Drazen Discusses Rise and Fall of Vioxx to Launch UCSF Lecture Series
by Jeff Norris
Jeffrey M. Drazen, editor-in-chief of the New England Journal of Medicine (NEJM), discussed the rise and fall of Vioxx and similar painkilling drugs at UCSF's Cole Hall Wednesday.
The key message in Drazen's presentation, which inaugurated the Chancellor's Health Policy Lecture Series: "You have to learn from the data - don't make the data fit your story."
Drazen's case study showed how the withholding and manipulation of key information about Vioxx slowed broader awareness of its dangerous side effects, which included deaths from heart attacks and strokes.
Similar side effects were later found with other drugs in the same class - the COX-2 inhibitors.
Both Vioxx and Bextra have been removed from the market. A third related drug, Celebrex, remains, but carries a "black box" warning for heart risks.
The NEJM published a key Vioxx study in November 2000. "We were publishing a safety paper, where it turns out we got only one of the safety outcomes in its entirety," Drazen said.
The Promise of Vioxx
COX-2 inhibitors were first developed to be painkillers. The hope was that COX-2 inhibitors would have fewer gastrointestinal side effects - bleeding, for instance - than other, nonsteroidal anti-inflammatory drugs (NSAIDs). Researchers and drug industry officials reasoned that people who took painkilling drugs regularly, such as arthritis sufferers, could benefit the most from COX-2 inhibitors. Early clinical trials quickly led to marketing approval for Vioxx and Celebrex before the new millennium. Additional trials were begun, sponsored by the National Institutes of Health and the drug manufacturers. A Trail of Deception
Members of Drazen's NEJM staff learned of a key deception in the publication only later, when they became involved in legal proceedings - against their will - in a Vioxx lawsuit. The journal editors were trying to protect - and succeeded, despite being dragged into court, Drazen said, in protecting - the privacy of peer review. In the course of legal proceedings, an editor was shown a company memo indicating that data on adverse cardiovascular side effects were known by at least two study authors before clinical trial results were published in the NEJM. However, those data were never submitted for publication. In addition, the trial's database for collecting information on cardiovascular side effects was closed prematurely, while the collection of information on gastrointestinal side effects continued. This was an important omission because heart attacks increased significantly after patients had been in the trial for several months. Arthritis patients in the published trial were assigned to take either Vioxx or naproxen. If all the data had been published, Drazen said, there would have been 21 fewer instances of serious bleeding reported for Vioxx, but 27 more heart attacks, strokes or peripheral vascular events. Lessons Learned
"We learned a lot," Drazen said. "First of all, we want to make sure that a trial is reported faithfully." The NEJM and several other leading medical journals now require that complete clinical trial protocols be registered and available online at www.clinicaltrials.gov. In addition, Drazen said, "We always ask . . . is this the entire adverse event database?" While precautions are helpful, Drazen asserted, "The only way to move things forward is for us to maintain a culture of honesty." Drazen advises study authors to "make sure the data submitted reflect the data available when the manuscript is finalized. "More importantly, this was an important piece of translational research. It turns out the theory was wrong. What this taught us is that COX-2 also has an anti-thrombotic effect." Failure to disclose adverse events "treats the people who participated in the trials with great disrespect," Drazen said. "The people who had the heart attacks, the GI bleeds . . . thought their data were going to be part of the public database, part of the public decisionmaking process, and they were excluded." The COX-2 story is not over, and some clinicians and researchers are rethinking assumptions about the safety of NSAIDs, as well. In addition, COX-2 inhibitors had shown preliminary evidence in studies for preventing colon cancer recurrence in patients who had earlier had polyps removed. Drug administration was stopped early in those studies due to concerns about cardiovascular risk. In fact, researchers who study how cancers arise are accumulating data that implicate COX-2 in several cancers. It is certain that researchers will want to continue to tease apart the different roles of COX enzymes, to learn more about health and disease - and to ensure that the next generation of drugs is better.
COX-2 inhibitors were first developed to be painkillers. The hope was that COX-2 inhibitors would have fewer gastrointestinal side effects - bleeding, for instance - than other, nonsteroidal anti-inflammatory drugs (NSAIDs). Researchers and drug industry officials reasoned that people who took painkilling drugs regularly, such as arthritis sufferers, could benefit the most from COX-2 inhibitors. Early clinical trials quickly led to marketing approval for Vioxx and Celebrex before the new millennium. Additional trials were begun, sponsored by the National Institutes of Health and the drug manufacturers. A Trail of Deception
Members of Drazen's NEJM staff learned of a key deception in the publication only later, when they became involved in legal proceedings - against their will - in a Vioxx lawsuit. The journal editors were trying to protect - and succeeded, despite being dragged into court, Drazen said, in protecting - the privacy of peer review. In the course of legal proceedings, an editor was shown a company memo indicating that data on adverse cardiovascular side effects were known by at least two study authors before clinical trial results were published in the NEJM. However, those data were never submitted for publication. In addition, the trial's database for collecting information on cardiovascular side effects was closed prematurely, while the collection of information on gastrointestinal side effects continued. This was an important omission because heart attacks increased significantly after patients had been in the trial for several months. Arthritis patients in the published trial were assigned to take either Vioxx or naproxen. If all the data had been published, Drazen said, there would have been 21 fewer instances of serious bleeding reported for Vioxx, but 27 more heart attacks, strokes or peripheral vascular events. Lessons Learned
"We learned a lot," Drazen said. "First of all, we want to make sure that a trial is reported faithfully." The NEJM and several other leading medical journals now require that complete clinical trial protocols be registered and available online at www.clinicaltrials.gov. In addition, Drazen said, "We always ask . . . is this the entire adverse event database?" While precautions are helpful, Drazen asserted, "The only way to move things forward is for us to maintain a culture of honesty." Drazen advises study authors to "make sure the data submitted reflect the data available when the manuscript is finalized. "More importantly, this was an important piece of translational research. It turns out the theory was wrong. What this taught us is that COX-2 also has an anti-thrombotic effect." Failure to disclose adverse events "treats the people who participated in the trials with great disrespect," Drazen said. "The people who had the heart attacks, the GI bleeds . . . thought their data were going to be part of the public database, part of the public decisionmaking process, and they were excluded." The COX-2 story is not over, and some clinicians and researchers are rethinking assumptions about the safety of NSAIDs, as well. In addition, COX-2 inhibitors had shown preliminary evidence in studies for preventing colon cancer recurrence in patients who had earlier had polyps removed. Drug administration was stopped early in those studies due to concerns about cardiovascular risk. In fact, researchers who study how cancers arise are accumulating data that implicate COX-2 in several cancers. It is certain that researchers will want to continue to tease apart the different roles of COX enzymes, to learn more about health and disease - and to ensure that the next generation of drugs is better.