Shadow FDA? Researchers Are Taking Approval Matters into their Own Hands

Drug and biotech companies will face an emerging trend of outside interference on pending drug reviews at FDA. The events surrounding the final outcome of FDA's review of Bristol-Myers Squibb and Merck's diabetes drug muraglitazar (Pargluva) signal that outside medical centers and prominent researchers will voluntarily play a larger role in drug safety oversight in the future.

Muraglitazar is dead after FDA asked for more safety data. Researchers at the Cleveland Clinic helped finish it off.

By Ramsey Baghdadi

It was so close. FDA had finished its review of Bristol-Myers Squibb Co. /Merck & Co. Inc. 's type 2 diabetes drug muraglitazar (Pargluva). The agency's Endocrinologic & Metabolic Drugs Advisory Committee voted almost unanimously to recommend approval of the drug. FDA issued an "approvable" letter asking for more cardiovascular safety data from ongoing or already completed studies, which sounded easy enough. The dual peroxisome proliferator activator receptor agonist for diabetes management was on track to be the first of its class to reach the market.

Now it may never get there.

On October 26, just over a week after the "approvable" letter, Bristol and Merck announced they are terminating their partnership for Pargluva. Although FDA delivered the final, official blow, the application may have been derailed by two prominent cardiologists who warned that the drug would turn into the next Vioxx. The two academics rushed to fill in what they saw as a horrible oversight in FDA's review process. They may not have influenced the agency's final decision, but they did deal a crippling blow to any strategies for salvaging the product. And the same thing is likely to happen again, with different drugs, because public perception of FDA's ability to ensure safety has eroded to the point where researchers who normally watch from the sidelines are starting to insert themselves into the game.

"Approvable" For Five More Years

In this case, the doctors were the enormously influential Cleveland Clinic cardiologists Eric Topol and Steve Nissen, who, along with Kathy Wolski, conducted an analysis of Pargluva compared to a combined control of Eli Lilly & Co. /Takeda Pharmaceutical Co. Ltd.'s pioglitazone (Actos) and placebo. The analysis, published two days after the FDA "approvable" letter in the Journal of the American Medical Association, found the drug carried double the risk of death, heart attack and stroke compared to control. The cardiologists found 35 deaths, heart attacks or strokes in 2,374 Pargluva patients compared to nine similar events in 1,351 control patients.

JAMA expedited its peer review of the study and published the results for public consumption at breakneck speed. Major media outlets, both television and print, highlighted the findings. For example, CBS News ran an Associated Press story on its website with the headline: "Risky Drug On Track For FDA OK."

The JAMA analysis appeared at the same time that Bristol was discussing the "approvable" letter with FDA. The company is understood to have outlined for the agency its plans to address the questions raised in the "approvable" letter. Apparently, FDA was not satisfied. Six days after the JAMA study was published, Bristol and Merck announced that FDA's questions on cardiovascular safety could take up to five years—and a lot of money for clinical trials—to answer.

For both companies, the loss hit hard. The drug was the most important near-term launch for both Bristol and Merck. Winning the right to co-promote the drug, after discontinuing its own dual PPAR in Phase III to side effects in 2003, cost Merck $100 million in an upfront fee and another $55 million in a filing milestone early this year [See Deal], not to mention its share of the development expenses.

One of the critical side effects to the Pargluva setback is that it allows competitors already on the market a substantial competitive advantage. For example, GlaxoSmithKline PLC 's PPAR agent rosiglitazone (Avandia) is already approved for type 2 diabetics. The drug carries its own cardiovascular toxicity concerns. Although GSK may have to ratchet up its efforts to defend the safety of the drug, it continues to generate revenue while it remains on pharmacy shelves.

A Race Against FDA

The wheels were set in motion for the Cleveland Clinic analysis after an FDA advisory committee recommended approval of the drug, about a month and a half before the agency was scheduled to complete its review. "We decided to do it after we saw the advisory committee shockingly vote 8-1 that the drug should be approved," Topol says. "We had heard there may be a cardiovascular safety question and there was no cardiologist that was participating on the panel."

After the panel, Topol, Nissen and Wolski gathered the data from the NDA published on FDA's website and conducted an analysis on the joint control group—something the sponsors had not done. "It was far worse than we had expected," Topol says.

What happened next is not completely clear, but it appears that at least one official within the Center for Drug Evaluation & Research became aware of the upcoming Cleveland Clinic publication. According to a source close to the situation, the official appropriately recused himself from any involvement in the Pargluva review at that point. However, the source says, the FDAer had expertise that would have been useful in the deliberations over the cardiovascular safety of the drug.

According to Topol, Nissen discussed the analysis with an agency contact. In an official statement, FDA said it is not aware of any conversations with Nissen regarding the JAMA editorial. "A rigorous peer-review process was undertaken to review the overall safety and efficacy of this compound. The agency concluded additional information is needed to more fully assess the cardiovascular safety profile of Pargluva," an FDA spokesperson said.

Nissen agrees with FDA's official account of the situation, saying he did not discuss the upcoming publication with anyone at the agency. He does, however, acknowledge that someone at FDA could have heard about the analysis through third parties prior to publication.

Do the actions of Nissen and his colleagues compromise the integrity of the review process? Topol says emphatically: "No." The review had already been completed, he says. FDA "had had the panel; there was nothing that had interfered with the review process. There was just a legitimate concern that the drug was on the cusp of getting commercially approved….Without any more trials, the drug appeared destined to get approved."

But Topol and Nissen did affect FDA's decision. A review is not truly complete until the agency issues a final decision to approve or reject a drug. Whether or not the researchers directly contacted the agency, they built a public case against approving Pargluva through one of the most respected peer-reviewed journals.

They may have had another indirect effect if, in fact, their work led an FDAer to stay out of a review that might have benefited from his perspective. That raises the question of how agency experts should react when they become aware of outside analyses of data involving pending applications: communities of medical specialists are tight-knit, and it is perhaps inevitable that individuals in FDA will learn of important research even without direct attempts at communication.

Topol says this may be the first time a drug was reviewed by unsanctioned external researchers for a product on track to receive formal approval. It probably won't be the last. "This is a new precedent and it wouldn't be at all surprising if we see this again in the future," he says.

That raises the possibility that prominent researchers and the prestigious medical centers they represent could become a new shadow FDA. "I hope it doesn't come to that," Topol says. "Obviously that's not our charge." But Topol will clearly be keeping a close eye on the agency. "If the FDA advisory committees and the internal staff there do a really good job then it wouldn't be a necessity to come up with an alternative view of the data."