David Brill
The events surrounding the withdrawal of rofecoxib shattered the widely held assumption that medical research serves to protect the public interest. In the first of a two-part series, Medical Tribune’s David Brill speaks exclusively to two of the central characters in the story of the key rofecoxib trial to find out what happened and what they have learned from the affair.
On December 29th 2005, the public’s fragile faith in medical research took another near-fatal blow. The withdrawal of the multi-billion dollar ‘blockbuster’ arthritis drug rofecoxib (Vioxx) the year before had been testing enough, but the latest revelation that safety data had been concealed in a key clinical trial of the drug was the final straw. The notions of objectivity and integrity that had lain at the very heart of research were exposed to an intense and damning scrutiny that continues to dominate both the academic and lay press to this day.
Rofecoxib, a selective cyclooxygenase-2 inhibitor, was voluntarily withdrawn in September 2004 by its manufacturers Merck & Co., which announced that ongoing research had shown definitively that the drug raised the risk of cardiovascular events. The news was shocking enough for the academic community – concerns had been repeatedly expressed but had largely gone unheeded by the manufacturers – but for the public the sheer magnitude of the decision was devastating. The drug had been on sale for more than 5 years in over 80 countries – leaving millions of people exposed to potentially avoidable heart attacks and strokes.
The affair left a profound legacy for the entire medical profession but for the members of the International Committee of Medical Journal Editors (ICMJE), it was a particularly bitter pill to swallow. Three years before the announcement they had warned in an editorial that the “precious objectivity” of medical research was under threat from the growing involvement of corporate sponsors in clinical trials which, they said, was allowing companies to increasingly dictate their own terms regarding the design, interpretation and publication of clinical trials. The gravity of the warning was clear but the timing of publication – September 10th 2001 – proved to be unfortunate when events in New York the following day pushed the discussion of medical ethics firmly to the bottom of the news agenda.
Objectivity and integrity are core values for all of the medical journals but for none is the weight of expectation greater than for the New England Journal of Medicine (NEJM). The journal – perceived by many researchers as the ultimate outlet in which to publish their work – has the highest impact factor of any medical journal and bestows upon its papers an inherent authority and prestige that only one or two other publications could claim to match. In 2007 alone the NEJM received some 6,000 original research papers, of which just 5 percent were published.
It was to the consternation of its editors, therefore, that the NEJM was to find itself sucked into the post-rofecoxib storm. As the legal process continued, attention turned to its November 2000 publication of the Vioxx Gastrointestinal Outcomes Research (VIGOR) trial, comparing the gastrointestinal toxicity of rofecoxib with naproxen in 8,076 rheumatoid arthritis patients. [23;343(21):1520-8] The plot had already begun to thicken in 2001 when an updated data set from the trial was presented to the US FDA, showing that three extra myocardial infarctions (MIs) in the rofecoxib group had occurred shortly after the cut-off point for the reporting of adverse events – results that, had they been included in the VIGOR analysis, would have increased the relative risk of MI from 4.25 to 5.00 for patients taking rofecoxib.
The NEJM editors were reassured that these were “late data” which had only been discovered after publication, but in November 2005 they learned that all was not as it had seemed. The extra MIs had, it emerged, been known about almost 5 months before the publication of VIGOR, giving the authors ample opportunity to include them in the 2000 paper. The NEJM announced the news by publishing an “Expression of Concern” on December 29th 2005, highlighting the “misleading conclusion” of the trial and questioning the integrity of its data. [29;353(26):2813-4]
For Professor Jeffrey Drazen, the journal’s editor-in-chief, the lesson of the story is clear. “We used to assume that everybody understood the rules. Now we no longer expect that people will play by them,” he explained on a recent visit to Singapore.
“When we discovered the extra events the authors said ‘well, we don’t know about this’. And in fact they didn’t.” It was hidden from them, says Drazen. “The violation of ethics, in my opinion, was not putting the material out in the open. So in retrospect, when it became clear that there was a discrepancy between what was in our pages and the data set sent to the FDA, we should have been more aggressive about querying the authors.”
The NEJM now takes a more direct interest in adverse event reporting when it comes to assessing the design of trials, according to Drazen. An unusual situation had arisen in VIGOR whereby the closing date for reporting GI bleeds – the study’s primary endpoint – was extended beyond the closure of the adverse event database. The trial continued for several weeks, but the additional cardiovascular events that occurred during this window were not included.
“They designed the trial in a sneaky way and never told anybody about it. Now we’re smart enough to ask about this. We want adverse event reporting up to the day of acceptance, not as of some prespecified day.”
Demanding accountability for the data is perhaps the most important lesson that the NEJM has learned from rofecoxib but now, Drazen says, they have “turned it around” by attaching a stronger sense of responsibility to research authorship. A study printed in the journal today may appear overtly similar to one published pre-rofecoxib, but a closer look at the methods section now reveals the key statement: “All authors had full access to the data and vouch for the accuracy and completeness of the data and the analyses.”
“The most important thing you have as an academic researcher is your reputation,” explains Drazen. “If you want to maintain your reputation you have to go to the company and say ‘I can’t put my name on this paper until I’m sure that the data I’m reporting are accurate and complete. I need to be sure that it’s my paper.’ Had we had such a statement from the authors we’d have been able to go back and tell them that they had committed perjury but we couldn’t do that.”
For Professor Richard Day, one of the academic authors of the VIGOR study, this is a lesson which had to be learned the hard way. The years that followed the NEJM’s Expression of Concern were difficult and distressing for him. Not only had his high hopes for rofecoxib been destroyed with the withdrawal of the drug, but now his name and those of his co-authors had been dragged down into the realms of scandal.
“It was a challenging time – there were lawyers engaged all over the place to try to put forward the positions of the various parties and it was all pretty unpleasant. I think we got pretty severely beaten up from all sides,” he says.
Day, a professor of clinical pharmacology at the University of New South Wales in Sydney, Australia, confirms Drazen’s assessment that the academic investigators were kept in the dark about the extra adverse events in VIGOR and feels that much of the “angst and uproar” could have been avoided had these data been discussed with them. He stands firmly behind the study itself, however, feeling that the issue of the extra MIs has been allowed to overshadow what was the largest trial of its kind and an exciting opportunity to tackle the huge problem of gastrointestinal bleeds.
“In my view it’s one of the best studies ever done in the field of rheumatoid arthritis. It has not been recognized for its quality because of this one issue. It’s an important issue, but it’s ruined it in lots of ways,” he says.
It was a strange and ultimately poor decision not to include the extra events, according to Day, but he notes that from a purist’s perspective it was technically correct since the prespecified trial protocol had demanded that they be excluded. The contentious extension of the GI events window was at the request of the academic authors, he adds, noting that it was an endpoint-driven study and that the researchers had simply wanted to accrue as much data as possible about the benefits of rofecoxib. Why the adverse events window was not also extended accordingly he does not know.
The complete story of rofecoxib and VIGOR may never be revealed but for now the scar on the public consciousness endures. Perhaps surprisingly, however, Day feels that despite the “collateral damage” suffered along the way, the long-term legacy of the affair may ultimately help to rebuild the integrity which it once destroyed.
“Clearly the light’s being shone more and more on this interaction between the clinical world and drug companies and that’s a good thing,” he concludes. “I’m sure there’ll be more problems but I think with every one of them we learn, tighten up and try to head them off so that in future we can believe what we see, read and hear, and feel that it represents a proper treatment of data and opinions which is independent and addresses conflicts properly.”
An edited version of this article appeared in Medical Tribune.
Dr. Jeffrey Drazen was in Singapore as part of the National University Health System’s Distinguished Editors Series.
Next month: David Brill looks at how the integrity of medical research can be restored through the lessons learned from rofecoxib.
No comments:
Post a Comment