Edward John Allera, chairman of the Buchanan Ingersoll & Rooney's FDA/Biotechnology Section and a managing shareholder of the firm's Washington, D.C., office, was quoted in an April 21, 2009, article published by Pharmawire. The article, titled "Dendreon's initial Provenge Phase III trials unlikely to contribute much to FDA submission package, biostatisticians say," reported on possible issues Dendreon Corporation — a Seattle-based biotechnology company focused on targeting cancer — will face when it resubmits overall survival (OS) data for the vaccine Provenge to the FDA.

As explained in the article, a number of biostatisticians surveyed say the OS data is from Dendreon's previous Phase III trials and will add "little, if anything, to the company's Provenge submission."

The article went on to say, "The data will be part of the package submitted to the FDA, along with data from the now-pivotal IMPACT trial, which was granted a special protocol assessment (SPA). Biostatisticians described the data from the previous, once-pivotal Phase III trials as exploratory, suggesting that it either not be used to influence the FDA's decision on Provenge, or that it be down-weighted. … Initiated in 1999, the first Phase III studies for Provenge both failed to meet their primary endpoint, time to disease progression, and also failed to demonstrate statistical significance on other pre-specified endpoints. When the data from the two trials was pooled and analyzed retrospectively, Dendreon found that there was a difference in OS between the active and placebo arms. The current IMPACT trial, which has enrolled 512 patients, is designed with OS as the primary endpoint in accordance with the positive OS data generated from the original Phase III trials. The previous trials enrolled 225 patients."

Biostatisticians interviewed for the article agreed that the FDA would likely disapprove of the pooled data because "there are so many ways to pool or combine the data and different ways to analyze the pooled data."

"The way they pick and choose data for meta-analysis, you would be ridiculed," said Allera, who previously served as associate chief counsel at the FDA.

Getting a SPA (special protocol assessment) is not a big deal, said Allera. "The real issue is how robust the data is and how that fits within the treatment armentarium," he said. "The reported data looks [like] very good data, but whether the FDA agrees that's the real number is the issue," Allera said.

"It all depends on how robust that 22 percent [reduced risk of death] is and whether its reproducible. That's the dilemma," said Allera. There is still a possibility the FDA could ask the company to run another follow on study, or even a post-marketing study, he added.