Dendreon and Provenge - The FDA Advisory Panel Meeting
On Thursday, March 29 I attended an FDA Advisory Committee meeting to review, approve or disapprove a new cancer treatment by Dendreon called Provenge. This treatment, if approved, would be the first therapeutic vaccine for cancer to come to market - Provenge is targeted at advanced metastatic hormone refractory prostate cancer.
The room was packed, and the most nervous group of people were from Dendreon (DNDN), the company proposing the treatment for approval. The Dendreon folks were even more nervous than the terminally ill cancer patients in the room to testify.
This hearing made history in many ways as Provenge was approved, not just as the first of its kind of treatment, but also the first approval by the FDA for a drug that failed to meet its clinical endpoint to stated goals in trial.
The actual vote by the 17 panel members at the end of the long meeting was actually going against the company when a man whispered into the ear of the person running the meeting and the world (at least Dendreon's world) changed, as did the vote, and a few minutes later the ninth yes vote was cast, the media relations person standing next to me from Dendreon began to show tears and some of the patients in the room began to think they might get a shot at living a bit longer.
Sounds overly dramatic? It was actually far more exciting than this - and, if you remember Mike Brown from FEMA, the man who went to dinner while the people of New Orleans drowned, I think he has a new home - the FDA - or perhaps at Dendreon. I am just joking but the level of screw ups by the FDA and Dendreon during the meeting was shocking, given the potential consequences - for DNDN, a business and billion dollars in market cap, for hundreds of thousands of patients, the possibility of a longer life.
Some background first. Provenge is a treatment, not a drug - an infusion of a patient's own dendritic cells, activated at a Dendreon facility by treating it with Provenge, and these cells in turn stimulate T cells in a patient's own immune system. Provenge is the first therapeutic cancer vaccine to be evaluated by the FDA. This created a very high bar for approval but that was only part of the problem.
Much worse was the paucity of data - Dendreon submitted data from two trials that had only 144 patients. But it gets worse than that - the trials failed to reach their stated goal or what is called clinical endpoint, reduction in the progression of the disease. And even worse than that (c'mon, this is fun), one trial had different results for the secondary endpoint - survival - than the other.
And therein lies the reason for approval - people did live longer. A secondary endpoint for the trial that has been called, by the FDA, the "gold standard" for cancer trials. On average, 4.5 months - but some have been alive for years and one appeared with his grandchildren at the panel meeting. The other reason for approval -- there are no other real treatments - the one approved treatment extends life less than three months and is highly toxic with terrible side effects on already weakened patients, side effects so bad many patients eschew treatment.
A third reason for approval was the panel itself - Provenge falls under the purview of Center for Biologics Evaluation and Research and was reviewed by the Cellular, Tissue and Gene Therapies Advisory Committee. The real gatekeepers of cancer therapies reside at the FDA in something called CDER - the FDA Center for Drug Evaluation and Research. They approve virtually all non-biologic cancer drug, are hard nosed statisticians and the two people on the panel who mostly work with CDER were the most vociferous opponents of approval.
A fourth reason was the composition of the panel - the FDA was very fair and had, I believe, seven immunologists on the panel, and these folks were eager to see some sort of treatment get to market. They also understood how hard it is to measure exactly how Provenge works - technically, the company cannot measure the real impact on the immune system in a meaningful way except through the extended lives of patients. The CDER people found this appalling; the immunologist readily accepted this as part of the newness of the technology. This lack of data on the actual mechanism of action of Provenge became an important part of the panel discussion - but actually turned in the company's favor due to the presence of so many immunologists.
The last reason was the testimony of the patients - all, technically dying except the grandfather who came with two grandchildren from California. They were all dying at different paces; most of them limped; one begged for approval for a dying father so he could see his grandchildren make their Bar Mitzvah. It may sound like a bad "made for television" movie but it was not. A very old hand at this, and maybe the best FDA analyst and watcher, Ira Loss of Washington Analysis Group, said this was standard at the hearings and had no impact on the panelists. But he spends all his time at CDER panels - two panel members at this meeting spoke of the impact of this testimony on their vote.
Yes, the vote - but before I get to the vote, let me take a moment to skewer Dendreon management. They were awful, with the exception of a consulting physician with experience in clinical trials. I subscribe to several fairly technical services and one went so far as to provide a biography and the past committee experience of each member. They also said there would be a spirited debate on the lack of data on the mechanism of action and this could hurt Provenge's chances.
Well, Dendreon management must not subscribe because they were caught flat footed at this discussion and did not prepare for questioning in this area. They played right into the hands of FDA statisticians and never once confronted, in their slides or formal presentation, the core issue - due to the small trial size and mixed results, by historical standards the treatment did not work. Thy were just about dead in the water when their consulting physician took the opportunity to seize the discussion during a Q&A session, steer the panel back to this issue, say Provenge worked and, politely explained that the patients were going to die anyway, what is the harm?
And that question - what is the harm - paved the way for the final, successful vote, sort of. Oh, before I get to that, let me tell you what your tax dollars buy you at the FDA.
Subscribes to my newsletter, ChangeWave Biotech Investor, besides having made an incredible pile of money on DNDN stock on Friday (I recommended the company in November of 2005) you would come to know I believe the FDA is the most undervalued, overworked and unappreciated of all government agencies --they need their own TV show, I guess, to get more attention from Congress. Their employees make one half to one quarter of what they could in the private sector and lobbying by Big Pharma, resulting in Congressional idiocy, has crippled them. For example, it now takes longer to get a generic drug approve din the US than a new drug because Big Pharma has convinced Congress to restrict funding for the staff that reviews generics. I estimate a twenty five million dollar increase in staffing would save Medicare and the Veterans Administration between $1.5 and $2.0 BILLION dollars a year.
Well, you get what you pay for - two leading figures in the discussion, senior biostatisticians who run divisions at the FDA, could not understand spoken English well enough to present proper testimony or answer questions adequately. They both had PhDs made in America but could not answer questions about the drugs they were evaluating, also made in America. When asked about the importance of survival, the agency's own gold standard for a trial, the official response was "they failed to meet their primary endpoint." A big help that. Their inability to hold a discussion could have had terrible consequences for thousands of patients.
But this problem is explainable - money - but the other mistake that almost derailed Provenge was inexcusable, a true "Mike Brown" moment, a near Katrina of drug evaluation.
Here is what happened - at the end of a tough eight hours, the panel votes on two issues - safety and efficacy. Here is how the questions were posed on the FDA website before the meeting and at the end of the meeting on slides.
• Does the submitted data establish that sipuleucel-T (APC-8015) is reasonably safe for the intended population?
• Does the submitted data establish the efficacy of sipuleucel-T (APC-8015) in the intended population?
The first vote went 17-0 in Provenge's favor - even the hard nosed folks more comfortable with CDER regulations and processes agreed the safety profile was very favorable, especially compared to the other possible treatment, a form of Taxotere. Then the fun began - the person chairing the session asked the second question.
I am not naming names - but imagine a serious, steady, super successful physician who knows patients are dying horrible, painful deaths with no possible treatments debating this question, out loud, with himself. "Yes, it looks like it works - but, established, I don't know..." and on and on and on and finally, after ten minutes, a no vote.
Panelist number two - same problem. They wanted to vote yes, but, well, I am a scientist, you asked if the efficacy is established the trials were small, and so on and so on and I must vote no.
Ditto for panelist number three.
At this point magic entered the room - maybe Frodo came in wearing the Ring of Power and invisibly touched the heart of the head of the committee. She asked it he panelists would like the question re-worded as they were having such difficulty with coming up with an answer they found suitable. And, let me sidetrack you again with something important.
During their testimony, Dendreon discussed a trial currently being enrolled with a target size of 500 patients, 400 already in the trial. Everyone agreed the statistical output of this trial would answer all the statisticians' questions - but the trial would not be over until the middle or end of 2010 and an application would not come up before the FDA until 2011. I did the math and that meant something like 380,000 patients might miss out on Provenge if they had to wait that long and most would die waiting. Panelists had asked if the company would provide unlimited access to the drug awaiting the trial outcome - which means providing free treatment for four years. The company hemmed and hawed, unprepared for this and finally said it was not feasible. So during the vote, this trial hung over everyone's head - it was what everyone wanted but four plus years is a long time.
Let's get back to the vote - just as the head of the committee asked if the question should be re-phrased, a slightly built, fiftyish man stood up from his seat next to the far wall, walked a few steps to her, whispered in her ear and handed her a slip of paper, and history was made. The chairman said the question posed by the FDA, was, well, not the language of FDA regs, and a new question should be posed and should read "Does the
submitted data provide sufficient evidence of efficacy.
As I wrote so eloquently in my newsletter, using the best scientific language "Ba Da Bing!"
With this new question in hand, the first three panelists reversed their votes and the countdown began. At the ninth yes vote, the media relations person for Dendreon standing next to me began to cry; the final vote was 14-3 although the FDA, true to form, recorded it as 13-4. A drug with a minuscule trial and a failed data set had garnered approval.
What next? The FDA has to formally approve the treatment no later than May 15. They could reverse the decision, they could do a lot of things. Tell me what you think, here, or read my letter as I handicap the possible outcome and cover this as news and opinion break. And please, write your Congressman and tell them to trade one Alaskan bridge to nowhere for an FDA staff that can read their own regs, speak English well enough to doe their job and maybe even run a meeting.
