Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Yes, unless someone can show me a trial in patients with known CVD or diabetics with multiple CVD risk factors that has a placebo rate lower than about 4.4%.
I got it now - you are talking about data totally separate from the poster they presented. Not sure if they have enough V clinical data yet in insurance databases to do FU analyses, but that would be an amazing finding if that is what they report.
JL - They have already released exactly that RWE information multiple times in the past 6 months (TGs increase risk of MACE events 25-36% in patients like R-It patients). This poster is nothing special.
AMRN has reported virtually identical data 3 times at conferences recently without a single conference call. The only difference with this poster (and posters are minor things at conferences) is that it focuses on PAD, the one R-It stratum 1 enrollment group not previously addressed by RWE data. Nice data but NOT worthy of a CC. It is something else. It MIGHT be to discuss ASCEND results, but what do they achieve by doing this? The PPS will increase without any further communication about this if (when!) R-It succeeds. I posted this morning the same thing Gabor did - I think they have the CC as a placeholder because they are analyzing the data and are not sure they will be ready by Monday. If the analyses are done, they announce them and nobody will complain that it was off topic.
Maybe they are analyzing results this weekend and are not sure they will be done by Monday morning. CC may be a placeholder that will let them discuss R-It topline if done, and if not, they have an opportunity to put the almost assuredly negative ASCEND results being presented tomorrow in context.
Ex - I agree with you on that.
Consonance dumping the rest of their shares since there has been no run up?
Just based on 8/14 disclosures. Should have said last quarter.
TTE - I think the smart money already has all the shares they want and is not buying more. I don't see that as a negative. Baker Brothers still hold all 22 million, Fidelity just bought 2 million, and even Consonance (who sold big) still holds 4-5 million shares.
Kiwi - I think a legitimate question re: applicability of JELIS to R-It results not talked about recently has to do with the MACE components in JELIS. If you look at the pub, the largest % of events by far were apparently "soft" unstable angina events. The p value in JELIS was driven heavily by this type of MACE event, and my understanding is that R-It has a stricter unstable angina definition. On the other hand, the other MACE components in JELIS that are in the R-It definition trended the same way, but the number of events were just too small to be significant individually. Numbers of these events in R-it will be much higher than in JELIS by definition (it is an event driven study). What I take from this is that JELIS is informative, but not conclusive My opinion, for whatever that is worth, is that JELIS supports R-It results being positive, but I am not expecting an RRR >30% because I think this would have been high enough for at least some key SEs to be significant at the 80% interim analysis (and would have led to a stop).
I think that "Phase 4" studies are trials done post-FDA approval.
Zip - DMC has no role anymore. AMRN biostats people will do the analysis. It is now all in house once data are unlocked.
Kiwi -
I work with people like Dr. Budoff all the time. "Supportive but not conclusive" is scientist-ese for "I believe there is an effect but I am unwilling to commit 100% because I don't want to look stupid if I am wrong."
Agreed that it is biased - it is the only possible comparison that is not confounded by crossover though.
AHP - I believe that the most interpretable pure OS analysis will be a true placebo-controlled analysis, comparing those assigned originally to DCVAX to those assigned to placebo who did NOT crossover. Good comparison but the drawback is low statistical power because of the small number of non-crossovers. The others possible are overall OS and milestone analysis (e.g., % surviving at 2 years, 3 years) in those originally assigned to DCVAX compared to historical controls. They might then also look at everyone who received DCVAX whether originally assigned this or on crossover, compared to the non-crossovers and to historic controls, but these analyses are more difficult to interpret. If DCVAX works in those originally receiving this treatment but is not effective for crossovers, this would make results easier to interpret. I disagree that analyzing crossovers as if they were receiving placebo the whole study is methodologically appropriate, and I don't think FDA advisors would think this is meaningful either. You analyze as assigned to address dropouts and noncompliance, but I don't think that applies to crossover designs.
Kiwi - I am on record for believing in the science (balanced evaluation of positives and negatives). My emotional reaction is in response to the implications of the scientific evaluation (because I have money invested). I wouldn't still be invested if I thought likelihood of failure was high.
Just my opinion, but I think V is not a sexy novel synthetic drug like many small biotechs are testing and which seem to trigger huge gambles and swings in price. It is "just" a high dose, high purity derivative of fish oil. I think the fish oil derivation of V makes some retail people discount the stock (until proven otherwise). However, sexiness of the drug has no bearing on its potential for success. Looking at the 13-F institutional holdings as of 8/15 is informative. Most of the recent downward pressure on the stock price is clearly from Consonance (the second largest institutional holder) unloading 13 million shares. I do not believe this is due to any inside information on R-It failure - they still have almost 5 millions shares which they clearly would not risk if they knew R-It would fail. The largest institutional holder, Baker Brothers, continues to hold over 22 million shares, so they do not expect failure either. Just eyeballing the 13-Fs, it seems that apart from Consonance, more shares were added than sold, with relatively large additions by Fidelity (+2 million shares), T Rowe Price (+1 million shares), Millenium (+1 million shares), Citadel (+750K shares), and Adage (+750K shares). There also seem to be many additions of calls (and substantially fewer puts). Taken together, I think this shows that "the street" remains interested in AMRN and at least somewhat confident in R-It results, but clearly has invested all they want to. Now, there are not enough retail buyers wowed by the sexiness of AMRN and V to bump up the price. Cannot wait until this proves to be a very bad decision
Chas - Thanks for your many interesting posts and the basis you describe for what you expect in R-It. I appreciate your optimism for outstanding RRR, but as a scientist I am always leery of expecting things to work out as "too good to be true." The real world rarely cooperates in my experience. Your line of argument and the one in my original post at least make me feel more confident that we will get a significant RRR, and if it happened, >20% would be icing on the cake.
Yes - Linear effect, so the fewer risk factors you have the less relative MACE risk you have (relative to a healthy population) if you are diabetic. Duuuuh....
Kiwi - Best argument for non-TG effects being important in R-It is from the JELIS data. I believe they found that MACE reductions with EPA were not correlated with degree of TG reduction. I think they even stated in the discussion that the study results implied that non-TG mechanisms might be responsible for EPA benefits. So I think that while the TG reduction story is important, it is not the whole story.
Kiwi - A couple of comments on this article (which I just looked at).
1) It was in Sweden, which likely has high fatty fish consumption (based on my experience working next door in Norway). May be high background O3 levels just like JELIS.
2) The abstract is a bit misleading. From text, it says:
The strongest predictors regarding the risk of acute myocardial infarction were the glycated hemoglobin level, systolic blood pressure, LDL cholesterol level, physical activity, and smoking (Fig. 2B). These risk factors showed a linear association
with the risk of acute myocardial infarction (Fig. 3B).
The results show a stepwise increase in the hazard ratios for each additional variable that was not within the target
range among patients with diabetes
Kiwi - Flip that around the other way and what you posted is consistent with what I posted previously from the RWE studies (i.e., increase TGs by 100mg/dL and you get a 16% increase in risk). In terms of hypothesized treatment effects, what you cite may be correct for pure TG lowering effects, but this ignores the known effects (and presumed CV benefits) of V on inflammation and other relevant CV risk biomarkers. We shall see.
GGW - FWIW - With the designs of the RWEs I mentioned, one of the studies showing TGs elevated MACE risk by 35% was based on a sample that was mostly risk category 2 (diabetic + another risk factor). I think that is a good sign.
Kiwi - The Kaiser study was LDL-C between 40-100 mg/dL. Others just said statin-controlled LDL-C.
Raf - I know some recent studies (not using TGs to select patients) showed a placebo rate around 5.2%, but I think I recall seeing one when I was scanning articles for this in the past that was about 4.9%. So I tried to take a little lower than the lowest I recalled seeing as the baseline for looking at TG risk effects (just as an example). It could definitely in reality be higher even without the added TG risk. It really was just a thought exercise to show what effect TGs might have on event rates in R-It, and show how much thought the company has put into their "model," which seems to be contributing to their confidence in the final results.
Kiwi - FWIW - Baker Brothers remains unchanged at over 22 million shares held as of 8/14. I am sure they have good experts as well. Clearly not all experts agree.
Regarding impact of TG selection criteria in R-It on RRR:
As has been discussed ad nauseum (with apologies to JL), we can hypothesize different RRR outcome scenarios based on estimated composite rate from PYs and total number of events, using different placebo rate scenarios. These rough models suggest that a 5.2% placebo rate (identical to assumptions in the study design) will be necessary to meet the target of 15% RRR. The actual placebo rate is the big unknown in these estimates. Looking at relevant trials, NONE of which include a specific TG level as an entry criterion (and which had median TGs < 150 I believe), I don't think the placebo rate has been lower than 4.8%. The question is how much would the R-It entry criterion for minimum TGs > 150 (later >200) increase the placebo event rate above these previous studies that did not require a particular TG level?
So, I looked again at the 3 recent RWE studies AMRN has presented recently, and it is clear that they have attempted to parallel R-It as much as they can. One study looked only at patients with DM (who satisfied stratum 1 or 2 R-it entry criteria), and was conducted only in the Kaiser database. Another looked at a national claims database, and looked at non-DM stratum 1 patients (existing CVD) and DM patients meeting stratum 1 or 2 entry criteria (the majority in the latter group, which Kiwi likes to consider lower risk). A third also looked at a national claims database, and looked at patients with existing CVD or DM (apparently irrespective of other risk factors). The first two compared those with TG<150 to those with TG = 200-499. The last compared those with TG<150 to those with TG>150. The increase in MACE event risk associated with elevated TGs across these studies using different databases was 26-35%. This suggests that a reasonable estimate of placebo event rate in R-It (regardless of risk stratum 1 or stratum 2 ) might be 4.8% + (4.8% x 26%) = 6%. Even if this estimate is high, it does suggest there is potentially quite a bit of leeway to exceed the minimum placebo rate of 5.2% for which the study was powered to be able to show significant results.
Interesting points JL
Raf - Screwed short term, but maybe not long-term. That bit of info would not come out until the publications come out, which would likely be at minimum 6-9 months post R-It topline. I hope that they use something like a cross-lagged panel analysis, where they could look at changes in TGs from baseline to some trial mid-point as predictors of decreased MACE risk from mid-point to end of the study. This can be interpreted to some extent as causation (TGs cause decreased MACE risk) rather than correlation. Hopefully they have creative statisticians.
Thanks for the correction.
Kiwi - My take on Budoff's comment was that JELIS was useful but not definitive because it was not placebo controlled, was lower dose EPA, and was not high risk patients (different sample). As a scientist, those would be the obvious things that would make me cautious about extrapolating. I agree with his comment which I recall suggested RRR in the 15% range as designed with high (85%) probability. I am leery of expecting anything higher because studies that I have been involved in rarely have outstanding effect sizes, but maybe that's just me.
Kiwi - If we assume V works best in higher risk patients, one factor in the failure to stop at interims might also be the initially lower TG levels required to participate. The higher risk patients were enrolled later in the study, and it is conceivable that the efficacy has increased over time as more higher risk patients were enrolled and had been in the study for extended periods.
Kiwi - I think the 2 risks you point out are reasonable. However, I doubt many patients were in the trial for 7 years (median about 3 years?).
Isaeed - Total number of adverse effects in JELIS: Placebo = 21·7%, EPA = 25·3%. "Most were regarded as mild." Agree that it is higher with EPA, but no idea whether this would be higher with V at higher EPA dose. I expect we could have a few more dropouts in the V group, but with a decent effect size, results should be good regardless (based on JELIS subgroup analysis). No way to tell for sure though.
Me - I think an important issue to consider is that if enrollees in R-It get a blood test, what marker would their MD be looking at to say there is something wrong and change treatment? All are already on potent statins to reduce LDL, increasing HDL does not work, and my impression is that there is little concern among MDs at present about controlling TGs in those without very high TG levels (based on the street's lack of belief in R-I success, statements from multiple MDs in print, mixed scientific evidence, and comments from a Harvard trained cardiologist friend of mine). R-It is trying to prove for the first time the benefits of TG reduction over and above statins in those with moderately elevated TG levels, an indication for which there is no currently approved drug. While I share some concerns about dropouts and confounds that have been expressed here, I think there is not really much an R-It patient's private MD would change based on off-study blood tests. They might try to improve DM medications, which could help reduce MACE risk some, but not much else. All of the above should also be considered in light of standard clinical trial requirements (likely in R-It also) that patient's sign up for when they consent to be in the study that they will not change meds that might alter what the study is targeting during their study participation. Some may do so anyway but I don;t think the % will be high.
That is what they get from the DMC, but not sure whether that means they do not have any access to the full blinded dataset. Others on the board seen anything on this?
TTE - While the company is blinded to treatment assignment, I have always wondered whether they had access to the full blinded dataset. If they did, it seems like they could look at TG reductions (or other biomarkers that V affects) over time as they relate to MACE events, and have a very good idea whether V is working. I know of at least one company that has actually published blinded data, so this does not seem impossible. I also worked as a consultant with a private pharma that seemed to have access to their blinded dataset before formal datalock in an FDA study. Wondering whether there are other precedents you know of in this regard? This would certainly explain the company's high confidence in R-it results if true.
And the V group takes longer to reach those no events
JL - True. Our lack of attention to high TG as a risk elevator in R-it stems (at least fro me) from the fact that no prior studies have selected specifically for this a priori (only have post-hoc subgroup analyses to look at, most of which suggest high TGs will be important). I agree that this is very important to factor in the "guess the P rate contest."
Agreed, and 1 extra criterion is the minimum - many patients may meet multiple criteria.