Texas' and Washington's Mifepristone Ruling
Did the FDA misuse accelerated approval?
Mifepristone (RU-486) is a synthetic steroid that is given in combination with prostaglandin (misoprostol) and can induce abortion. On Friday, two legal decisions put the fate of mifepristone in question in the United States.
I have read both court decisions, many supporting documents, pulled old drug labels, and tried to summarize the regulatory issues, as best I can. This essay will be divided into 5 parts.
The regulatory history of mifepristone
What are the regulatory issues at stake in this case/ How does it affect the FDA?
Should the FDA be second guessed by judges?
Bonus: one interesting legal question
I attach the full text documents of the opinions
I: The Regulatory History of Mifepristone
Mifeprisone received FDA’s Accelerated approval in 2000 for, “the medical termination of intrauterine pregnancy through 49 days’ pregnancy” The drug was supported by data from three uncontrolled studies (2 french trials and 1 US study), where consecutive women took the drug. 92% of 800+ US women had a complete abortion, and, per FDA package label, there were “65 women (7.9%) who received surgical interventions: 13 (1.6%) were medically indicated interventions during the study period, mostly for excessive bleeding; five (0.6%) interventions occurred at the patient’s request; 39 women (4.7%) had incomplete abortions at the end of the study protocol; and eight (1.0%) had ongoing pregnancies at the end of the study protocol.”
Soon thereafter, FDA added mifepristone to REMS, and required 3 office visits (one before, one to give the medicine, and one after) to get the drug.
In 2016, the FDA:
Increased the time where it could be given from 7 wks (49 days) to 10 weeks (70 days)
Eliminated the requirement to report non-lethal safety events
Permitted non-doctors to prescribe the product
Allowed a second dose of the medicine
Allowed the drug to be given outside the office
Allowed for sooner misopristol administration afterwards
Reduced the required office visits from 3 to 1.
In 2019, the FDA modified the REMS requirements.
In 2021, the FDA allowed mail order pharmacies to dispense the medicine by mail. The plaintiff has a list of concerns of all points of this regulatory history.
II. What are the regulatory issues at stake in this case?
The plaintiff argues that the FDA’s 2000 approval violates Accelerated Approval.
The language of AA states that a drug must:
Have been “studied for [its] safety and effectiveness in treating serious or life-threatening illnesses.” 21 C.F.R. § 314.500.
And it must also “provide [a] meaningful therapeutic benefit to patients over existing treatments.”
FDA concedes that pregnancy itself is not a serious illness, but points out that unwanted pregnancy can cause depression and depression is a serious or life-threatening illness. The Texas judge notes that by this logic nearly anything can be a serious of life threatening illness if it causes depression.
The Texas judge argues that the condition must be in itself life threatening, and the life-threatening condition cannot be not a downstream consequence of it. But, this is an odd distinction. A medical issue/ concern is linked to all downstream consequences. For instance, one argument against widespread PSA testing is it turns people who would not die of prostate cancer into patients, and this has harms. For instance, a diagnosis of prostate cancer may increase suicide. As such, all cause mortality is a better endpoint of screening trials. If an over-diagnosis leads to suicide, you don’t get to dismiss the suicide by saying that is not a consequence of prostate cancer, per se.
Similarly, if you sold a test for Huntington’s disease that was totally inaccurate, a lot of people may end up depressed (you just told them they have Huntingtons!). You can’t say you don’t own that consequence since it was not Huntington’s per se. After all they don’t even have it! At the same time, I do think Accelerated approval is generally overused and we have detailed that in cancer medicine here.
Of course, some pregnancies can be dangerous for the mother— and this would clearly be life-threatening— but restricting to those scenarios would limit prescribing, and we cannot always tell by week 10 which pregnancy will be life-threatening.
The second argument is tricky. Mifepristone must “provide [a] meaningful therapeutic benefit to patients over existing treatments.”
The Texas judge argues that medical abortion was (a) not tested against surgical abortion and (b) must be inferior to surgical abortion b/c surgical abortion is more successful. Ergo, even if one defines abortion as a meaningful benefit (The Judge allows this hypothetical even though he rejects it), the pill fails to beat the procedure.
The problem with the judge’s interpretation can be clarified with a hypothetical trial design. Imagine a 3 arm study including women who want to abort in the first 10 weeks:
medical abortion only,
surgical abortion only, or
medical abortion with surgical abortion back up.
The judge’s point is surgical abortion (arm 2) will beat medical abortion (arm 1), and that might be true. But, it is entirely possible the 3 arm does the best in terms of quality of life, pain, discomfort and success rate and beats arm 2. Ergo, there is a meaningful benefit to having the medical abortion available, as part of a sequence strategy. The judge’s error is thinking this is either-or and not a daisy chain strategy.
However, in support of the Judge’s point, no one has actually ran this trial, so it is speculation. In fact, had they ran this trial, the drug might have regular approval— and skip all this accelerated approval debate.
One unfortunate aspect of the Judge’s opinion is reliance on anecdotes and not data, but I suppose that is why these are lawyers and not doctors. The Judge makes the additional argument that there are adverse events and some can be bad with mifepristone. Sure, everything in life has adverse events, but you have to 1. quantify them and 2. compare them to the rates of adverse events by not having the intervention. If women without any access to abortion still seek illegal abortions in high enough number, the adverse events (from back-alley abortions) could dwarf mifepristone.
Should judges second guess the FDA?
Joe Ross makes this point that this case means that the FDA can be questioned endlessly. Should this be permitted? Is this bad or good?
I need to sleep on this question some more. Ultimately the FDA was created by laws— the purview of politics— although it enjoys wide statutory freedom. When it comes to cancer and rare diseases, I and many doctors (probably including Joe) believe the FDA is out of control. They have abused Accelerated approval and surrogate endpoints.
They rubber stamp dozens of drugs with no evidence they help Americans. Just think about aducanumab, and exondys, and belantumab, and melflufen, selinexor (boston trial horrible), and olaparib (polo trial is disgraceful). I have written two books critical of FDA.
If the legal system cannot say the FDA violated the law in misusing their regulatory authority, who can? Is the FDA unaccountable to anyone who doesn’t work there? That can’t be. Meanwhile, we probably don’t want a system where the only person who can shape the FDA is the commish— who is appointed by the president. We need a system that can say: FDA’s civil servants should have a lot of authority, but when Peter Marks abuses EUA authority to approve a bivalent booster in toddlers who already had COVID19— shouldn’t he be able to be challenged?
I am not sure what the solution is, but the answer: whatever the FDA says cannot be questioned even if it bends the laws that gave it authority in the first place— just can’t be the right answer. Ultimately, the FDA is accountable to the American people— either through congress, the executive branch or the courts. My immediate reaction is: I need to think about it more. What is the best system to make sure the FDA correctly uses their legal authority? It is clear to me— that one does not have even have to mention abortion— to think of dozens of examples of the FDA misusing their power. The FDA cannot be above the law, but who should oversee it?
IV: One interesting legal argument
Does the Comstock act prohibit mailing of mifepristone?
The act apparently prohibits “[e]very article or thing designed, adapted, or intended for producing abortion, or for any indecent or immoral use; and [e]very article, instrument, substance, drug, medicine, or thing which is advertised or described in a manner calculated to lead another to use or apply it for producing abortion, or for any indecent or immoral purpose.”
Will be interesting to see how this shakes out in courts.
IV. Attached documents
Thanks for the crisp, clear summary!
Does accelerated approval vitiate the normal requirements for full licensure?
Seems like a more robust clinical trial could and should have been completed over the past couple decades. There’s an obvious benefit to medical abortion, but short term data on hundreds of women isn’t ideal. By now we should have long term controlled trial data on tens of thousands of women. What’s the regulatory argument otherwise?