solubilized ACE2 is a decoy receptor for the virus that competes for the membrane-bound ACE2, reducing the infectivity of the virus. We only pointed out the most obvious reasons why we conclude that the paper is wrong out of a large number of questionable observations.

    1. Is that opinion or fact though? There are hundreds of biological processes, especially within cell biology, that we still don’t have a full understanding of. Just because there’s strong evidence suggesting that ACE2 directly impacts COVID’s ability to infect cells, this doesn’t mean that it’s the only mechanism responsible.

      I agree the paper is not without flaws but in my personal opinion there is more harm in retracting it than benefit. Even if the results are not replicable, it is likely that a research group will at least attempt to replicate it and it’ll lead them to other unexpected discoveries.

      The published content that deserve this kind of scrutiny is the drivel put out by Peter Mcculough and the likes, who are focused on a political agenda and not science

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.