We contacted the authors of the papers in question. They replied that “The first publication in a Supplement of the German Journal ‘Zeitschrift für Rheumatologie’ was just a summary of an oral presentation that the first author … had given during an international symposium on glucocorticoids held in Bad Nauheim in the year 2000 … [H]ere only preliminary results of the study were presented, which included only some aspects of the radiographic analysis of the per protocol population and no results of the intention to treat analysis. Results of clinical findings and the complete analysis of side effects were not included.” The concerned reader counterargued that the results would appear duplicated in, and potentially distort, a Cochrane Collaboration analysis. He further pointed out that a 7-page paper should not be considered an abstract, and that the abstract had not been referenced in the full paper that appeared in Arthritis & Rheumatism. In this particular instance, the fact that the “abstract” is identifiable in a PubMed search lends credence to the concern of the reader.
Ordinarily, abstracts presented a public meetings do not constitute prior, hence duplicate, publication when later published in full detail in a peer-reviewed journal. We agree with the concerned reader, but we recognize that we had not provided a formal definition regarding what does or does not constitute an abstract. We also agree that the “abstract” should have been identified as a reference or other citation in the published paper in order to prevent confusion.
The Editorial Board has looked into this issue and has concluded that abstracts that are more than 2 published pages and/or are cited in PubMed will in the future be considered prior publication, precluding further publication of the same data. Furthermore, authors should always indicate prior abstract publication, in text or reference, in the full (definitive) publication.