"Bear forgive my ignorance on the subject and you guys most likely have talked about this before. But for us latecomers, whom actually sets the primary endpoint. RVX? Steering committee? FDA? Combination of two or three?"

Probably a combination of the three in most cases.  

"Second question. If data shows it was just shy does the FDA still see it as a fail? There must be some wiggle room?  Why even such a precise number. Why not a range? We are not splitting an atom ....."

I don't have a great answer. Only a few good references as a starting point. I agree that the cut-off for statistical significance is arbitrary. There is minimal difference between p=0.04 and p=0.06 other than one is below and one is above the 0.05 cut-off. 

When the Alpha is the Omega: P-Values, “Substantial Evidence,” and the 0.05 Standard at FDA

The Primary Outcome Fails — What Next?

Statistics in Medicine — Reporting of Subgroup Analyses in Clinical Trials

A side note of two recent trials for the diabetes drugs empagliflozin (SGLT2 inhibitor) and oral semaglutide (GLP1R agonist). These drugs were already approved for glucose lowering, but needed to show cardiovascular safety by proving to be non-inferior to placebo and possibly superior to placebo. And if they showed superiority for cardio risk reduction, a very valuable expanded label to include cardio indication would be the prize.

EMPA-REG OUTCOME for the SGLT2 inhibitor empagliflozin achieved a p<0.001 for non-inferiority but only a p=0.04 for superiority for 3-point MACE primary endpoint. Empagliflozin now has a cardiovascular indication label. But what if it would have had a p=0.06 for their primary endpoint? Would this have invalidated their finding of 38% RRR (p<0.001) in cardiovascular death (one of the three 3-point MACE components) that got a standing ovation when these trial results were first presented? Would the medical community have deemed this drug a failure? I don't know. Just sharing this example of barely making the cut-off.

PIONEER-6 for the oral semaglutide is still a work in progress. Novo Nordisk had previously completed the cardio outcomes trial SUSTAIN-6 for injectable semaglutide with impressive results. PIONEER-6 was a smaller trial that seemed to be just designed to prove non-inferiority to placebo, which it did. However, it did not achieve statistical significance for superiority (see here). What does this mean? Yet to be determined. There were mixed results between SUSTAIN-6 and PIONEER-6 that makes cardio label uncertain in my opinion. Oral in PIONEER-6 did not show superiority (non-significant 21%RRR); most effects on CV death & stroke. Injectable in SUSTAIN-6 had significant 26% RRR but no effect on CV death (mostly stroke & MI). Both small CVOTs. Novo has already filed for a CV risk reduction claim using both the SUSTAIN-6 and PIONEER-6 datasets. Decision expected early next year.

BDAZ