Retraction Watch

Tracking retractions as a window into the scientific process

Lawsuit couldn’t stop four retractions for diabetes researcher

with 2 comments

Mario Saad

Mario Saad

Four expressions of concern in the journal Diabetes have turned into retractions for Mario Saad, a move which he had tried to stop with a lawsuit.

Last August, a judge dismissed Saad’s suit against the American Diabetes Association, which publishes Diabetes, concluding that the expressions of concerns on the papers were not defamation, but part of an “ongoing scientific discourse.” Now, after an investigation at the University of Campinas in Brazil, where Saad is based, and an assessment from an ADA ethics panel (which overturned some of Unicamp’s recommendations), the journal has added to that discourse by turning the EOCs into retractions — and flagging two more of Saad’s papers with EOCs.

Together, the retracted papers have been cited more than 600 times.

As the retraction notes explain, after the journal issued the expressions of concern, it asked an investigative commission, appointed by the University of Campinas, to investigate whether the papers contained issues such as duplications and splicing.

The investigation concluded that some of the concerns were valid, but not all errors affected the integrity of the rest of the article in which they appeared. According to the retraction notices, the university recommended that one article be kept as-is, one be retracted, one be partially retracted, and one be corrected.

However, after the American Diabetes Association’s Panel on Ethical Scientific Programs (ESP) reviewed the university commission’s report and recommendation, it decided to retract all four papers.

Here are the papers that are now retracted, with some specifics from their notes to explain the issues with the papers and the discrepancy between the University’s decision and that of the journal.

According to the retraction note for “Physical Exercise Reduces Circulating Lipopolysaccharide and TLR4 Activation and Improves Insulin Signaling in Tissues of DIO Rats,” Unicamp concluded the paper included duplications, but recommended a partial retraction of just one figure.

The note explains why the journal decided to retract the entire article:

Despite the university’s assessment, the ESP still had concerns about the integrity of the quantification and analysis described in Fig. 3D. According to the original publication, the analysis was performed 6–8 times, but it is not possible to confirm this because no other analyses were offered. In addition, the ESP determined that a partial retraction would not serve as a clear and appropriate update to the publication status of this article, as the selective retraction of a key portion of the data would only raise more questions about the overall reliability of the research presented in the article.

The note explains that there were additional concerns about the paper, as well:

It should be noted that the ESP was informed by readers of additional concerns involving other data presented in this article. The investigative commission appointed by the University of Campinas did not identify or address these additional concerns in its investigative report.

That paper has been cited 51 times, according to Thomson Reuters Web of Science.

For “Loss-of-Function Mutation in Toll-Like Receptor 4 Prevents Diet-Induced Obesity and Insulin Resistance,” which has been cited 479 times — earning it a “highly cited” designation — the retraction note explains:

The university commission concluded that modifications (e.g., brightness and contrast adjustments, reordering of lanes, changes to the size of the bands) had been made to several images but that the presented data are “basically correct.” These images include Figs. 3A, 3D, 3E, 5B, 6E, 7A, and 8E. The university commission also concluded that bands 1, 3, and 7 in Fig. 3B are duplicates. According to the university commission report, the duplication in Fig. 3B is the result of an unintentional mistake. To address these issues, the university commission recommended publishing a corrigendum with replacement images to alert readers to manipulations made to Figs. 3A, 3B, 3D, 3E, 5B, 6E, 7A, and 8E.

And here’s why the journal wanted to retract, not correct, the paper:

Despite the university’s recommendation, the ESP does not agree with the publication of a corrigendum to report manipulations (whether deliberate or inadvertent) to several key portions of data in an individual study. The practice of doing so underscores the unreliability of the study as a whole without appropriately and responsibly updating the literature record. The ESP believes that this is especially relevant where issues of data duplication are concerned.

There were additional concerns with this paper too, the note explains — a duplication flagged by readers that Unicamp did not investigate.

For “Effect of Captopril, Losartan, and Bradykinin on Early Steps of Insulin Action,” cited 99 times, Unicamp reviewed the allegations and concluded that two bands were not duplicated, and the paper should be left as-is. The journal disagreed, as the retraction note explains:

Despite the university commission’s assessment, the American Diabetes Association’s Panel on Ethical Scientific Programs (ESP) remains concerned that bands 2 and 3 in Figs. 7B and 8C are duplicate images, particularly because the evidence submitted by the corresponding author does not appear to match the images presented in the article. The presence of error bars in Fig. 8, which is based on two experiments, magnifies the Panel’s concerns about the reliability of the quantification and analysis described in the article.

For the three articles above, the retraction notes conclude:

On the basis of these concerns, the ESP believes that the only responsible course of action for updating the status of [the paper] to issue a full retraction. The American Diabetes Association, the publisher of Diabetes, approved the Panel’s recommendation.

There was one paper that both Unicamp and the ADA agreed should be retracted: “Exercise improves insulin and leptin sensitivity in hypothalamus of Wistar rats,” cited 84 times, on which Saad is the second-to-last author. The retraction note explains:

According to the university commission, the corresponding author of this article admitted that the pY images in Figs. 2C and 3C are duplicates, that the PI3K images in Fig. 3A and C are duplicates, and that lanes 1 and 2 in Fig. 3B are the same as lanes 3 and 2, respectively, in Fig. 5A. The university commission, therefore, recommended that the article should be retracted.

Readers also raised concerns that part of a figure had been previously published and then republished, which:

raises further doubt about the reliability of the data presented in the study and substantiates the need, as the University of Campinas recommended, to retract this article. The American Diabetes Association, the publisher of Diabetes, approved the Panel’s conclusion.

In addition, Diabetes has added two expressions of concern to Saad’s papers:

The EOCs both note:

On the basis of the recommendation of the American Diabetes Association’s Panel on Ethical Scientific Programs (ESP), the American Diabetes Association, the publisher of Diabetes, is issuing this expression of concern to alert readers to questions about the reliability of the data in the above-cited article.

In both papers, there are concerns with “potentially duplicated images,” as well as images — or parts thereof — that may have been republished in other papers, sometimes representing different experiments and with adjustments to size, brightness, contrast, or orientation.

The expressions of concern outline the the steps the journal is taking to decide whether to retract the papers:

The Panel has contacted the corresponding author to inform him of these concerns, and the corresponding author’s institution, the University of Campinas (São Paulo, Brazil), has been asked to undertake an investigation of these issues. Diabetes will make a final decision on the publication status of this article after the journal obtains more information on the reliability of the data and conclusions presented in the article.

We’ve reached out to Saad and his lawyers, as well as to the ADA to see if they have any additional comments. We’ll update this post with anything else we learn.

Update, 5:15 PM EST: Saad sent us the following statement, noting that he hopes to republish the results in his papers in another journal:

I would like to start by explaining that I am a correspondent author- PI of three out of four retracted papers from Diabetes. This process is a huge injustice and the MOST UNFAIR that I have ever seen in Science. The accusations in these papers are about splicing and repetition of bands of western blots. I’ve proved them that at the time I sent the papers, Diabetes had no clear instructions for splicing. I also showed with original gels that most of the accusations of repeated bands were false, because the bands were similar in spite of coming from different gels. For example, in general, all beta actin coming from the same tissue look similar. I assume that in two of these three papers there was one mistake in each one, but they are editing mistakes (not fraud) and should be corrected by an erratum. In addition, the University here investigated these papers with two different commissions and showed that there is no fraud but editing mistakes that could be corrected by an erratum. There was a repeated band in Diabetes and Plos One, and this mistake was corrected by erratum in Plos One. This mistake does not exist anymore. The University did not suggest retraction for any of these three papers that I am responsible for. It is important to mention that in the manuscript of 1997 there is no mistake and the commissions confirmed that. Diabetes insists that there is a mistake, even though they can not prove it, and retracted also this paper. Diabetes disregarded the University’s first report, asked for a second investigation that was performed with foreign scientists , and the conclusion was that there was no fraud but minor mistakes that do not needed retraction.

I am convinced that it  is unfair and a persecution, because we can see splicings in most of the papers published before 2010 and, interestingly, also in a paper published in 2008 in Diabetes by the main editor of the journal (this paper was not retracted but corrected by an erratum). As you can see in Pubpeer, there are dozens of accusations of repeated bands in Diabetes’ papers from important labs from US, and the journal takes no action against them. It is probably easier to take a very rigorous action against a researcher from Brazil and from a Brazilian University. Retractions should be used to correct science which is not the case.

I would like to emphasize that all the science we presented in the three papers are correct and reproducible, but it seems that Diabetes does not care about this. It seems to me that they want to demonstrate that they are rigorous with bands and they need to find a scapegoat, and for sure it is easier and coward also, to find one from South America. For sure, I will publish again these results, showing that the science is correct.

Diabetes does the following process: every time they send us a decision, they also send another accusation, and before our response, they decide to publish an expression of concern. Of these new accusations, I have all the original gels because it is a paper from 2013, and I will prove them that the accusations are false, but I can guess that they will not care.

I would like to mention that I was the Dean of a Medical School in Brazil for 8 years, and candidate to be a president of the University. My political enemies are probably feeding diabetes with false accusations and they are taking it seriously and thinking that it is scientific.

Last year I went to court against Diabetes, in order to stop the expression of concern. However, at this moment, I cannot support a lawyer, and the only decision I’ve made is to publish again these results in another journal.

Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our new daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.

  • fernandopessoa March 24, 2016 at 9:29 pm

    With the new expressions of concern, only one Diabetes paper discussed at Pubpeer from these authors still remain untouched by ADA:

    Yet ADA/Diabetes should be congratulated overall.

  • Dave Fernig March 29, 2016 at 11:17 am

    The PI’s comments about the lack of guidelines regarding splicing are quite remarkable. There should be no need for such guidelines. It is not allowed and never has been – butting gel bands and moving them with a scalpel in pre-western blot days was wrong. Doing this electronically is equally wrong, albeit easier.

    ADA/Diabetes are indeed to be congratulated, particularly on two counts.
    One for resisting the lawsuit.
    Two, because they have not taken the easy way out and accepted a replacement figure (aka a correction). We never know where these originate and in some other instances documented here on RW, corrections have merited further corrections, which strongly suggests an absence of data.

  • Post a comment

    Threaded commenting powered by interconnect/it code.