Final results from the drug company’s own clinical trial, known as the RECORD trial, were published last year. While confirming an increase in heart failure with the drug, it showed no overall hike in cardiovascular disease or deaths.
“The totality of the evidence, from RECORD and other long-term studies, demonstrate no increase in the overall risk of these untoward outcomes,” the drugmaker emphasized in a report to the FDA in advance of the panel meeting.
Theoretically, a clinical trial—in which patients are assigned at random to a treatment or comparison group and followed—could be far more definitive than meta-analyses and studies plumbing databases of patient records. But it’s around the design and conduct of its trial that Glaxo has received perhaps the most damning criticism.
In preparation for this week’s advisory committee meeting to decide Avandia’s fate, the FDA conducted an unusually rigorous review of RECORD’s raw data. One reviewer, FDA scientist Thomas A. Marciniak, M.D., issued a scathing report to the committee. His central concern was that investigators and certain Glaxo staffers knew which patients in the trial were getting Avandia and which were not—and, whether deliberately or by unconscious bias, might have misclassified or dismissed cases unfavorable to the drug. Only 15 additional heart attacks among Avandia users would have changed the results so that they would have pointed to a statistically significant increase in risk.
“We count 25 patients with a potential MI [heart attack] and questionable handling favoring rosiglitazone,” Marciniak’s report said.
Last week, the New England Journal of Medicine published an editorial saying RECORD’s commercial sponsor “went around” the study’s data and safety monitoring board—the group charged with guarding patient safety—in order to rush out interim findings. Such “manipulations,” the editorialists ventured, call into question “the integrity of the whole clinical-trial enterprise.” The editors called for a new approach in which safety monitoring committees are established not by the trial sponsor, whether public or private, but by an independent public body.
Sen. Chuck Grassley (R-Iowa), long a gadfly to the pharmaceutical industry, has suggested similar reforms—chiefly the establishment of an independent body to monitor drug safety—at the FDA, which has been getting heat from all sides in the Avandia debate, and whose own staff has been sharply divided on the issue.
FDA employees who study a drug’s side effects after it has been on the market and used by the general population “should be able to express their opinions in writing and independently without fear of retaliation,” Grassley told a House subcommittee in April. “Instead, the FDA physicians and scientists committed to post-market monitoring of drugs have sometimes been suppressed. In the case of Avandia, it appears that they have been ignored.”
But by this week the views of Avandia critics both inside and outside the FDA had been published widely and received a great deal of attention.
Before the FDA advisers convened this week, David Juurlink, M.D., a University of Toronto specialist in the epidemiology of drug effects, predicted that whatever the panel decided, the FDA would have some tough questions to answer. If it doesn’t pull the drug from the market, he said, that would raise doubts about whether the agency is capable of protecting the public from unacceptable risks. “If they decide to remove the drug from the market,” Juurlink said, “the question they’ll have to answer to the public is, why did it take so long?”
Katharine Greider lives in New York and writes about health and medicine.