As Rosiglitazone Review Looms, FDA Documents Slam RECORD and Find Significant Risks for the Drug

Shelley Wood

July 09, 2010

Updated July 9, 2010 (Gaithersburg, Maryland) — The storm brewing over rosiglitazone (Avandia, GlaxoSmithKline) is finally making its way to Maryland, poised to touch down Tuesday when an FDA joint advisory committee meets to assess the cardiovascular safety of rosiglitazone. Briefing material from the FDA was posted on an FDA website Friday morning, as was that from the sponsor.

The bulk of the FDA material overwhelmingly points to an increased risk of adverse cardiovascular events with rosiglitazone over pioglitazone (Actos, Takeda Pharmaceuticals) and other diabetes drugs--with one FDA review stating bluntly that the drug should be taken off the market and that the "unethical" Thiazolidinedione Intervention with Vitamin D Evaluation(TIDE) trial be halted. By contrast, the GlaxoSmithKline documents conclude that the "totality of the data . . . continue to support the overall positive risk profile of rosiglitazone as an important medicine for type 2 diabetes" and repeatedly underscore the need for a randomized controlled trial.

The nearly 1000 pages of documents now posted online will no doubt add torque to the tornado. Journalists now brooding over those materials were invited to a unique premeeting press conference yesterday and heard from agency staffers that the two-day meeting is generating an unprecedented media interest from reporters the world over.

Two Long and Action-Packed Days

The two-day meeting, starting at 7:45 AM Tuesday, will follow a tight schedule, starting with an introduction and background information from the FDA's Center for Drug Evaluation and Research. That will be followed by a series of presentations from GlaxoSmithKline, including its own 52-study meta-analysis, which, according to the briefing documents, shows no significantly increased risk of myocardial ischemic events with rosiglitazone vs control. The sponsor will also present a review of observational studies that suggest mixed results for MI vs other diabetes drugs but no increased risk of CV mortality or all-cause mortality. In "the majority" of studies comparing rosiglitazone and pioglitazone, the sponsor's review concludes there is no increased risk of MI with rosiglitazone. Finally, Dr Philip Home (Newcastle University, UK), chair of the RECORD steering committee, will present results from that trial, which according to the briefing documents demonstrate the noninferiority of rosiglitazone compared with metformin and sulfonylurea on the combined end point of CV death and CV hospitalization.

Familiar Names

In the second half of the morning, some of the most familiar names in the rosiglitazone saga will take the floor. Dr Steven Nissen (Cleveland Clinic, OH), coauthor on the notorious meta-analysis that first blew the rosiglitazone safety debate wide open, will lead off with a 30-minute "critical overview." He'll be followed by two additional analyses of the RECORD trial, conducted by FDA physicians.

Following a truncated lunch break (already imperiled by an ambitious schedule), FDA physicians will present a medical and statistical review of RECORD, in turn followed by a report from the FDA's Good Clinical Practice Branch of the agency's inspection of the RECORD trial.

The briefing documents offer a peek at what these presentations will look like. In one, Dr Thomas Marciniak, from the FDA's Division of Cardiovascular and Renal Products, concludes that RECORD (touted by GlaxoSmithKline as supporting rosiglitazone's safety) "was inadequately designed and conducted to provide any reassurance about the CV safety of rosiglitazone." He concludes, "The results do confirm and extend the recognized concerns regarding increased heart failure and HF deaths with rosiglitazone. They also suggest that rosiglitazone increases the risk for MI, although the confidence intervals for estimated MI hazard ratios are wide and include no risk, while biases in the study suggest that the true risk could be higher."

In another, however, Dr Ellis Unger, from the same FDA division, concludes that while the primary end-point results (time to first cardiovascular death or hospitalization) are "not entirely conclusive," the all-cause mortality findings appear to favor rosiglitazone. The results "can and will be debated," Unger concludes, "but aside from the known risk of heart failure, the study does not appear to demonstrate harm."

The rest of the first day will be devoted to presentations by FDA physicians already well-known to those following the rosiglitazone story: Dr Kate Gelperin will present an updated overview of observational studies, Dr David Graham will present the results of his Journal of the American Medical Association Medicare analysis as well as a "personal perspective on the TIDE trial"--a study he has long believed should be halted.

In the briefing documents, Gelperin and Graham are highly critical of RECORD, which they call "unreliable and uninterpretable," and of TIDE, calling any head-to-head trial of rosiglitazone vs pioglitazone "unethical and exploitative."

Their stinging conclusion reads: "The risks of rosiglitazone use are serious and exceed those for pioglitazone. Rosiglitazone confers no unique and medically important benefit that distinguishes it from pioglitazone. The risks of rosiglitazone use exceed its benefits compared with pioglitazone. Rosiglitazone should be removed from the market."

Rounding out the day are further presentations by FDA scientists, including meta-analyses conducted in-house, plus presentations from non-FDA scientists, who will review rosiglitazone data from the BARI-2D and VADT trials. Time for questions from the committee members are scheduled throughout the day.

Day Two Devoted Mostly to Difficult Questions

Day two will open with a presentation on the TIDE trial by one of the study's primary investigators, Dr Hertzel Gerstein (McMaster University, Hamilton, ON), who recently argued in an American Heart Journal editorial that the study should be allowed to continue [1]. A National Institutes of Health statistician will speak about the strengths and limitations of different types of studies, followed by a report from the Institute of Medicine (IoM), to be presented by Dr Ruth Faden. This report, made available on the IoM website, forms part of the report requested of the IoM by FDA commissioner Dr Margaret Hamburg and deals specifically with the ethics and science surrounding trials aimed at establishing safety of already-approved drugs in the face of a signal of harm.

Only after lunch will committee members get down to the business of discussing the issues and questions at the heart of the rosiglitazone debate. These include, briefly:

  • The strengths/weaknesses of the ischemic CV risk data (MI, stroke) on rosiglitazone vs pioglitazone and vs other diabetes drugs.

  • Whether or not rosiglitazone increases the risk of ischemic CV events in patients with type 2 diabetes relative to non-TZD antidiabetic agents and relative to pioglitazone.

  • The strengths and weaknesses of the mortality risk data on rosiglitazone vs pioglitazone and vs other diabetes drugs.

  • Whether or not rosiglitazone increases the risk of mortality in patients with type 2 diabetes relative to non-TZD antidiabetic agents and relative to pioglitazone.

  • The relative benefit/risk profile of rosiglitazone as compared with other diabetes drugs.

  • A choice of five possible actions the FDA should take, ranging from allowing continued marketing with the warning labeling removed (for ischemic CV events) to withdrawing the drug from the market.

  • A determination of what to do about the TIDE trial.


The FDA's Dr Janet Woodcock, director of the Center for Drug Evaluation and Research, emphasized that while some material predates the last FDA rosiglitazone review in 2007, all of the material has been updated, and new information is available. "In other words, there are additional data of different types that may bear on [the safety] question."

Stay tuned for breaking news on both days of the FDA meeting next week, or follow us for updates throughout each day on Twitter .

Rosiglitazone: A Timeline

June 28–29, 2010 Early release papers in the Journal of the American Medical Association, Archives of Internal Medicine, including the Nissen-Wolski "reanalysis," suggest higher risk with rosiglitazone; FDA reviewer says his data were leaked as an "act of sabotage." One day later, BARI 2D data find no increased risk with rosiglitazone.
June 10, 2010 FDA reviewer's data are leaked to the media.
April 28, 2010 US Congressional committee that holds jurisdiction over FDA holds rosiglitazone review to discuss Senate Finance Committee report.
March 30, 2010 FDA responds to Senate Finance Committee report on rosiglitazone safety; announces plan to reassess continuation of TIDE.
February 24, 2010 GlaxoSmithKline rebuts Finance Committee report.
February 22, 2010 Senate Finance Committee releases report concluding "serious health risks" with Avandia; criticizes FDA handling of concerns.
June 5, 2009 RECORD trial presented at ADA, published in the Lancet: concludes no "overall" risk of rosiglitazone
January 14, 2009 Emails from GlaxoSmithKline indicate the company had concerns about the CVD risk of rosiglitazone in 2005–2006.
December 17, 2008 FDA announces new recommendations that all diabetes drugs have data proving CV safety.
July 2, 2008 FDA's Endocrinologic and Metabolic Drugs advisory committee recommends CV risks of diabetes drugs must be specifically studied.
March 25, 2008 FDA letter reprimands GlaxoSmithKline for failing to report all postapproval rosiglitazone data.
January 30, 2008 Nature reports that New England Journal of Medicine peer reviewer Dr Steven Haffner leaked Nissen-Wolski meta-analysis to GlaxoSmithKline weeks ahead of publication. Haffner later calls his decision to fax the article to GlaxoSmithKline a "mystery" and says he "wasn't feeling well" when he did it.
November 20, 2007 Senate Finance Committee releases report summarizing GlaxoSmithKline's efforts to "intimidate" and "silence" concerns of Dr John Buse about rosiglitazone, dating back to 1999.
November 14, 2007 FDA announces decision to allow rosiglitazone to stay on the market, but with beefed up warning labeling; announces request for large, long-term trial of rosiglitazone vs pioglitazone (eventually titled the TIDE trial).
August 15, 2007 Black-box warnings on heart-failure risk added to both rosiglitazone and pioglitazone labels, at FDA's behest.
August 7, 2007 "Reanalysis" of Nissen-Wolski meta-analysis questions findings, methods.
June 6, 2007 US Congress House Committee on Oversight and Government Reform hearing ends with no clear answers on rosiglitazone.
June 5, 2007 RECORD "unplanned interim analysis" shows no significant differences in CV risk between rosiglitazone and control group.
May 21, 2007 Nissen-Wolski meta-analysis published in the New England Journal of Medicine, showing significantly increased risk of MI with rosiglitazone.
May 10, 2007 Dr Steven Nissen secretly records a meeting with four GlaxoSmithKline executives, who don't disclose they already know the results of his meta-analysis, leaked to them by Haffner.
December 4, 2006 ADOPT published online in the New England Journal of Medicine.
September 15, 2006 DREAM published online in the Lancet.
July 16, 1999 FDA approves pioglitazone (Actos)
May 28, 1999 FDA approves rosiglitazone (Avandia)