FDA and Evidence-based Drugs
If you are a person diagnosed with breast or ovarian cancer, or a family member or close friend, you probably are at least somewhat aware of what the Federal Food and Drug Administration does on our behalf. The FDA is charged by Congress with approving pharmaceuticals that treat important diseases. These drugs may be offered by prescription or in hospitals, clinics or doctors offices. Generally if a drug is approved it is approved for particular diseases, known as the "indication(s)."
Regulation is essential to ensure that drugs offered do what their manufacturers claim they do, and that they don't cause serious harms to persons taking them. Many horror stories can be told about drugs that killed people while purporting to help them. Without regulation, patients would be involuntary guinea pigs for drug manufacturers.
Generally, a manufacturer seeking FDA approval for an oncological drug -- that is, a cancer treatment -- submits the results of well planned clinical trial(s) with a large number of patients diagnosed with the target disease. Patients are randomly assigned to the set that gets the drug, and the control set that does not get the drug. Trial data must show a statistically significant clinical benefit. In the best case scenario, the drug will prolong survival in substantially more patients than might occur by chance. This Overall Survival (OS) end-point of a clinical trial is the most desirable one, for obvious reasons. If trials cannot show better OS, the substitute end point is longer time before a disease worsens, known as Progression Free Survival (PFS).
Sometimes a manufacturer will seek Accelerated Approval (AA) for a drug. AA began decades ago because people diagnosed with AIDS had no drug that would help them, and manufacturers were racing to market drugs that would at least stem the pace of this killer. AIDS patients and their friends and families became ardent advocates, and convinced the government that it made no sense for them to wait for standard clinical trials, since the disease was killing so many so fast. The results of AA were successful for AIDS treatment.
Soon, manufacturers began to submit AA applications for drugs directed at other diseases. These applications were sometimes also directed at problems for which there was no known effective treatment. Thus they clearly targeted an unmet need, which was the criterion established for an application for AA. But often, they were intended as alternate or additive treatments for conditions for which several drugs already existed.
Beginning early in February, procedures for granting AA for new drugs will be up for debate at the FDA. Although the first meeting, which will take place on February 8, does not involve breast or ovarian cancer drugs, its goal is important for us: "the optimization of the accelerated approval process with a focus on decreasing the amount of time to confirm (or fail to confirm) clinical benefit while continuing to provide early availability of promising oncology products."
Very much on the minds of participants in this discussion will be the case of Avastin, the Genentech drug that received AA for use for breast cancer patients, which approval was subsequently withdrawn in late 2010. Withdrawal occurred because the additional trials that are required for permanent approval failed to show the same clinical benefits of the initial trial for AA, and because serious harms were revealed. Genentech is planning to appeal the withdrawal.
There are two important considerations surrounding the upcoming debate. One, clinical trials are very expensive, and confirmatory trials once a drug is on the market for an indication are not very attractive to manufacturers, who pay for them. The second is, once a drug is approved for an indication, patients are far less likely to take part in a trial in which they cannot have that drug, so confirmatory trials become more difficult.
Crucial issues that are at stake include:
What is the definition of "unmet need" required for drugs eligible for AA? The FDA could decide that AA is appropriate only for diseases for which no existing treatment is effective.
If unmet need is broadly defined to indicate people who remain ill regardless of many previous treatments, then it would be important to know why certain patients respond to the new drug and others do not. Right now, trials are not designed to yield this information.
What is the threshold benefit that must be shown for a drug to earn accelerated approval? If it is fulfilling an unmet need as defined above, the clinical benefit could arguably be less than if the drug is an alternative or an additive.
But to earn permanent approval for a particular indication, should drugs approved under AA have to meet a higher standard, and show improved OS in further clinical trials following accelerated approval? Again, the answer would seem to depend on the number of available drugs for that indication. The fewer there are, the more relaxed the standard could be. If there are many drugs available, then a contribution to OS would be more important. If the drug is an additive to another proven drug, then a very high additional clinical benefit would seem to be a justifiable requirement. That said, it must also be admitted that a great many trials fail to show better OS.
What harms if any must be taken into consideration for AA? Harms are and must be studied in initial clinical trials, and reported. If they are great, AA should probably not be granted. But if they are not highly significant, that should not relieve the manufacturer of carefully studying harms for several more years, because many bad side effects tend to take several years to emerge.
The FDA will have to address the very real problem that if AA is granted, and few side effects are known, many patients will be using the drug when the results of confirmatory studies come in. Four categories of outcome are possible, each with a theoretically appropriate action by FDA:
Drug improves OS, with minimum safety concerns. Action: grant permanent approval.
Drug is less effective than initially shown, and it has major safety problems. Action: deny permanent approval.
Drug improves OS, but has major safety problems. Action?
Drug is less effective than initially shown, has minimum safety concerns. Action?
There are question marks after the last two outcomes because it is very difficult to tell a person taking a drug approved under AA that she can no longer get it, or that because it is not approved, insurance providers will not cover it. Some people are willing to risk the most serious side effects if they believe their cancer will be controlled. Some people are willing to take any drug in the hopes that for them, even if for no one else, it will work.
This makes no sense for individuals, for money is being spent to no avail, and they may actually be shortening their lives. It makes no sense for society at large which pays one way or another for useless treatments. The only real beneficiaries are drug companies. That's why, with all patients' interests at stake, the FDA is the public protector of evidence-based treatment.
Read more about FDA and AA
Posted January 26, 2011.
« Next Post | » Previous Post
While os is certainly the gold standard, there is no reason to think that pfs can't be as important in cancer as it is in other diseases, like heart disease. Stable disease which allows a person to live out their life with little or no side effects is a goal we should at least consider in cancer therapy. The idea that people can live with cancer as they do heart disease is still in its infancy but may actually be a really good strategy, especially if the side effects are minimal and if the person is older. The "silver bullet" is as elusive as the holy grail and maybe it's time we considered alternatives that give people both quality and time rather than a cure which sometimes is so debilitating that patients don't even want to try therapy.
— Diane Paul