乌鸦传媒

WASHINGTON -- Rosiglitazone (Avandia) should be pulled from the market because of the "serious" cardiovascular risk the diabetes drug poses, which exceeds that of competitor pioglitazone (Actos), FDA staff reviewers again recommended.

Rosiglitazone is no more effective at improving glycemic control in type 2 diabetes than pioglitazone and provides no known unique health benefits, and while both drugs increase the risk of congestive heart failure, rosiglitazone's effect is "substantially greater," FDA reviewers wrote in a massive briefing document released Friday.

"Our review of the literature finds no evidence of a unique or meaningful health benefit from rosiglitazone that is not also provided by pioglitazone and there is substantial evidence that rosiglitazone increases cardiovascular risk compared with pioglitazone," wrote David Graham, MD, and Kate Gelperin, MD, both epidemiologists with the FDA.

Graham and Gelperin had reached the same conclusion in a report requested by the Senate Finance Committee in February.

Whether a joint panel of outside medical experts will agree with their assessment remains to be seen. The panel will meet July 13 and 14 to make their recommendations about rosiglitazone, among them whether the drug should be pulled from the market.

The FDA does not have to follow the advice of its advisory committees, but it usually does.

Advice Redux

Next week's meeting will be the second time an advisory panel was tasked with guiding the FDA on what to do about rosiglitazone.

In 2007, a panel voted 20 to 3 that the drug increased cardiovascular risk, but then 22 to 1 that the benefits outweighed the risks.

In their current review, the FDA staff blasted the logic in the 2007 decision and said the panel failed to prove what, exactly, the benefits of rosiglitazone were. Since then, the benefits of the drug are still just as unclear, they said.

"In the three years since that advisory committee meeting, the Office of New Drugs has not answered the question of what are the unique health benefits of rosiglitazone that have justified its continued marketing," the reviewers wrote.

They criticized the FDA mindset that "rosiglitazone is safe and remains so until 'definitive proof' forces reconsideration."

"In our opinion, this approach is inconsistent with public health policy that places patient safety first," Graham and Gelperin wrote.

On the RECORD

Since much of the rosiglitazone debate has been fueled by a seemingly endless array of dueling meta-analyses and observational data -- famously initiated by a 2007 meta-analysis by Steven E. Nissen, MD, of the Cleveland Clinic -- many observers have expressed hope that the FDA would gain access to new, unpublished data rather than simply rehashing what is already in the public space.

And indeed, the briefing documents released today included an analysis of the controversial RECORD trial, which GlaxoSmithKline has leaned on as one of its main defenses on the cardiovascular safety of rosiglitazone.

That open-label trial, which was requested by and designed in conjunction with European regulators, 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).

The trial also met its primary outcome of noninferiority in the composite of cardiovascular death and cardiovascular hospitalizations with a hazard ratio of 0.99 for rosiglitazone versus control (95% CI 0.85 to 1.15).

When the company sent the FDA its completed results of RECORD, it also submitted a proposed change to the labeling to remove mention of risk of myocardial ischemia.

But staff reviewers led by Thomas Marciniak, MD, of the cardiovascular and renal products department, scathingly criticized the trial on design, conduct, and results.

"Our review of RECORD leads us to conclude that its design and execution were biased in a manner that consistently favored not finding a cardiovascular effect if it was present," they wrote in the document. "In our view, it provides no credible evidence of rosiglitazone's cardiovascular safety."

For example, most errors in handling case report forms on cardiovascular events appeared to be in favor of rosiglitazone.

In the FDA's review of a random sample of these reports, all four cases of report mishandling, all eight cases of failure to refer events for adjudication, and nine of 14 missed events were in patients given rosiglitazone.

"While these numbers may seem small compared with the size of the trial, note that about 15 more MIs in the rosiglitazone arms are needed to change the GlaxoSmithKline MI results to a relative risk of 1.4 and a P-value of 0.042," the reviewers wrote.

Furthermore, they noted, the bias of the study design was toward the null, the primary endpoint was inappropriate (major adverse cardiovascular events would have been better), and there were substantial errors in both endpoint determinations and the censoring date for patients without endpoints.

Also, the final results of the trial were reported in an intent-to-treat manner, ignoring the extensive crossovers that arose as patients and physicians -- who knew which drug they were on because of the open-label design -- grew concerned over the media firestorm after the initial Nissen meta-analysis.

The FDA reviewers believe that intent-to-treat was not a valid approach and instead they censored the data once good follow-up fell below 90% with 80% of patients still on treatment as randomized. That point occurred at 3.4 years rather than the five to seven years used by GlaxoSmithKline.

In the FDA reviewers' analysis, the hazard ratios with rosiglitazone versus placebo were:

• 1.42 for MI (95% lower confidence limit 0.93, 95% upper confidence limit 2.16)
• 0.96 for strokes (95% LCL 0.60, 95% UCL 1.54)
• 0.95 for cardiovascular death (95% LCL 0.57, 95% UCL 1.60)
• 1.13 for major adverse cardiovascular events (95% LCL 0.85, 95% UCL 1.50)

These data, added to additional evidence from the trial, suggest that the elevated MI risk is real, the reviewers concluded.

"One does not have to be a mathematician or to perform the calculations to estimate that incorporating our RECORD results into a meta-analysis similar to Nissen's will produce a risk for MIs that is highly statistically significant," they wrote in the briefing documents.

FDA Staff Blast Postmarketing Study

Following the 2007 FDA advisory committee meeting, FDA required GlaxoSmithKline to conduct a head-to-head outcomes trial comparing rosiglitazone to pioglitazone. That trial -- called TIDE -- was hampered by sluggish enrollment of patients -- possibly because of safety concerns.

Graham and Gelperin called TIDE "unethical and exploitative," arguing that GlaxoSmithKline never proved rosiglitazone is equivalent to pioglitazone.

"That is, there must be equal evidence favoring both therapies. But that is not the case here because no one has argued that rosiglitazone is safer than or preferable to pioglitazone," FDA reviewers wrote. "In our view, the TIDE trial is unethical because it subjects human beings to unnecessary risks without any possibility of a meaningful, unique health benefit from rosiglitazone."

They also said the informed consent document for TIDE was misleading.

"Given the weight of evidence regarding cardiovascular risks with rosiglitazone, we believe that TIDE is an unethical study and that it was unethical before it was started," Graham and Gelperin wrote.

At next week's meeting, the Endocrinologic and Metabolic Drugs and the Drug Safety and Risk Management Advisory Committees will also consider the TIDE study, and may vote to halt it, or modify it in some way.

In the company's briefing documents, GlaxoSmithKline said it remains committed to finishing the TIDE trial.

The FDA hopes to avoid some of the sticky ethical questions washed up by the TIDE trial. It asked the Institute of Medicine (IOM) to come up with a framework for obtaining informed consent when comparing two drugs, one of which is believed to carry an increased risk.

On Friday, the IOM published a number of recommendations for the FDA when planning a postmarketing, randomized safety trial, including that the agency order such trials only when a responsible policy decision cannot be made without the new evidence, and that the informed consent process continues throughout the trial to include relevant findings that might influence a "participant's willingness to accept the risks associated with the trial.

Panelists at next week's meeting will likely place great weight on an analysis based on data from the Centers for Medicare and Medicaid Services which was published online two weeks ago.

That observational retrospective cohort study of 227,571 Medicare patients 65 and older from July 2006 to June 2009 found that rosiglitazone was associated with a significantly elevated risk of all other endpoints including stroke, heart failure, death, and a composite with MI of all four endpoints.

Company Defends the Drug, Europe Investigates

A spokeswoman for GlaxoSmithKline said the company stands behind the safety and efficacy of rosiglitazone.

"Avandia has a positive benefit-risk profile. When you look at the science and the totality of the data -– especially the most rigorous scientific studies -– Avandia has been shown to be a safe and effective diabetes medicine when used appropriately according to its labeling," said Mary Anne Rhyne, director of U.S. Media relations for GlaxoSmithKline. "Avandia is an important medicine for the treatment of type 2 diabetes, who often need two or three medicines to help maintain their blood sugar levels."

Following the lead of the U.S. agency, Europe's version of the FDA will also launch a review into rosiglitazone.

The European Medicines Agency announced Friday that its Committee for Medicinal Products for Humans will review the safety of rosiglitazone when it meets July 19 to July 22.

The committee may decide to suspend marketing for rosiglitazone or remove it from the market.

Comment