The 2 draft questions are as followed:
Within the briefing documents provided by the FDA, there appears to be 2 key issues that will dictate the discussion on October 16th. The first issue raised was whether or not the mineral oil placebo directly led to an adverse performance in the changing lipid/biomarker parameters (TG, LDL, etc...) over the course of the ANCHOR trial in the control arm. The second, and, in our opinion, only important issue was whether or not the FDA should wait to see positive results from the REDUCE-IT trial (the Cardiovascular Outcomes Trial) before they consider approval of Vascepa in the high TG (>200 to <500 mg/dL) population. Recall that Vascepa was already approved in the severe TG (>500 mg/dL) population, which was supported by the MARINE trial results.
REDUCE-IT. To wait or not to wait, that is the question.
As we were writing this article, another was posted by Ross Klosterman called "Will Cardiovascular Concerns Jeopardize Amarin's Upcoming FDA Advisory Panel?" It explores the issues surrounding whether or not the high TG indication can be approved before the completion of the REDUCE-IT trial. We cannot do it any better, and so we encourage investors to read his evaluation.
The competing hypotheses
The focus of this article is to explore concerns regarding the placebo used in all of the Vascepa trials, but first, a brief introduction to Occam's razor. In short, Occam's razor states that among competing hypotheses, the simplest hypothesis should be selected. We feel that this is an important principle for panelists to keep in mind when they discuss the mineral oil placebo issue.
The FDA wants to make light of the possibility that the mineral oil placebo is not inert, and that it may have contributed to the adverse performance in the lipid/biomarker profiles of the control arm. Indeed, this possibility drives much of the short thesis leading into the advisory hearing.
However, an active placebo is not the only possible explanation for the statistical differences in the lipid/biomarker parameters among placebo patients. When we look at those other possible explanations, we are convinced that the panelists will sufficiently dismiss the FDA's concern.
Bulls argue that unknown factors could have adversely affected lipid and biomarker performances amongst all the treatment arms. This would have negatively affected the control and treatment arms equally.
When considering the eventual approval of Vascepa for the severe TG indication, the FDA noted similar adverse lipid/biomarker performances for the mineral oil placebo arm in the MARINE trial. Given the absence of any evidence that mineral oil was the cause, the FDA concluded that the likely culprits were unknown factors that were likely evenly randomly distributed amongst all treatment arms. Such factors could be that some patients changed their diet, physical activity, statin therapy, or independently noted lacking benefit from treatment, etc...
However, in the absence of supporting evidence regarding those previously mentioned factors, the FDA has exhibited reluctance towards extending such considerations towards their assessment of the ANCHOR trial.
Fortunately, there are other possibilities, which are all mentioned in the briefing documents. One scenario arises from the non-random selection of patients to be randomized into the trial. 70% of prospective patients failed in the "approval" process. The mean lipid/biomarker values for all the prospective patients (those that were allowed to enroll, and those that weren't) may have led towards a regression towards the mean among all the enrolled patients. In other words, given the trial requirements for enrollment, there may have been a higher proportion of patients with "high" outliers for TG (as opposed to "low" outlying), and "low" outliers for LDL-C (as opposed to "high" outlying). The theory is that the skewed representation of outlier patients could potentially lead to greater magnitude changes in the control arm. This would suggest that, perhaps, the 4-week stabilization period was too short for this patient population. Regardless, all the treatment arms would have been negatively affected equally.
Another possible scenario to consider is that the trials enrolled different patient populations, and/or had differing conditions, than those examined by the FDA in similarly run trials. Biotech investors are always reminded to be hesitant when comparing placebo results from different trials, and we see no reason why that principle should be abandoned when assessing Vascepa's clinical results.
On the other hand, the bears suggest that concomitantly ingesting mineral oil with statins could possibly inhibit the absorption of the statins. The clinical data supporting this possibility can only be found in the early 1940s, when a trial concluded that mineral oil may possibly hinder the absorption of fat-soluble vitamins. However, the FDA investigator quickly reminds us that such findings were contradicted by later and far more recent studies, some of which utilized large quantities of mineral oil (up to 150 mL/day). For comparison, ANCHOR applies mineral oil placebo up to 4 mL/day over a period of 2-4 separate doses per day.
Another bear argument has been that mineral oil itself led to the negative control arm performance during the trial. This possibility has been discredited by numerous historical studies that showed zero statistically significant changes in lipid/biomarker parameters in trials that used mineral oil. To suggest that a small amount of mineral oil has just now started to impact those parameters is contradictory to a mountain of clinical data. It is highly unlikely that mineral oil would ever find itself to be considered inert, and sufficient for use as a placebo, by the FDA if it actually had an impact on lipid/biomarker parameters. In the remote event that mineral oil placebo were shown to be the culprit, it will be due to interference with the absorption of statins, and not from any other clinical activity.
Aside from a dated 1940s trial, there has been zero biological or chemical rationale provided by the FDA for an interaction between the statins and mineral oil ... Nevermind that it could get to the point where the latter could hinder the absorption of the former. Anyone who genuinely thinks this argument should be considered anything more than a "possibility" is attempting to create a causal relationship between both of these entities on the basis of correlative data alone.
We do not think the FDA is going to base their decision on this mineral oil placebo anyways. They are just doing their job by scrutinizing the data, but it's important that the panelists are not misled into thinking ill of the ANCHOR trial results by the misleading hypothesis. The treatment arm performances alone defends the efficacy of Vascepa in a patient population that has no other alternatives to reducing TG without dramatically increasing LDL-C.
We feel that applying Occam's razor to the mineral oil issue sufficiently justifies the explanations given by the FDA, with regards to their assessment of the MARINE trial, and by Amarin Corp, with regards to the non-random selection process. Even in the unlikely event that the panel were to reject the efficacy data seen in the ANCHOR trial, due to the mineral oil placebo issue, we do not think it would impact the FDA's decision with regards to approval since they have already previously approved Vascepa based off another trial that used the same placebo.
Note: We do not think the SPA is relevant in this issue. Sure, the FDA approved the placebo, but they are well within their boundary to "change the goalpost" in light of new evidence. The FDA does not have this evidence, and thus we suspect the mineral oil placebo will have no more of an impact on the approval decision for the ANCHOR indication than it did for the MARINE indication.
Given that the negative market reaction to the briefing documents were heavily dependent on the 2 issues mentioned in this report, we feel that the sell-off was unjustified but has now provided a relatively handsome risk:reward ratio leading into the advisory committee date. Therefore, we initiated a long position on October 14th to capitalize on the opportunity.