"Let's assume we do not pass the 3 point MACE test but we excel with the 5 point MACE (best on the planet), and do well with the CKD and AD. What will happen with stock price, and will FDA allow us to market drug for 5 point MACE."

I'm not sure if 5-point MACE is even a secondary outcome in BETonMACE. I think it is, but the wording isn't explicitly clear. The trial listing reads "Secondary Outcome Measures: Time to first occurrence of adjudication-confirmed broadly defined MACE.....CV death, non-fatal MI, hospitalization for CVD events, or stroke." But in a recent corporate update, slide 18 indicates "Secondary endpoints include: hospitalization for unstable angina, emergency revascularization procedures, and all-cause mortality."

Let's assume that 5-point MACE is indeed a secondary outcome in BETonMACE and it includes the strict 3-point MACE (CV death, non-fatal MI, non-fatal stroke) and the additional 2 soft events of hospitalization for CVD event/unstable angina and emergency coronary revascularization. If the placebo and apabetalone groups are statistically different for time to first occurence of 5-point MACE, but not 3-point MACE, this would probably mean that an overwhelming majority of the 5-point MACE reduction was driven by hospitalization for CVD event/unstable angina and/or emergency coronary revascularization. It would also probably mean that difference between the groups for the collective measure of CV death, non-fatal MI and non-fatal stroke was more modest. 

Is this a killer? On the surface, yes this would appear horrible! But keep in mind that: 1) The 3-point MACE metric is a composite of CV death, non-fatal MI, and non-fatal stroke; 2) Cut-offs for statistical significance, while important, are somewhat arbitrary; and 3) There are many pre-specified analyses of sub-groups for the primary outcome. 

I remember in EMPA-REG OUTCOMES for empagliflozin (see NEJM article here) that it achieved a 14% RRR in 3-point MACE with a p-value for superiority of p=0.04. If I understand correctly, this JUST squeaked by with statistical significance. Looking at the individual components of the 3-point MACE, this 14% RRR was driven by a non-significant (p=0.16) 24% increase in non-fatal stroke, a non-significant (p=0.22) 13% RRR in non-fatal MI, but an amazing 38% RRR in CV death (p<0.001). It was this latter data on CV death reduction that got a standing round of applause when it was presented in fall of 2015. 

What will happen if BETonMACE is at the margins of statistical significance for 3-point MACE? It may depend on which side of the arbitrary p-value cut-off line BETonMACE lies and what the individual components fo the 3-point MACE look like. No one is going to get excited about reductions in hospitalization for CVD event/unstable angina or emergency coronary revascularization. But in my opinion a huge victory in CV death, despite a disappointing 3-point MACE composite, would still be a remarkable achievement.

Now, would the FDA approve apabetalone if 3-point MACE fails to meet statistical significance but CV death reduction was  remarkable for the entire BETonMACE population? Would FDA approve apabetalone if 3-point MACE for pre-specified subgroup(s) (i.e. rosuvastatin + apabetalone works MUCH better than atorvastatin + apabetalone) was statistically significant but 3-point MACE for entire population was not? What impact would failure to achieve statistical significance on 3-point MACE have on the future of the cognition and renal programs, assuming the latter two sub-studies show promising results? All great questions that I don't have an answer to.

BearDownAZ