Much has been said and written regarding the FDA Advisory Committee briefing documents released last Friday regarding Amarin Corporation's (AMRN) upcoming meeting regarding their triglyceride lowering medication, Vascepa. While some were informative, and others more of the "entertaining" nature, I have not read any new literature that focuses on more than one or two specific concerns brought up in the documents, or takes certain sentences out of context. I will do my best to offer a complete analysis on the most relevant topics here.
Luckily for Amarin longs, there were no new or unexpected adverse events brought up in the documents, save for one small issue regarding the placebo choice as a Vascepa comparator. I devote some of my writing to this, in my opinion, non-issue at the end of the article.
The primary issue Amarin will have to address this Wednesday is whether or not ANCHOR data, coupled with large CV (cardiovascular) outcomes study data both new and old, merit approval for the expanded use of Vascepa to patients with high triglycerides (200-499 mg/dL). The FDA has some very legitimate concerns with this question, as recent studies have seemed to indicate that the added supplement of a TG lowering drug to existing statin therapy.
Before continuing, I will mention that I am not going to discuss safety data or statistical methodologies, as these have been discussed in the past and - particularly with safety - are of little to no concern at this point.
Cardiovascular Outcomes and the Risk to Amarin
The FDA makes it quite clear what they are going to hit at during the majority of Wednesday's panel. All one has to do is look at the two notes the panel is asked to respond to:
1. Please discuss the efficacy results from the ANCHOR trial, including the clinical significance of the observed changes in lipid/lipoprotein parameters and your level of confidence that these changes will translate into a meaningful reduction in cardiovascular risk among the target population.
2. Taking into account the described efficacy and safety data for Vascepa, do you believe that its effects on the described lipid/lipoprotein parameters are sufficient to grant approval for co-administration with statin therapy for the treatment of patients with mixed dyslipidemia and CHD or CHD risk equivalent prior to the completion of REDUCE-IT? Please provide the rationale underlying your recommendation.
The panel members are merely discussing one point and voting on more or less the same talking points. I read question two as the FDA leaning quite pointedly toward approval. The issue is whether that approval will come "prior to completion of REDUCE-IT" or after. As I outline below, the FDA antagonist case will be the recently completed CV outcomes studies showing no statistically significant CV benefit. On the surface, this ammo appears pretty strong. However, based on further analysis of the cited studies, I believe Amarin should be more than capable of defending Vascepa's sNDA.
For over a decade, the FDA has approved all lipid altering drugs based on the presumed fact that lowering lipids = lowered CV risk (see below). After some time, statin based studies solidified this presumption as a statistically significant fact. So, too, this thought has been applied to omega-3 fatty acid compounds. These studies are what paved the way for OTC fish oil to dominate store shelves and for Lovaza, Vascepa, and a multitude of other clinical products to undergo development.
This is where Vascepa hits its primary road block. The ANCHOR indication for which this sNDA is based is for the use of Vascepa "as an adjunct to diet in the treatment of adult patients with high triglycerides (TG ≥200 mg/dL and < 500 mg/dL) with mixed dyslipidemia." Studies published in the last five years have called into question the effectiveness of supplementing a statin with a TG lowering drug. Again, the FDA points this concern out.
Unbeknownst to most, if not all, investors was that during Amarin's pre-IND meetings with the FDA in 2008 regarding their trials, the FDA explicitly expressed concern regarding the lack of controlled, non-open label studies demonstrating the benefit of combo statin + second TG drug. Based on the language below, the FDA planned to take into account some specific ongoing outcomes trials (ACCORD-Lipid, AIM-HIGH, and HPS2-THRIVE, all discussed later) to assist with their indication preference for Vascepa + statin. As such, the FDA wanted certain data as well as a CV outcomes study underway - both of which Amarin has provided.
Amarin Cardiovascular Outcomes Study - REDUCE-IT
The CV study known as REDUCE-IT is nearly fully enrolled (over 6,000 of 7,990 as of late September) and is under a special protocol assessment (SPA) with the FDA much as ANCHOR was. A quick note on SPAs - unlike what many folks are saying, an SPA can be broken. Although infrequent, Amarin outlines it quite nicely in their 10K below. Investors should note that SPAs are "generally binding" unless substantial new information is presented relating to safety or efficacy. The recent CV outcomes studies do, on the face of it, sound as though they may fit the "substantial effect" efficacy claim of a statin + EPA combo. I discuss these in the next section.
I personally believe that Amarin has crafted their REDUCE-IT trial to their advantage. This merits further study; however, unlike the FDA cited failed CV trials, the REDUCE-IT patient profile is substantially different enough to give me faith that they will see added CV benefit with a statin + EPA combo. Much of the result depends on how Amarin doses the statin and what "optimized" levels actually translate into.
FDA Cited CV Trials - A Discussion
Up to this point, I have sounded rather downbeat regarding the studies the FDA cites in the briefing documents. Let me be clear in that while the FDA raises legitimate concerns, I do not believe them substantial enough to merit waiting for REDUCE-IT results to approve the ANCHOR indication. The FDA language regarding these concerns is highlighted below.
There are two "groups" of studies which the FDA speaks to. The first is omega-3 fatty acids and cardiovascular outcomes, and the second includes the ACCORD-Lipid, AIM-HIGH, and HPS2-THRIVE study outcomes listed above. I begin with the second group.
The ACCORD-Lipid study was published in 2010 in the New England Journal of Medicine. The study looked at a ~5500 person sample of "middle-aged or older people with type 2 diabetes who are at high risk for having a cardiovascular disease event, does a therapeutic strategy that uses a fibrate to raise HDL-C/lower TG levels and uses a statin for treatment of LDL-C reduce the rate of CVD events compared to a strategy that only uses a statin for treatment of LDL-C." It is important to note that all three studies included in this second group of studies used statin supplements that were not EPA, DHA, or any combination of the two.
The ACCORD study group utilized fenofibrate, a drug compound currently marketed under the name Trilipix. Fibrates lower lipid levels much to the same effect of EPA, but in a completely different mechanism. Simply, this drug stimulates the production of an enzyme that "eats" the lipids (lipolysis) in one's plasma. The primary outcome was the first occurrence of a nonfatal myocardial infarction, nonfatal stroke, or cardiovascular death. There was no statistical difference between the placebo + statin and fibrate + statin group.
Fenofibrates were also studied in the Fenofibrate Intervention and Event Lowering in Diabetes (FIELD) trial (of which ACCORD was partially based) in which fenofibrate did not significantly reduce coronary heart disease death or nonfatal myocardial infarction, both endpoints in that trial. It is not clear from my limited documentation of the FIELD trial, but upon performing subgroup analysis of the ACCORD-Lipid study, a statistically significant improvement in the MACE (nonfatal MI, nonfatal stroke, and CHD death) primary outcome did occur in the population most similar to ANCHOR's indication. This is quite interesting, especially when you consider that this sub-group was 15% of the overall population of ~5500, which is a respectable sample of nearly 900. The FDA reviewer does point this out, albeit with caveats.
The FDA clearly recognizes the possibility here, but obviously cannot base a decision on sub-group analysis without a proper study. The EMDAC panel that looked at the ACCORD data recommended against any label changes, but as one can see in the last sentence recommend the FDA take a stronger stance for lipid altering drugs relying on CV outcomes studies. I believe it was this specific recommendation that has resulted in Amarin's Advisory Panel for ANCHOR. This last sentence is also what I believe is giving investors the most concern. Remember, though, that Vascepa is not a fenofibrate, so any significant reliance on this study would be misguided by the panel and the FDA in my opinion.
The second study referenced in the ad-com documents is the *clears throat* Atherothrombosis Intervention in Metabolic Syndrome with Low HDL/High Triglycerides: Impact on Global Health Outcomes or "AIM-HIGH" (like they did with this acronym). Instead of fenofibrates, AIM-HIGH employed a niacin + statin combination with primary endpoint of first event of the composite of CHD death, nonfatal MI, ischemic stroke, hospitalization for an acute coronary syndrome, or symptom-driven coronary or cerebral revascularization. Niacin functions quite different from both EPA and fenofibrates, so again the ad-com and FDA need to heed this fact and determine whether they can make serious enough comparisons to warrant going against the in-place SPA and require Amarin to wait for REDUCE-IT results. Thestreet.com author Adam Feuerstein has been critical of my comparing Lovaza and Vascepa to each other, so I do hope he has viewed these CV studies with the same critical eye, especially considering he is negative on the FDA panel.
The trial was stopped 3 years in for lack of efficacy. As with ACCORD, a subset of patients with TG profiles more closely resembling ANCHOR showed indications of combination benefit. Again, the reviewer correctly notes that these data cannot be relied upon without further trial. The results remain very interesting nonetheless.
The final study included here is the HPS2-THRIVE study, in which patents either ingested specially formulated extended-release niacin combined with the anti-flushing agent laropiprant or placebo on background simvastatin therapy. The FDA spends little wording on this study, as the full results from the study have not been published. From some interim analysis presented at a recent conference, it appears as though the results will be as ineffective as the other two studies discussed here as well as no indication of sub-group benefit. For the third time, however, the compound in question has no similarity to EPA in form or function. I simply do not see how the FDA can make the jump to placing significant reliance on CV studies of completely different compounds, especially when an EPA outcomes study exists, albeit a poorly controlled study in comparison.
I do not believe this study will be referenced much in the panel. Instead, it will be ACCORD and AIM-HIGH as well as JELIS and GISSI-P, which is up next for discussion.
Omega-3 Cardiovascular Outcome Studies
The omega-3 fatty acid (FA) group of studies the FDA refers to are below. Note that only one study, JELIS, uses pure EPA dosages in their study and in any dosage amount close to Vascepa's REDUCE-IT/ANCHOR/MARINE (all had 4g dosages) trials. The DOIT study doses at 2.4g a day, but does not report vital data such as LDL-C and TG levels to be very helpful in comparison. This is why Amarin has focused on the JELIS study in nearly all presentation materials and will surely try to keep the focus on JELIS come Wednesday.
Just as there is a marked drop in efficacy for the 2g dosage of Vascepa vs. 4g, one would expect further degradation of results at 1g, which nearly all of the other studies dose at. In fact, at 1g EPA/DHA mix, it is reasonable to believe that no statistical CV benefit would exist at all. And, in fact, one generally does not. GISSI-P is the exception to this rule, but as noted below, very little patients were on statins at baseline measurements, so any omega-3 effect would most probably be amplified.
A multi-study comparison done by Kotwal and associates assessed CV outcomes for total population, and unsurprisingly found no statistically significant correlation as a whole. However, upon further sub-group analysis, the same trend as was mentioned earlier is noted in the following table under "Mean Trig." Patients with over 1.70 mmol/L (~150 mg/dL) mean triglycerides clearly saw benefit with treatment.
As the FDA points out below, JELIS is the only trial listed that demonstrated a positive treatment outcome when added to a low-dose statin. The FDA offers that this could be due to bias introduced due to the open-label nature of the study. I believe the positive results are due to the following:
- JELIS is the only trial to use pure EPA as well as dose it in some form close to 4g.
- While open-label, the patient sample is 18,645, so I find it hard to imagine that much bias could persist among the entire group for such an extended period of time.
- JELIS patients were on low dose statins, so some EPA specific benefit would be expected over completely optimized statin dosage, strengthening the idea that EPA does have CV benefit.
The bottom line here is that the only leg to stand on for the FDA is the CV outcome question. From my analysis above, I do not believe there is compelling enough evidence to warrant a substantial efficacy question, which would otherwise force the FDA to wait for REDUCE-IT data. Some will read into the data in the exact opposite way from me, but I do not see how relying on studies of completely different compounds or similar compounds at ineffective dosages can pose enough of a concern so as to discredit a large and much more similar, albeit open-label, study in JELIS and force a renege on an agreed upon SPA.
Mineral Oil Used as Placebo
Another question raised in the FDA briefing documents concerns the placebo choice for ANCHOR. The ANCHOR placebo group noted a marked increase in lipid levels, more than what would be expected. This does not mean that the data are erroneous, but simply calls them into question. Amarin chose mineral oil versus other oils due to the inert nature of mineral compared to vegetable oil, as explained below.
Looking back at different omega-3 based trials, however, one notes variability in placebo results such that the ANCHOR results appear reasonable. The amount of mineral oil ingested in the placebo group would not be anywhere near the levels needed to cause any physiological effects. The FDA agreed on the oil as placebo, and its common alternative use is as a laxative in dosage ranges more than 4 times higher than taken by placebo patients. I see no cause for concern as to the use of mineral oil and the effect on the ANCHOR trial.
As an aside, in my research as to whether or not mineral oil could be absorbed into the body and affect lipid profile, I stumbled upon this gem from the 1975 issue of the South African Medical Journal. It has a bit of dark humor to it - a man does die from ingesting too much mineral oil - but "he was then noted to be excessively preoccupied with his bowel function." Lesson to all - do not ingest what had to be enormous amounts of mineral oil, no matter how much you want to "cleanse."
There is little doubt that Wednesday's ad-com panel is highly anticipated. The best one can do is trust in his or her research, and that is what I am doing now. I wish everyone luck going into the panel (although I wish longs a bit more luck than shorts!).
Additional disclosure: I initiated a hedged LONG position on 10/14/13 in Amarin and intend to hold through the FDA Advisory Panel. The thoughts herein are not an endorsement to buy or sell, simply my thoughts disclosed publicly.