The authors of a paper in the New England Journal of Medicine are retracting it, after being unable to find data supporting a table that required corrections.
Here’s the notice:
As the authors of the article entitled “CPAP for the Metabolic Syndrome in Patients with Obstructive Sleep Apnea”1 published in the Journal on December 15, 2011, we regret to report that transcription errors occurred in the Supplementary Appendix, available with the full text of the article at NEJM.org. There were multiple errors in the table on pages 18 and 19 of the Supplementary Appendix concerning data on the accumulation of abdominal fat as assessed with the use of computed tomography and on carotid intima–media thickness as assessed with the use of ultrasonography. These errors, in turn, changed some values in Table 4 of the article. Although these changes do not alter the conclusions of the article, the primary data could not be located to verify corrections made from secondary tables. Accordingly, we have no way of confirming the correct data and, with regret, wish to retract the article.
The paper, by Surendra K. Sharma of the All India Institute of Medical Sciences and colleagues. has been cited 69 times, according to Thomson Scientific’s Web of Knowledge.
In an exchange of letters published in the journal in March 2012, Harvard Medical School’s Suzanne Bertisch and a colleague questioned the findings. Bertisch tells us she was surprised by the retraction:
Our letter focused on the selected methodology, but I had been surprised by some of the results, particularly the weight loss, as well. Though I was surprised, I hadn’t considered/suspected fabrication or something else that would lead to a retraction. In our letter, we had asked for previously unreported data, and the authors produced it. In fact the data they produced gave credence to our concerns, and I was somewhat surprised how fully they disclosed these data. It suggested to me that they didn’t fully understand our concerns. While I perplexed by their weight loss findings, and wondered what the explanation was scientifically, I didn’t think to question the validity of the results. I also assumed the data were somewhat vetted by NEJM, though I was surprised that the data we asked for was not already reported. It’s disappointing, particularly from the perspective of a sleep researcher, as there just a few NEJM papers on sleep. It’s a step back for our field.
Sharma did not respond to our requests for comment. Pfizer, which supported the study, told us:
Pfizer provided funding for this independent research study pursuant to our independent grant policy that provides support for investigator-initiated research to advance medical and scientific knowledge. To maintain independence, we did not participate in the actual study and had no oversight or control of the study or the analysis of the results other than to provide this funding.
Retractions are rare at the NEJM, although the journal tops the Retraction Index. The last retraction in the journal was in 2011, of a paper by Anil Potti.
Note: Ivan also reported on this case for MedPage Today, where he is vice president and global editorial director.
The wording by Pfizer was absolutely hilarious, in a tragic-comedy sort of way. Are they saying that they throw money at research without caring about experimental design, the validity of data or the ethical implications of a retraction? Hey Pfizer, throw some at me, I need a financial boost for research. How can a pharma giant have such a hands-off approach to bad research ethics? Surely, the correct thing it should have done was first to publically state that the practice of not keeping data by Sharma et al. was wrong and secondly, to demand a full refund from Sharma. What concerns me even more is if any of the 63 studies that may have used this key study should now also be retracted. If the base of a house is flawed, then the whole house is flawed.
Someone was asleep at the wheel.
This is an embarassment to the sleep research field and NEJM. “Although these changes do not alter the conclusions of the article…”….who are they kidding ? Why didnt the NEJM or reviewers look at the erroneous supplementary table ? One hopes that these authors are investigated by their own institution.
The NEJM has to fall on its sword here. This was a typical “too perfect” result from a group who had never published any sleep research in major journals before. There seems to have been multiple repeat entry of data with lines of exactly the same data to the second decimal point in the supplementary table. Instead of a detailed apology from the journal, it allows a wishy washy retraction that uses the transcription error excuse instead of admitting much worse. This paper has been held up by the CPAP industry and their acolytes in academia as “proof” sleep apnea is a cause of diabetes. NEJM will gleefully accept misleading and potentially fraudulent data like this while avoiding publishing negative studies because the positive studies make better press releases.