Three journals belonging to the American Heart Association are dealing with a data manipulation case involving a Japanese scientist who collaborated with some of the United States’ most prominent cardiac specialists.
Circulation has retracted a 2008 article after the researchers said the lead author’s sloppy record-keeping prevented them from reproducing their experiments. We think there might be a bit more to the story.
Here’s the notice:
The authors of the following article have requested that it be retracted from publication in Circulation:
Kawakami A, Osaka M, Tani M, Azuma H, Sacks FM, Shimokado K, Yoshida M. Apolipoprotein CIII links hyperlipidemia with vascular endothelial cell dysfunction. Circulation. 2008;118:731–742.
The corresponding author, Dr Akio Kawakami, reported to the editors that the authors of this manuscript have raised concerns related to the accuracy of the data presented in this article. In the process of following up the findings reported in this study, they were unable to reproduce some experiments due to Dr Akio Kawakami’s negligence in keeping proper original records. This information was reported to the editors by Dr Kawakami directly. The authors apologize to the readers of Circulation for any inconvenience caused by this retraction.
The paper, according to the authors, carried substantial promise:
Our results suggest that apoCIII is a crucial link between dyslipidemia and insulin resistance in vascular endothelial cells with consequential deleterious effects on their atheroprotective functions.
It has been cited 37 times, according to Thomson Scientific’s Web of Knowledge.
We generally don’t like to read between the lines — unless we’re forced by opacity and lack of cooperation, anyway — but we think it’s worth noting that the paper includes a data supplement of figures and a Word document describing, in detail, the research methods.
So, a few things could be true. The data supplement might be woefully inadequate to guide the reconstruction of the study — in which case, perhaps the reviewers and editors might have figured that out prior to publication. Or, the supplement contains enough information, but Kawakami’s results don’t jibe with the follow-up findings for other reasons.
According to his most recent publication, Kawakami is with the department of Geriatrics and Vascular Medicine at Tokyo Medical and Dental University.
FM Sacks is Frank Sacks, of Harvard’s Channing Laboratory. He and Kawakami have published at least six papers together on apolipoprotein CIII, three of which appeared in Circulation, including the 2006 article “Apolipoprotein CIII in apolipoprotein B lipoproteins enhances the adhesion of human monocytic cells to endothelial cells.” Their most recent title appears to be a 2010 article in Arteriosclerosis, Thrombosis, and Vascular Biology, which, like Circulation, is a title of the American Heart Association.
We’ve left a message for Sacks, and will update with anything we hear back.
Update, 2:30 p.m. Eastern, 2/15/12: Meanwhile, readers have pointed us to another Kawakami paper, in Circulation Research, that has been retracted. The notice is not live on the journal’s website, however — we’re told it will be going up on Thursday afternoon — and we don’t know what it says. We’ll update this post when we learn more.
Update, 4:45 p.m. Eastern, 2/15/12: We’ve heard from the AHA that the Kawakami paper in Arteriosclerosis, Thrombosis, and Vascular Biology also will be retracted, although the notice is not yet available. According to an AHA spokesperson, the journals first learned about a problem with the articles in December.
In addition to Sacks, another prominent Harvard co-author on the papers in Circulation Research and Arteriosclerosis, Thrombosis, and Vascular Biology is Peter Libby, chief of cardiology at Brigham and Women’s Hospital. (Note: Libby and Kawakami appear on a different paper in Arteriosclerosis, Thrombosis, and Vascular Biology from the one that is being retracted, as well as two other articles in Circulation, here and here, neither of which to our knowledge is being retracted.)
In 2011, Libby won the AHA’s Basic Research Prize for his work on illuminating the role of inflammation and other triggers of heart disease.
Update, 5:20 p.m. Eastern, 2/15/12. Headline, first paragraph changed to reflect new information. We finally have the retraction notices for the two other papers.
Here’s Circulation Research‘s statement:
The authors of the following article have requested that it be retracted from publication in Circulation Research:
Kawakami A, Osaka M, Aikawa M, Uematsu S, Akira S, Libby P, Shimokado K, Sacks FM, Yoshida M. Toll-like receptor 2 mediates apolipoprotein CIII–induced monocyte activation. Circ Res. 2008;103:1402–1409.
The corresponding author, Dr Akio Kawakami, admitted to the editors to improperly handling the collection and presentation of data in this article such that the authors can no longer verify the authenticity and accuracy of the data presented. These errors include, but may not be limited to, the blots in Figure 2A, Figure 4D, and Online Figure III originating from unrelated experiments of the corresponding author, and the incorrect reporting of “n” in Figures 5 and 6, which are less than indicated. As such, data in those figures are not verifiable.
All co-authors involved in this study other than the corresponding author, Dr Kawakami, had no knowledge of any scientific impropriety related to the collection, analysis, or presentation of data in this article. Dr Kawakami takes full responsibililty for this.
And here’s ATVB‘s notice:
The authors of the following article have requested that it be retracted from publication in Arteriosclerosis, Thrombosis, and Vascular Biology:
Abe Y, Kawakami A, Osaka M, Uematsu S, Akira S, Shimokado K, Sacks FM, Yoshida M. Apolipoprotein CIII induces monocyte chemoattractant protein-1 and interleukin 6 expression via toll-like receptor 2 pathway in mouse adipocytes. Arterioscler Thromb Vasc Biol. 2010;30:2242–2248.
The corresponding author, Dr Akio Kawakami, reported to the editors his concerns related to the authenticity and accuracy of the data presented in this article. Because the number of experiments was inflated in some studies, and the corresponding author presented some data from unrelated experiments, the results are not verifiable. All co-authors involved in this study other than the corresponding author had no knowledge of any scientific impropriety related to the collection, analysis, or presentation of data in this article. Dr Kawakami apologizes for any adverse consequences that may have resulted from the article’s publication and any inconvenience and wasted effort that this may have caused the scientific community and readers of the journal.
I did a research at Tokyo Medical and Dental University, and the place does high quality research and has a very high integrity, thus I think that Kawakami found something odd in the results he could not clarify retrospectively, and decided to be honest about it, and rather retracted a paper than to be one of those “it just does not work in other labs” zanies. I think it is an exemplary approach by a PI…
sorry, he probably was not the PI on this
Here is another retraction of Kawakami’s paper.
http://circres.ahajournals.org/content/103/12/1402.full
This one is Circulation Research.
another image manipulation by Kawakami
http://tmdu-kawakami.blogspot.com/2012/02/pitavastatin-inhibits-remnant.html
AHA should retract this article.
Another image manipulation by Kawakami (Circulation. 2003 Nov 25;108(21):2679-88.)
http://tmdu-kawakami.blogspot.com/2012/02/remnant-lipoprotein-induced-smooth.html
Likewise, AHA should retract this article.
Now, the total number of problematic articles by Kawakami is 5.
Article No.1 → Circulation. 2003 Nov 25;108(21):2679-88.
Article No.2 → Arterioscler Thromb Vasc Biol. 2005 Feb;25(2):424-9.
Article No.3 → Circulation. 2008 Aug 12;118(7):731-42.
Article No.4 → Circ Res. 2008 Dec 5;103(12):1402-9.
Article No.5 → Arterioscler Thromb Vasc Biol. 2010 Nov;30(11):2242-8.
In Circ Res. 2008 Dec 5;103(12):1402-9, “total” bands (Flip vertically) in Figure2A are similar to “IB:NF-kB p65” bands in online Figure III.
http://tmdu-kawakami.blogspot.com/2012/02/toll-like-receptor-2-mediates.html
The retraction in Circulation is behind a paywall, but the Circulation Research paper is available in its entirely, headlined “Retracted.”
Here is the factual part for Circulation Research:
“The corresponding author, Dr Akio Kawakami, admitted to the editors to improperly handling the collection and presentation of data in this article such that the authors can no longer verify the authenticity and accuracy of the data presented. These errors include, but may not be limited to, the blots in Figure 2A, Figure 4D, and Online Figure III originating from unrelated experiments of the corresponding author, and the incorrect reporting of “n” in Figures 5 and 6, which are less than indicated. As such, data in those figures are not verifiable.”
And what do you know, ladies and somewhat gentle men, them figures in question again contain cut up into strips Western blots… I am going to start collecting signatures for petition to outlaw Western blotting in universities, public places and within 1000 feet from schools and daycare centers.
11jigen = integrity ninja!
I quote: “a few things could be true. The data supplement might be woefully inadequate to guide the reconstruction of the study — in which case, perhaps the reviewers and editors might have figured that out prior to publication”
Unfortunately, this is almost certainly the case, and not exceptional to this retraction – the methods descriptions in the vast majority of publications are incredibly short and incomplete, and truly reproducing the results of almost ANY paper requires either A): A huge amount of un-cited, assumed knowledge and technique is shared between the lab publishing and the lab reproducing the paper, or B) you email the authors and ask them the details of what they really did.
Besides the obvious deviations from the scientific method, the problem with A is that what one researcher assumes is a standard practice and then therefore describes in a single sentence often has key details which vary from lab to lab and experimenter to experimenter. The problem with B is willingness and time, and spam filters 😉
I love this website, and I think it’s great that retractions and scientific misconduct are getting more attention – academic science functions on “the honor system” after all. However, I think there are large, systematic problems such as the one I mention which are considered completely acceptable by the community as a whole, and yet drastically reduce the worth of the science which we produce. For every western blot that is fixed in photoshop, there are 5,000 which are simply not done in replicate. For every altered light microscopy image there are 10,000 which are not quantified but simply presented as “see, look, w’ere right! We even put little arrows to show you were to look to see we’re right!” For every incorrect conclusion derived form falsified data, there are hundreds which are simply not logically supported by the presented results, or for which key control experiments are missing.
In the end, the incentives in the academic research system only very indirectly reward reproducible, sound science which can be built upon and is truly valuable (partly because this is so hard to measure). They do directly and immediately reward quick, flashy publication followed by lofty claims in the discussion section.
I agree with most of what you say, but I think you have to look deeper into one of the root causes… bad peer review. This does not happen because people are “lobbing” papers to their friends. It happens because peer reviewers are un-paid, and underappreciated. In the current system, you will be VERY hard pressed to find a university faculty member who has more than 1 hour (maybe 2) to devote to the thorough peer review of a manuscript – it’s just not worthwhile to spend any longer (the situation may be different for more junior folks who see it as a CV builder). Even making peer-review paid will not solve the problem, since many academic departments do not recognize it as a valuable activity for promotion and tenure. Imagine saying to your chair, “hey, I didn’t get my grant funded, but I reviewed 100 papers in the past year, surely that counts for something?”. Yes, diddly squat. Until peer-review is either compensated, or recognized by institutions as a valid exercise (or both), these types of abuses will continue to fall through the cracks.
The irony of it all, is that the journal publishing business remains one of the most profitable games in town, and the profits of the big houses have been virtually untouched by the financial crisis. Diversion of a tiny percentage of those profits to ensuring better proof-reading and hiring more paid scientists to edit manuscripts, would go a long way to increasing the validity of what gets published.
I agree completely. As you say, there simply isn’t any incentive for the faculty to do the kind of quality review that the science deserves and the academic system requires. The rewards do not encourage the desired behavior, and so the desired behavior is increasingly hard to find.
What a country. The government and corporations fund research that gets turned into content for free and you publish it and get rich.
2 suggestions.
A. In the digital age all data and detailed methods can be published. Electrons are cheep!
B. editors should subject a few papers per year to a detailed audit (real review). This kind of works for the IRS,
Ok 3 suggestions
Numbers appear in a poison distribution unless there is cheating. (1 appears more than 2 which appears more than 3. You can run programs to audit for this.
David egilman Editor IJOEH.COM