Doing the right thing: Authors retract PNAS paper when new experiments show “conclusion was incorrect”

pnascoverResearchers in Sweden and Australia have retracted a paper in the Proceedings of the National Academy of Sciences (PNAS) after follow-up experiments disproved their conclusions.

Here’s the notice for “Dominant suppression of inflammation by glycan-hydrolyzed IgG,” which is signed by all nine of the paper’s authors:

The authors wish to note the following: “Using studies of IgG hydrolyzed by the streptococcal glycan hydrolyzing enzyme EndoS, we found that treatment of mice with hydrolyzed IgG blocked antibody mediated arthritis. As an explanation for this observation, we suggested that EndoS-hydrolyzed IgG per se dominantly blocks local immune complex formation.

“With new data from our own follow up experiments, we have now found that this conclusion was incorrect.

“Our new data shows that injection of EndoS is much more potent in vivo than we could logically anticipate, as i.v. injection of doses containing less than 0.1 μg EndoS mixed with IgG suppressed arthritis using the same model as the one reported in the initial paper (collagen antibody-induced arthritis). We previously excluded the possibility that contaminating EndoS could play a role, as this contaminating amount was not detected using standard methods in the hydrolyzed IgG fraction we used in the experiments. Furthermore, much higher doses of EndoS injected in the same mouse strain as a control experiment did not affect collagen induced arthritis in earlier experiments. The correct interpretation of our collective data is that EndoS operates very potently in vivo on an immune complex-mediated disease, possibly by accumulating within immune complexes. Because this interpretation is different from our major conclusion of the published paper, the authors have unanimously decided to retract this paper to be able to publish the data connected with a correct interpretation. We sincerely apologize to readers of this paper, who might have been misled by our earlier interpretation.”

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

19 thoughts on “Doing the right thing: Authors retract PNAS paper when new experiments show “conclusion was incorrect””

  1. I wonder how many researchers would do the same instead of withholding that information as some kind of “white lie”.

  2. Is it a good thing to retract research simply because new research contradicts the old research? I mean, on the face of it, you have 1 paper saying “yes” and 1 paper saying “no.” How do they know that the old research was invalid while the new 1 isn’t? 1 paper doth not a meta-analysis make…

    1. I agree with saijanai. The fact that today a theory is right, but tomorrow it could be wrong does not mean that we have to retract everything that was every previously published, unless explicit fraud was involved. The nature of science is experimental, and an outcome is based on a set of experimental procedures which most likely are more accurate now then they might have been 10 or 20 years ago, simply because the techniques used might be more precise, or more accurate. In this particular case, could it be related to the purity of the immunoglobulins used? There are some real cases that deserve to be retracted. But I feel this is not one of them. By having this murky list of retractions that are all clustered under the same umbrella of misconduct + disproving, we risk clouding the literature with doubt about what should or shouldn’t be retracted: I sense that in 100 years from now, there will be a sense of pity in the air, possibly even of bitterness, at what is taking place now.

      Just because, after about the 17th Century humanity learnt that the world was round doesn’t mean that we have to erase all the literature about the flat earth theory [1]. It is precisely because the flat earth theory existed in the first place that people could then challenge that thought, through discovery, and science. Even today, the fact that magnetic north and magnetic south are constantly changing [2], doesn’t mean that when finally they do actually change that we should erase all literature that stated, until that point, that magnetic north was in the Arctic.

      Similarly, just because PNAS paper 1 produced a NO result while the subsequent study produced an apparently contradictory YES result should not mean that it should be retracted, because this is such a fundamental step in science: testing, building, disproving, reproving. And I really don’t like that last sentence “We sincerely apologize to readers of this paper, who might have been misled by our earlier interpretation.”, which sounds so lawyer-doctored. Is this possibly a phase in the history of science and science publishing that is the antithesis of a Renaissance?

      [1] http://en.wikipedia.org/wiki/Flat_Earth
      [2] http://www.huffingtonpost.com/2014/07/10/earth-magnetic-field-weak-flip_n_5574638.html

    2. The way I read the line “…the authors have unanimously decided to retract this paper to be able to publish the data connected with a correct interpretation.” is that they are retracting the paper so that they can use the information therein for a new paper supporting their new hypothesis. It’s not that the old research is invalid but it will be used in a new context to support a different conclusion.

      1. But they could have published a corrected interpretation without retracting the original. That’s kind of just how science works. I agree with most of the other comments here that this doesn’t seem to warrant a retraction, and I’m not sure the authors deserve a pat on the back for “doing the right thing.” (Although, certainly they didn’t do anything wrong, and I think they do deserve a pat on the back for being so transparent about their incorrect conclusions).

  3. I admit to have zero knowledge of this area of research, but I, too, don’t see why retracting this paper is the right thing to do. What would be the harm in having the authors publish their new augmented data with the new analyses and interpretation as long as it is made crystal clear to the reader that the new data also include the old data reported in this paper? Would other journals, even PNAS, be less likely to publish such a paper because some/much of its data have already been published? I think not.

  4. I think we lose something from science when we attempt to hide incorrect conclusions. At some point, nine well-trained experts in a given area interpreted their data incorrectly. Doesn’t that tell us something about how the process of science works? Reanalyzing data happens all the time in the physical sciences, primarily by theoreticians, but there is nothing to stop experimentalists from getting someone’s data and publishing it with a different fit or whatever, leading to different conclusions. Seems like it would be even easier for the same authors to reanalyze their data.

    1. There is no hiding. I previously had similar doubts, but someone explained me that a retraction is just a notice, you can still download and read the article (the paper version in naturally also still in all libraries).

      That being said, I see no reason to retract in this case. Many studies are found to be wrong later on. That is scientific progress.

  5. IMO the retraction may have more to do with the patent and possible therapeutic application rather than a decision of conscience to correct the interpretation. As per the foot note “Hansa Medical AB has filed patent applications for using EndoS-modified IgG for treatment of arthritis and K.S.N., M.C., and R.H. are listed as inventors”.
    Science is constantly evolving. With newer techniques and methodologies we have been learning that what we believed to be true was not necessarily true. Therefore it makes no sense to retract for the reason stated.

  6. Is it as simple or simpler to determine if a paper’s proclaimed results could not be reproduced than if it was retracted? If not, then retraction might be the practical solution.

  7. I think this highlights a problem with the current “retraction system.”

    We have a set of experiments, with apparently nothing wrong with them except their poor interpretation and now, if someone wants to cite this paper, it will be a little fishy.

    If the authors are allowed to republish this data along with a better interpretation and a bit more data, it sets a dangerous precedent. Just publish quick, and if you learn a little more, retract and republish.

    Personally, I would feel better about a situation like this being handled by correction. I’m a bit surprised that PNAS went along with a retraction here… why not just force the authors to make a correction and if they refused, issue an expression of concern?

    1. No correction or EoC necessary here, though. As long as the methods and data are honest, make whatever conclusions you want and put it out there. Let the scientific community decide if the conclusions are justified based on the data. That’s science. Interpretations/conclusions are subjective. Methods/data are not.

      1. Yeah, you’re probably correct… an EoC threat is probably overkill… I was concerned that the assumptions underlying the very idea of the experiments were really flawed to the point where the reviewers probably should have said something but failed.
        Also, thinking on this more, perhaps the ideal situation would be something of an addendum… not really a correction, but rather as statement of “we thought this, but we were wrong here and here because of X.”
        If this addendum could be attached below the paper, that would really be a great benefit to the community and to science. That’s also something that I really wish would occur:
        I see no reason at all not to include ALL subsequent comments/adds to a publication in the official PDF: The original paper, any correction notices, a retraction notice, a related letter to the editor, etc. Much more efficient and ensures that everyone reading (at least the newest) version of a pdf has the whole story, rather than making them follow a bunch of links.

    2. I agree, QAQ, an EOC or erratum could have easily sufficed, or a letter to the editor, which could easily have included additional data or analyses. I think scientists do not have to feel embarassed about making errors, or about interpretations changing over time. Were these authors feeling so guilty about their own analyses, or did soeone perhaps pressure them to retract? We come back to the same points repeatedly in recent times. Who has the power to retract? The authors, the publisher, the EIC, the lawyers that oversee the journal? There are too many fuzzy gaps in the backgrounds of far too many retractions, and their wording, and given that RW is the de facto leader in this debate on the web, I think that there should be greater pressure on authors, their institutes, editors and/or the EIC as well as journals and/or their legal or managerial representatives to provide comment in all cases here at RW, and even at PubMed Commons and PubPeer, to better elucidate the scientific public. Commenting on retractions in full should form part of the new editorial portfolio, I believe. Otherwise, we are left guessing and coming up with some really bizarre explanations and theories when, very possibly, the actual reason might be much simpler.

  8. I second all the voices saying that such a situation does not warrant a retraction, more likely an EOC of sorts, but frankly – a completely new paper would be best, citing and expanding on the original work. If we wanted to remove from the body of scientific literature all the studies that were shown to be wrongly interpreted, then we’d be lost with very little, I’m afraid. As Lawrence Krauss (I believe) once said, 90% of the works published in theoretical physics are wrong. As someone said here earlier – knowing where we came from, knowing our mistakes, actually helps us progress.

  9. You are missing the point of this: in the original paper the authors stated that contaminating EndoS could not play a role because they did not detect it, and they therefore assumed that it was not there. But in reality it was there at very low levels, and it probably synergized with immune complexes. Therefore the original paper reported something that was not true (no EndoS) which in the end turned out to be the critical driver. I totally agree with the actions of the authors and wish to congratulate them.

  10. I appreciate _a reader’s_ clarification of the problem with this paper. My own lack of knowledge of the area notwithstanding, I still wonder whether a retraction was the right call in this case. Surely there are other published studies across the biomedical disciplines that are similarly plagued by critical confounds that are subsequently clarified by further experimentation. On the other hand, if the consensus of those who do research in this area is that a retraction is the most appropriate approach, then perhaps we should all trust their judgement.

  11. The interpretation is wrong, but much more importantly, the samples used were fatally contaminated rendering the study highly misleading at best. Think not taking this action would be in the same category as not withdrawing oncology studies that have inadvertently used the wrong cell line – the observations are ‘correct’, the conclusions are ‘wrong’, but the whole study is flawed. I vote retraction and good on them!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.