Cervical disc arthroplasty: do conflicts of interest influence the outcome of clinical studies?

Abstract

Cervical disc arthroplasty (CDA) is an emerging technique for the treatment of cervical degenerative disease. Multiple studies have investigated the outcomes of CDA, particularly in comparison with cervical arthrodesis techniques such as anterior cervical discectomy and fusion (ACDF). As many entities have financial interests in CDA implants, it is imperative to consider the influence of conflicts of interest on the results of studies investigating the efficacy of CDA. This study aimed to determine if there is an association between the presence of conflicts of interest among study authors and the reported outcome of studies involving CDA. This is a systematic review of clinical CDA publications until October 2016. The outcome measures are presence of conflicts of interest, level of evidence, and outcome for all included studies. PubMed and MEDLINE databases were searched for articles presenting clinical, radiographic, and cost outcomes of CDA. Data extracted from each article included title, authors, publication year, level of evidence, prosthesis type, number of operative levels, presence of conflicts of interest, and outcome. Conflicts of interest were determined by the presence of any conflicts for any author within manuscript disclosure sections or through Open Payments reporting. Outcomes of each study were graded as either favorable, unfavorable, or equivocal. The presence of conflicts of interest was tested for an association with the level of evidence and study outcome using Pearson chi-square analysis, Fisher exact test, or logistic regression for categorical variables. The authors report no conflicts of interest directly related to this work, and have not received any funds in support of this work. A total of 98 articles were included in this analysis. In total, 44.9% (44) of articles had the presence of a conflict of interest, whereas 55.1% (54) of articles did not. Conflicted studies were more likely to present level I evidence and less likely to present level IV evidence than non-conflicted studies (p<.001). Furthermore, conflicted studies were more likely to report favorable outcomes after CDA than non-conflicted studies (90.9% vs. 74.1%, p=.040). The results of this study suggest that the majority of conflicted and non-conflicted studies report favorable results in patients undergoing CDA. However, conflicted studies were also more likely to report favorable outcomes compared with non-conflicted studies. Individual clinicians must critically review published studies for potential conflicts of interest before incorporating CDA into their practice.