Warnings, Not Ban, Urged for Diabetes Drug

A pair of Food and Drug Administration advisory panels called yesterday for new warnings for the widely used diabetes drug Avandia because of evidence that it significantly raises the risk of heart attack, but they stopped short of recommending that the drug be pulled from the market, as some FDA officials had urged.

The 22 to 1 vote to allow continued sales of the drug caps years of controversy among regulators, drug company officials, physicians and advocates for patients, who have offered dueling interpretations of contradictory safety studies. The pills are taken by more than 1 million Americans and racked up $3 billion in sales last year for their maker, GlaxoSmithKline of North Carolina.

Some analyses by experts inside and outside the agency have concluded that Avandia increases patients' heart attack risk by 30 percent or more. Several experts have also concluded that it makes no sense to wait for results from other studies because they were not designed in a way that will settle the question.

"Cardiovascular disease being the leading cause of death of people with diabetes, having a treatment that causes that is something that doesn't make sense to me," said Gerald Del Pan, director of the FDA's Office of Surveillance and Epidemiology.

But a majority of the advisers sided with Glaxo, which offered hundreds of pages of evidence suggesting that Avandia poses no greater risk than other diabetes drugs. It encouraged the agency to wait for findings from studies that are due to produce results next year.

Many diabetic patients, physicians and outside experts, who filled to overflowing a large Gaithersburg hotel ballroom where the committees met, said they found fault with the statistical accuracy of recent studies that indicated Avandia is risky. They urged the panels to keep Avandia on the market so that patients could have a better chance of finding a medicine that works for them.

"Let's not throw out the baby with the bathwater," Philip L. Hooper, a physician at the University of Colorado Health Sciences Center, said in one of many such written comments submitted to the panels. "We need all the resources that we can to fight this disease."

The FDA is not required to follow its advisory committees' recommendations but usually does. FDA officials would not predict how long it will take to make a final decision about the marketing and labeling of the drug.

Beyond the impact the decision will have on the nation's 20 million diabetics -- and on Takeda Pharmaceutical and Eli Lilly, which jointly market Avandia's prime competitor, Actos -- yesterday's decision is sure to send reverberations across Capitol Hill. Many lawmakers have been growing increasingly agitated about what they believe has been a failure by the FDA to match its accelerating rate of drug approvals with equally aggressive tracking of those drugs' long-term risks.

As more evidence came to light about Avandia's risks, and as the agency's consideration of the data dragged on for more than a year, the drug became a poster child for those who believe that the agency is too beholden to the pharmaceutical industry.

Chris Viehbacher, Glaxo's U.S. president, praised the vote. "We see the vote as a strong positive for patients," he said, adding that the company intends to work with the FDA to help clarify what kinds of warnings may be appropriate.

Several experts noted that people taking insulin or certain drugs for chest pain or high blood pressure appear to be at particular risk of heart attack from Avandia.

Glaxo's value has slumped by billions of dollars in the two months since an analysis in the New England Journal of Medicine found that patients taking Avandia had a 43 percent higher rate of heart attacks.

Diabetes has reached epidemic proportions in the United States, with more than 4,000 new diagnoses every day. Diabetes costs the nation more than $100 billion a year and is expected to cost $200 billion by 2020.

The job of reviewing all the relevant data was "extraordinarily complex," said Karen Mahoney, an FDA medical officer, in part because all diabetes drugs cause potentially serious side effects. Actos, for example, can lead to congestive heart failure and may cause bladder cancer, and other drugs can cause dangerously low blood-sugar levels or other problems.

Mahoney said she personally reviewed more than 100,000 pages of documents. The data sets were so complicated, she said, that some FDA computers had to be upgraded to handle the load.

Adding to those complications, said some FDA reviewers, was the fact that some studies sponsored by Glaxo were designed in ways that favored the company's goals of keeping Avandia on the market. The company's highly touted, ongoing, 4,000-plus-patient Record study, the largest ever launched to look at the heart risks of Avandia, so far has not found an increased risk of heart attacks.

But an analysis by David J. Graham of the Office of Surveillance and Epidemiology concluded that because of some of the definitions that Record used, the lack of an untreated placebo group and other design shortcomings, the study will not be able to detect the increased heart risk that others studies have shown. Graham called the Record study "biased."

Similarly, FDA reviewers Chuck Cooper and Yu-Te Wu of the Office of Biostatistics concluded that other continuing studies that Glaxo says could settle the issue are statistically "questionable." The only way to get convincing results, they concluded, would be for Glaxo to start over with "a well-designed study to address the issue."

Graham's analysis indicated that since Avandia was approved in 1999, about 80,000 patients have died from the drug's side effects. He also pointed to recent evidence, largely accumulated by Takeda, that Actos reduces the risk of heart attack by perhaps 20 percent.

Actos "certainly is not increasing the risk, and if anything it may be decreasing the risks," Graham said.

But other experts warned that many of the new data on Actos have yet to be independently analyzed.

"The data were given to FDA on the fly . . . yet we're asked to put that in our decision-making," said Annette Stemhagen, vice president at United BioSource in Ambler, Pa., a nonvoting member of the review committee representing the drug industry.

Perhaps more than anything, yesterday's meeting made starkly clear how complicated it is to quantify the risks of a drug whose metabolic effects are as complex as the underlying disease -- especially when so much of the data are generated by drug companies. It is well documented that such studies are more likely to favor the company's position than if the research were done independently.

"This may be an example where the quality of the analysis far exceeds the quality of the data," said Nancy Geller, director of the Office of Biostatistics Research at the National Institutes of Health.