How a sleuth’s email turned a correction into a retraction

On Sept. 2, 2021, a professor at the University of Iowa, in Iowa City, emailed a biochemistry journal asking to correct a paper she had published the previous year. An experiment had “unintentionally” been omitted from a figure, Isabella Grumbach explained, and a comparison of experimental groups contained “a minor error in the degree of statistical significance.” A correction ensued. 

But the problems with the article, “Inhibition of CaMKII in mitochondria preserves endothelial barrier function after irradiation,” appear to have been more deep-rooted than the email suggested. An anonymous commenter on PubPeer had first raised concerns about the article, which had appeared in Free Radical Biology and Medicine (FRBM), in July 2021, more than a year after it was published. The commenter claimed error bars between two figures were vastly different, even though they were meant to be related data points. 

Grumbach, the corresponding author of the paper, promised to investigate. A month later, she wrote to Rafael Radi, an editor at the journal, saying she had omitted an experiment from the paper by mistake, according to documents Retraction Watch obtained through a public records request. But, she noted, “the data interpretation or overall conclusion from these data are not impacted by this error.” 

Grumbach, who is interim chair of internal medicine at the Carver College of Medicine at the University of Iowa, said the repeated analysis had strengthened the paper’s conclusions and requested a corrigendum. The correction was published in November 2021. 

But that didn’t make the problem go away. Within a month of the correction, another comment appeared on PubPeer claiming the revision didn’t explain the errors the first comment had flagged. Again, the commenter requested the authors share the raw data. This time the authors did not respond to the post. 

Then, in January 2022, a sleuth contacted the editor-in-chief of FRBM, Kelvin J. A. Davies, claiming: 

All figure in this paper have data and statistic method that is manipulated or falsified. Paper say “independent experiment” but honestly data not from independent experiment. Actually data show replicate reading of same sample from same experiment. This false statistic analysis make very small error bar so statistic significance is reached. If data points from independent experiment analyze then there is big variation and no statistic significance so major conclusions of paper not valid.

Check original data and see how statistic falsified to reach desired conclusion.

The email was appended with the comments from PubPeer. 

A month later, the chair of the journal’s ethics committee, Giovanni E. Mann, wrote to Grumbach saying the committee had reviewed the case following the sleuth’s email. Mann noted the authors did not respond directly to the commenter’s request for raw data. He also flagged statistical concerns and added, “We do need to see the individual data points on which these figures and statistics were based.” 

Grumbach said she would perform a comprehensive review and prepare a reply to the queries. In her email, she also wrote, “we have seen a very unusual increase in activities regarding the manuscripts from a select group of investigators in the Dept. of Internal Medicine.” 

Grumbach did not respond to our queries about the “unusual increase in activities.” She is listed as an author on two other papers that were called out on PubPeer around the same time. One was issued an erratum and another an expression of concern, both related to the inclusion of duplicate data. 

A spokesperson for Elsevier, the publisher of FRBM, said the editors were not aware of the PubPeer comments before the sleuth’s email, and made the decision to publish a correction “based on mistakes raised by the author.” 

In April 2022, Grumbach officially requested to retract the paper, which has been cited 14 times, according to Clarivate’s Web of Science. She told Davies: “I analyzed these data and inaccurately pasted the data into the analysis software, leading to an incorrect grouping of the data. I apologize for the inconvenience. Please advise me on the next steps.” 

The retraction notice, published online in May 2022, stated: 

This article has been retracted at the request of the Authors and Editor-in-Chief.

Some of the data presented in Figure 6C, F and G of the above-titled paper were reported incorrectly in the published article. After being contacted by the Journal, the authors discovered an unintentional error in how the original data were analyzed that could affect the accuracy of the subsequent analysis. The raw data were incorrectly grouped in the analysis software, thereby altering the comparisons. Therefore the authors wish to retract the paper and will recollect and reanalyze the data appropriately. The authors apologize for any inconvenience.

The notice added that the authors would “recollect and reanalyze the data appropriately.” 

Grumbach declined to be interviewed for this story but a spokesperson passed on a statement from her: 

I take research integrity very seriously, which is why I asked for a retraction of this paper when I realized there were errors in the data and analysis after I had submitted an erratum.

Since its retraction, the paper has also been included among the references in another paper by the group, although that “manuscript was drafted and submitted for initial review before the retraction,” according to the university spokesperson.

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].