ARTICLE IN BRIEF
Editors at neurology journals support the establishment of a registry of all drug trials as a measure to ensure that both positive and negative findings are accessible for publication and review.
Robert C. Griggs, MD, Editor-in-Chief of Neurology, did not feel smug when he learned of the embarrassing situation in which another medical journal recently found itself. Research came to light contradicting a positive study the journal published on the effects of an antidepressant in adolescents. Dr. Griggs knew it could have happened to him.
Neurology “absolutely” could fall victim to a similar incident, Dr. Griggs, also Chairman of the Department of Neurology at the University of Rochester in New York, told Neurology Today. Even with all the checks and balances in place, the use of multiple reviewers, and the disclosures required by authors, no editor or author can be all-knowing about the integrity of the papers submitted for publication, he said.
That problem involved the American Journal of Psychiatry, with the publication in June of a double-blind study by University of Texas researchers that showed the selective serotonin reuptake inhibitor (SSRI) citalopram (Celexa) reduced symptoms of major depression in 174 patients aged 12 to 17 (Am J Psychiatry 2004;161(61):1079–1083).
But the journal, and apparently the authors, did not know about unpublished findings from an eight-year trial in Europe showing that citalopram was no more effective than placebo (see sidebar, Drug Company Response on Citalopram). The European study was not cited in the article. This incident occurred against the backdrop of a swirl of questions about the effectiveness of some SSRIs in children, including a lawsuit filed by New York's attorney general alleging that a different drug maker had also suppressed publication of results.
INFLUENCE BY DRUG COMPANIES
“The problem here arises if the company has other data on a topic that they may not have disclosed and that provided a negative result for a drug,” said Dr. Griggs. “Drug company sponsors of one trial are not generally under orders to disclose information about other studies.”
Dr. Griggs and editors of other neurology journals say they worry about the influence of drug manufacturers on clinical trial research, and support the establishment of a public registry that would contain data on all trials.
They plan to discuss the registry idea and other ways to strengthen safeguards for published research in October when the editors meet at the American Neurological Association (ANA) Annual Meeting in Toronto.
“We do not have everything in place that we can do,” said Vladimir Hachinski, MD, Editor of Stroke. Dr. Hachinski, who will be hosting the editors' meeting, is a Professor of Neurology and former Chairman of the Neurology Department at the University of Western Ontario in Canada.
“The registry and other changes may be needed because dealing with a potential conflict of interest is no longer the primary issue,” said Richard T. Johnson, MD, Editor of Annals of Neurology and a Professor and former Neurology Department Chair at Johns Hopkins University in Baltimore, MD. “That's easy enough. You declare it. The much more difficult thing is the protection of authors against abuse by the commercial sponsors.”
REGISTRY TO SUPPLEMENT DISCLOSURES
Any new requirements would augment the ones medical journals have in place today.
For many years journals have asked authors to disclose financial arrangements with study sponsors, such as whether they have received honoraria or have invested in the firm developing the particular drug or device.
For example, if a study submitted to Stroke is “supported by – or any income is received from – a commercial sponsor,” then the authors are required to specify “the input of the sponsoring company in study design, collection, analysis, and interpretation of data, and writing the reports,” according to authors' instructions. “It is important to specifically address whether the sponsors could have suppressed publication if the results were negative or detrimental to the product they produce,” the instructions state.
“We have turned down probably 10 percent of papers because of some concern about the independence of the investigator,” Dr. Griggs said of work submitted to Neurology. “Clinical trials that come from a drug company are suspected of having a bias; and clinical trials that come from investigators who are sponsored by drug companies are suspected of having a bias. We are always suspicious of bias.”
REQUIREMENTS FOR PUBLICATION
Since 2002, 14 neurology-related journals have insisted that authors meet certain conditions before their papers are considered for publication. Specifically the journals require that principal authors declare in writing that they will take full responsibility for the data, the analyses and interpretation, and the conduct of the research; that they had full access to all of the data; and that they had the right to publish any and all data, separate and apart from the attitudes of the sponsor (Neurology 2002;58:347–348).
Roger N. Rosenberg, MD, Editor of Archives of Neurology, Dr. Griggs, with Neurology, Dr. Johnson, with Annals, and Dr. Hachinski, with Stroke, were among those who signed the policy statement.
The declarations were prompted by similar requirements put in place by the large general medical journals, including the Lancet and the New England Journal of Medicine (N Engl J Med 2001;345(11):825–827).
However, the access to data requirement refers only to the findings submitted at the time – not to any others that might concern the same drug or device that is the subject of the paper. That loophole is what can lead to a situation like the one at the psychiatry journal.
“That's why we need to establish the registry, so no one can cut corners. It's like testifying in court,” said Dr. Rosenberg, Chairman of the Department of Neurology at the University of Texas Southwestern Medical Center in Dallas. “You swear to tell the truth, the whole truth. I ask authors to tell the truth; the registry will tell the whole truth.”
SUPPORT FOR NATIONAL REGISTRY
A national registry of clinical trials also has the support of the American Medical Association (AMA). Citing “growing concern about the influence commercial support of drug trials may have on this publication bias,” the AMA in June issued a statement saying it had asked the Department of Health and Human Services “to establish the new registry that would ensure all trials, with negative or positive results, are made publicly available. Each trial would be provided a unique identification and ensured placement on an electronic database of all results from registered trials.”
Dr. Griggs terms the registry “an extremely good idea.” Without it, it is difficult, if not impossible, for researchers to learn on their own all the data that has accumulated about a drug or device. “There are simply too many trials,” he said, estimating that approximately 300,000 trials have been done.
REVIEW PROCESS IN PLACE
All journal editors contacted by Neurology Today described a similarly rigorous peer review process in place today for papers they consider for publication, and said they give extra scrutiny to drug company-sponsored research. At least two take special steps for these kinds of papers.
At Neurology, for example, an average of three reviewers go over papers about clinical trials, and “it is not uncommon to have five or more for large trials,” said Dr. Griggs. “For a big clinical trial we usually have a statistician review the paper as well,” he added.
Archives Editor Dr. Rosenberg has an independent, PhD-trained statistician review all drug company-sponsored research papers – a policy enacted before the author requirements signed by the 14 editors were implemented.
Dr. Rosenberg recounted three instances in his seven-year tenure in which the statisticians kicked out some research as flawed, either because the calculations were incorrect or because the wrong methods of measuring statistical significance were employed.
At Annals, if a reviewer raises questions about the “validity” or “authenticity” of a paper, Dr. Johnson sends it back to the author with a copy to the dean of the affiliated medical school or to some other appropriate oversight office, stating that he will not consider the paper unless the university conducts “an audit of the data.”
Said Dr. Johnson: “You'd think you'd get the paper back saying the audit has been done. That has never happened.”
Editors rely heavily on their peer reviewers to be objective and to spot problems with studies. Stroke reviewers don't even know the names of the authors who wrote their papers, at least until a revision process begins. But Dr. Hachinski said in reality most reviewers will recognize the work because the research community is small.
“If you write a paper on the neurological complications of West Nile virus, there are about eight of us who can review that, and two of us will get it. It's a small world and we all know each other,” added Dr. Johnson. “People think I have my post-docs reviewing papers. I have Nobel laureates reviewing papers.”
Dr. Johnson shared an incident in which the same reviewer, working for two different journals, received a paper twice. The first time he rejected it because some critical tests had not been done. When he received the paper a second time from another journal a short time later, the tests results were there. He was suspicious that the new data had simply been filled in. The journal rejected the paper and sent the letter to the author with a copy to the dean of the medical school.
Through the use of author requirements and sharp-eyed reviewers, the integrity of published papers in neurology journals is probably increasing, Dr. Hachinski said. He also believes editors should more openly encourage the submission of “negative” papers, which could help to reduce instances of suppressed research.
“The most important thing is that it is a good study, even if it is negative,” he said.
DRUG COMPANY RESPONDS TO CLAIMS OF SUPPRESSING DATA
Were data on unpublished findings from an eight-year trial in Europe – showing that citalopram (Celexa) was no more effective than placebo in treating children or adolescents for depression – suppressed from publication by Forest Laboratories, the company that manufactures it in the US? The company responded in a statement – posted online on the company Web site, www.frx.com – that they were not.
“Although no published peer-reviewed reference of the European citalopram pediatric study is currently available, the results of this study were disclosed on several occasions to the scientific community in various formats prior to the June 2004 publication of the US study [in the American Journal of Psychiatry,]” according to the company announcement.
The company cited presentations that were made on both US and European studies last year at the Annual Meetings of the American Academy of Child and Adolescent Psychiatry and the Scandinavian College of Neuropsychopharmacology.
Citing studies by H. Lundbeck A/S, the Danish company that developed citalopram – which did not show efficacy versus placebo – the statement from Forest noted that these findings were disclosed to “US and UK regulatory authorities, both of which posted summary results on the Internet.”
[The FDA report is posted here: www.fda.gov/ohrms/dockets/ac/04/briefing/4006B1_03_Background%20Memo%2001-05-04.doc.]
H. Lundbeck A/S also has advised Forest that the principal investigator of the European stud is preparing a report for publication.