Keep Rosiglitazone on the Market, FDA Panel Says
July 14, 2010
GAITHERSBURG, Md. -- An advisory panel voted 20-12 to recommend that the FDA allow rosiglitazone (Avandia) to stay on the market, but most panelists want to see the controversial diabetes drug carry tougher warnings on its label.
Wednesday's vote marks the second time an FDA advisory panel has essentially endorsed rosiglitazone. In 2007, a panel voted that while the drug appears to carry a higher risk of cardiovascular events, it should continued to be marketed.
Since then, concern over the health risks of the drug has continued to grow, so the FDA convened the current panel of diabetes and statistical experts to revisit the issue. The 33-member panel was comprised of the Endocrinology and Metabolic Drugs and the Drug Safety and Risk Management Advisory Committees.
And once again the committee confirmed that it thinks there is a safety risk with the drug, but that rosiglitazone should remain an option for patients with type 2 diabetes.
The FDA does not have to follow the advice of its advisory committees, but it usually does.
Not Just an Up or Down Vote
Of the 20 panel members who voted that rosiglitazone should stay on the market, only three voted that there be no changes in the product labeling; 17 voted to recommend that GlaxoSmithKline be allowed to continue marketing the drug -- but with several caveats.
For instance, seven panelists said the company should be required to revise its current label to include contraindications for certain patient populations -- although they didn't clarify which patients might fall into the high-risk subgroup. They also favored adding a warning to the label to recommend rosiglitazone's use only in patients who have already tried another antidiabetic agent.
The 2007 advisory committee voted that a black box warning be added stating the drug should not be used by patients requiring insulin, those with a history of coronary heart disease, those with congestive heart failure, and those who are long-time users of nitrates.
According to Steven Nissen, MD, of the Cleveland Clinic -- whose meta-analysis sparked the firestorm that culminated in this week's advisory committee meeting -- if the FDA adopts the tougher restrictions, 95% of rosiglitazone use will disappear. (Use of rosiglitazone also declined following the 2007 advisory committee meeting).
"Effectively, the drug is gone," Nissen told reporters after the meeting. "Obviously, I wanted a clearer decision," he added. "I wanted withdrawal."
Ten panelists voted that the FDA should go further than requiring a tougher warning -- to also restrict prescribing to physicians who have special education in rosiglitazone.
"I do not think the preponderance of evidence was strong enough to suggest rosiglitazone be taken off the market," said Allison Goldfine, MD, assistant director of clinical research at the Joslin Diabetes Center in Boston. "But it should be restricted to physicians who have better education who could better educate patients."
The 12 panelists who voted that the the drug be yanked from the market argued that there isn't a need for rosiglitazone when patients can take pioglitazone (Actos), which works as well as rosiglitazone but hasn't been found to present excess ischemic events.
"There clearly are adverse signals [with rosiglitazone]," said Marvin A. Konstam, MD, a cardiologist at Tufts-New England Medical Center in Boston, who voted to pull the drug. "I just don't see any substantive benefits for rosiglitazone versus pioglitazone."
Votes on Safety Went Against Rosiglitazone
The panel agreed with Konstam that rosiglitazone is riskier than non-thiazolidinedione (TZD) diabetes drugs and pioglitazone. Members voted 18-6 (with nine members voting that they were unable to make a decision) that there is sufficient evidence that rosiglitazone increases the risk for ischemic cardiovascular events compared with non-TZD drugs.
Panel members then voted 21-4 (with eight member unable to decide) that rosiglitazone is less safe than pioglitazone.
"I do think there are strong safety concerns for rosiglitazone compared with pioglitazone," said Thomas Fleming, PhD, a biostatician at the University of Washington in Seattle.
But, perhaps forecasting the later vote on whether the drug should stay on the market, the panel did not think there was enough data to show that rosiglitazone increases the risk of mortality.
The committee reached its decision after two days of presentations and discussions on GlaxoSmithKline's trial data. While panelists didn't agree on the fate of rosiglitazone, they all agreed the data was exceedingly difficult to interpret.
Heavy Focus on RECORD Trial Data
The panel spent hours poring over several interpretations of GlaxoSmithKlines's controversial RECORD trial, which the company has leaned on as one of its main defenses on the cardiovascular safety of rosiglitazone.
It was "a massive amount of very confusing material," said Janet Woodcock, MD, the director of the FDA's Center for Drug Evaluation and Research.
Moreover, the company's studies were far from perfect.
"None of the studies are without mild, moderate, or notable flaws," said committee chairman Kenneth Burman, an endocrinologist at the Washington Hospital Center.
RECORD, an open-label trial, was requested by and designed in conjunction with European regulators. It showed no elevation in the combined endpoint that includes death, myocardial infarction, and stroke compared with active control (hazard ratio 0.93, 95% confidence interval 0.74 to 1.15, P=0.5).
In briefing documents released in advance of the meeting, an FDA staff reviewer -- Thomas Marciniak, MD -- slammed the trial for its design, its findings, and GlaxoSmithKline's interpretation of the findings. According to Marciniak, RECORD was "biased in a manner that consistently favored not finding a cardiovascular effect if it was present."
Marciniak's report, and his passionate presentations over the past two days, convinced a number of panel members.
"I don't put a lot of weight on the RECORD trial, frankly," said Michael Proschan, PhD, a statistician with the National Institutes of Health. "I don't trust it."
Despite the overall positive vote for rosiglitazone, there weren't many good things said about RECORD over the two-day meeting.
TIDE Trial Should Continue, Panel Says
Another trial that came under fire -- but was ultimately exonerated -- during the meeting was the TIDE trial, which is an ongoing head-to-head outcomes trial comparing rosiglitazone to pioglitazone. The FDA ordered GlaxoSmithKline to conduct the trial after its 2007 advisory committee hearing.
The panel voted 19 to 11, with one member abstaining (and one ducking out early), to keep TIDE open.
Marciniak had argued that TIDE -- which was created to examine whether the safety signal with rosiglitazone is real -- should be shut down.
He called it "unethical and exploitative," to enroll patients in a trial that was created to detect harm. He also said the informed consent form for the trial obscures the risks of rosiglitazone while focusing too heavily on a more innocuous portion of the study -- a piece designed to detect vitamin D's affect on cancer.
But the majority of panelists said that if rosiglitazone stays on the market, then a head-to-head trial is necessary in order to keep an eye on potential safety signals. Several members favored retooling the trial's consent form.
Woodcock said the agency would make a decision "as soon as possible."
GlaxoSmithKline said its trials prove that rosiglitazone does not increase the overall risk of heart attack, stroke, or death.
"Following today's recommendations, we will, of course, continue to work with the FDA in the best interest of diabetes patients who face this chronic and serious disease," said Ellen Strahlman, MD, the company's chief medical officer in a prepared statement.
"Patients taking Avandia should speak with their physician about their treatment and any questions they may have regarding the safety of the medicine," Strahlman said.
Rosiglitazone was approved by the FDA in 1999 to treat type 2 diabetes. Controversy over the drug began three years ago when Nissen's 2007 meta-analysis of data from 42 clinical trials found a 43% increase in relative risk of myocardial infarction among patients treated with rosiglitazone.