Two years ago, the FDA approved Avastin to combat breast cancer, even though an advisory panel determined that risks such as high blood pressure and death outweighed the benefit of slowing the spread of tumors. The agency, however, acted under its accelerated approval program and the move pumped up sales of a Roche drug that is also used to treat brain, lung and colon tumors.
Now, though, Avastin is on trial again. Another FDA panel today will decide whether use of the $50,000-a-year med should be continued, expanded, or halted after two studies - which were undertaken as a condition of approval - found patients given Avastin and chemotherpay didn't survive longer than those given chemo alone. And Avastin patients also suffered more serious side effects.
This is the second time in recent weeks that a drug approved as part of the FDA's accelerated approval program is being scrutinized anew. Last month, Pfizer withdrew Mylotarg, a drug used to treat acute myeloid leukemia, or AML, at the agency's behest after a study, which was begun four years after approval, found a lack of clinical benefit and an unexpected number of deaths (back story).
These back-to-back episodes are casting a spotlight on the accelerated approval program, which some say is working. "I think it has done what it set out to do: get drugs to market faster with less evidence than is typical and confirmed a drug's risk-benefit profile once more evidence becomes available," says Ramsey Baghdadi of Prevision Policy. "Every so often, that evidence turns out unfavorable for the drug in question...overall, I think it has achieved what it was supposed to do."
Nonetheless, we asked the FDA about the extent to which the program is being reviewed. In response, Richard Pazdur, who heads the Office of Oncology Drugs, wrote to say the agency will be asking drugmakers to provide required confirmatory studies, not just accelerated approval studies, (see here) and completion dates, as well. And since these trials are expected to be under way at the time of approval, any delays would have to be explained.
All this should occur "definitely before or at the end of Phase II meetings. These plans should include populations to be studied, planned enrollment, and estimated completion dates and any anticipated effect that the accelerated approval would have on the completion of confirmatory trials," he writes, adding drugmakers "strongly" encouraged to have more than one study serve as a confirmatory trial. And they should be prepared to add trial sites, if necessary, to meet their promised schedules.
But what if a drugmaker fails to pass muster? What kind of stick does the FDA have to work with? The agency could impose penalties and withhold approval, according to Tamy Kim, associate director regulatory affairs in the OCD. "We could make it a requirement that, if they don’t have a confirmatory trial or even plans under way, it would be a condition of approval," she tells us. "...Since the FDAAA was passed in 2007 and implemented in 2008, there’s a provision that states if a sponsor doesn’t conduct a confirmatory trial under subpar H with due diligience, it would be violation of the post-marketing requirement...These give us greater teeth than we had before." She adds, however, this is just now being considered.
As for the carrot, Pazdur notes the agency has exercised "regulatory flexibility" in allowing confirmatory trials to be conducted in "a different setting" than the approved indication. So if a trial turns out positive, it would not only satisfy the subpart H (or subpart E for biologics) requirement, but also effectively serve as a new marketing application for a supplemental NDA or BLA.
Just how fast the OCD will move is unclear. Pazdur says he plans to hold oncology drugs advisory committee meetings (that's meetings, not a meeting) to discuss the subpart H commitments, and the possibility of a yearly meeting to review confirmatory trials is being discussed. But no, accelerated approval is not being eliminated.
3 Comments
Pazdur says Accelerated Approval will not be eliminated, but in some ways he has already dramatically scaled it back. Prior to 2003, many accelerated approvals were granted based on compelling data from single arm Phase II trials, and the randomized trials were run (often in other countries) after the approval. The very large majority of those drugs became mainstream, usefull and iften very important cancer drugs. The emphasis, consistent with Congressional intent, was to meet patient's unmet needs (i.e., provide them their chance for access to progress and continued life before they died waiting for FDA's glacially-slow process to run its full course). However, beginning in 2003, as part of a Phase IV trial enforcement initiative started by Pazdur, the bar for accelerated approval was raised so close to that previously required for regular approval, the Congressional intent of the AA program was nullified, and the bars for both AA and regular approval were raised. It was clear then, and is even more clear now, that the only thing that will matter at the FDA is meeting its rigid statistical standards. The result is it takes a lot more time and money to get a safe and effective cancer drug approved today, fewer of them are being approved, progress against cancer has slowed, and patients are waiting longer for fewer chances at continued life, tens of thousands of them dying while they wait. Now we have Pazdur pulling drugs off the market for purely statistical reasons. In the case of Avastin in the setting being discussed today, FDA is spinning the confirmatory trials as not confirming the earlier results, but they did, with advantages in both progression free survival and overall survival; but the advantages in these larger trials are not as great, and the FDA is making a judgment here that the average statistical benefit shown in these new trials just isn't good enough to be "clinically meaningful." That is a value judgment the FDA should not be making, especially since the average response of a population (e.g., comparisons of the survival of the median patient in the treatment group versus the median patient in the control group) doesn't begin to describe what happens to the subsets of patients who respond dramatically. Pazdur is a rigid thinker and regulator with enormous power to control every aspect of cancer clinical research, drug approvals and withdrawals, and quite directly the practice of oncology. He does all those things in a very narrow-minded way and more than a bit too zealously. The truth is, Dr. Pazdur has no signal accomplishments as an oncologist in his career, save becoming the cancer czar at the FDA ten years ago. Since that time he has aggressively defended the scientific status quo in oncology during a time when dramatic change in response to rapdily advancing science and knowledge of the biology of disease should have been driving major change in the way his division does its job. Instead he has been a stickler for ever more perfect statistics from clinical research, enforcing his requirements by delaying and denying approvals,and by pulling drugs that obviously help some patients off the market. We need better treatments and cures, and patients need options, not perfect enforcement of obsolete regulatory standards. Is he the right person in a time when the FDA should be focusing on modernizing its regulatory tools and approval standards? We can't do better than a cancer cop at FDA? Really?
"especially since the average response of a population (e.g., comparisons of the survival of the median patient in the treatment group versus the median patient in the control group) doesn’t begin to describe what happens to the subsets of patients who respond dramatically."
It's very simple, identify the subset patients first before you move to large trial, then apply for approval on those subset patients.