Portola’s experimental anticoagulant betrixaban failed a large, pivotal study among at-risk patients back in March. But the biotech says it believes the data are still good enough to win an FDA approval.
That’s going to be a difficult pitch, but the biotech today spelled out the data it plans to use to win over regulators in a study published in the prestigious New England Journal of Medicine. A regulatory filing for the treatment, which will try to go on to become the first oral Factor Xa inhibitor anticoagulant to make it to the market for preventing venous thromboembolism in medically ill patients, is now planned for the second half of this year.
By mid-morning Friday, Portola’s shares ($PTLA) were up by just a few cents.
The way this study was designed, investigators had to see a statistically significant advantage for their drug arm in a cohort of patients with elevated D-dimer levels--which is used to evaluate patients--in order to gain a clear understanding of the therapy’s benefit in a second cohort including older patients and then the entire patient population.
That didn’t happen, but the researchers went on to do an “exploratory” analysis of the data.
“Because there was no significant between-group difference in the primary outcome in cohort 1 (P=0.054), the protocol specified that all subsequent analyses were considered to be exploratory,” states the report in the NEJM. “However, taken together, these additional analyses provide support for the interpretation that the risk of venous thromboembolism was lower with betrixaban than with enoxaparin.”
The big argument in Portola’s favor was that it was close to success. In the overall patient population, there was a drop in the rate of deep vein thrombosis; 7% in the control arm and 5.3% in the betrixaban arm. There was also virtually no difference in the rate of major bleeding episodes.
“In a pre-specified subgroup of medically ill patients who were D-dimer positive, extended- duration betrixaban demonstrated a reduction in VTE events approaching statistical significance. In the pre-specified exploratory analyses of central lab D-dimer values and in progressively larger cohorts, including all study patients, the data demonstrated a consistent and significant reduction in VTE with betrixaban with no statistical difference in major bleeding between the betrixaban and enoxaparin arms,” said C. Michael Gibson, senior author of the NEJM publication, in a statement.
“The data is clearly showing that the event rates are decreased in the betrixaban groups,” Alexander Cohen, co-chair of the trial’s executive committee, told CardioBrief's Larry Husten. The P-value of 0.054 is the “sticking point,” he said, adding that “the level of uncertainty compared to what we normally accept is very minor.”
Investigators suspect that the trial failed because of variations in the way D-dimer levels were measured, Husten reports. Whether the company can make that case stick with the FDA, without first proving it in a trial, is uncertain.
In the meantime, Portola CEO Bill Lis is contesting the notion that the pivotal study failed at all, though it clearly fell short of the key endpoint.
"Firstly, the investigators do not state that the trial failed, and there’s no interpretation in NEJM or in the presentation at ISTH that this is the case," Lis said in a message to me, after the story was posted Friday morning. "The investigators believe that the dat are strong based on consistency of the result across cohorts, including central lab-based definition of Cohort 1. Importantly, there was a 20-30% relative risk reduction in the composite endpoint. Across all the cohorts there was a 30-45% reduction in symptomatic VTE and these benefits were seen without an increase in major bleeding. For the 80 mg dose, which the majority of the patients received, the results were positive in both Cohorts and in the full study population. Furthermore, there’s biological plausibility because the results are consistent with those seen in prior trials of NOACs except that net benefit is better. As you know, these large clinical trials are always analyzed on multiple criteria including the full dataset."
- here's the study