Portola Pharmaceuticals Inc (NASDAQ:PTLA) just held a conference call to discuss its latest earnings, and on the call, reiterated its intentions to submit a new drug application (NDA) to the FDA in an attempt to get approval for its experimental anticoagulant betrixaban. The drug failed to achieve statistical significance in a key part of the phase 3 on which the NDA will be based, and when news hit earlier this year of this failing, Portola took a real hit to its market capitalization. The company is still down on its pre-release valuation, and we believe there may be an opportunity to take advantage of wider market misinterpretation of this so-called failure and get in at a discount ahead of what could be an unexpected FDA decision. Here is why.
First, a quick look at the drug in question. As mentioned, it is called betrixaban, and it is what’s called a Factor X inhibitor. Factor X is an enzyme that plays a key role in the coagulation cascade, which is the process through which blood coagulates, or clots. By inhibiting this enzyme, betrixaban reduces the blood’s ability to clot, and in doing so, is theoretically an effective target against a condition called thrombosis, which is life-threatening blood clotting. Portola is trying to get betrixaban approved for thrombosis in acute medically ill patients – i.e. patients with very severe medical conditions such as stroke, heart attack or infection.
So what happened in the trial?
The company enrolled more than 7500 patients across 450 clinical sites worldwide, and set up to pitch its betrixaban candidate against a current standard of care treatment called enoxaparin. The goal was to demonstrate that betrixaban could reduce venous thromboembolism (VTE) events to a higher degree than could enoxaparin, and it did, when looked at across the entire trial population. However, Portola had set up its trial so that it would address two groups of patients within the overall population – those with elevated D-dimer levels and then those with elevated D-dimer levels who are also 75 years old or older. In both groups, betrixaban beat out enoxaparin, but in the first group, the age irrelevancy group, the trial came out with a P value of 0.054. In biotechnology, and statistics in general, 0.05 is generally considered the gold standard for statistical significance. Anything above this, and a trial’s outcome is regarded as not statistically significant. The overall population and the age restricted population both scored P values of lower than 0.05, and so the reduction in VTE events induced by betrixaban in these groups was considered statistically significant. However, due to a quirk of the trial set up (and a mistake for which Portola management are now likely kicking themselves) the first cohort needed to demonstrate statistical significance in order for the second cohort to count as anything but exploratory. It didn’t, and so the efficacy data goes on record as exploratory.
The question is, is this a drug fail, or just a trial design fail? We believe it is the latter. Without the enriched setup, the drug would have met its primary endpoint and markets would probably be buying into Portola ahead of the planned NDA submission. Instead, markets have sold off on the company, and it has somewhat fallen off the radar. There’s also the fact that while 0.05 is considered the gold standard, 0.054 is not all that far off, and these P values are not firm thresholds. They are set before an investigation is carried out, and for whatever reason 0.05 seems to have become the most widely regarded as indicative of statistical significance. Just because a trial as a P value of 0.05 doesn’t necessarily mean the drug in question didn’t cause the measured response.
All this – in our eyes – translates to an opportunity. Markets are looking the other way, and haven’t really read into the data other than interpreting the negative headlines as indicative of a drug’s failure. For a speculative biotech investor looking to get in at a discount, this is a great opportunity. Yes, there’s still risk involved. The FDA may request further data, and this would involve cost and time, but as a speculative allocation, it looks intriguing.