Amarin Corp Plc (AMRN)

[IgnoranceIsBliss]

OK thanks all for the memory jogging...

Here's some math for your consideration.

Three things:

- If V reduces primary events overall by 25%, it's reasonable to assume it also reduces secondary events by 25%.

- The longer ramp (4.9 vs 2.2) should increase the proportion of secondary events to primary from FOURIER (as might other factors you mentioned in terms of CV risk profile). Let's assume that instead of a FOURIER ratio of 4266 / 2907 = 146.7%, let's assume we have a ratio of 180%... and 180% of 1612 is 2902 total SE events.

- If there are more secondary events, they are more likely to be hard events, because things tend to get worse, not better. So let's now say about 55% of all events are now hard MACE, instead of a little less than 50%.

At 25% RRR, the primary events split 691 V vs 921 placebo. If we assume the 691 V's are 25% less likely to have a second event than the 921 placebos, then the additional (2902 - 1612) = 1290 events are distributed 465 to V and 825 to placebo.

That would result in an overall RRR, inclusive of additional events, of....

Wait for it...

Wait for it...

Almost 34%!!!!

Now are you ready to see the p-values if the 34% occurs uniformly across the different events?

Events RRR P-value
CVD 378 34% 0.0000531554290926
MI 853 34% 0.0000000012941479
Stroke 361 34% 0.0000782409205244
Revasc 1026 34% 0.0000000000283591
Angina 283 34% 0.0004773709900376

Blended 2902 34% -
Hard MACE 1593 34% 0.0000000000000002
All Death 670 34% 0.0000000747022977

There are some oversimplifications in this, because RRR isn't just about event counts, it's also about time (later improves RRR as well). And there is more nuance to calculating each SE than this.

But still -- if this is in the right neighborhood?

What do you think about THAT, people?