FDA panel recommendation creates mixed opinions about future of rosiglitazone

The FDA advisory committee recommendation that rosiglitazone stay on the U.S. market as a treatment option for patients with type 2 diabetes but with restrictions on use and revised labeling has many physicians questioning the future of the diabetes drug as well as how to rank medical evidence.

“This was not just a discussion about the safety of rosiglitazone, it was a discussion about which type of scientific evidence is more important,” Yehuda Handelsman, MD, medical director of the Metabolic Institute of America, Tarzana, Calif., told Endocrine Today.

The FDA advisory committee was presented with conflicting evidence about the cardiovascular safety of rosiglitazone from randomized controlled trials, such as the Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycemia in Diabetes (RECORD) study, to observational studies to meta-analyses. Much of the discussion focused on RECORD, with criticisms about its open-label study design, biased mishandling of CV events and overall conclusions.

“There remains great uncertainty that the results of the studies to date have been inconclusive from a scientific standpoint,” Richard G. Bach, MD, professor of medicine at Washington University School of Medicine, said in an interview. “The degree of inconsistency and discordance between studies raises great questions.”

Ranking medical evidence

While many were focused on the decision itself, the quality and ranking of scientific evidence was the star of the 2-day meeting. Fraught with statistical analysis, the meeting focused on the debate about which level of evidence has more weight: randomized clinical trials or observational studies and meta-analyses.

“I saw an attack on the scientific basis for medical practice,” Alan J. Garber, MD, PhD, chief medical editor of Endocrine Today, said of the meeting.

“This is of great concern because in the last 30 years, we have built up an edifice called evidence-based medicine,” he said. “Now we are seeing these randomized clinical trials — the gold standard for evidence-based medicine — being attacked because what was previously thought to be inferior levels of evidence are now being elevated, like database analyses.”

On July 30, 2007, the CV safety concerns associated with the drug were discussed before a joint advisory committee, which voted 20-3 that available data suggest rosiglitazone increases cardiac ischemic risk in type 2 diabetes. The committee voted 22-1, however, that the overall risk-benefit profile of rosiglitazone supports its continued marketing in the United States.

The safety of rosiglitazone has been questioned since 2007, when a meta-analysis by Steven E. Nissen, MD, and Kathy Wolski, MPH, of the Cleveland Clinic, showed that rosiglitazone increased the risk for MI by 43%. The findings were published in The New England Journal of Medicine.

The FDA panel was also presented with results of a nationwide, observational, retrospective study by David Graham, MD, FDA drug safety scientist, and colleagues that looked at outcomes of about 230,000 Medicare patients who were treated with rosiglitazone or pioglitazone. The study examined a combined endpoint of MI, stroke, heart failure or death and found that patients taking rosiglitazone had an 18% increased risk compared with those taking pioglitazone. The data were published in the Journal of the American Medical Association in June 2010.

“We have to deal with the fact that many of the clinical interventions we provide for patients are supported only on the basis of either observational data or retrospective analyses,” Bach said. He said it is difficult to appreciate the caveats and question marks that remain about the way studies such as RECORD, observational studies and meta-analyses were conducted.

Alan J. Garber

At the FDA meeting, Thomas Marciniak, MD, of the FDA division of CV products, criticized the study design, conduct and results of RECORD. He said the open-label trial was inadequately designed to determine the safety of rosiglitazone and had biased mishandling of CV events, suggesting a higher number of MIs than were reported.

None of the studies discussed by the FDA panel is without problems, Garber said.

“In looking at the data on rosiglitazone, the level of evidence is not to that of our highest gold or platinum standard,” Bach said. “In fact, even amongst lower levels of evidence, such as observational studies and meta-analyses, there are weaknesses within those studies available.”

Inconclusive data

At the end of the meeting, panel member Lamont G. Weide, MD, PhD, said he remained concerned about removing a drug from the market that provides benefit to some patients.

“I would hate to take away a drug without definitive evidence from the few patients who need it,” Weide, chief of diabetes and endocrinology at University of Missouri-Kansas City, said of his vote to keep rosiglitazone on the market.

Voting member Arthur J. Moss, MD, PhD, said he could “absolutely not vote for withdrawal” based on the inconclusive data presented to the panel.

“I wanted to send a message to the FDA because a proper trial [on the safety of rosiglitazone] should have been done going all the way back to 1993 — and that trial still needs to be done,” Moss, professor of medicine and cardiology at University of Rochester, N.Y., said at the meeting.

Some members voted for withdrawal and stricter restrictions on the rosiglitazone label because alternative therapies to rosiglitazone are available, such as pioglitazone (Actos, Takeda), the other thiazolidinedione on the U.S. market.

“We can still take care of diabetes,” if rosiglitazone is banned by the FDA in its final vote, said Abraham Thomas, MD, MPH, panel member and division head of endocrinology, diabetes, bone and mineral disorders at Henry Ford Hospital.

Future of TIDE

The Thiazolidinedione Intervention with Vitamin D Evaluation (TIDE) trial is designed to examine the comparative CV safety of rosiglitazone and pioglitazone in patients with type 2 diabetes. The postmarketing trial was mandated by the FDA after the 2007 safety debate.

When asked whether to recommend continuation of TIDE, the panel voted 19-11 that the trial should continue, with two abstentions and one non vote.

Hertzel C. Gerstein, MD, principal investigator of TIDE, said the randomized controlled TIDE trial data are essential to resolve the CV safety issue surrounding rosiglitazone.

“We can be seriously misled by the types of data that were presented and discussed [at this meeting],” he said during his presentation before the FDA.

Panel member, Moss, who voted that the trial should continue, said “a definitive study needs to be done and this is the first time that the FDA can demand [as] such.”

TIDE is an ongoing, event-driven, multicenter, international, randomized, double blind, placebo-controlled trial designed to evaluate the effects of rosiglitazone, pioglitazone and placebo added to background antidiabetes therapies in more than 11,000 patients with type 2 diabetes. The trial has two co-primary objectives: a non-inferiority comparison of rosiglitazone vs. placebo performed after a total follow-up of 4.5 years and a superiority comparison of thiazolidinediones (TZDs) vs. placebo performed about 1 year after the non-inferiority comparison. A secondary objective is to demonstrate noninferiority between rosiglitazone and pioglitazone on the primary composite of major adverse CV events. This is the only CV outcomes trial including a direct comparison between the two TZDs.

Study critics say the design is unethical and is designed to establish proof of harm. In a presentation before the panel, Graham said the trial must be stopped, adding that “the emphasis of TIDE is really shifted from its true CV purpose.”

Nissen said his calculations, based on current enrollment, estimate that the much-anticipated final answers from TIDE will not be available until 2020.

“Are we willing to wait that long for a drug with this hazard?” he asked the panel.

In late July, the FDA informed GlaxoSmithKline that TIDE trial has been placed on partial clinical hold. Under the partial clinical hold, no new patients may be enrolled into the trial until further notice from the FDA. Patients already enrolled in the trial will be allowed to continue to participate.

In a statement, the FDA instructed GlaxoSmithKline to update investigators, institutional review boards and ethics committees involved in the TIDE trial regarding new safety information presented at FDA advisory committee meeting, along with information regarding the deliberations and votes of that meeting.

According to a GlaxoSmithKline release, the company will work with the TIDE Steering Committee to send a summary of recent safety data and a summary of the FDA advisory committee meeting to all investigators and institutional review boards to ensure they have the latest information for patients.

“This pause in enrollment will give clinical trial investigators and patients time to learn about the data presented to the FDA advisory committee and the committee’s recommendations,” Ellen Strahlman, MD, chief medical officer of GlaxoSmithKline, said in the release.

Question marks remain

For some, the 2-day meeting provided no new insight since the first safety debate in 2007.

“I am not persuaded by the data to take any strong positive action. As many others have said, we do not know any more now than we did in 2007,” Garber said.

Others, such as Nissen, view the meeting as a “tremendous victory.”

“If you look at the vote, there were five options. The two most draconian options got the vast majority of the votes and, of the people who voted to keep it on with severe restrictions, most said they were leaning toward withdrawal but wanted to have some back-door mechanism for a few people to get the drug.

“I think that Avandia is now effectively gone,” Nissen said in an interview.

The clinicians Endocrine Today interviewed interviewed expressed concern about the lack of variety of experts on the joint Endocrinologic and Metabolic Drugs Advisory Committee and Drug Safety and Risk Management Advisory Committee.

Yehuda Handelsman

“As a clinician, I am somewhat amazed that these kinds of decisions are being made by so many epidemiologists and statisticians. At the end of the day, which drug to use should be a clinician’s choice,” Handelsman said.

On a similar note, regarding the attention on ranking of medical evidence, Handelsman said he does not deal with statistics as a clinician.

“I do not have three out of 1,000 patients that something happens to,” he said, adding that he has not observed the “dire effects” of rosiglitazone in patients he has treated for more than 10 years.

Even with the panel’s recommendation, the amount of scrutiny the drug has received may lead some diabetes patients who currently take rosiglitazone to want to stop taking the drug. The Endocrine Society, American Diabetes Association (ADA) and American Association of Clinical Endocrinologists (AACE) released a joint statement urging diabetes patients to not make any changes to their medication use without discussing their treatment with their physician.

“The worst outcome would be to not treat diabetes properly, thereby risking its complications,” Daniel Einhorn, MD, president of AACE, said in the release. “This unintended consequence has happened with past inquiries into diabetes medications, and we very much want to avoid it happening again.”

Following any decision about rosiglitazone from the FDA, The Endocrine Society, ADA and AACE will provide detailed information interpreting FDA action for both health care professionals and patients with diabetes. The American Heart Association issued a similar release about the drug.

“We, as clinicians, have to make judgments based on more than just the P value,” Bach said.

Final vote to come

As of press time, the data for the FDA vote was not announced. While the agency is not required to follow the recommendations of the advisory committee, it usually does.

Janet Woodcock, MD, director of the FDA Center for Drug Evaluation and Research, said the agency will evaluate the advisory committee proceedings and available regulatory options and come to a decision “as soon as possible.”

Richard G. Bach

“It is always nice to say, ‘Here is the answer.’ But that is not realistic in this circumstance,” Bach said.

“The medical community and FDA are learning a lot from this experience. The problem of surrogate endpoints in clinical trials, the type of data that need to be acquired and so on requires a level of scientific rigor and careful attention to detail in trial design and conduct that will probably improve the way we study and approve drugs in the future. It was a difficult path to get here, but the lessons learned will be important in the future,” Bach said. – by Katie Kalvaitis

How do you rank and use medical evidence?

For more information:

• Graham DJ. JAMA. 2010;304:doi:10:1001/jama.2010.920.
• Home PH. Lancet. 2009;373:2125-2135.
• Nissen SE. N Engl J Med. 2007;356:2457-2471.