By Ivan Oransky

Authors have retracted a JAMA article summarizing the evidence behind the benefits of a supplement, after the systemic review upon which it was based was withdrawn.

The 2014 paper, “Oral Zinc for the Common Cold,” drew from a 2013 Cochrane Review, considered the gold standard for rigorous analyses of clinical treatments. That Cochrane review was withdrawn last year, a decision that the editors upheld this past September. Both were co-authored by Rashmi Ranjan Das, of the All India Institute of Medical Sciences, in Bhubaneswar, and Meenu Singh, of the Post-Graduate Institute of Medical Education and Research, in Chandigarh, India.

JAMA editor in chief Howard Bauchner stated that this week’s retraction followed an investigation by the journal:

"…as is always part of our process we contacted the authors. Initially they were not in favor of retraction and so that led to some investigation by us; and ultimately the authors agreed to a retraction.

As you well know our processes are thorough and we always start these discussions with trying to understand the facts – this can always take time."

The JAMA paper has been cited twice. (One of those citations was by a letter in JAMA by Harri Hemila see more on him below.) The Cochrane Review has been cited eight times.

The Cochrane Review was first withdrawn in April 2015, as we reported last October, following concerns about plagiarism and errors raised by Harri Hemila, a public health researcher at the University of Helsinki who had written a 2011 review in The Open Respiratory Medicine Journal, “Zinc Lozenges May Shorten the Duration of Colds: A Systematic Review.” In September of this year, the Cochrane editors added an update, which reads, in part:

The Cochrane Acute Respiratory Infections Group, which maintains the review, withdrew the review in April 2015, pending an assessment of the errors reported, and the group referred the allegations of plagiarism to the Editor in Chief. The Editor in Chief notified the authors of the concerns, and followed the Committee for Publication Ethics (COPE) guidelines.

Replication of text was identified in the Cochrane Review. This was limited to copying of short phrases and was acknowledged by the authors. The level of text plagiarism was minor and at a level that would be addressed by a correction. The Editor in Chief carried out further investigation into the alleged plagiarism of data, with the co-operation of the review authors, who provided supplementary information in support of their work. The allegations related to the derivation of means and standard deviations of data from some of the included studies. Although the authors acknowledge and cite the Hemila 2011 review, the Editor in Chief considered that the authors’ explanation regarding some similarities in presented data between the two reviews was not conclusive.

This version of the review will therefore remain withdrawn.

Hemila stated he did “not disagree with” the updated withdrawal:

There are lots of further problems in that Cochrane review, which I described previously.

However, in my experience, sloppiness is rarely a reason for retractions and I did not expect discussion of the errors in calculations etc in the withdrawal notice.

In particular, the Cochrane format makes it possible to correct various calculation errors in the updates, and in that respect that forum is more robust to sloppiness than ordinary journals.

The plagiarism of the text sections in the Cochrane review was silly and embarrassing for Singh and Das, but I understand the Cochrane comment that it was not the basis for withdrawal.

Plagiarism of data with statements that “we calculated the data from the available information”, and …“We extracted data …” is much more serious violation of scientific processes and I am not at all surprised that Cochrane decided that that was basis for withdrawal.

The authors of the JAMA article have not yet responded to a request for comment.

Update, 4 p.m. Eastern, 1/13/17: Cochrane editor in chief David Tovey:

"As far as I am concerned, the issue is closed. We found evidence of plagiarism across some ‘non-critical’ content (background, etc.), and we had a high suspicion of at least one example of duplicated content in relation to data presented. The authors denied that this had been copied, and they sent us additional data in support of their statement. Our independent statistician (employed part time by Cochrane, but independent of the review author team and the editorial group) concluded that the additional data did not provide evidence of innocence, and, in demonstrating an error that was duplicated, perhaps even argued the reverse – although not conclusively."