No confidence vote on sepsis paper data leads to Blood retraction

The journal Blood has retracted an article after the authors determined that they could not longer trust in the validity of the data.

The paper has been cited 22 times, according to Thomson Scientific’s Web of Knowledge. From the retraction notice:

Niessen F, Furlan-Freguia C, Fernández JA, Mosnier LO, Castellino FJ, Weiler H, Rosen H, Griffin JH, Ruf W. Endogenous EPCR/aPC-PAR1 signaling prevents inflammation-induced vascular leakage and lethality.

Blood. 2009;113(12):2859-2866.

The authors retract the 19 March 2009, paper cited above. Recently, the authors discovered that some primary data presented in this paper could not be independently reproduced. All coauthors concur with the retraction of the paper and apologize to the readers, reviewers, and editors of Blood for publishing these invalid data.

The nature of the data problem isn’t clear from that passage. But Blood editor Cynthia Dunbar told us that her journal

was reassured that it did not leak over into a second Blood paper published in 2007 that involved some of the same authors. Regardless, this other paper did not involve the technique/experimental design that was problematic to replicate.

That paper has been cited 15 times.

We spoke with the corresponding author of the retracted paper, the prominent Scripps immunologist Wolfram Ruf, who added some details. Ruf said he became aware of a potential problem with figures (see image here) — which he described as permeability stains involving the administration of Evans blue dye into the abdominal tissues of mice — after failing to reproduce the findings during subsequent experiments.

He brought his concerns to his co-authors on the paper, and everyone agreed that retraction, rather than an expression of concern or some other step, was warranted. Everyone, that is, except the first author, former post-doc Frank Niessen, who left Ruf’s lab in 2009 and has not been heard from (by his former colleagues, at least*), since:

We could not get clarification from the first author, who performed these experiments

We asked Ruf if he thought Niessen had falsified his figures or merely bungled them. He demurred:

We have not come up with the clear explanation why it didn’t reproduce. It’s very difficult to prove what he actually did.

Was is possible that Niessen, faced with a technique he didn’t understand well, was out of his depth?

In retrospect, that was my impression.

Scripps had been investigating Niessen’s work, but Ruf said he did not know if the inquiry was ongoing.

Niessen and Ruf collaborated on at least seven published papers, many of them appearing in prestigious journals, during a rather productive two-year period. In late February 2008, the pair made a splash with a Nature paper on sepsis, titled “Dendritic cell PAR1-S1P3 signalling couples coagulation and inflammation” and cited 63 times since.

At the time it was published, Scripps issued an enthusiastic press release  claiming that the two researchers, whom it identified by name, had “uncovered a connection between blood coagulation and the immune system that may have important implications for people with sepsis.”

Those findings, according to Ruf and the release, might lead to new targets for the treatment of sepsis:

This defines a crucial point where the immune system spirals out of control to cause severe sepsis and where there is an opportunity for therapeutic intervention.

Ruf said that article still stands.

This paper did not involve the permeability measurements that were of concern in the retracted paper.

*Update, 3:20 p.m., 6/1/11: Blood tells us that they did hear, independently, by email, from Niessen, who agreed to the retraction, as the notice makes clear. It was his co-authors to whom he never responded. The journal’s Anna Trudgett said:

We also heard that Dr. Ruf had trouble contacting Dr. Niessen. We did some e-mail searching, since it’s our policy to contact all authors and get their consent (otherwise we state the names of those who did not consent or could not be reached) and we received an email agreeing to retract.