To the Editor:
The editorial staff of Clinical Orthopaedics and Related Research (CORR) has expressed concern that the two sets of papers discussed below represent duplicate publication as defined by the COPE (Committee on Publication Ethics) Guidelines on Good Publication Practice.3 While we acted in good faith and believed at the time that the submissions of these papers were done within the boundaries of ethical behavior, in light of our recent review of the COPE guidelines and our communications with the editors of the journals, we recognize that there is an overlap in these papers inconsistent with the COPE guidelines. In addition, in the first set of papers Figure 1 (a photograph of the prosthesis) was identical in both papers; we were in error by not obtaining permission from the publisher of CORR before publication of the second paper. We wish to express our regrets to the editorial staff of the journal, and to its readership. To insure that there is a clear demarcation between clinical studies submitted for publication in the future, an oversight committee in our department has been established to vet manuscripts before submission for publication. Given the circumstances presented below, we will pay particular attention to papers that are submitted in response to an invitation to participate in a special symposium or are submitted for inclusion in proceedings from subspecialty meetings, especially if they involve a patient cohort that has been the subject of previous publications. If there is any question about potential overlap, this will be brought to the attention of the author who will address this in the cover letter to the journal. This will insure transparency for the editors and reviewers.
In the November 2004 issue of CORR we published a paper entitled, The Progression of Patellofemoral Arthrosis After Medial Unicompartmental Replacement: Results at 11 to 15 Years,2 and in May 2005 we published a paper in the Journal of Bone and Joint Surgery (JBJS) entitled, Results of Unicompartmental Knee Arthroplasty at a Minimum of Ten Years of Follow-up.1 The JBJS paper was initiated by us and was written and submitted first. The CORR paper was requested by the Knee Society to be included in their 2004 proceedings; this paper, submitted second but published first, is an update of the JBJS paper. There are three differences, which in our view at the time justified separate publication: (1) these two articles include slightly different cohorts; (2) the two articles contain different data from two different followup intervals separated by 2 years; and (3) the CORR paper reports progression of disease in the patellofemoral compartment which was discovered after the additional 2 years of followup.
The article in JBJS1 includes an unselected group of all unicompartmental knee arthroplasties done at our institution from 1987 to 1993 (n = 62 unicompartmental knee replacements in 51 patients). The article in CORR2 includes only the more common medial unicompartmental replacements done at our institution during the same period. The followup of these two cohorts is different. For the JBJS article, the study period closed in mid 2001. This resulted in the followup for this unselected group from 10 to 13 years, with an average followup of 11.8 years. There were two failures in this group at 7 and 10 years. The survivorship analysis was carried out to the end of the followup (13 years), and the paper was finished in early 2002. It was submitted to JBJS shortly thereafter.
As a direct result of an invitation by the Knee Society to participate in the scientific program of the 2004 Open Meeting in San Francisco, a study was done to update the results to 15 years followup on a group of selected unicompartmental replacements, including just the more common medial unicompartmental replacements (n = 59 of the 62 unicompartmental knee replacements in 48 of the 51 patients reported in the JBJS paper); the less common lateral replacements (n = 3 unicompartmental knee replacements in three patients) were excluded. The study period closed in mid 2003, 2 years after the study period for the JBJS paper. All patients who were still alive and were able to travel were seen for followup, were examined, and had new radiographs taken. The average in followup for this group of patients was 2 more years than in the previous paper on the entire population of patients who had medial and lateral unicompartmental replacements. However, because some patients had died, the average followup only increased by 1.5 years to 13.3 years; all patients had between 11 to 15 years followup. This analysis was completed and the paper was submitted to CORR in early 2004 as requested by the Knee Society to be included in their proceedings. The survivorship analysis was redone and was carried out to the end of the followup, 15 years. Because there were no new failures since Years 7 and 10, the survivorship was unchanged after 10 years to 15 years. In addition, the failures reported in this cohort were the same as those reported in the JBJS paper. Figure 4 in the JBJS paper and Figure 5 in the CORR paper are different because the x-axis in the JBJS paper is carried out to 13 years, whereas the x-axis in the CORR paper is carried out to 15 years.
In addition, the CORR paper focuses on the progression of patellofemoral disease which was discovered after the additional 2 years followup. In the JBJS paper, the prevalence of radiographic changes was 34%, with nearly ⅔ of these having Grade 1 changes. In contrast, the CORR paper reported a prevalence of radiographic changes in the patellofemoral joint of 54%, with less than ⅓ having Grade 1 changes and more than ½ of these having Grade 2 changes. We believe this is a clinically relevant finding.
It is unfortunate that the paper with longer followup was published before the paper with shorter followup; this can occur with the vagaries of the review process. We did not cite the JBJS paper in our CORR paper, because it was not yet published.
In the March 2004 issue of CORR, we published an article entitled Cementless Acetabular Reconstruction in Revision Total Hip Arthroplasty4 as part of the symposium on Revision Total Hip Arthroplasty sponsored by the Hip Society. This invited article presented detailed descriptions of the surgical technique we presently use, a review of the published literature on the topic, and clinical results of a cohort of 138 patients who had cementless acetabular reconstructions at the time of their revision total hip arthroplasties who were followed for a minimum of 10 years.
Subsequently, we submitted an original scientific article to JBJS entitled Revision of the acetabular component without cement after total hip arthroplasty5 which was published in August 2005. This was a followup note updating previous reports in JBJS now at a minimum of 15 years; the same cohort of 138 hips was studied. This report included information regarding periacetabular osteolysis and the fate of the femoral components; two important facets of this report that were not investigated in the CORR publication. Further, at this later time (minimum 15 years as opposed to 10 years), there were two reoperations for wear and osteolysis that were not recognized previously and an additional acetabular component failure. The article published in JBJS did not reference the work in CORR as it had not been published at the time of submission.
We believed at the time of submission that the publication of these two separate articles was justified as they had a different scope and intent (an invited review article with a review of the literature and a detailed description of the surgical technique as compared with an original scientific article), reported different clinical information (eg, periacetabular osteolysis and the outcomes of the femoral components), and despite studying the same cohort of patients had differing criteria for entrance into the study (10 years as compared with 15 years minimum followup), and a mean followup that differed by more than 2 years.
We appreciate this opportunity to clarify the nature and intent of these two sets of papers and will be vigilant in future submissions regarding adherence to the COPE3 guidelines.
Richard Berger, MD
Craig Della Valle, MD
Joshua J. Jacobs, MD
Mitchell B. Sheinkop, MD
Aaron G. Rosenberg, MD
Jorge O. Galante, MD
Department of Orthopaedic Surgery
Rush University Medical Center
1. Berger RA, Meneghini RM, Jacobs JJ, Sheinkop MB, Della Valle CJ, Rosenberg AG, Galante JO. Results of unicompartmental knee arthroplasty at a minimum of ten years of follow-up. J Bone Joint Surg Am
2. Berger RA, Meneghini RM, Sheinkop MB, Della Valle CJ, Jacobs JJ, Rosenberg AG, Galante JO. The progression of patellofemoral arthrosis after medial unicompartmental replacement: results at 11 to 15 years. Clin Orthop Relat Res
3. COPE. 2003 Reports. Committee on Publication Ethics (COPE): guidelines on good publication practice. Available at: http://www.publicationethics
4. Della Valle CJ, Berger RA, Rosenberg AG, Galante JO. Cementless acetabular reconstruction in revision total hip arthroplasty. Clin Orthop Relat Res
5. Della Valle CJ, Shuaipaj T, Berger RA, Rosenberg AG, Shott S, Jacobs JJ, Galante JO. Revision of the acetabular component without cement after total hip arthroplasty: a concise follow-up, at fifteen to nineteen years, of a previous report. J Bone Joint Surg Am