Obviously, a near 5% (4.7%) level is one that would be unacceptable to the FDA.
I've seen published thresholds on what constitutes an event (i.e. >3xULN), but I have never seen absolute thresholds on rates. That said, I expect you are right that the FDA would not approve a dosage of a drug for a non-life-threatening illness with an existing treatment option if that dosage created 5% >3xULN. (But they might if showed no Hy's Law issues and it was to treat something life-threatening with no existing treatment options.)
Does RPRX submit 50 mg data for a NDA? I doubt it. If this is the case, the number of patient exposures has decreased markedly. This will set them back considerably.
I consider it an absolutely done deal that the company will not ask for approval of the higher dose - but they will submit the data because they are obligated to. And the data will almost certainly count towards the safety database except where there is reason to believe that there is an inverted safety vs dose curve and the existing database is not large enough. My *guess* is that only the bleed issue has evidence of inverted safety vs dose database - and I suspect that is not the main reason for the FDA's desire for a large database. I.e. I think there is some chance you are correct that it will impact safety database - but it isn't a done deal IMO.
Market Data 
Markets 
AI
