Retraction Watch

Tracking retractions as a window into the scientific process

NIH/Harvard team loses aging study to manipulated data

with 14 comments

agecoverAge has retracted a 2012 article by a group of scientists from the National Institutes of Health and Brigham and Women’s Hospital in Boston after an NIH inquiry turned up evidence of data manipulation in the work.

The article, “Aging decreases rate of docosahexaenoic acid synthesis-secretion from circulating unesterified α-linolenic acid by rat liver,” came from the lab of Stanley Rapoport, chief of the brain physiology and metabolism section of the National Institute on Aging.

As the abstract explained:

Docosahexaenoic acid (DHA, 22:6n-3), an n-3 polyunsaturated fatty acid (PUFA) found at high concentrations in brain and retina and critical to their function, can be obtained from fish products or be synthesized from circulating α-linolenic acid (α-LNA, 18:3n-3) mainly in the liver. With aging, liver synthetic enzymes are reported reduced or unchanged in the rat. To test whether liver synthesis-secretion of DHA from α-LNA changes with age, we measured whole-body DHA conversion coefficients and rates in unanesthetized adult male Fischer-344 rats aged 10, 20, or 30 months, fed an eicosapentaenoic acid (EPA, 20:5n-3)- and DHA-containing diet. Unesterified [U- (13) C]α-LNA bound to albumin was infused intravenously for 2 h, while [(13) C]-esterified n-3 PUFAs were measured in arterial plasma, as were unlabeled unesterified and esterified PUFA concentrations. Plasma unesterified n-3 PUFA concentrations declined with age, but esterified n-3 PUFA concentrations did not change significantly. Calculated conversion coefficients were not changed significantly with age, whereas synthesis-secretion rates (product of conversion coefficient and unesterified plasma α-LNA concentration) of esterified DHA and n-3 DPA were reduced. Turnovers of esterified n-3 PUFAs in plasma decreased with age, whereas half-lives increased. The results suggest that hepatic capacity to synthesize DHA and other n-3 PUFAs from circulating α-LNA is maintained with age in the rat, but that reduced plasma α-LNA availability reduces net synthesis-secretion. As unesterified plasma DHA is the form that is incorporated preferentially into brain phospholipid, its reduced synthesis may be deleterious to brain function in aged rats.

But here’s the notice:

This article has been retracted by the authors as they were unable to reproduce some of the data and therefore consider them unreliable.

To the editor of Age:

The NIH found that Dr. Fei Gao engaged in research misconduct by fabricating and/or falsifying data in Figures 1-7 and Table 2 in “Aging decreases rate of docosahexaenoic acid synthesissecretion from circulating unesterified α-linolenic acid by rat liver. Gao F, Taha AY, Ma K, Chang L, Kiesewetter D, Rapoport SI. Age (Dordr). 2012 Mar 3,” and therefore I request a full retraction of this paper. Please note, none of the other authors were implicated in any way.

Stanley I. Rapoport

The paper has only been cited twice, according to Thomson Scientific’s Web of Knowledge.

We’re guessing that because NIH pays the salaries of most of the co-authors on the study, including Rapoport, the Office of Research Integrity has been, or will be, involved in this case. We found a dozen articles on which Gao and Rapoport share a byline, although the Age article is the only retraction so far. Gao’s Harvard email bounced. We’ve sought comment from Rapoport and will update this post when we learn more.

This is the second retraction in a week involving Harvard’s Brigham and Women’s Hospital.

Hat tip: Rolf Degen

Like Retraction Watch? Consider supporting our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, and sign up on our homepage for an email every time there’s a new post.

Comments
  • Alan Price April 15, 2014 at 12:21 pm

    You are correct, Adam, the ORI would have been notified at least at the start of an NIH intramural investigation, and ORI would review the NIH report and decide on pursuing HHS administrative actions against the person who falsified the results, such as supervision and certification of future NIH-supported research or debarment from receiving federal grants, contracts or fellowships, for a period of years.

    • david hardman April 15, 2014 at 1:05 pm

      Another problematic Harvard n-3 polyunsaturated fatty acid paper.

      http://retractionwatch.com/2012/11/20/image-correction-in-current-biology-for-harvards-sam-lee/#comment-46910

      https://pubpeer.com/publications/578F85A1DDAAA96599C26C88EA7BE1

      Cancer Res. 2001 Feb 15;61(4):1386-91. PMID:11245439
      Polyunsaturated Fatty Acids: Mediation through Cyclooxygenase-independent Pathways 1

      Mary D. Boudreau, Kyung Hee Sohn, Sang Hoon Rhee, Sam W. Lee, Jay D. Hunt, and Daniel H. Hwang.

      – Author Affiliations

      Pennington Biomedical Research Center, Louisiana State University, Baton Rouge, Louisiana 70808 [M. D. B., K. H. S., S. H. R., D. H. H.]; Stanley S. Scott Cancer Center, Louisiana State University Health Sciences Center, New Orleans, Louisiana 70112 [J. D. H.]; and Department of Medicine, Beth Israel Hospital, Harvard Medical School and Harvard Institute of Medicine, Boston, Massachusetts 02215 [S. W. L.]

  • Dan Zabetakis April 15, 2014 at 1:47 pm

    It’s not clear what the problem with the data is. All the figures are graphs or tables. No westerns unfortunately.

    I guess that the numbers were made up, or else altered from the primary data.

    • littlegreyrabbit April 15, 2014 at 9:34 pm

      One guesses that it being unlikely anyone repeated this experiment then it fell to an internal whistleblower. It is shame we never get more detailed reports about such matters or how responsibility was determined. Westerns are an easy thing to check, but they are only ever revealing the incompetent. If I wanted to cheat on a western for publication it would be blinding easy to make it undetectable at the level of the journal paper (even at the level of the lab book it wouldn’t be too hard to make it difficult to determine cheating beyond reasonable doubt).

      This one
      http://www.ncbi.nlm.nih.gov/pubmed/23336521
      from the same lab with Rapoport as an author but not Gao, reports a Western in panel D of this figure showing massively upregulated Bax protein levels
      http://www.ncbi.nlm.nih.gov/core/lw/2.0/html/tileshop_pmc/tileshop_pmc_inline.html?title=Click%20on%20image%20to%20zoom&p=PMC3&id=3606672_nihms437525f4.jpg
      If you magnify it using the plus symbols on the right hand side you will see that the two lanes of this tiny western are of different height dimensions, also the lanes seem to have differing background intensities.

      Because of the poor resolution of the western, it is difficult to be absolutely confident of a splice line, but the irregular dimensions would lead a reasonable person to assume that the chances of this being cobbled together with photoshop is not insignificant.

    • littlegreyrabbit April 16, 2014 at 12:13 am

      This paper also with Rapoport and Rao but not Gao
      http://www.ncbi.nlm.nih.gov/pubmed/22621398
      Westerns on this figure
      http://www.ncbi.nlm.nih.gov/core/lw/2.0/html/tileshop_pmc/tileshop_pmc_inline.html?title=Click%20on%20image%20to%20zoom&p=PMC3&id=3464147_1471-2202-13-50-1.jpg
      I would place a question mark on Panel A CD11b
      Panel C and E clearly have the same actin bands. Given they are quite far apart in size that might be theoretically possible, but the actin bands show a definite slope and this is not visible in the 17 kDa IL-1B bands of Panel E – so I don’t think it is reasonable to allow that these came from the same blot

      Panel B and F show actin bands of 45 kD and proteins of 49 and 50 kD respectively. These are pretty hard to separate – particularly when either the bands get very fat (eg Panel B high LPS) or when the actin bands show a lop-sided smile (Panel F) – in this case it should a) merge with the band above and b) the band above it should also have a one sided smile – when it is flat as a ruler.
      I don’t see how these actin bands can have come from the same blot.

  • Albert Donnay April 15, 2014 at 2:12 pm

    I’m disappointed that the editor allowed authors to say “Please note, none of the other authors were implicated in any way.”

    The other authors clearly want to absolve themselves, but aren’t co-authors accepting responsibility for the entire paper when they put their name on it? If they weren’t willing to accept full responsibility, they should have not agreed to be a co-author in the first place and should have requested only an acknowledgement.

    They are at least implicated in being duped, as were the original editor, reviewers, and presumably also the study’s funders, who could have required some basic fraud and plagiarism checks before publication.

    Disappointed also that there is no mention of whether the funders will be repaid by anyone involved?
    Are reporters even asking this question?

    Funders could help by requiring that they be repaid by the institutions they fund–including overhead– for any studies they fund that are subsequently retracted for any type of scientific misconduct.

    And if funders offered to share some percentage of these recovered funds with those who first reported the suspected misconduct–as Qui Tam law already provides to whistleblowers who report fraud involving federal govt funding–there’d be a lot more whistleblowers.

    Whether this might reduce future research fraud is an open question, but it would at least stop institutions from profiting by such fraud.

    • DocMartytn April 15, 2014 at 5:33 pm

      If he was responsible for all the extractions and chemical analysis and the others did the animal work, then they are completely in the clear. His publication record suggest a chemistry bent.
      The really strange this is the why did he do it.

      • JAdS April 15, 2014 at 11:56 pm

        Actually, I agree fully with Albert’s assessment and comments. In any other day, I would almost say that those comments were made by me because I share the exact same concerns. The issue of responsibility will depend on the authorship definition and responsibilities adopted by Springer. I believe that Springer adopts the ICMJE’s definitions, which would make Albert’s comments pertinent, and thus assign responsibility to ALL co-authors, even if they did not have their fingers in the pie.

    • Alan Price April 16, 2014 at 2:45 pm

      Albert, when I was an ORI official 1990-2006 and since then as a private consultant, I have always encouraged institutional officials and editors to include such statements in retractions — label them as “research misconduct” by the person(s) who committed the falsification, fabrication, or plagiarism, and clear the names of the coauthors who were unaware or innocent victims of that action (“duped” was your word). There is a serious “taint” to one’s reputation to having one’s name on a retracted article; as I observed in ORI, and Fang et al. published in PNAS 109: 17028 (2012), at least two-thirds of retracted articles are associated with misconduct cases.

      Students have asked me how they should consider such a misconduct-retraction on their CVs. Can they exclude it, since it was retracted? If so, might a new employer do a PubMed search and question why that paper was excluded, see the link to the retraction and the ORI misconduct finding, and thus question their honesty? If they still include it in their CV, should they explain who committed the misconduct, and they they were unaware then? Of should they just say in their CV that the data was not reproducible? It seems like a “Catch 22″ no matter how they handle it, as a “taint” on their reputation, no matter how innocent they were.

      • Albert Donnay May 20, 2014 at 10:04 am

        Alan,

        Thanks for your inside ORI perspective.
        So what did you recommend to the students with this great question?

        I’ve not been a professor myself but was raised by two and live with one:

        I think they’d all agree that transparency is always the best policy, esp. in the internet age.

        Co-authors should keep retracted papers on their resume and add a note to any that have been retracted, saying when and why and even providing a link to the published retraction statement or providing some explanation of their own role.

        Not all retractions are due to misconduct, after all, and my parents considered it a badge of honor for authors to voluntarily retract a paper that they or others later discovered to be fatally flawed in some respect and without possibility of correction.

        Anyone checking a job applicant’s resume is going to check their publication list and find most if not all of their retracted papers still listed on PubMed or other databases but flagged as such, so it would be better for job applicants to disclose this themselves.

        Can you tell us what ORI’s position was in the 1990s on the question of co-author responsibility for the contents of their papers? [distinct from their responsibility for misconduct, which I realize ORI always sanctioned author-by-author]

        Did ORI ever have a blanket policy specifying that all co-authors either do or don’t share full responsibility?

        • Alan Price May 20, 2014 at 8:34 pm

          Albert, I share your view that transparency and honesty is the best policy — nonetheless, it is sad that the taint of retraction has an impact on the honest coauthors who did not commit research misconduct, but have to figure our how to explain it on their CV when the are such a coauthor.

          ORI has authority only over research misconduct issues. ORI has no authority over “whether all coauthors share responsibility for a publication” – even one that contains material that was fabricated, falsified, or plagiarized (by others). Thus, ORI has no “policy” on such coathor-responsibility, and ORI cannot have it,

          • Albert Donnay May 21, 2014 at 1:41 pm

            Thanks for that clarification. I agree it is very sad, but at least also an unforgettable lesson. We all need to think very carefully about what papers you choose to co-author and with whom!

            So did any of the investigations you worked on result in a recommendation from ORI that a paper be retracted?

            And if so, to whom did ORI usually direct this request?
            The editor? or the authors? both?

          • Alan Price May 31, 2014 at 12:56 am

            Sure, Albert. We made many findings and settlements with respondents found to have committed research misconduct that required them to submit retractions, with our review and approval — we also notified their institutional research integrity officer of our findings and requirement for retraction — we also notified the editor to expect such a retraction request, noting that we would provide more details on request if the respondent failed to submit the retraction, so that the editor could take action directly and editorially retract the paper.

            The first such case that I handled in ORI involved such a disputed paper by James Abbs at Wisconsin —
            http://www.gpo.gov/fdsys/pkg/FR-1996-04-09/html/96-8685.htm — when he did not agree with the standard retraction language used by medical editors, that editor (Robert Daroff) retracted the paper himself, citing the ORI findings and decision on the need for the retraction [ http://www.ncbi.nlm.nih.gov/pubmed/8927237 ]

  • Shi V. Liu April 18, 2014 at 8:13 am

    Over the years I have seen many retractions being made as results of discovering data manipulation-type misconduct. However, there are many other papers which I think contain solid data but totally incorrect interpretations and thus erroneous conclusions. I say these papers should be retracted because they distorted the truth but they are still immune from even solid criticisms because they were published in some top journals and, more significantly, fitting the “convention wisdom”.
    For example, a research group in Harvard University published a paper in Nature in Jan. 15th, 2012 “Multi-isotope imaging mass spectrometry quantifies stem cell division and metabolism
    ” which claims a strong evidence for random DNA segregation (Nature 481: 516, 2012; http://www.nature.com/nature/journal/v481/n7382/full/nature10734.html ). However, I conclude that the data presented in this paper actually supports regular DNA strand segregation that fits very well with my hypothesis submitted to but rejected by Nature in 2005. My hypothesis, as published in Logical Biology (5: 51, 2005), a journal I founded in 2000, predicts retention of older DNA template strand by mother cell and reception of younger DNA template strand by daughter cell when a mother cell reproduce its daughter cell(s). This cell reproduction-based hypothesis thus not only resolved an outstanding controversy on DNA template strand segregation but also provided an organic linkage between molecule aging and cell aging and some deep insights on respective contributions of genetics and epigenetics to heredity/stability and differentiation/adaptation.
    Thus, on March 9th, 2012 I sent Prof. Lechene, the corresponding author of the Nature paper, an email containing my Communication Arising (CA) on his Nature paper and asked him and his coauthors to “carefully examine my analysis and argument and then give me a collective formal response (within 10 days as Nature usually requires) so that I can submit my CA to Nature in a timely fashion”. However, I received no response, despite two follow-up emails were sent to him on March 12th and 15th, 2012, respectively. Then I submitted my re-analysis to Nature on March 23, 2012. However, Nature rejected without peer review on March 28, 2012. Nevertheless this criticism and re-analysis of the Nature paper was published in Logical Biology on March 29th, 2012 (12:14, 2012;). On the same day of its publication, a full version which can be seen free of charge at http://im1.biz//LB2012V12N1A3_DNA_segregation.pdf was sent to Prof. Lechene by Logical Biology to invite his group for a rebuttal. But no response has been received from the Nature author.
    Meanwhile I have seen the authors of this Nature paper continuing to spread what I think is their wrong interpretations and erroneous conclusion. And their claim of random DNA template strand segregation still holds strongly in the mainstream field of cell division-based life science, despite the publication of my Open Letter to CNS (see a complete copy at http://im1.biz/albums/userpics/10001/LB2012V12N1A4_OpenLetter2CNS.pdf).
    However, science is about truth. Thus any distortion of the truth should be retracted from scientific publication. It is too sad that some retractions have to be made until research misconduct is exposed. But, if free scientific debate was allowed in a more open and transparent manner, not only creativity would be enhanced but also dignity of scientists could be maintained (see Nature 403, 592, 2000; http://www.nature.com/nature/journal/v403/n6770/full/403592a0.html).

  • Post a comment

    Threaded commenting powered by interconnect/it code.