JAMA urges journals to toughen data review
■ Controversy over GlaxoSmithKline's diabetes drug Avandia prompts calls for requiring independent analysis of drugmakers' trial data.
- WITH THIS STORY:
- » External links
Medical journal editors should require independent analysis of industry-sponsored trial data by an academic statistician before publishing results, according to an editorial published in the March 24/31 Journal of the American Medical Association.
The call comes in response to internal GlaxoSmithKline documents revealed as part of a February Senate Finance Committee report investigating the company's handling of data related to its diabetes drug Avandia (rosiglitazone).
A meta-analysis published June 14, 2007, in the New England Journal of Medicine found that Avandia increases heart attack and cardiovascular death risk by 43% and 64%, respectively. The company responded by unblinding and publishing the results of an ongoing safety study known as RECORD: Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycemia in Diabetes.
The internal documents show debate over whether the RECORD interim analysis manuscript, published July 5, 2007, in NEJM, might misrepresent the study's results and raise the question of whether the academic investigators involved had adequate access to trial data to draw their own conclusions.
U.S. sales of Avandia dropped dramatically after the Food and Drug Administration added a black-box warning about heart attack risk in 2007. An FDA advisory panel meeting scheduled for July will consider whether the drug should be removed from the market.
It is in these kinds of cases that journals requiring an independent academic statistician to review the data analysis could help ensure that physicians and patients are not misled about industry-sponsored trial results, said Catherine D. DeAngelis, MD, MPH, editor-in-chief of JAMA.
Imposing such a requirement before publication "assures that if there is any question about the veracity of statistical analysis, the editor can call the dean of the institution where the statistician is a faculty member and ask for a review of the statistical analysis," Dr. DeAngelis said. "If there is a problem, the article would be corrected or withdrawn and the dean would deal with the faculty member if needed."
JAMA and the Archives journals published by the American Medical Association are the only major journals to have adopted such a requirement, Dr. DeAngelis said. At this article's deadline, the International Committee of Medical Journal Editors planned to discuss, during an April meeting, whether to incorporate such a standard in its uniform requirements.
The ICMJE is composed of 12 of the leading journals, including NEJM, JAMA, Annals of Internal Medicine and Lancet. Hundreds of journals already follow the committee's uniform requirements for manuscripts, which lay out conventions on authorship credits, peer review and how to report research results.
NEJM spokeswoman Jennifer Zeis said the journal "employs four biomedical statisticians who review the appropriateness of statistical approach and analysis plan in all manuscripts before they are published" and often seeks clarifications, revisions or additional statistical tests from study authors. Zeis said editors published GSK's interim analysis because it "provided additional data on the cardiovascular toxicity of rosiglitazone."
By requiring stricter academic oversight of company-sponsored trials, medical journal editors might give drugmakers reason to think twice about misleading the public, said Steven E. Nissen, MD, lead author of the NEJM meta-analysis on Avandia's cardiac risk and chair of the Dept. of Cardiovascular Medicine at the Cleveland Clinic Foundation.
Dr. Nissen, who also authored a March 24/31 JAMA commentary on the Avandia case, said other steps should be taken, such as making the trial protocol and statistical analysis plan public.
GSK said in a statement that the JAMA editorial and commentary "unjustly challenge the motives of physician scientists seeking to inform diabetes treatment and patient safety, and mischaracterize the RECORD study on rosiglitazone." The pharmaceutical company said appropriate measures were taken to ensure appropriate oversight, and that the company's analyses were "independently verified at the London School of Hygiene and Tropical Medicine."
A March 18 online study published in BMJ found that of the 180 authors writing 202 articles reviewing Avandia's cardiac risks, those with financial ties to GSK or other drugmakers were much likelier to write favorably about the drug's safety.
Stricter medical journal standards can help prevent deception but will not be a panacea, said Roy Poses, MD, president of the Foundation for Integrity and Responsibility in Medicine and clinical associate professor of medicine at Warren Alpert Medical School of Brown University in Rhode Island.
"It means there would be a somewhat less biased expert at least looking at the analysis and making sure there was nothing really funny going on with it," Dr. Poses said. "Is it sufficient to prevent all the sorts of problems we've seen with commercially sponsored research? Clearly not."
Every step of the product development process, from study design to postmarket analysis of safety data, is vulnerable to manipulation, he said.