Clinical trials are conducted to advance medical knowledge and thereby improve patient care. The results of clinical trials are generally published in the peer-reviewed medical literature, which provides physicians with easy access to this important information. Physicians can then assess the trial results themselves, decide how to incorporate them into patient care, and plan future research. It is through this peer-reviewed collection and dissemination of information that our collective medical knowledge advances. If trial results are not published in the peer-reviewed literature, then this process of advancing knowledge breaks down. We, therefore, find it disturbing when the results of industry-sponsored clinical trials of neurointerventional therapies are not published in the peer-reviewed literature.
Industry sponsors have an inherent conflict of interest regarding publication of trial results. They will tend to encourage publication of favorable results and discourage publication of unfavorable results. This tendency is known as publication bias and has been well documented as a real phenomenon.1 Unlike an industry sponsor, physicians and patients can benefit from access to all trial results, regardless of whether the results are favorable or unfavorable. If we are unable to access information from trials with unfavorable outcomes, then it is difficult to analyze the results and incorporate them into decisions about patient therapy and future research.
There are varying degrees of unavailability of trial results. For example, trial results are sometimes presented at an open meeting of physicians but then not subsequently published. The Schneider Wallstent trial is such a trial.2 This trial randomized patients with symptomatic carotid stenosis to undergo carotid angioplasty and stent placement or carotid endarterectomy. The trial was stopped early for safety concerns because of excessive complications in the angioplasty and stent-placement arm. Obviously, this was not a favorable result for the corporate sponsor. While some important information is released when results are presented at a meeting, it is not an adequate substitute for publication in a peer-reviewed journal. The peer-reviewed literature is the foundation of our collective learning experience in medicine. Research results are subjected to the peer-review process to provide a theoretically unbiased referee who will attempt to identify any serious bias or unsubstantiated claims or conclusions. The peer-review process allows editors to identify which articles should be published and to make improvements in these articles via revisions. Once published, the articles are then permanently accessible to physicians through libraries.
Another example of an unpublished trial is the Acceleration of Connective Tissue Formation in Endovascular Aneurysm Repair (ACTIVE) registry, which was a multicenter study of the Matrix detachable coil (Boston Scientific, Natick, Mass), which occurred immediately after the product was released on the US market in 2002 following US Food and Drug Administration (FDA) clearance. A newsletter from Boston Scientific did publicize some of the ACTIVE results.3 Obviously, such a newsletter would not be expected to be very critical in its analysis of the results, nor would the newsletter be as accessible to the community as would a peer-reviewed journal article. One might reasonably speculate that the ACTIVE results were not published in the peer-reviewed literature because a positive interpretation of the results would not stand up to the peer-review process.
More recently, trial results not published in the peer-reviewed literature have appeared in the “Instructions for Use” included in the packaging for neuroendovascular devices. Some results of the US Multicenter, Randomized, Controlled Study Comparing the Performance of Onyx (ethylene-vinyl alcohol copolymer; ev3, Irvine, Calif) and Trufill (n-butyl cyanoacrylate [n-BCA]; Cordis, Miami Lakes, Fla) in the Presurgical Embolization of Brain Arteriovenous Malformations (AVMs) are available in this format. This trial was conducted to evaluate the safety and effectiveness of Onyx compared with n-BCA for the presurgical treatment of brain AVMs. The trial was completed in 2003, and Onyx was approved by the FDA in 2005. If nearly 6 years have passed since the completion of this study, it seems highly unlikely that it will be published in the peer-reviewed literature.4 The US Multicenter Onyx Aneurysm Study is another study for which results are available only in the “Instructions for Use.” Onyx HD-500 was approved by the FDA in 2007 as a humanitarian use device for cerebral aneurysm treatment. Perhaps the results of this trial will be published eventually, but as more time passes, publication probably becomes progressively less likely.
To help ameliorate the problem of unpublished results, the FDA Amendments Act of 2007 now mandates that results of trials be disclosed through a data base that will be publicly accessible via the internet at www.clinicaltrials.gov.5 This legislation will certainly help to improve public access to data of interest to physicians and patients. Hopefully, publication of clinical trial results in peer-reviewed literature will also occur more consistently in the future because the peer-reviewed publication process is essential to complete and critical evaluation of the data.
References
- Copyright © American Society of Neuroradiology