I do research with an expectation that my findings will offer an objective review for potential FDA approval. Not just for me, but for those who might be interested in what I do. So, in turn, I give both sides of the coin as fairly as possible. And despite all of our earnest research, we still see promising drugs fail. I constantly ask myself what exactly the FDA wants in their drug trials. What makes a good drug?
My final conclusion/reasoning may not suffice for some of you out there, and in different shoes perhaps, I would feel the same. But hopefully my thought process is clear and concise enough to leave at least a parting thought.
Before I talk about Portola Pharmaceuticals (NASDAQ:PTLA), I want to talk about study design. Here's some more reading for those interested in the topic.
It takes a good study to power results that are statistically significant. For some drugs, this may require thousands of volunteers, and for others, significantly less. But significance is akin to meeting the GPA requirement to apply to your favorite school. The FDA has and will continue to deny approval for things much more nuanced than statistical analysis.
Clinical endpoints drive relevance. Endpoints for research papers have quite a bit of leeway in this regard because they're meant to teach and discover; not cure. Curing a patient effectively and consistently is the only thing the FDA wants to know, and it wants to know it without a doubt.
After going through these initial entry steps, we move onto things like the numbers behind the data. Sure, the drug's numbers might be statistically significant, but do they show unequivocal regard towards an actual treatment?
While reading the rest of the article, I ask you to keep in mind this one question: Does this drug, without a doubt, provide treatment?
Portola's drug trial for Andexanet alfa (AAlfa) has a very strong showing of statistical significance. Without a doubt, AAlfa reverses coagulation inhibitory factors during the time of clinical action. This action period is very sensitive because the drugs Apixaban and Rivaroxaban can cause fatal internal bleeding, which must be resolved within minutes.
Before I continue expanding on the above statement, let's go over the data. I honestly don't know if I'm allowed to post images from the study because it specifically says for personal use only. So... I'm going to cite information without actually showing it. Here's a link to the study, which you need special access to read. The quotes and data I talk about in here are all accurate as per what the study has written.
AAlfa does indeed reduce Anti-Factor Xa activity "within 2 to 5 minutes after administration" to extremely high statistical significance. I expressed some initial concern over how much the drug stimulates an exaggerated overcompensation effect and increases Anti-Factor Xa activity when compared to the placebo. This only occurs during the 3-4 hour mark, which can be argued is well beyond the moment of urgency AAlfa specifically poses to address.
Overcompensation (compared to placebo) does not seem to occur in a significant way when AAlfa is given to patients who took Apixaban. In the Rivaroxaban leg, however, it does seem to be of significance. This overcompensation (for Rivaroxaban) does not seem to occur in a convincing manner in the "Bolus only" treatment leg, but almost assuredly does in the "Bolus + Infusion" leg. The problem with this though is that the "Bolus + Infusion" leg for both Apixaban and Rivaroxaban offers much stronger reduction in Anti-Factor Xa activity during the period of clinical significance.
Both of the "Bolus only" legs see a linear increase in Anti-Factor Xa activity right after the treatment is administered. This value approaches the placebo much quicker in the Rivaroxaban leg, but does so nonetheless for both drugs. We can therefore safely conclude that "Bolus + Infusion" is a much stronger treatment protocol than "Bolus only" because "Bolus + Infusion" shows an almost complete flat-line in Anti-Factor Xa activity before the one-hour mark.
The study doesn't talk about this overcompensation effect. Aside from the graphs, there is no mention of potential risks associated with increasing Anti-Factor Xa after the initial two-hour treatment window.
Portola Pharmaceuticals chose an interesting endpoint. Going back to my first paragraph, yes, AAlfa is very significant in reducing what the study tests for. But here's a very important quote that should explain where I'm going with this:
"In this report, we do not present data on the efficacy and safety of andexanet in patients who require urgent reversal of factor Xa inhibitor activity because of bleeding or for emergency surgery."
Given all of our data, we can assume a lot of things. Yes, AAlfa reduces Anti-Factor Xa activity by a huge margin. And, yes, it does so in the correct time period it is needed. But does this reduction directly correlate to a reduction in bleeding? In essence, the true clinical endpoint for this study is a statistically significant reduction in bleeding.
Directly from the study:
"Furthermore, the high frequency of coexisting conditions among patients with acute bleeding makes it challenging to determine whether possible complications, such as thrombosis or death, are related to the reversal treatment or the underlying medical illness. Although the extent and duration of factor Xa inhibitor reversal that is required to achieve hemostasis in patients is unknown, studies in animals of the use of andexanet to control bleeding have shown that..."
To note, animal models don't count. I'm confused why the study even mentioned that in there. It's completely pointless for human trials.
So now we're left with these points:
- With no prior knowledge, can we conclude that reversing Anti-Factor Xa activity reduces bleeding.
- In patients who require urgent treatment (which is the demographic this drug aims to provide treatment for) is Anti-Factor Xa activity the only and most significant cause of bleeding.
- Does the previously mentioned overcompensation effect of Anti-Factor Xa activity have a negative impact on patients in urgent care.
- If we choose to utilize the "Bolus only" treatment leg, does the linear increase in Anti-Factor Xa activity during the period of clinical significance present a negative impact for patients in urgent care.
Now go back to the question I first posed. Does this drug, without a doubt, provide treatment?
In my opinion, no it does not. The study leaves a lot in wanting. Think of FDA approval like a court case. You have to prove without a doubt that someone is guilty. In this case, the company has to prove that a drug provides the treatment it's applying for. Portola Pharmaceuticals hasn't shown that to me.
Disclosure: I am/we are short PTLA.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.