By Heidi Ledford

Patients who take the controversial diabetes drug Avandia are more likely to have a stroke or heart failure, or die, than those who take a rival drug, a survey of more than 200,000 insurance records has revealed.

The finding, published June 28 by JAMA, The Journal of the American Medical Association (JAMA), is the latest salvo in a continuing battle over Avandia (rosiglitazone).

It arrives just weeks before a U.S. Food and Drug Administration (FDA) advisory panel meets on July 13 and 14 to evaluate whether Avandia, which is produced by the London-based pharmaceutical giant GlaxoSmithKline (GSK), should be removed from the market.

The possible health risks of Avandia were brought to light in 2007, when a meta-analysis by Steven Nissen, a cardiologist at the Cleveland Clinic in Ohio, suggested that the drug could cause heart attacks. This is a particular concern given that diabetics already have a high risk of cardiovascular disease.

Since then, Avandia sales have fallen from £1.6 billion (U.S. $2.4 billion) in 2006 to £771 million in 2009. In the interim, controversy surrounding the drug continued to build. The U.S. Senate launched an investigation into the matter, and concluded this February after a two-year inquiry that GSK was aware of the adverse effects of Avandia on the heart well before the risks were made public.

Heated battle

In a paper published June 28 in Archives of Internal Medicine, Nissen and his colleague, statistician Kathy Wolski, repeat their 2007 meta-analysis, incorporating new clinical-trial data and addressing some criticisms of their earlier work. They again found that Avandia boosts heart-attack risk--by up to 39 percent. "A drug that increases the rate of myocardial infarction by a third in a vulnerable population represents a huge public-health burden," says Nissen.

Meanwhile, in the study in JAMA, David Graham, a drug-safety researcher at the FDA, and his colleagues analyzed the insurance records of 227,571 patients who were treated with either Avandia or Actos (pioglitazone)--a similar drug produced by Takeda Pharmaceuticals, headquartered in Osaka, Japan. The team found that patients taking Avandia were 18 percent more likely to suffer from a stroke, heart failure, heart attack or death than those on Actos.

"The study by Graham et al. is very convincing," says Corinne de Vries, an epidemiologist at the University of Bath, UK. "I can't fault it, and I suspect it might be the nail in the coffin for rosiglitazone."

Less than two days after Graham e-mailed a pre-publication draft of the article to his colleagues at the FDA, he says he received a phone call from a reporter, seeking to verify the authenticity of the manuscript they had received. The manuscript was posted in its entirety on the popular pharmaceutical blog Pharmalot.

Graham says he believes that someone leaked the draft paper in the hope that releasing the results early would discourage JAMA from publishing the work. (Many journals, including JAMA, reject articles if authors have already released the results to the media.)

"This had all the hallmarks of a dirty trick," he says. "I was outraged."

In a statement issued in response to both Nissen's and Graham's papers, GSK maintains that six clinical trials completed since 2007, taken together, show no increase in heart attack, stroke or death. Furthermore, the company points to its own meta-analysis, which uncovered no increase in heart attack. Nissen disputes these findings.

After Nissen's 2007 study, the FDA added a warning label to Avandia, but decided to leave it on the market. At the time, the FDA advisory panel was concerned that pulling the drug might indirectly implicate Actos, says Clifford Rosen, an endocrinologist at the Maine Medical Center Research Institute in Scarborough who served on the panel. As a result, the panel was loath to take an action that could remove an entire class of medications, he says.

The JAMA study could ease these fears because it suggests that Actos is safer than Avandia but just as effective at lowering blood sugar. The committee will still face a tough decision, says Rosen, "but there's not a strong rationale for continuing to keep rosiglitazone around."