Amarin Corp Plc (AMRN)

[iwfal]

Me: Q: if the same AE shows up in 3/4 of trials, but is stat sig in none, do you dismiss it? A: absolutely not. It is probably real.

You: Why? If not stat sig, means results were random, could have occurred by chance - just because multiple trials (likely with huge differences in trial design and patient population) might show what appears to be a pattern, that does not mean the pattern is real. You could probably find other invalid patterns if you dug deep enough.

I am not sure how to answer this without being a little pedantic, so I apologize in advance.

Question: If you saw 8 trials and 6 of them were p=0.06 for mortality harms (and the remaining 2 were trended for harm (i.e. HR>1) but p>0.5)?

How about if you had 24 trials and 18 were p=0.06 for mortality?

Etc

My point is that p>0.05 is not black and white. It depends on the p value(s), the totality of trials, whether it is a one-of-many endpoint (or one that is always the most important secondary endpoint - mortality), what the base rate is, ... .

As I've noted I think that because the two biggest Omega 3 randomized trials (Jelis and OMEGA) both trended meaningfully for a mortality harm due to treatment it is odd to dismiss it. Is it 'proven'? No, of course not. But I'd happily take 1:2 odds that the REDUCE-IT trial show >5% excess mortality for Vascepa. Will it be stat sig against Vascepa? I think that unlikely - just because there aren't enough events. Would the FDA make note of the excess mortality when creating the Vascepa label if it shows >5% excess mortality in REDUCE-IT? If the p value is less than 0.3, then I'd guess yes, because of Jelis and OMEGA. (they did something similar with the bleeds already - i.e. they referenced other Omega-3 bleeds data in the Vascepa Review).

(Personal note: If REDUCE-IT shows excess mortality I sure as heck wouldn't take it unless there were a significant quality of life benefit. Too much risk of dying before my time.)

PS I will respond to your other post. But it will require time since there are nuances etc.

PPS To everyone else on the board: if you are actually genuinely interested in what I've said and have a thoughtful/nuanced response - then I will respond. But if your only interest is immediately 'disproving' it then I will not. No malice is intended - but I have other things I need to get done.