FDA Faces Tough Choice On New Diabetes Drug

Does a promising new class of diabetes drugs boost the risk of a rare thyroid tumor that may take decades to develop?

That's the confusing and almost unanswerable question Food and Drug Administration regulators find themselves facing after a federal advisory panel on Thursday split its vote on whether to approve Novo Nordisk's promising new diabetes drug liraglutide.

The drug is one of a new class of injected compounds called GLP-1 drugs that lower blood sugar levels while also causing weight loss.

Liraglutide would compete with Eli Lilly's ( LLY - news - people ) and Amylin Pharmaceuticals ( AMLN - news - people ) rival drug Byetta. Doctors like the drugs because they provide a potent new option for treating patients whose disease cannot be controlled with existing diabetes pills. If approved, Liraglutide might steal market share from Byetta, as it can be injected just once a day, compared with twice a day for Byetta.

But the panel was worried about animal studies showing that liraglutide causes a rare type of thyroid cancer called medullary thryroid cancer in both mice and rats. Novo Nordisk told Forbes that there is no evidence that it causes such cancers in humans. But the panel of doctors said that Novo had not ruled out the possibility. It deadlocked six to six on whether liraglutide should be approved. The 13th doctor on the panel was so confused by the issue he abstained on the key vote.

At the hearing, the FDA said it was "very rare" for a drug that causes cancer in multiple animal species to be approved.

The thyroid concern threatens to spill over and hit Amylin Pharmaceuticals and Eli Lilly, which are also testing a long-acting formulation of Byetta. Some members of the panel were worried that any thyroid risk might apply to the whole class of drugs. "It certainly sounded from what I heard today that this may be a class effect for any of the longer-acting agents," said Peter Savage of the National Institutes of Health, who voted against approval.

Novo Nordisk shares declined 8.3% at midday Friday, while shares of Amylin Pharmaceuticals also dipped 8.7%. In a report downgrading Amylin, Lazard Capital Markets analyst Matthew Osborne called the rodent thyroid cancer problem "completely unanticipated" and said that the thyroid findings "raise concerns" that the long-acting version of Byetta could be delayed as well, or face restrictions that limit its sales potential.

"There is currently no clinical evidence that thyroid cancer is a GLP-1 class effect. There is no identified association between thyroid C-cell carcinoma and [Byetta] based upon clinical trial and post-marketing data," Amylin Pharmaceuticals said in a e-mailed statement in response to a Forbes query.

A cloud over the injected GLP-1 drugs would be good news for Merck ( MRK - news - people ), which sells a rival pill drug Januvia. It also could help for Bristol-Myers Squibb ( BMY - news - people ) and AstraZeneca ( AZN - news - people ), whose experimental diabetes pill saxagliptin sailed through the same advisory panel on April 1 with no problems.

Novo contends the mice and rat data are irrelevant to humans. It hopes to get the drug approved by promising to carefully study the thyroid issue in a big 9,000-person safety trial that will also look at cardiovascular safety. "In my mind there is not a human health hazard based on what has been shown to date," says Novo Nordisk Chief Scientific Officer Mads Krogsgaard Thomsen. Liraglutide does not boost thyroid cancer risk in monkeys even when they are given enormous doses, he says.

But he admits that the intense discussion about the rodent thyroid cancers took Novo Nordisk by surprise--as it did Wall Street. "We knew we were going to discuss the issue, but we did not know [it was going to be] the overarching dominating thing," says Thomsen.

Going into the meeting, the main focus was expected to be cardiovascular safety. This has been a big subject of scrutiny for all diabetes drugs ever since Avandia from GlaxoSmithkline was linked to heart problems in 2007. But cardiovascular concerns turned out to be a comparatively minor point of discussion on Thursday.

Instead, the debate centered on the unsettling mouse and rat studies. Committee members were not convinced by Novo's contention that the rodent findings would not apply to humans. "I felt the animal data were worrisome and didn't see sufficient human data to feel reassured," panel member Peter Savage said at the hearing. "I am not sure the benefits of adding this drug at this time outweigh the trade-offs."

The problem is that medullary cancer is so rare and develops so slowly over decades, that there is be no easy way to rule out the possibility that the drugs might boost cancer risk decades after people start taking them. "We are not talking about something that is going to suddenly pop up and kill someone in six months or a year," explained panel member Michael Tuttle of Memorial Sloan-Kettering Cancer Center. "The risk is, are they going to develop [thyroid tumors] years down the road." He voted for approval because he thought there was no obvious study that could be done to rule out the risk.

Nothing is likely to answer the question definitely short of approving the drug, letting millions use it, and waiting to see whether more people get medullary thyroid cancer. "We are facing a modified Hobson's choice," said committee chair Kenneth Burman of Washington Hospital Center. "Should we expose the entire population to the potential risk even though it is probably low?" He voted against approval but said more human data showing no thyroid risk could easily change his mind.

Novo Nordisk may already have some of the data Burman wants. Thomsen says new data from recently completed two-year human studies of liraglutide show no hints of thyroid problems in roughly 1,500 patients. These data are so new they weren't part of the company's original submission to the FDA.

Cardiologist Marvin Konstam of Tufts Medical Center, the potential deciding vote on the panel, abstained. "I can think of no better question for a cardiologist to abstain with," he said. "I don't come away with a clear opinion what to do."

The panels' non-decision puts the question squarely back in front of the full FDA. Should it approve the drug right away but with restrictions and demands for further studies? Should it demand additional, longer-term human studies of liraglutide before approving the drug that could reassure people that the risk is low? Should the drug be rejected entirely? And if the FDA thinks thyroid cancer is a real risk, does it also apply to other drugs like Byetta?

At the end of the hearing, the FDA itself seemed somewhat confused as to what to do next. "I am somewhat comforted that you have struggled" with the thyroid concern, FDA division director Mary Parks told the panel. Now, she said, "we will take on the difficult task of next steps."

The GLP-1 drugs are the result of decades-old breakthroughs in understanding blood-sugar regulation. In the late 1970s, Harvard Medical School researcher Joel Habener serendipitously discovered a hormone in monkfish called GLP-1 (glucagon-like peptide). It stimulated the body to secrete just the amount of insulin needed to control blood sugar, but not too much.

But GLP-1 turned out to be an impractical medicine, as it degrades quickly once injected into the body. But in the early 1990s endocrinologist John Eng at the VA Medical Center in the Bronx found a chemical in venomous Gila monster saliva that mimicked the effects of GLP-1. When he injected it into diabetic mice, it controlled their blood sugar all night long. The Gila monster finding led to Byetta, a synthetic version of what is found in the lizards.

Novo's Liraglutide is a modified version of the human GLP-1 that lasts longer in the bloodstream.

Source : www.forbes.com