Nature Medicine retracts MS paper with ghost data by former GSK researcher
Nearly six months after first expressing concern about the validity of a 2010 paper on multiple sclerosis, Nature Medicine has retracted the article for containing “erroneous” data — which in this case don’t seem to have existed, making them more fabricated than wrong.
The paper, “Crucial role of interleukin-7 in T helper type 17 survival and expansion in autoimmune disease,” came from a group led by Jingwu Zhang, who at the time ran GlaxoSmithKline’s Research and Development Center in Shanghai.
Here’s the retraction notice:
The above manuscript was authored by scientists from the GlaxoSmithKline (GSK) Research and Development Center in Shanghai, China, and a researcher from Baylor Medical College who later became a GSK employee. Following anonymous reports of inaccuracies in this study, GSK conducted an investigation into these allegations.
The investigation established that the data depicted in Figure 6 and in Supplementary Figure 7 were erroneously attributed to experiments at Baylor Medical College with blood cells from patients with multiple sclerosis. In fact, no data from experiments with blood cells from patients with multiple sclerosis and no data from experiments at Baylor Medical College were included in the paper. GSK has therefore concluded that the paper contains erroneous data and requests that it be retracted.
Nature Medicine wishes to acknowledge that the two graphs at the bottom left-hand corner of Figure 2a are identical and that this was an error introduced by the journal during the production of the article. Below are the correct graphs as submitted by the authors (Fig. 1).
All authors agreed to the retraction of the paper with the following exceptions and clarifications. Xuebin Liu and Stewart Leung declined to sign the retraction and stand by the conclusions of the paper. Chunxia Wang, Xia Qin and Limin Lu did not respond to Nature Medicine’s requests for comment on the retraction. Ji Wang, Lei Fang, Bing Wan, Jian Hong and Hongtao Lu could not be reached by the journal for comment on the retraction. However, Chunxia Wang, Xia Qin, Limin Lu, Ji Wang, Lei Fang, Bing Wan, Jian Hong and Hongtao Lu signed an initial version of the retraction submitted to the journal by GSK.
The glaring omission here is Zhang, but a spokesperson for Nature Medicine tells us he agreed to the retraction. The paper has been cited 74 times, according to Thomson Scientific’s Web of Knowledge.
Update 12/6/13, 11:00 a.m. Eastern: GSK has issued a statement regarding the study, which you can read here. It says, in part:
In June we concluded an investigation of the publication and, regretfully, established that certain data were indeed misrepresented. We shared our conclusion with Nature Medicine and recommended that a retraction was appropriate.
An early clinical study in which healthy volunteers were receiving an investigational drug targeting the same signaling pathway** was also stopped in June as a precautionary measure. This study had been intended to lay a foundation for eventual testing in MS patients but given the misrepresented data, MS is no longer a focus and no clinical studies are taking place at the moment. GSK still believes the investigational drug and the signaling pathway may have potential in other disease states.
Following the investigation GSK took appropriate action. Five R&D employees who were among the named authors of the publication have left GSK.
The company asked a retired Cambridge professor, Sir Patrick Sissons, to lead a “data review team” to review the research. It found:
There was no evidence that patient safety had been compromised and found no reason to suspend or stop any clinical research.
Data misrepresentation was found in only one publication – the retracted Nature Medicine article.
Two other unpublished manuscripts, involving the same small research group that drove the Nature Medicine publication, contained potentially purposeful misrepresentation of data.
Opportunities exist to further strengthen our existing culture and processes around data management.
The recommendations of the review team are being fully implemented and will be embedded as standard practice.