Getting Together on Genomics

ﾩ 3D4MEDICAL.COM/GETTY

In a twist to the usual approach to clinical trials, the pharmaceutical company Wyeth was the guinea pig this time. The company had agreed to a videoconference with the US Food and Drug Administration (FDA) and the European Medicines Agency (EMEA) last winter as part of the process for submitting pharmacogenomics data jointly to the agencies. The data weren't from a particular clinical trial, but were instead part of Wyeth's efforts to lay the groundwork for regularly integrating genomics data into future trials. Before the FDA and EMEA could issue the new guidance this spring, says Allen Rudman, associate director of the FDA's office of clinical pharmacology in Silver Spring, Md., "we had to try it out."

For its part, Wyeth had been collecting genetic samplings from 400 healthy volunteers in the United States, Canada, and Germany over the course of a year, broken into 18 subgroups by such factors as gender, race, age, and whether the volunteers smoked. This was not a typical clinical trial, with some of the volunteers taking the test drug and the rest on placebo. "There was no drug intervention at all," says Ronald Salerno, director of global regulatory affairs at Wyeth. Rather, the company was aiming to use the data to create a baseline of biomarkers that it could then refer to in future drug trials.

These data, collected five times over 12 months, would show certain naturally occurring genetic variations in blood cells and how they changed over time. Later, the data could be compared with the genetic makeup of volunteers who have a disease under study, and also with those taking a drug. "You can find out whether the gene expression could be expressive of any disease state," Salerno says.

Wyeth was eager to get the regulators' opinions on this approach, because the use of biomarkers is so new. "Our perception of the data and utility of it may be different from the FDA's. They have their own attitude," Salerno says. "The purpose was to understand the perception and the scientific acumen of the reviewers who use this information."

What did they learn? For one thing, both regulatory agencies were concerned about "the variability coming from different laboratories processing the blood samples," Salerno says, and Wyeth is now trying to figure out how to eliminate the variability. In fact, Wyeth received similar feedback from both agencies regarding the trial design and the data analysis. That's encouraging, because it could show that drug companies won't have to submit different sets of data to each agency - a major concern.

If this meeting hadn't been held, Salerno says, Wyeth might well have done the same research, but "it would be all on risk. We could end up with a diagnostic that could be unusable." Salerno says this particular set of data was not collected with any specific drugs in mind, but Wyeth has separately collected pharmacogenomic data on patients with Alzheimer disease, asthma, and renal cell cancer.

While the FDA's Rudman won't comment about the meeting specifically (the FDA comments only on drugs that have been approved), he says the agency sees this as an opportunity for the FDA, EMEA, and companies to understand each other's issues. If regulators don't understand what's actually involved, "they might ask things that are not realistic" when they review the drug application, says Mathias Hukkelhoven, global head of drug regulatory affairs at Novartis, who is based in East Hanover, NJ. "It's a free, extra, reality-check," says Lois Hinman, director of regulatory affairs at Roche's US headquarters in Nutley, NJ.

Drug companies, of course, are traditionally reluctant about sharing data not directly focused on gaining regulatory approval. For a company submitting data it happened to discover during a trial that wasn't part of the original trial protocol, "their nervousness is that we take information and find some very small safety questions," says Steven K. Galson, director of the FDA's Center for Drug Evaluation and Research. That's why, says Galson, the regulators specify that such incidental genomic data will not be considered in that drug application unless it is directly relevant - for instance, if the company plans to target the drug to a genetic subpopulation.