Tired of waiting for a university, a publisher commissions its own investigation — and retracts two papers

The journal Diabetes has retracted two 2006 papers by a group of researchers in Germany whose work has long been the subject of concerns about image duplication and manipulation. 

The first author of the articles is Kathrin Maedler, a prominent diabetes specialist at the University of Bremen, where she’d been a named professor but lost the title over the affair. Maedler’s group now has four retractions resulting from problematic figures. 

The University of Bremen in 2016 found insufficient evidence that Maedler committed research misconduct, but concluded that she was negligent. Maedler at the time told us: 

During the investigations, we did find mistakes and identified their origin, which could have been avoided with a better strict and fully organized laboratory data documentation, saving and presentation strategy. In some cases, I requested repetitions of the experiments-in-question from lab members, who originally were not co-author of the published papers and were not involved in the original experiments- for me to independently show that all data are correct.  All repetitions fully confirmed the published results.

Although the retractions are a long-time coming — the first correction in Diabetes appeared in 2014, and the first retraction, in the Journal of Biological Chemistry, a year later — the ADA has handled the case about as well as one might hope. So well, in fact, that in 2017 we singled them out for our inaugural Doing the Right Thing award in STAT. 

Here’s the retraction notice for “Aging correlates with decreased beta-cell proliferative capacity and enhanced sensitivity to apoptosis: a potential role for Fas and pancreatic duodenal homeobox-1”:

The above-cited article has been retracted by the American Diabetes Association (ADA), the publisher of Diabetes. This article was previously the subject of an expression of concern in the November 2018 issue of the journal (Diabetes 2018;67:2478–2479. DOI:10.2337/db18-ec2018a. PMID: 30213821). 

In September 2018, ADA’s Panel on Ethical Scientific Programs (ESP) contacted the University of Zurich [where Maedler did a post-doc], the lead and corresponding authors’ institution at the time the research was conducted, to request an institutional investigation of concerns related to image manipulation and duplication in the article cited above. The University’s review is ongoing. 

The ESP has since coordinated an independent review of the anomalies cited in the expression of concern. On the basis of its review, the ESP has determined that the following instances of potential image duplication compromise the overall reliability of the study. 

In the “Actin, 30 h” panel of Fig. 3B, lanes 4 and 6 appear to be duplicates.

In Fig. 4A, lanes 3 and 4 of “PDX-1, 96 h” panel appear to be duplicates.

In Fig. 4A, lane 2 of the “Actin, 96 h” panel and lane 5 of the “PDX-1, 96 h” panel appear to be duplicates.

Lanes 3–6 of the “Actin, 96h” strip of Fig. 3D of this article and lanes 3–6 of the “Actin 42 kDA” strip in Fig. 3B of Diabetes 55:2713–2722, with horizontal rotation and contrast and size adjustments, appear to be duplicates. 

The second retraction involves “Low concentration of interleukin-1beta induces FLICE-inhibitory protein-mediated beta-cell proliferation in human pancreatic islets”. After an identical preamble, the notice states: 

On the basis of its review, the ESP has determined that the following instances of image duplication compromise the overall reliability of the study. 

In a 2009 PLOS ONE article (PLoS ONE 2009;4:e4394. DOI: 10.1371/journal.pone.0004394), lanes 1 and 2 of the “b-Actin” panel in Fig. 1A appear to be the same as lanes 2 and 3 of the “Actin 42 kDa” panel in Fig. 3G of this article. It should be noted that a correction to replace Fig. 1A in the PLOS ONE article was published in May 2015 (PLoS ONE 2015;10:e0117766. DOI: 10.1371/journal.pone.0117766).

Lane 4 of the PDX1 panel in Fig. 1E of a 2011 article in the Journal of Biological Chemistry (J Biol Chem 2011;286:17144–17155. DOI: 10.1074/jbc.M110.210526) closely resembles lane 1 of the FLIP 55KDa panel of Fig. 3D of this 2006 Diabetes article. In addition, lane 8 of the PDX1 panel of Fig. 1E of the Journal of Biological Chemistry article closely resembles lane 5 of the FLIP 55KDa panel in Fig. 3D of this Diabetes article. The Journal of Biological Chemistry article was retracted in November 2015 (DOI: 10.1074/jbc.A110.210526). 

Lanes 2, 4, and 6 of the “Actin 42 kDa” panel of Fig. 3D appear to be duplicates. 

Lanes 3–6 of the “Actin 42 kDA” strip of Fig. 3B of this article and lanes 3–6 of the “Actin, 96 h” strip in Fig. 3D of Diabetes 55:2455–2462, with horizontal rotation and contrast and size adjustments, appear to be duplicates. 

Maedler told us that she and her colleagues:  

We fully disagree with the two retractions and requested immediate retraction of these online publications and a discussion with the ADA on that matter.

Maedler denied that the paper on islet cells contained manipulated images: 

The western blot bands are not identical. They don’t even look similar.

Maedler also objected to the fact that ADA had prepared two versions of the retraction notices, but that she and her colleagues did not receive the final until it was published. 

‘University’s review is ongoing’

We asked the ADA about the gap between its last action — the expressions of concern in late 2018 — and the retractions. Christian Kohler, the group’s associate publisher for scholarly journals, told us: 

The length of time between the publication of the expressions of concern and the retraction notices does indeed have to do with the university’s investigation. ADA had hoped to cite the results of the university’s investigation in updated amendments to these papers, but the university’s review is ongoing. ADA determined that a reasonable amount of time had passed to complete the investigation and that the issues mentioned in the expressions of concern needed to be finally addressed. Therefore ADA’s Panel on Ethical Scientific Programs coordinated an independent analysis of image anomalies presented in the papers. The Panel determined that the evidence presented in the analysis is consistent with image duplication and manipulation in a manner that undermines confidence in the papers’ data and conclusions. Based on the Panel’s assessment and recommendations, ADA has chosen to retract the publications.

Kohler also acknowledged the two notices, but dismissed Maedler’s complaint. 

As Dr. Maedler pointed out, the former version does not mention the later republications in PLOS ONE and JBC. The ADA-ESP decided to include these later instances of image duplication in the final versions of the statements because they were previously cited in the expression of concern and the Panel believes that these instances contribute to the overall lack of confidence in the data and conclusions presented in the article.

We reached out to Rita Stöckli, the Secretary General at the University of Zurich, where the studies were conducted, to find out the status of the investigation and will update this post if we hear back.

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 [email protected].